==============
Page 1/1
==============

WHAT MAKES THEM CLICK?
APPLYING THE RATIONAL CHOICE PERSPECTIVE
TO THE HACKING UNDERGROUND

by

MICHAEL BACHMANN
M.A. University of Mannheim, 2004

A dissertation submitted in partial fulfillment of the requirements
for the degree of Doctor of Philosophy
in the Department of Sociology
in the College of Sciences
at the University of Central Florida
Orlando, Florida

Summer Term
2008

Major Professor: Jay Corzine

ABSTRACT
The increasing dependence of modern societies, industries, and individuals on
information technology and computer networks renders them ever more vulnerable to
attacks on critical IT infrastructures. While the societal threat posed by hackers and other types of cyber-criminals has been growing significantly in the last decade, mainstream criminology has only recently begun to realize the significance of this threat. Cyber-criminology is slowly emerging as a subfield of criminological study and has yet to
overcome many of the problems other areas of criminological research have already
mastered. Aside from substantial methodological and theoretical problems, cybercriminology currently also suffers from the scarcity of available data. As a result, scientific answers to crucial questions, such as who exactly the attackers are and why they
engage in hacking activities, remain largely fragmentary.
The present study begins to fill this remaining gap in the literature. It examines
survey data about hackers, their involvement in hacking, their motivations to hack, and
their hacking careers. The data for this study was collected during a large hacking convention in Washington D.C. in February 2008. The theoretical framework guiding the
analyses is the rational choice perspective (Clarke & Cornish, 1985). Several hypotheses about hackers are derived from the theory and some of its models are transposed
into the context of hackers. Results suggest that the rational choice perspective is a viable theory when applied to cyber-criminals. Findings also demonstrate that the creation
of more effective countermeasures requires adjustments to our understanding of who
hackers really are and why they hack.
ii

To my wife and family, who of all that walk the earth, are most precious to me and
whose continuing support and love enabled me to write this dissertation.

iii

ACKNOWLEDGMENTS
I would like to deeply thank some of the persons who, during the several months
of this endeavor, provided me with the necessary support and helpful assistance. Without their care and consideration, this dissertation would not have matured in the same
way.
In particular, I would like to thank the members of my dissertation committee for
their encouragement, guidance, and their feedback during all phases of this project.
One of my wisest decisions as student was to select Dr. Jay Corzine as my dissertation
chair. He was an excellent and committed supervisor of this project, and he also proved
to be an invaluable help during my job search. I would further like to thank him as well
as Dr. Jana Jasinski and Dr. James Wright (in alphabetical order) for providing me with
the opportunity to even become a Ph.D. student at University of Central Florida, and to
do so at a unique moment in time—during the start of a new program. All of them provided me with an excellent education that significantly altered my life-course trajectory
and turned me into the person I am today. I would further like to extend my gratitude to
my external committee member Dr. John Jarvis, whose contacts and expertise were exceptionally helpful for this project.
Second, I would like to especially thank my colleague Dr. Monica Mendez for her
offer to review and edit my drafts. Without her nuanced understanding of English language, this dissertation would not have been written in the same way. I cherish the time
we spent together and hope we will be able to stay in touch despite the large distance
that will be between us.
iv

TABLE OF CONTENTS
LIST OF FIGURES ........................................................................................................ viii
LIST OF TABLES ............................................................................................................ix
LIST OF ACRONYMS/ABBREVIATIONS ....................................................................... x
CHAPTER ONE: INTRODUCTION ................................................................................. 1
CHAPTER TWO: THE MEANINGS OF HACKING AND HACKERS.............................. 7
The Controversial Meanings of Hacker and Hacking ........................................ 7
The Etymological History of the Term Hacker ................................................... 9
The Current Controversy Over the Meaning of Hacker ................................... 12
The (Mis)representations of Hackers in the General Media ............................ 16
CHAPTER THREE: THE GROWING THREAT OF CYBERCRIME ............................. 24
The Co-Evolution of the Internet and Cybercrimes ......................................... 25
The Novelty of Cyber-Criminology .................................................................. 29
The Novelty of Cybercrime.............................................................................. 30
CHAPTER FOUR: THE CHALLENGES IN RESEARCHING CYBERCRIME .............. 35
Methodological Problems ................................................................................ 35
The Problem of Defining and Classifying Cybercrimes ............................... 35
The Problem of Measuring the Scale of Cybercrimes ................................. 38
Theoretical Problems ...................................................................................... 42
The Problem of Isolating Structural Correlates ............................................ 42
The Problem of Isolating Offender Characteristics ...................................... 43
The Proposed Theoretical Approach .............................................................. 45
v

CHAPTER FIVE: THE RATIONAL CHOICE PERSPECTIVE ...................................... 48
The Six Basic Propositions of the Rational Choice Perspective...................... 49
The Rationality of Criminal Acts .................................................................. 50
The Limitations of Rationality in Decision Making Processes ...................... 51
The Importance of Situational Adjustments ................................................. 55
The Differences between Involvement and Event Decisions ....................... 57
The Three Different Stages of Involvement ................................................. 58
The Reconstruction of Crime Events as a Decision Sequence ................... 66
CHAPTER SIX: RESEARCH DESIGN .......................................................................... 69
The Difficulty of Sampling Hackers ................................................................. 70
The ShmooCon Convention ............................................................................ 75
The Pretest and IRB Approval ........................................................................ 79
The Data Collection Process........................................................................... 82
CHAPTER SEVEN: THE SURVEY INSTRUMENT ....................................................... 85
The Measurements of General Hacking Activity ............................................. 87
The Measurements of Risk Propensity and Rationality................................... 99
The Measurements of Sociodemographics ................................................... 101
CHAPTER EIGHT: METHODS AND RESULTS ......................................................... 103
The Sociodemographic Composition of the Sample ..................................... 103
The Descriptive Statistics of Continued Involvement .................................... 108
The Initiation Phase ................................................................................... 108
The Popularity of Different Hacking Methods ............................................ 114
The Prevalence and Success Rates of Different Kinds of Attack Methods 117
vi

The Developments During the Habituation of Hacking .............................. 123
The Risk-Avoidance Efforts and Desistance Considerations..................... 129
The Assessment of Scale Validity and Reliability.......................................... 133
The Exploratory Factor Analysis................................................................ 134
The Test of the Study Hypotheses ................................................................ 137
The Test of Hypotheses H1 and H2 .............................................................. 139
The Test of Hypotheses H3 and H4 .............................................................. 142
The Test of Hypotheses H5 and H6 .............................................................. 145
The Test of Hypotheses H7 .......................................................................... 147
CHAPTER NINE: DISCUSSION, IMPLICATIONS, AND LIMITATIONS ..................... 150
The (In)Accuracy of the Common Hacker Stereotype................................... 150
The Suitability of the Rational Choice Perspective ........................................ 154
The Policy Implications of the Study Findings ............................................... 157
The Limitations of the Present Study and Suggestions for Future Research 161
APPENDIX A: THE SURVEY QUESTIONNAIRE ....................................................... 168
APPENDIX B: THE SURVEY CONSENT FORM ........................................................ 177
APPENDIX C: THE IRB APPROVAL LETTER ........................................................... 180
GLOSSARY ................................................................................................................ 182
LIST OF REFERENCES ............................................................................................. 199

vii

LIST OF FIGURES
Figure 1: Model for the Initiation of Hacking .................................................................. 61
Figure 2: Model for the Habituation of Hacking ............................................................. 63
Figure 3: Model for the Desistance from Hacking ......................................................... 64
Figure 4: One Stage of the Crime Event: The Selection of a Target ............................. 66

viii

LIST OF TABLES
Table 1: Sociodemographic Characteristics of Sample Respondents ......................... 104
Table 2: Motivations for Interest in Hacking and First Hack ........................................ 105
Table 3: Details of the First Hacking Attempt .............................................................. 112
Table 4: Methods Used in First Hacking Attack ........................................................... 113
Table 5: Measurements of Hacking Activity – Technical Intrusions............................. 117
Table 6: Measurements of Hacking Activity – Social Methods .................................... 119
Table 7: Measurements of Hacking Activity – Malware Distribution ............................ 121
Table 8: Developments during Hacking Career ........................................................... 123
Table 9: Target Preferences........................................................................................ 125
Table 10: Anonymizing Methods and Desistance Considerations............................... 128
Table 11: Risk Estimates and Carefulness.................................................................. 131
Table 12: Personality Scales, Item, Factor, and Index Analysis .................................. 135
Table 13: OLS Regression Coefficients for Estimated Effects of Rationality
and Risk Propensity on Hacking Success.................................................. 136
Table 14: OLS Regression Coefficients for Estimated Effects of Rationality
and Risk Propensity on Perceived Risk Involved in Hacking ..................... 140
Table 15: OLS Regression Coefficients for estimated Effects of Rationality
and Risk Propensity on Preparation Time.................................................. 146
Table 16: OLS Regression Coefficients for estimated Effects of Rationality
and Risk Propensity on Total Amount of Hacking Attacks ......................... 148

ix

LIST OF ACRONYMS/ABBREVIATIONS
ARPANET

Advanced Research Projects Agency Network

BJS

Bureau of Justice Statistics

CGI

Common Gateway Interface

CMC

Computer-mediated communication

DDoS

Distributed Denial-of-Service

DNS

Domain Name System

DOJ

U.S. Department of Justice

EFA

Exploratory Factor Analysis

ICMP

Internet Control Message Protocol

I2P

Invisible Internet Project

IP

Internet Protocol

IRB

Institutional Review Board

IT

Information Technology

NCSS

National Computer Security Survey

NCSD

National Cyber Security Division (U.S. Department of Homeland Security)

NIBRS

National Incident Based Reporting System

OS

Operation System

P2P

Peer-to-peer network

PDP

Programmed Data Processor

REI

Rational-Experiential Inventory

RPC

Remote Procedure Calls
x

TCP

Transmission Control Protocol

TOR

The Onion Router

UCR

Uniform Crime Report

UDP

User Datagram Protocol

UNIX

UNiplexed Information and Computing System

URL

Uniform Resource Locator

xi

New technologies do not determine the particulars of human fates;
they alter the spectrum of potentialities within which
people may act (McClintock, 1999, p. 3)

CHAPTER ONE:
INTRODUCTION
Estonia, April 26, 2007. In retaliation for the removal of a World War II-era statue
of a Soviet soldier, pro-Russian hackers launched a month-long campaign that has become known as the first war in cyberspace. Using a technique known as distributed
denial-of-service attack (DDoS) on a hitherto-unprecedented scale, the attackers managed to effectively shut down vital parts of Estonia’s digital infrastructures. In a coordinated effort, an estimated one million remote-controlled computers from around the
world were used to bombard the Web sites of the president, the prime minister, Parliament and other government agencies, Estonia’s biggest bank, and several national
newspapers with requests. The attacks were so massive that NATO rushed a cyberwarfare team of international security experts to assist the Estonian government, and
the country’s defense minister, Jaak Aaviksoo, who described the attack as a national
security situation, requested that the European Union classify it as an act of terrorism
(Landler & Markoff, 2007). In reference to the events in Estonia, Suleyman Anil, the
head of NATO’s incident response center, later warned attendees of the 2008 E-Crime
Congress in London that “cyber defense is now mentioned at the highest level along
1

with missile defense and energy security.” According to Anil, “we have seen more of
these attacks and we don’t think this problem will disappear soon. Unless globally supported measures are taken, it can become a global problem” (Johnson, 2008, p. 1).
Turkey, May 18, 2006. A Turkish hacker who calls himself ‘iSKORPiTX’ commits
the biggest hacking incident in web-hosting history. He successfully hacked and defaced at least 21,549 websites simultaneously and another 17,000 websites a few hours
later (Lemos, 2006). For the time being, this incident marks the largest single defacement attack in the fast growing area of website hackings, but it is merely one of many
cases. According to Zone-H, the Internet’s watchdog on website defacements, iSKORPiTX alone has defaced close to 189,000 websites, and the top 50 attackers hacked a
total of approximately 2.5 million websites across the globe (Zone-H, 2007). These
numbers are supported by the exponential increase in website defacements found in
the CSI/FBI Computer Crime and Security Survey (2005). In 2004, 5% of the respondents experienced 10 or more website incidents, while in 2005 that figure went up to
95% (L. A. Gordon, Loeb, Lucyshyn, & Richardson, 2005).
The above examples are merely two incidents of what have become a long series of high-profile hacking attacks (Aguila, 2008). Although warnings of the societal
threat posed by cyber-attacks on critical network infrastructures have been heralded
since the 1980s, it is only in recent years that the problem has made it onto the radar of
governments. Partly due to the experiences of Estonia, countries around the globe are
now reassessing the security situation of their key information systems. They are enacting new security measures to better protect their critical network infrastructures, and
they are increasing their readiness to respond to large-scale computer incidents
2

(NCIRC, 2008). In Britain, for example, Conservatives have recently proposed the creation of a new position for a cyber-security minister and a national hi-tech crimes police
squad to better combat the “growing and serious threat to individuals, business and
government […] that will continue to escalate as technology changes” (Johnston, 2008,
p. 1).
The implementation of effective countermeasures against hacking attacks is exceedingly facilitated by the vast amount of knowledge that has already been accumulated in numerous computer science research projects (cf. Chirillo, 2001; Curran et al.,
2005; J. Erickson, 2008). Several studies conducted by computer scientists and computer engineers have closely examined the technical details of the various attack methods and have produced a significant body of information that can now be applied to
help protect network infrastructures (Casey, 2004). Unfortunately, the guidance provided by these studies is limited to only the technical aspects of hacking attacks and,
sharply contrasting from the substantial amount of knowledge already gathered about
how the attacks are performed, answers to the questions of who exactly the attackers
are and why they engage in hacking activities continue to remain largely speculative.
Today, the persons committing the attacks remain mysterious for the most part, and
scientific information about them continues to be only fragmentary.
The present lack of information concerning the sociodemographic characteristics
and the motives of cybercrime offenders can be attributed to a number of causes. One
of the main reasons can be traced back to the unfortunate circumstance that, until recently, mainstream criminology has underestimated the potentially devastating societal
impacts of cybercrimes and has diverted only limited attention to this relatively new type
3

of criminal behavior (Jaishankar, 2007; Jewkes, 2006; Mann & Sutton, 1998). Cybercriminology is only now beginning to evolve as a distinct field of criminological research,
and it has yet to overcome many methodological and theoretical problems that other
areas of criminological research have already solved (Yar, 2005b, 2006).
A particular challenge for researchers in this young field of study arises from the
various methodological obstacles entailed in the sampling of cyber-criminals. As a result
of these difficulties, available data sources are scarce, and quantitative studies are limited to surveys of cybercrime victims. At this point, only a few qualitative case studies
(eg. Mitnick & Simon, 2005; Schell, Dodge, & Moutsatsos, 2002; Taylor, 1999, 2000)
and biographies (eg. Mitnick, Simon, & Wozniak, 2002; Nuwere & Chanoff, 2003) exist
that examine individual hackers, their motivations, preferences, and their hacking careers. While such studies are well suited to provide in-depth insights into the lives of a
few individuals, they are unfit for generating generalizable information about the population of hackers at large. Yet, just “like in traditional crimes, it’s important to try to understand what motivates these people to get involved in computer crimes in the first place,
how they choose their targets and what keeps them in this deviant behavior after the
first initial thrill” (Bednarz, 2004, p. 1). This comment, stated by Marcus Rogers, an associate professor at Purdue University and head of the cyber-forensics research in the
department of computer technology, accurately describes the task cyber-criminologists
have to accomplish. The aim of the present study is to undertake this task and to begin
filling the remaining gap in the criminological literature on hackers and the hacking
community by providing the first quantifiable insights into the hacking underground.
Such insights are needed to create a more profound understanding of the nature of the
4

threat and a more complete assessment of the problem and its solutions. The identification of the reasons and motives for an attack helps to better identify the actors’ behaviors, to develop better countermeasures, and to ultimately make systems safer. To accomplish this goal, the study is structured as followed.
First, a detailed introduction into the fiercely contested dispute over the definitions of the terms “hacker” and “hacking” is provided, and the exact usage of both terms
in the study is specified. Second, the typical perceptions of hackers in the general public
are discussed. Most people do not have direct contacts with hackers, but derive their
knowledge of the hacking scene and the digital underground from mass media or stereotype-laden, personal prejudices. Unveiling the common (mis)perceptions about
hackers is an important part of the present study because a clear, scientific vision of the
characteristics of the hacking community has to be established against the context of
the predominant perceptions of who hackers are and why they hack.
Third, the present study is located within the current state of cyber-criminology.
The developments that have led to the establishment of cyber-criminology as a distinct
field within criminological research are traced, and the justifications for this establishment are discussed. Then, detailed descriptions are provided of the current methodological and theoretical challenges with which cyber-criminologists are confronted. When
studying hacking activities, it is important to bear in mind that hacking is a form of cybercriminal activity, and, as such, it is subjected to the various obstacles challenging all
studies of criminal activities that are taking place in the virtual worlds of interconnected
computer networks.

5

Fourth, the theoretical framework guiding the study is introduced. The study follows the rational choice perspective (Clarke & Cornish, 1985), a popular criminological
version of rational choice theory. A rationale for the selection of this particular theoretical framework is provided and its core components are discussed. The central hypotheses of the study are derived from the main propositions of the rational choice perspective, and Clarke and Cornish’s (1985) central models of the decision-making
processes are modified to fit the context of hackers.
Following the introduction of the theoretical framework, the details of the research design are specified. A rationale for the selection of a particular sampling strategy is provided, and the steps involved in the processes of creating the survey instrument and collecting the data are elaborated. The discussion of the various aspects of
the research design is concluded by a separate chapter on the different sections and
individual items of the survey instrument.
The last two chapters of this study present and discuss the findings of the research project. The results paint a detailed picture of distributions of a range of sociodemographic characteristics within the hacking community. They further specify the
relative importance of various motivations, and they assess several components of the
suggested theoretical frameworks. The study concludes with a discussion of the implications these findings hold for our understanding of who hackers are and why they
hack, for the design of more effective countermeasures, and for the appropriateness of
the rational choice perspective as a theoretical framework for the study of computer
hackers. The limitations of the current project are considered and suggestions for future
research are given.
6

CHAPTER TWO:
THE MEANINGS OF HACKING AND HACKERS
An indispensable step in every scientific research project is the explicit operationalization of the main study concepts. The first important step within this process is the
specification of central terms and core concepts with precise definitions. In the context
of hackers, this step is of crucial importance. Devising exact definitions of the terms
hacker and hacking is particularly problematic because both terms are fiercely contested within the hacking community itself and between hackers and the general public.
Thus, a description of the various meanings ascribed to both terms is necessary for a
complete understanding of the various definitions and the disputes between them.

The Controversial Meanings of Hacker and Hacking
Applied to the context of computer technology, the two terms “hacker” and “hacking” were originally used only in small circles of computer experts, and familiarity with
both terms was limited to a few specialized computer scene insiders. The popularity of
the two terms slowly began to increase during the late 1980s and 1990s, when they became more commonly adopted in mainstream media reports of several high-profile cyber-trespasses that sometimes involved shocking exploitations of computer and telecommunication technologies (cf. Aguila, 2008). In recent years, the more frequent media coverage of such computer abuses has turned these formerly uncommon terms into
the everyday vernacular of the mainstream public discourse (K. Erickson & Howard,
2007; Schell & Martin, 2004).
7

Today, the vast majority of the general population has at least some vague idea
about what hackers are and what they do. The typical understanding of a hacker is that
of a perpetrator who illegally invades other people’s computer systems with the intent to
destroy, disrupt, or carry out illegal activities on that system. As Taylor notes in his
prominent study on hackers and the hacking community, the public perceives hacking
as “the unauthorized access and subsequent use of other people’s computer systems”
(Taylor, 1999, p. xi). Unfamiliar with the specific details of hacking, many members of
the general public even tend to imprecisely subsume any illegal abuse of a computer
system or network under the term hacking. This inaccurate perception of hackers is
hardly surprising because hackers and hacking have become the center of present-day
public debates about cybercrimes. In fact, mainstream media reports oftentimes use the
terms hacker and hacking synonymously with cyber-criminals and cybercrimes (K.
Erickson & Howard, 2007; Yar, 2005a). Moreover, most media coverage of hackers
tends to reinforce their stereotypical perception as prototypes of mysterious and dangerous cyber-criminals. As a result, hackers are commonly seen as personified representations of the threat cybercrimes pose to society (K. Erickson & Howard, 2007).
This common conceptualization of hackers and their activities, while unanimously
accepted and unchallenged within the public discourse, is actually deceptive because it
oversimplifies the meaning of hackers and hacking in a twofold way: It overlooks the vicissitudinous etymological history of the two terms, and it ignores that both terms remain deeply contested today. To familiarize the reader with the historical and contemporary controversies surrounding the meanings and connotations of hackers and hacking,
both aspects are discussed in the following section.
8

The Etymological History of the Term Hacker
Since its first appearance in Yiddish language, where it was used to denote
someone as inept as to make furniture with an axe (Schell & Martin, 2004), the term
“hacker” has undergone many changes. According to the Oxford English Dictionary
(2008), the first emergence of the term in English language dates back to 1480, when it
was used to signify a tool used for hacking, chopping wood, or breaking up the earth. A
century later, the first reference to a “hacker” as a person was recorded in English language. Back then, the word was used to identify someone as a cutter, a cut-throat, a
notorious thief, or generally as a bully or a rogue. This meaning was again slightly
changed about 30 years later, when Thomas Cartwright (1618) used the term as a characterization of somebody who mangles words, senses, and meanings. Around the
same time this shift in meaning took place, a second, completely different connotation of
the word “hacker” emerged. This new version signified a farm laborer who hacks or cuts
more than half an acre of ground in a day with a hack (Oxford English Dictionary, 2008).
The etymological history of the term “hacker,” especially the double entendre that existed during the 17th century where it referred to both malevolent outlaws and persons
who excelled in their legitimate profession, is particularly interesting because both connotations have again resurfaced in the contemporary debate surrounding the meaning
of the word.
After it had almost fallen into oblivion in intervening years, the term “hacker” resurfaced again in the context of computer technology during the 1960s. It was reintroduced as a neologism into the specialized and confined language of computer techni9

cians and programming experts, who used it as a positive label for somebody who was
particularly skilled in developing highly efficient, creative, and compact programs and
algorithms. Rather than simply using existing computer technology, “hacker” enthusiasts
were united in their passion for technological innovations and by their playful and individualistic quest to satisfy their intellectual curiosity. “Hacker” in this understanding denotes someone who is obsessed with designing computer software and hardware, a
superb technician who possesses a substantial degree of skill and competence and
spends much of his time writing computer programs. Respectively, “hacking” originally
referred to the continuous improvement of computer program codes and algorithms, of
software applications and hardware components.
It was hackers who first realized the true potential behind the steel facades of
computing machines which—as a side note—were originally developed to better calculate the trajectories of ballistic projectiles. Among the many inventions attributed to
hackers are, for example, the first computer chess applications and even the Apple
computer, one of the first personal computers (Schrutzki, 1989, p. 166). All early contributors to the advancement and expansion of computer technology, all innovators who
developed new computer-based solutions to a multitude of problems, all entrepreneurs
who pioneered and fostered the “computer revolution” (Naughton, 2000, p. 313), and all
those who paved the way for today’s superhighways of the Internet, were considered
prototypical hackers in this original understanding of the term (Levy, 1984).
The original hacker community formed a subculture shaped by the ideals and
moral concepts of the zeitgeist. As did many other sub- and counterculture movements
at the time, the early hacking community was characterized by a fundamental distrust
10

against governmental and military monopolies, power, and authority. Early hackers defied corporate domination of culture and rejected traditional and conservative values,
norms, and lifestyles. Instead, they genuinely adhered to the enlightenment ideals of
human emancipation and self-fulfillment through rational thought. They advocated the
freedom of information, knowledge, and intellectual thought, and promoted the idea that
information and knowledge should be accessible for everyone without restrictions
(Thomas, 2002). Many members of the hacking community were idealists who advertised the usage of computer technology for the higher goals of intellectual discovery, for
creating art, beauty, and for improving the overall quality of life (Schrutzki, 1989). Still
today, programming and other activities that support these views are oftentimes referred
to as “hacktivism” within the hacking subculture to emphasize their political nature (Taylor, 2004).
While their exploratory quests for new information and data frequently included
unauthorized accesses to remote computer systems, traditional hackers undertook such
accesses without criminal intent. Instead, they were carried out to investigate and better
understand the intricacies of different system setups, to utilize existing computing resources (which at the time were very costly), to detect security breaches and weaknesses, and to ultimately enhance the security of computer protections. The vast majority of members of the original hacker community adhered to a “Hacker Ethic,” a set of
rules that was introduced by Steven Levy (1984) to describe the values of the hackers
at the MIT Artificial Intelligence Laboratory. The main principles of this Hacker Ethic are
that: 1) access to computers and anything which might teach something about how the
world works should be unlimited and total; 2) all information should be free; 3) authority
11

should be mistrusted and decentralization promoted; 4) hackers should be judged solely
by their hacking, not ascribed criteria such as degrees, age, race, or position; 5) art and
beauty can be created on a computer; and 6) computers can change life for the better
(Levy, 1984). One important implication of this Ethic is that any form of damage to remote computer systems, be it intentional or as a result of incompetence, is principally
objectionable and contemptible.

The Current Controversy Over the Meaning of Hacker
As was mentioned earlier, the original positive meanings of the terms hackers
and hacking became gradually substituted with negative connotations in the 1980s and
1990s. The increasingly mission-critical nature of computer networks for many industries and the expanding popularity of electronic financial transactions began to interest
many people in breaking into computer systems, not in an attempt to understand them
or make them more secure, but to abuse, disrupt, sabotage, and exploit them. Angered
by what from their perspective appeared to be a misrepresentation and a denunciation
of the hacking community, traditional hackers reacted to this development by introducing the new label “crackers” for unethical and menacing hackers, from whom they attempted to distinguish themselves. It is important to note that in the context of the debate about hackers, the term “crackers” has no racist connotation. Rather, it is derived
from the activity of cracking, or breaking into, a safe and it refers to people who breach
(or crack) security measures on a computer system, a network, or an application with
the intent to damage or exploit the target or to steal information from it. Hackers who
engage in these kind of malicious activities and for whom the label “cracker” was in12

tended, on the other hand, largely reject it because cracking typically involves programming software applications specifically designed to discover and exploit weaknesses. In their logic, the ability to create programs that are able to circumvent or
breach defensive security measures is proof of their ability to write superior code.
Hence, they prefer to refer to themselves as hackers.
The fiercely contested battle over these two labels created considerable linguistic
confusion. Making matters even worse, however, is the circumstance that the distinction
between hackers and crackers is not the only controversial differentiation. Similar to the
distinction that exists between “hackers” and “crackers,” and just as contested, is the
differentiation between “white hat” and “black hat” hackers. Again, these two terms do
not have a racist background, but are derived from old black-and-white western movies,
in which the sinister villain typically wore a black hat, and his law-abiding or lawenforcing counterpart a white hat. Applied within the context of computer technology,
“white hat” denotes hackers who abide by the Hacker Ethic and hold its rules in highest
regard, whereas “black hat” hackers do not commit themselves to the same ethical
standards.
Today, white hat hackers are oftentimes employed or contracted by computer
companies, governments, and financial institutions. Their job is to identify security
weaknesses and holes in computer systems, networks, or newly-developed software
applications and hardware components. Ethical hacking, or “penetration testing,”
emerged as a byproduct of the increasing need for all corporations, organizations, and
social institutions to be permanently connected to the Internet. In an attempt to avoid
being victimized by black hat hackers, many corporations and agencies decided to hire
13

or contract white hat hackers, who secure their networks by legally attempting to break
into them (Damsell, 2003).
Whereas many penetration testers practice ethical hacking as a legal profession,
the term “ethical hacker” remains an oxymoron for many black hat hackers. For them,
true hacking implies compromising the security of a system without permission from an
authorized party. Although there continues to be a controversy over when the term
black hat was coined and to whom it was first applied, many reports identify John Draper (a.k.a. Captain Crunch) as the first cracker to whom the label black hat hacker was
applied, even though he was de facto a phreaker, a person who hacks telephone systems (Schlegel & Cohen, 2007). Draper used a plastic whistle from a cereal box to generate the exact 2,600 Hz tone American Telephone and Telegraph (AT&T) and other
telephone providers used at the time to indicate the availability of long-distance phone
lines. By mimicking the whistling of a dial tone, Draper tricked the controlling connection
switch into stopping all billing because the switch assumed the call had ended. With the
help of their cereal box whistles, Draper and his friends could call each other for free—
at least until irregularities in the account billing were discovered and Draper was sentenced to prison (Schell & Martin, 2004).
Draper and his friends are just one example of the countless young, creative
crackers who experiment with security mechanisms in an attempt to break them and
who learn from each other in the computer underground. While the majority of hackers,
especially the masses of unskilled “script kiddies,” who merely download and execute
preconfigured attack applications and routines, can adequately be subsumed under the
label of black hat hackers (Twist, 2003), the contested nature of the terms “hacker” and
14

“hacking” within the scene is important to bear in mind when studying hackers. Studies
that do not distinguish between white hat penetration testers and criminal black hat
hackers will inevitably introduce a bias that produces distorted and inaccurate results.
Notwithstanding the resistance to the criminal label within the traditional hacker
community, large parts of the general public are either unaware or ignorant of the distinctions between hackers and crackers, white hats and black hats. Similarly, the mainstream media as well as law enforcement agencies and computer security industries
generally do not subscribe to these distinctions. Instead, they subsume any type of
hacking activity under the currently predominant definition as an inherently negative,
criminal activity. They usually also equate hackers with cyber-criminals, merely because
this is how the vast majority of people outside of the community understand hacking and
hackers (Twist, 2003). Thus, the definitions of what constitutes a hacker and hacking
activities are not only deeply contested within the hacking underground, but also between the general public and the members of the hacking community, to whom the
criminal label is universally applied. While many black hat hackers accept or embrace
this label, self-proclaimed white hat hackers consider themselves misrepresented by it
and continue to challenge and reject the label.
Although the above section shows that awareness of the different meanings associated with the two terms “hacker” and “hacking” and of their contested nature is important for the study of hackers, subsequent uses of both terms will refer to the understanding shared in the general public. This usage is not intended to discriminate against
traditional hackers or penetration testers, but has a twofold reason. First, the focus of

15

this study lies solely on illegal hacking activities. The second reason is simply pragmatic
brevity.

The (Mis)representations of Hackers in the General Media
While hacking as the unauthorized intrusion into computer systems with the intent to abuse or exploit them is undoubtedly at the center of commonly held stereotypes
about hackers, their conventional representations are far more complex. They involve
several additional assumptions, many of which are problematic because they are stereotypical misrepresentations (Thomas, 2002). The distorted picture of hackers is partly
a result of the circumstance that only very few people have direct personal contacts to
actual hackers in their social networks. Without direct personal experience, the vast majority of people derive their knowledge about hackers from representations in popular
fiction and the media. Thus, an examination of popular hacker representations has to
proceed from reviewing their depiction in mainstream media and culture.
The representation of hackers in the general media can be best understood
against the wider context of societal responses to technological innovations and transformations (Taylor, 1999, 2000). In the past, many such transformations have provoked
both technophobic fears and technophilic fascinations in equal measures. Historically,
the introduction of any new technology with the potential to transform social structures
and interactions has routinely created a tension between an anticipation of desirable
improvements in everyday life and an anxiety that revolves around unintentional or unforeseeable negative consequences (Simon, 2004). Numerous examples of such societal reactions exist in modern history, be it after the introduction of steam engines,
16

trains, electricity, telephones, automobiles, airplanes, television, nuclear bombs, computers, or the Internet (cf. Simon, 2004). Many famous novels and fictional stories owe
their success to their effective play on such technophobic fears. In Mary Shelley’s Frankenstein, a classic within this body of literature, the fictional scientist Dr. Victor Frankenstein uses electricity to revive a monster he composed of dead body parts. Despite
his good intentions, his ignorance exceeds his knowledge and he ultimately pays for his
attempt to play God with his life (Shelley, 1995). The tale of Dr. Frankenstein exemplifies more than just the fears technological transgressions of unseen boundaries oftentimes evoke. It also hints at the religious dimension of technophobic fears. Several classic tales, such as the Greek tale of Prometheus, indicate that the misgiving that mankind as the creator of fundamental accomplishments interferes with the prerogatives of
the gods and ultimately must pay the price for this sacrilege exists since ancient times.
Throughout the long history of this fear, its prominence appears to be highly correlated
to the significance of technological inventions (Edgerton, 1995). The multitude of tremendous inventions during the last century and the fundamental societal changes they
have wrought have turned the peoples’ fears of mad scientists, who unintentionally unleash their destructive and monstrous technological creations on us, into a dominant
theme in contemporary culture (Tourney, 1992; Tudor, 1989).
Since their invention in the mid-20th century, computers, in particular, have
changed almost all aspects of society and human life in unprecedented ways (Negroponte, 1996). This tremendous societal influence has turned them into today’s primary
projection screen for technophobic fears and anxieties surrounding technological innovations. Exacerbating societal fears of computers is the rapidity of their continuously
17

advancing technology. The exponential growth of computerized calculating capacities
(Kurzweil, 2005) is daunting to many people who are afraid that mankind, as creation’s
crowning glory, will eventually be challenged by a superior artificial intelligence. The
contemporary prominence of fears that mankind will be rivaled by new forms of computerized artificial intelligence is visibly reflected in the popularity of this recurring theme in
popular fiction. In Stanley Kubrick’s masterpiece 2001: A Space Odyssey (1969), for
example, the initially philanthropic spaceship computer HAL 9000 begins killing the crew
members when faced with the prospect of disconnection. The artificial intelligence system Proteus IV in Donald Cammell’s Demon Seed (1977) even rapes its creator’s wife
in an attempt to conceive a human-machine hybrid successor who is no longer subjected to any of its boundaries (Kozlovic, 2003). Since Demon Seed, attempts of artificial computer intelligence to subdue and enslave mankind have been a recurring theme
in numerous blockbuster movies, such as the Terminator and its accompanying sequels
and The Matrix trilogy (Dinello, 2005).
Against this background, many hacker movies and stories can be characterized
by their distinctive play on the fear of unleashed and uncontrollable technological catastrophes. In the classic hacker movie War Games (Badham, 1983), a young hacker exploits a back door into a central military computer. Not realizing that he is actually
logged into a computer system that has control over the United States’ nuclear arsenal,
he starts playing a game of global thermonuclear war and unknowingly brings the world
to the brink of nuclear annihilation. Such blurring of the line between simulation and reality is a common theme in hacker movies. In many such films, the fate of real persons,
organizations, countries, or even the world is decided by actions that take place in a
18

contested virtual space. In eXistenZ (Cronenberg, 1999), the line between video game
and reality becomes blurred in a shocking and gruesome way when a video game designer creates an artificial reality game that is played directly in peoples’ minds. Similarly, the data trafficker Johnny Mnemonic (Longo, 1995) stores sensitive information directly in a cybernetic storage system within his brain and is forced to hack his own brain
when parts of the secret message become lost. Already in the classic movie Tron (Lisberger, 1982), a hacker is transported into the digital universe inside a computer, where
he must survive combat as a cyber-gladiator in order to stop the artificial intelligence
that has taken over his computer company. The implicit suggestion of these and many
other fictional and non-fictional hacker stories is that the preoccupancy of hackers with
computer technology and cyberspace dehumanizes them.
Such dehumanizing depictions further reinforce the common stereotype of hackers as antisocial, reclusive, introverted, and dysfunctional hermits, whose computercentered escapism from the social worlds of interpersonal interaction eventually endows
them with superior technological skills. Similar stereotypical portraits of hackers characterize them as young male misfits and fledgling teenagers who, rather than being intimidated by the latest computer technology, embrace this technology instead. They are
seen as “geeks,” juvenile technophiles to whom computer technology provides an opportunity for self-realization in an environment that is not laden with the complexity and
unpredictability of human interactions. By excelling in their mastery of this technology,
hackers supposedly compensate for their alleged social shortcomings. Along these
lines, the negative stereotype of hackers displays them as teenagers who are initially
driven to their computers by social deficits and who, through their preoccupation with
19

computer technology, become even further detached from reality. In some instances,
hackers are even seen as juveniles who are being driven by feelings of contempt and
disdain toward a social environment that excludes them and who use their hacking skills
as a resource for taking revenge against the persons or institutions they scorn. As one
commentator notes, “the fact that hackers are also invariably young in popular perceptions is also not a coincidence – the apparent ease with which a ‘younger generation’ is
able to engage with the realm of computer technology, a technology that many older
people continue to see as mysteriously daunting, merely serves to sharpen the sense
that all manner of extravagant things may be possible for those with the know-how”
(Yar, 2006, p. 25). Thus, the portrayal of hackers as exclusively young seemingly mirrors technophobic fears that are particularly prevalent among older generations. Many
members of older generations feel disconnected, left behind, and disadvantaged by the
spreading of a technology that provides younger generations with the advantage of instantly accessible and virtually limitless information (Dowland, Furnell, Illingworth, &
Reynolds, 1999; Negroponte, 1996). Hackers, in particular, stand out due to their mastery in manipulating and utilizing this seemingly incomprehensible computer technology,
an ability that in the perception of many borders on wizardry. It is this uncertainty and
weariness about the ominous capabilities of hackers that provokes technophobic fears
and paints the hacker image negative.
Notwithstanding the many unfavorable components of commonly held beliefs
about hackers, their public representations are not unanimously negative, but can best
be described as intrinsically ambivalent. Just as technological inventions cause anxiety
and fascination in equal measures, persons who master these inventions also evoke a
20

particular fascination and admiration. This admiration is also reflected in many movie
portrayals of hackers, in which they are displayed as socially maladjusted yet brilliant
heroes who use their extraordinary computer expertise to save the world from evil viruses, rampaging corporations, nuclear Armageddon, or even invading aliens. Contrasted with classic action heroes, they are no longer muscular one-man armies, but
technologically sophisticated protagonists, “keyboard cowboys” (Softley, 1995), who
solve the world’s problems with keystrokes rather than guns.
Even though many fictional hacker characters might have had some trouble with
the law in the past, they are typically displayed as essentially good-natured characters
with whom viewers can identify. As Thomas (2002) points out, the criminality of hackers
is oftentimes contextualized, negotiated, and relativized by the narratives themselves. In
many storylines, their illegal actions are justified as being forced by others, accidents,
or—most commonly—unavoidable means to expose horrible secrets, injustices, or to
fight a greater evil. Thus, while seen as principally threatening because of their selfimmersion into abstract, virtual, and dehumanized worlds, they are also portrayed as
“freedom fighters of the 21st century” (Kovacich, 1999), heroes who either reclaim human authority over rivaling technological entities (Taylor, 2000) or fight against unjust
domination by corrupt governments and megalomaniac corporations. Particularly the
latter is a dominant theme in many fictional hacker stories. In Hackers (Softley, 1995),
for example, a clique of juvenile hackers accidently discovers and prevents a largescale corporate conspiracy that would have otherwise led to an ecological disaster. Similarly, a computer hacker becomes suspicious of his corrupt employer, a multinational
computer company CEO that is loosely based on Bill Gates, in the movie AntiTrust
21

(Howitt, 2001). His investigation into company procedures confirms his suspicion and he
eventually unveils a ruthless and murderous plot to dispatch the company’s anti-trust
problems.
Summarizing, fictional accounts and portrayals of hackers can be characterized
as being inherently ambivalent in their reflection of technophobic fears and technophilic
fascinations. This cultural ambivalence is not only mirrored in the positive and negative
associations evoked by the terms “hacker” and “hacking,” but also in surveys that examine public opinions about hackers (Dowland et al., 1999; Voiskounsky, Babeva, &
Smyslova, 2000). Dowland et al. (1999) found principally ambivalent attitudes toward
hackers. The study also demonstrated that the division between rather positive and rather negative opinions runs alongside generational lines. Whereas members of older
generations expressed primarily anxious and negative attitudes, younger respondents
tended to see hackers in an overall positive light (Dowland et al., 1999). One year later,
Voiskounsky and his colleagues again found the same patterns expressed in a study
conducted in Russia. In Voiskounsky’s survey, young Russians expressed a particular
admiration for the intriguing cleverness and intelligence they ascribed to hackers
(Voiskounsky et al., 2000, p. 72-3).
Thus, the ambivalence between fears and admirations in the perception of hackers as “a schizophrenic blend of dangerous criminal and geeky Robin Hood” (Taylor,
2000, p. xii) appears to be a universal pattern in mainstream culture as well as public
opinions. Yet, this ambivalence is not the only one in the common perception of hackers. While both fears and admirations are sparked by the ingenious computer skills typically ascribed to hackers, such skills are commonly seen as having been developed as
22

means for compensating social isolation and underdeveloped social skills. The admiration for hackers is undoubtedly caused by the inability of most people to comprehend
the techniques hackers employ to manipulate computers. The explanation of such superior computer skills as a compensation for deficient social skills, however, presents an
unvalidated and questionable assumption in the representation of hackers and should
be scrutinized.
Both the various contested meanings of the terms “hacker” and “hacking” and the
perceptions of hackers in the general public are important elements that have to be
considered in a scientific study of hackers. The disputed nature of the two terms implies
that researchers have to be explicit when informing study subjects about what exactly
they mean by hacking. The common perception of hackers, on the other hand, has to
be considered because it likely exerts an influence on the self-perception of hackers.
More importantly, it is also the broader background against which a scientific understanding of hackers has to be established.
Both considerations about the meaning of hacking and hackers, while presenting
contextual aspects that are necessary for an accurate understanding of hackers, are
only two elements of a more complex operationalization process. Hacking as a form of
cybercriminal activity is carried out in a unique environment—cyberspace. This environment exhibits several criminogenically relevant features that render it distinctively
different from any other environment and need to be addressed in cybercrime studies.
Thus, the next section outlines the development of the societal threat posed by cybercrimes and the distinct challenges the cyberspace environment holds for cybercrime researchers.
23

CHAPTER THREE:
THE GROWING THREAT OF CYBERCRIME
Riyadh, November 28, 2005. During a conference on information security in the
banking sector, Valerie McNiven, a former e-finance and e-security specialist for the
World Bank and current U.S. Treasury advisor on cybercrime, estimated that in 2004,
financial harm from global cybercrime eclipsed the revenue generated from drug trafficking in the U.S. for the first time. According to her estimates, the revenue gained from
illegal cybercrime activities was over $105 billion.
Even though from a scientific standpoint, McNiven’s statement that global cybercrimes are becoming more profitable than drug trafficking, which is currently the highest-grossing illicit activity in the U.S., simply cannot be substantiated, it nevertheless
hints at substantial developments in the evolution of cyber-criminality. The statement
cannot be sustained because McNiven, aside from including illicit activities such as corporate espionage, child pornography, stock manipulation, phishing fraud, and copyright
offenses in her estimate of total losses, also failed to recognize cross-national differences in the legal definition of what constitutes a cybercrime. Moreover, consider the fact
that the estimates for many of the activities included in McNiven’s definition are provided by interest groups who gladly inflate statistics to the limits of credibility. For example, the recording industry, whose estimated losses comprise a large portion of McNiven’s $105 billion estimate, equates every illegally obtained copy with a direct loss of
unsold merchandise, a calculation whose inflation is inherently obvious (Leyshon,
Webb, French, Thrift, & Crewe, 2005; Marshall, 2004). Gauging the financial harm of
24

the crimes included in McNiven’s cybercrime definition is fraught with problems, and the
empirical evidence upon which McNiven’s claim is based is clearly preliminary, thus
rendering its validity questionable.
In spite of the methodological flaws of the estimate, McNiven‘s statement is nevertheless important because it points to significant developments in the area of cybercrimes. FBI statistics from recent years suggest that the financial damage caused by
cybercrimes increases between 20% to 35% each year (L. A. Gordon et al., 2005; L. A.
Gordon, Loeb, Lucyshyn, & Richardson, 2006), thereby likely surpassing the drug trade
in terms of growth rate. The prolific growth of the Internet in terms of the number of users, functionality, and efficiency of electronic business procedures has convinced many
industries to move their operations, and most importantly their money, online. Online
shopping, online banking, online stock broking, credit card gateways for wire transfers,
and e-commerce all provide opportunities to make business processes more efficient.
On the other hand, the use of the Internet for the transfer of financial resources has
drawn many criminals to this new medium (Newman & Clarke, 2003). The temptation
toward the profit that can now be gained online has completely changed the face of cybercrime (Schell et al., 2002). In order to provide a better understanding of cybercrime
phenomena as they exist today, the following section provides a brief summary of the
co-evolution of the Internet and cybercrimes.

The Co-Evolution of the Internet and Cybercrimes
Misuse of computer systems is not a new phenomenon. In fact, the abuse of
computer systems is as old as the technology itself, but cybercrime was a phenomenon
25

of no significance before computer networks were developed. The reason is a practical
one: When information is stored locally without being networked, it is much more difficult
to gain access to than it is to intercept information that is transferred across physical
distances within a network. Another advantage is that access to one element in the
network can possibly provide access to other vulnerable computers in the same network. However, even in the early days of computer networks, cybercrime was still in its
infancy and of little societal significance. In these times, computers were rare, very expensive, and required specialized knowledge to be operated successfully. All these factors inhibited widespread use and misuse.
The importance of the Internet, and with it, the significance of cybercrime began
to slowly grow as computing became easier and less expensive. In the 1970s, the first
affordable personal computers appeared and ARPAnet, the predecessor of our current
TCP/IP based Internet, was created (Hafner & Lyon, 1998, p. 40-41). Nevertheless,
even as Internet use became more widespread, connecting computers to it was quite a
difficult undertaking. Existing operating systems provided no software to connect to Internet Service Providers, which at the time were few and very expensive. Access to the
Internet was a privilege reserved for a few researchers, military personnel, government
agencies, and technology experts who used it primarily to exchange highly specialized
information (Vallee, 2003). Computer crime first drew the attention of legislators after
ISPs such as AOL, CompuServe, and Prodigy began to provide software that made it
much easier to connect to their services (Naughton, 2000). In 1986, alarmed by the rising number of cyber-trespasses, the U.S. government passed the Computer Fraud and
Abuse Act (U.S.C., 1986), the first legislative act that made breaking into computer net26

works illegal. It is worth mentioning that this first legislative effort did not apply to juveniles, a circumstance that is inconceivable today. Hackers at the time were skilled computer experts motivated primarily by curiosity to explore other systems’ weaknesses, to
reveal their configurations and operations, and to take advantage of the technical possibilities of networked computing (Graham, 2004). Many of them followed an ethical code
that disallowed them to inflict damage on the hacked system. Early hackers who followed this code distinguished themselves from so-called “crackers,” hackers motivated
by malevolent intentions aimed at exploiting the hacked system (Schell & Martin, 2004).
In the 1990s, Internet prices dropped dramatically and the Internet began to grow
exponentially. Today, computers are omnipresent in our daily lives. A modern society
without computers or the Internet has become unthinkable. The Internet has developed
into an essential, mission-critical infrastructure for government agencies and organizations, almost all private business industries, and financial institutions (Li, 2006). Among
many other aspects, manufacturing routines and processes, water supplies, electric
power grids, air traffic, and stock markets are essentially operated, managed, and controlled through computer networks (Painter, 2004). E-commerce is growing with yearly
increase rates of 20%, reaching a total volume of $116 billion today (Richtel & Tedeschi,
2007). Today, 44 million citizens in the U.S. have online bank accounts and the online
banking population in the U.S. grew by 9.5% in 2006 (Warner, 2007). Also, a third of the
U.S. workforce is online, an important factor since more than half of all e-commerce
transactions are work-related (Alaganandam, Mittal, Singh, & Fleizach, 2005). The financial value of Internet activities and the wealth stored on computer systems has
reached unprecedented dimensions and attracts criminals that are no longer motivated
27

by technical curiosity, but rather by financial motives. Today’s societal dependence on
the Internet, and particularly the use of the Internet for banking and commercial transactions, has fundamentally changed the face of cybercrime. The early hacker population
of highly skilled computer experts motivated primarily by technical curiosity has given
way to a wide array of criminals with varied skills, knowledge, resources, authority, and
motives. Profiles of today’s offenders range from teenagers to disenchanted employees,
traditional criminals, organized racketeers, or even foreign intelligence agents. The
present-day dimensions of the term ‘cybercrime’ encompasses a vast scale of multifaceted phenomena ranging from small-time pranksters to organized crime circles as
well as nationally funded spies (Bednarz, 2004).
The many dimensions and different methods of today’s cybercrimes and the multitude of offender profiles, in combination with modern society’s dependency on computer technology, necessitates an accurate assessment of the risks and threats in order
to answer the question of how to combat computer criminals effectively. Protection
against the various methods of cyber-criminals and the ability to trace, capture, and
prosecute them require research in order to provide an accurate assessment of the current landscape of cybercrimes and the threat they pose. Empirical research is vital for
the design of effective countermeasures to combat not only hackers, but also other
newly emerging types of cyber-criminals. Only then is it possible to minimize the potential for harm in the online environment. The salience of cybercrime stems from its increasing societal relevance, and mainstream criminology is slowly starting to grasp the
extent of the problem.

28

The Novelty of Cyber-Criminology
Researchers have observed the emergence of cyberspace as a new realm for
criminal activity since the 1990s (Littlewood, 2003; Thomas & Loader, 2000; Yar,
2005b), yet mainstream criminology has long been negligent towards the problem of
cybercrime and has failed to recognize its societal relevance (Jaishankar, 2007;
Jewkes, 2006; Mann & Sutton, 1998). The study of cybercrime has only recently gained
widespread attention among criminologists. Several indicators demonstrate that the
study of cybercrimes is only now catching up with the developments and methods of
cyber-criminals. The first conference devoted entirely to cybercrime, the “Cyber Criminology and Digital Forensics Initiative Conference”, took place in Spokane Valley, WA,
USA in October 2006 (Jaishankar, 2007). Probably the strongest indicator of the recent
establishment of cyber-criminology as a distinct subsection within criminology is that the
first interdisciplinary research journal dealing exclusively with cybercrime, the International Journal of Cyber Criminology (IJCC), was first published in January 2007 (Jaishankar, 2007). Another indicator that the field is just now gaining popularity is the
growing number of scholarly publications addressing the various aspects of cybercrime
(Furnell, 2002; Yar, 2006). Empirical publications require data to analyze and explain
cybercrime phenomena. The small number of publications is partly attributable to the
small number of surveys and datasets available to researchers. For example, the first
nationally representative survey, the National Computer Security Survey (NCSS), which
was sponsored by the U.S. Department of Justice (DOJ), Bureau of Justice Statistics

29

(BJS), and the U.S. Department of Homeland Security National Cyber Security Division
(NCSD), was fielded in 2006.
Taken together, these indicators suggest that criminological studies of cybercrime are in their early stages and that more attention must be devoted to this pressing
new form of crime. In other words, “A new, radical discipline named cyber- criminology
is the need of the hour to explain and analyze the crimes in the Internet” (Jaishankar,
2007, p. 1). In addition to the relative lack of publications and empirical data, this new
discipline faces a range of problems and difficulties. Some researchers even question
whether cybercrime is a qualitatively new form of crime justifying the establishment of a
distinct subsection within criminology. Thus, the headmost question is whether cybercrimes constitute a unique type of crime.

The Novelty of Cybercrime
Some researchers argue that from a legal standpoint, cybercrimes are largely the
same as ‘old-fashioned’ non-virtual crimes committed with new tools and nothing else
than the “same old wine in new bottles” (Grabosky, 2001, p. 243). However, there is
broad consensus among the majority of researchers within this newly-developing subsection of criminology that society is confronted with a new type of crime that has distinctly different qualities from ‘terrestrial’ crimes. Most cybercrime researchers, particularly criminologists with a sociological background who examine the social-structural
conditions of the environment in which the crimes are committed, insist that it presents a
radically new and distinguishable form of crime. Social interactions, in large part, are

30

determined by the possibilities imposed upon or restricted by the structures of the environment they are taking place in.
Cyberspace, a term coined by William Gibson in his famous science-fiction novel
Neuromancer (1984) and later defined as “there’s no there, there” (Gibson, 1988, p. 40),
presents an electronic universe of information resources available through computer
networks. Social interactions taking place within this world of connected computers are
markedly different than in the ‘meatspace’ (Pease, 2001, p. 23). The Internet, as the
largest of these electronic information networks, “variously ‘transcends’, ‘explodes’,
‘compresses’, or ‘collapses’ the constraints of space and time that limit interactions in
the ‘real world’” (Yar, 2006, p. 11). It represents the most important element for the
‘time-space compression’ of globalization (Harvey, 1989), in that it allows nearly instantaneous interactions between spatially distant actors.
Furthermore, the Internet offers various degrees of automatization in interaction,
thereby creating ever-new forms and possibilities of social exchange (Shields, 1996).
Among various other aspects, automatized interactions within computer-mediated
communication (CMC) networks extend both the scope and scale of offending, as the
aforementioned examples of the Turkish hacker and the attacks on Estonia’s digital infrastructure demonstrate. Many-to-many communications inexorably alter the relationships between offenders and victims, as well as the possibilities of criminal justice systems to counteract or resolve such crimes (Capeller, 2001). Internet users are instantaneously targetable by the largest known pool of potential offenders from all over the
world. Hackers are not bound by the limits of physical proximity and can strike by
launching highly automated attack routines. A wide variety of specialized software ap31

plications such as network crawlers, sniffers, and scanners exist that can be used to
execute such attack routines. These programs perform automatic scans of wide ranges
of IP addresses for vulnerabilities. CMC allows individual users to reach, interact, and
affect thousands of other users simultaneously. Thus, the possibility for potential offenders to target large quantities of other users and their property is multiplied to an unprecedented degree. Electronic communication networks are unique ‘force multipliers’
that can be abused to create enormous negative effects with only very limited resources
and efforts. One example of the misuse of the Internet as a force multiplier is the distribution of viruses, worms, trojan horses, root kits, and other malicious code; another is
the daily flood of mass-distributed ‘spam’ emails. The research institute IDC estimates
that in 2007, 40 billion of the nearly 97 billion emails sent daily will be spam messages
(Levitt, 2007). According to their report, the worldwide volume of spam email sent is expected to exceed the volume of person-to-person emails sent for the first time in 2007.
Another criminogenically relevant feature of CMC networks is that they allow for
easy creation, alteration, and reinvention of the social identity. Internet users can create
arbitrary virtual avatars, electronic representations and identities that are not required to
be analogous to their ‘real world’ identities (Turkle, 1995). The ability to disguise social
identity is of high criminological relevance because it allows potential offenders to remain anonymous (Snyder, 2001). The increased anonymity in electronic communication
networks lowers the offender’s perception of the risks involved in the commission of cybercrimes. As with terrestrial crimes, anonymity in cyberspace ultimately raises the likelihood that a cybercrime is committed (Joseph, 2003).

32

From the above, it can be concluded that three features of these electronic networks create new forms, patterns, and structures of illegal behavior: the irrelevance of
tempo-spatial distances, the unlimited potentiality of many-to-many connectivity, and the
possibility to disguise oneself behind arbitrary, anonymity-granting identities. These features of the cyberspace environment render all interaction taking place within it, including cybercrimes, original and distinctly different from interactions in any other social environment known before.
The above characteristics of digital crimes justify the establishment of cybercriminology as a distinct subsection within criminology. They also pose new challenges
for both researchers and law enforcement alike which require innovative responses.
The challenges for the criminal justice system arise in particular from the global dimensions and the borderless nature of the environment. Disparate criminal laws and geographical boundaries of jurisdictions pose important limitations for law enforcement efforts because they have no direct equivalent in cyberspace (Koops & Brenner, 2006).
Many hackers take advantage of this fact by deliberately exploiting sovereignty limits
and cross-jurisdictional differences. Though internationally cooperative efforts to harmonize Internet laws, such as the groundbreaking Council of Europe’s Convention on
Cyber Crime, are being undertaken, large gaps remain in the realm of international cybercrime legislation.
In addition to the legal complexities resulting from the global nature of the problem, there are challenging technical intricacies. These complications include the volatility and intangibility of evidence owing to the lack of traditional forensic artifacts such as
fingerprints, DNA, or eye-witness accounts associated with these types of offenses
33

(Shinder & Tittel, 2002). Hackers enjoy a high level of anonymity stemming from anonymizing services, encryption algorithms, ‘spoofing’ tools, and the increasingly widespread availability of Internet cafés and public wireless networks (Computer Crime and
Intellectual Property Section, 2002). To make matters worse, legislative efforts to force
ISPs to keep log files documenting user data long enough to conduct criminal investigations have not yet been successful. Each of these technical obstacles and time constraints raises the cost of cybercrime investigations far above the cost of traditional
crime investigations. Presently, law enforcement does not have the required resources
and experts with specialized technical knowledge necessary to track and apprehend a
large portion of online offenders. One indication for this lack of resources is the estimate
that the chance for being prosecuted for hacking in the US is one in 10,000 (Bequai,
1999). Hence, even in jurisdictions with appropriate laws to counteract digital crimes,
the implementation of sanctions remains restricted by logistical and financial limitations.
The preceding discussion delineated the problems that the new features of cybercrimes pose for law enforcement. Central to this study, however, are the unique
challenges that arise in cybercrime research. The qualitatively new features complicate
the research endeavor in several ways. These complications are addressed in the following chapter.

34

CHAPTER FOUR:
THE CHALLENGES IN RESEARCHING CYBERCRIME
The valuable information gatherable only through criminological research on cybercrimes is required for addressing and counteracting this growing societal threat, but
the developing field of cyber-criminology has yet to overcome a series of distinct methodological and theoretical problems other research areas have already mastered. The
foremost methodological problem is that, currently, no universally agreed on criminological definition of the term cybercrime exists. This problem is fundamental because of its
implications for the operationalization of the concept and is, therefore, discussed at the
beginning of the current chapter.

Methodological Problems

The Problem of Defining and Classifying Cybercrimes
The investigation of cybercrimes is a relatively new field in criminology and, as
such, it is confronted with several methodological problems that other areas have already overcome. One main problem in the operationalization of the concept is the absence of an widely agreed on criminological definition of what constitutes a cybercrime.
The problem is further exacerbated by the fact that the complex concept cybercrime encompasses multiple dimensions and various manifestations. The term cybercrime denotes a wide range of phenomena ranging from an attack against essential national infrastructure systems, such as the National Grid, to fraudulent manipulations of penny
35

stocks in chat rooms and bulletin boards. This vast variety of phenomena subsumed
under the term makes it difficult to provide a precise definition in terms of aggregate incidence and impact. Wall also reminds us that even though the term cybercrime has
considerable currency in discourses within the media, politics, academics, and the public, it still has “no specific referent in law” (Wall, 2001, p. 2). The term is often understood as signifying an array of illegal acts whose commonality is that they are conducted
through electronic information and communication networks. An example of such a
general definition is the one proposed by Thomas and Loader, who summarize all
“computer-mediated activities which are either illegal or considered illicit by certain parties and which can be conducted through global electronic networks” (Thomas & Loader, 2000, p. 3) under the term cybercrime. Under this definition, the novelty of cybercrime is that it is conducted in the ‘virtual space’ of worldwide networks of information
exchange through interconnected computers, particularly the Internet (Castells, 2002).
The term ‘conducted’, in this context, can refer to computers and networks functioning
as the agent, facilitator, or the target of criminal activity.
Within this broad conception of cybercrime, certain types of activities included in the term are commonly distinguished. Typically, types of cybercrimes are categorized by the function technology has in their commission. Furnell differentiates between ‘computer-assisted crimes’, such as fraud, theft, sexual harassment or pornography, in which “the computer is used in a supporting capacity, but the underlying crime or
offense either predates the emergence of computers or could be committed without
them,” and ‘computer-focused crimes’, such as hacking, website defacing, spamming,
or launching of virus/worm attacks, “cases in which the category of crime has emerged
36

as a direct result of computer technology and there is no direct parallel in other sectors”
(Furnell, 2002, p. 22). A similar distinction is proposed by S. Gordon and Ford (2006),
who isolate three main dimensions of the term cybercrime: (1) The computer and hardware device dimension, i.e. the role technology plays in the conduct; (2) the technology
– human element continuum, a dimension that ranges from cybercrimes that are primarily technological in nature to those that have pronounced human elements, like for example cyber-stalking; and (3) the ‘crimeware’ focus, a term S. Gordon and Ford use to
denote software whose primary application is the commission of cybercrimes. Unfortunately, Furnell’s as well as S. Gordon and Ford’s classifications, while being useful in
analyzing interdependency processes in social and technological developments, are of
limited criminological value.
One alternative is suggested by Wall (2001, pp. 3-7), who translates four existing
criminal law classifications into their cybercrime equivalents. He distinguishes between:
(1) ‘(cyber)-trespass’, the unauthorized crossing of the boundaries of computer systems
and/or the causation of damage to those systems or their possessors; (2) ‘(cyber)deceptions/theft’, stealing of money or property, e.g. credit card fraud, identity theft, or
violations of intellectual property a.k.a. ‘piracy’; (3) ‘(cyber)-pornography/obscenity’, all
pornographic, violent, hate-filled, racist, or generally offensive activities breaking obscenity and decency laws; and (4) ‘(cyber)-violence’, the infliction of psychological harm,
or the incitement of physical harm, e.g. hate speech, ‘flaming’, harassment, stalking, or
the distribution of information assisting in dangerous activities such as bomb-building.
By emphasizing the origins of cybercrimes in traditional crime categories, Wall pursues
a conservative approach. Each of his four categories subsumes rather broad and di37

verse ranges of phenomena, a circumstance that calls the validity of his typology into
question.
A practical solution to the challenge of defining cybercrimes is offered by Ralph
D. Clifford (2006). Instead of continuing the debate with just another proposed definition,
Clifford and the team of legal experts he assembled question the practicality of attempting to subsume the vast scale of multifaceted phenomena that is encompassed by the
term cybercrime under one legal definition. In this regard, Clifford’s approach presents
an important conceptual break that calls into question the appropriateness of the very
debate surrounding the legal definition of the term cybercrime. Instead of continuing the
academic search for one overarching legal definition, the approach taken in this second
edition is to examine the practical implications for investigators, prosecutors, and defense attorneys of the various cybercrime-related statutes (Bachmann, 2008). Despite
Clifford’s suggestion to end the debate about an overarching definition this debate will
likely continue, and eventually, a sound criminological definition will be proposed. Until
then, Clifford’s approach seems to be the most practical solution because it is best fitted
to address the vast variety of criminal activities subsumed under the term cybercrime.

The Problem of Measuring the Scale of Cybercrimes
Obtaining an accurate assessment of the scope and severity of cybercrimes is a
difficult undertaking. Unfortunately, this is true for many of the types of offenses studied
by criminologists. Official crime statistics, for instance, do not objectively measure incident trends and distributions, but are socially constructed. Official datasets include only
crimes that have been reported, and there are several reasons why crime victims can
38

be reluctant to report offenses. These reasons range from perceptions of the offense as
a private matter or as trivial to fears of retaliation, unawareness of the victimization, embarrassment due to the victimization, or lack of faith in an effective response. Especially
corporate actors oftentimes also fear the potential damage the reporting of their victimization can have for their public reputation. Another major problem is that even when offenses are reported, they may be excluded from the Uniform Crime Report (UCR) because only the primary charges in each incident are recorded. Fortunately, this problem
has been resolved in the National Incident Based Reporting System (NIBRS), which includes all charges in the incident report as well as several situational variables. A related under-reporting problem that remains unresolved in both national datasets arises
from the circumstance that they rely on police reports which are submitted on a voluntary basis. In addition to the many obstacles challenging cross-sectional measurements,
longitudinal datasets oftentimes also have to resolve issues resulting from changing legal classifications.
While the aforementioned problems are faced by all quantitative criminological
studies, they are magnified with respect to hacking attacks and, to varying degrees, to
all other forms of cybercrimes. As was stated previously, the intangibility of evidence
and the lack of traditional forensic artifacts make online offenses more difficult to detect
than terrestrial crimes. Complicating matters further is the remaining lack of knowledge
of many as to what exactly constitutes a cybercrime, and consequently, whether reporting of a particular incident is appropriate (Howell, 2007). Moreover, the global nature of
cybercrimes and the high level of offender anonymity in the online environment are two
additional aspects that discourage both victims and law enforcement from reporting
39

such crimes because they decrease the perceived chance of apprehending the offender. As a result, many police stations prioritize reporting of local problems (Wall, 2001)
occurring on their “patch” (Lenk, 1997, p. 129). Taken together, the above factors justify
the conclusion that cybercrimes are greatly under-reported in official statistics. In 2000,
Detective Sergeant Clive Blake from the Metropolitan Police Fraud Squad estimated
that as few as 5% of computer crimes are actually reported (Blake, 2000).
Crime and victimization surveys offer an alternate assessment of crime levels.
Victimization surveys are often used by criminologists because of their ability to encompass offenses that are typically under-reported in official statistics. Despite their advantages, crime and victimization surveys cannot completely eliminate all of the difficulties
faced by official measurements. To begin with, it is self-evident that an undetected
crime cannot be reported. Of higher importance for the quality of survey data, however,
are systematic errors and the bias they introduce. Systematic errors can result from
many different sources, such as incongruities in the definition of what constitutes a
crime between interviewer and interviewee, various other interviewer effects, the presence of third persons, sponsorship-biases, or the so-called response set of the participant to name but a few. Survey researchers have long recognized that even the highest
possible optimization of survey instruments will never completely eliminate survey errors
(cf. Groves et al., 2004). Despite their shortcomings, victimization surveys are especially
relevant for cybercrime studies because official data on computer offenses remains
scarce. Unfortunately, survey-related problems are again exacerbated when measuring
cybercrimes. Cybercrime victimization surveys typically have decidedly selective populations and study samples. The majority of surveys, including the annual CSI/FBI Com40

puter Crime and Security Survey, measure only corporate or organizational victimization
and exclude private computer users. Wall (2001) also hints at the various existing intersurvey inconsistencies with regard to methodologies and classifications. These inconsistencies complicate meaningful comparative analyses and aggregations. Yar lists
three more reasons for under-reporting by corporations and organizations that are subject to public scrutiny. He states that public actors “may prefer not to acknowledge victimization because of (1) fear of embarrassment: (2) loss of public or customer confidence (as in the case of breaches relating to supposedly secure e-shopping and ebanking facilities): and (3) because of potential legal liabilities” relating to their violation
of data protection responsibilities (Yar, 2006, p. 14).
All the above mentioned difficulties suggest that more studies and improved
measurement techniques are needed, and they should lead cybercrime researchers to
be cautious about the validity of their data. However, they do not imply that researchers
should refrain from using all available data. To the contrary, more current data are
needed for a greater understanding of the limitations and for the refining of methodological techniques to better address them.
Eventually, meta-studies of official data and victimization surveys will be able to
provide a reasonably adequate picture of the threats Internet connected computer users
are facing. When pursuing this approach, however, one has to be cautious in drawing
conclusions about the offenders because the high degree of anonymity and inaccessibility granted by the Internet environment conceals many relevant offender characteristics to the victims, and the low apprehension rate prevents accurate estimates of systematic differences between offenders who get caught and those who do not. In addition
41

to these measurement issues, cyber-criminologists face certain theoretical problems
that are addressed in the following section.

Theoretical Problems
Attempts of criminology to uncover the underlying causes of criminal behavior in
cyberspace are inevitably based on data and theories relating to terrestrial crimes. Differences in the cyberspace environment with regard to structural and social features call
into question the transferability of traditional criminological theories to cybercrimes. Presently, the applicability of mainstream criminological theories is yet to be determined.
Two particular problems arise from questions relating to who commits cybercrimes and
where are they committed.

The Problem of Isolating Structural Correlates
The first problem arises from the circumstance that many criminological theories,
in varying degrees, rely on ecological and environmental propositions. These theories
explain crimes as phenomena that occur within particular settings exhibiting specific social, cultural, and economic characteristics. For example, routine activities theories focus on the time-space convergence of motivated offenders with suitable targets and the
absence of capable guardians, thereby implying certain environmental settings in which
offenses can occur (Cohen & Felson, 1979; Felson, 2000).
Disorganization and other theories that focus on ecological factors have initiated
crime mapping projects as well as several measurement and prevention programs
aimed at the removal of criminogenic factors from the environment (Akers & Sellers,
42

2004). Nonetheless, such approaches are problematic when transposed to the context
of cybercrimes. The virtual cyberspace environment within which cybercrimes are committed has no equivalent to easily distinguishable terrestrial locations. Rather, cyberspace has to be understood as “fundamentally and profoundly anti-spatial” (Mitchell,
1995, p. 8), an environment in which there is no physical distance between locations.
The irrelevance of spatial distances within cyberspace renders all criminological theories relying on assumptions specific to terrestrial environments handicapped for the explanation of cybercrimes.

The Problem of Isolating Offender Characteristics
A different group of criminological theories examines the reasons why some individuals repeatedly involve themselves in criminal activity while others abstain. These
theories correlate specific categories of crimes with certain social characteristics of the
offender. An example of these corresponding characteristics can be illustrated using
property crimes and violent crimes. Such offenses are primarily committed by persons
at the lower end of the income distribution. Other social characteristics predisposing individuals to commit property or violent crimes might include lack of educational attainment or involvement in deviant subcultures. There is a significant amount of empirical
evidence within the criminological literature substantiating this correlation between economic disadvantage and engagement in these types of crimes. The same correlation,
however, cannot be established in the case of cybercrimes.
To the contrary, the aforementioned correlation is reversed with regard to computing skills and Internet access, in that the more disadvantaged segments of the econom43

ic and educational distribution possess them the least (Yar, 2006). At the same time,
individuals with the economic and intellectual resources necessary for the commission
of cybercrimes usually come from more socially advantaged backgrounds. Internet offenders typically exhibit characteristics that are substantially different from the attributes
of the majority of other criminals (Wall, 2001).
The two theoretical problems associated with the identification of ecological and
individual characteristics of cyber-offenders render many traditional criminological theories that focus on environmental and individual factors unable to explain of Internet
crimes. This circumstance has led some cyber-criminologists to suggest that the appearance of cybercrimes and the inability to simply transpose traditional empirical assumptions and explanatory concepts to them may require considerable theoretical innovation. Criminology may need to develop new analytical tools to adequately assess
these qualitatively new crimes (Yar, 2006).
Recognizing these new conceptual challenges, Ronald Clarke suggested a rather radical solution. His memorable and highly controversial suggestion is to completely suspend criminological inquiries into the motivations of cyber-offenders, and instead
focus solely on the crime reduction strategies provided by the ‘crime science’ approach
(Clarke, 2004). Diverting from the attempts of criminology to unveil causational crime
factors, crime science merely attempts to reduce it through suitable prevention and detection solutions. Crime science utilizes computer science, engineering, and the physical sciences as much as the social sciences, and resorts to the fields of statistics, environmental design, psychology, forensics, policing, economics and geography to exclusively study the characteristics of crime incidents, while ignoring the traits and motiva44

tions of criminal actors. Crime scientists substitute the analytical tools of traditional criminology such as cohort studies, criminal career studies, and regression analyses with
methods that include crime patterning, hot spot analyses, and crime mapping studies.
Clarke demands that criminologists focus on concrete and manipulable factors of crime
events and accept the remainder as immutable facts. This suggestion is in line with his
prior work in the area of situational crime prevention and rational choice approaches
(Clarke, 1992; Clarke & Cornish, 1985, 2001; Clarke & Felson, 1993; Cornish, 1994;
Cornish & Clarke, 1986), because these theories consider criminal behavior to be indifferent from any other behavior place, and their central focus is on the reduction of opportunities inducing criminal activities.

The Proposed Theoretical Approach
The approach taken in this research project is based on the argument that first,
crime science and the attempts of traditional criminological theories to explain crime
causation are not mutually exclusive. To the contrary, they can be reciprocally beneficial. Crime science models striving to reduce crime are necessarily based on at least
implicit assumptions of causational factors and characteristics of criminals. Thus, the
knowledge and insights gathered through traditional criminological theories can increase the validity of crime science models and make them more accurate. Crime
science models, on the other hand, have the potential to reveal insights that allow inferences and conclusions about crime causation factors, and can thereby contribute to
advancing criminological theories.
45

Second, it is argued here that, given the current absence of a criminological
theory specifically designed to meet the characteristics of the online environment and
the challenges arising from them, modeling decision-making processes which lead to
the commission of criminal acts seems to be a promising approach. Such models are
developed mainly in the context of rational choice approaches which attempt to reconstruct decision making processes of criminals in order to isolate manipulable factors
with the potential to reduce crime. A few examples already exist in the criminological
literature that demonstrate how rational choice and situational crime prevention models
can be successfully applied to the area of cybercrimes (D'Arcy, 2007; Newman &
Clarke, 2003).
For a number of reasons, rational choice and situational crime prevention models
appear to be well suited to the application of cybercrimes. Unlike many other criminological theories, they do not rely on the particular assumptions of disadvantaged social environments or individuals, and thus are not affected by the associated problems mentioned earlier. The only assumption they make about the offending persons is that they
deliberately choose to commit the criminal act because they consider it beneficial. The
sole proposition of rational choice models with regard to human actors is that they use
their cognitive resources to weigh the potential benefits of an action against its potential
costs, and that this calculation can be manipulated through changes in the parameters
of estimated risks and costs.
In the case of cyber-criminals, it is reasonable to assume that offenders premeditate their actions before they decide to commit a criminal act. All different types of cybercrimes require detailed planning in different stages of their commission and have to
46

be committed in a decidedly rational and deliberate manner in order to be successful.
Carrying out cyber-trespasses, cyber-thefts, deceptions, or frauds, requires the offender
to carefully consider and weigh alternative routes of action, select the techniques and
tools appropriate for the attack, and identify exploitable weaknesses, to list but a few
decisions. As was delineated in the first chapter, cyberspace exhibits several features
that can lead a rationally-acting person to the conclusion that committing crimes in this
environment is a viable option. Among them are the reduced risk of being apprehended,
the multitude and instant accessibility of suitable, unprotected targets, and the high degree of anonymity. Clearly, crimes in the digital worlds of information exchange are the
result of reasoning processes of rational calculation, purposive self-interest, and lowcost opportunities. All these aspects suggest that the engagement in cybercrime activities can appropriately be modeled as a series of decision-making processes, and that
rational choice and situational crime prevention models are particularly well suited for
application to cybercrimes.

47

CHAPTER FIVE:
THE RATIONAL CHOICE PERSPECTIVE
As a theoretical concept, rational choice revolves around the notion that crime is
chosen because it appears as a promising and rewarding alternative to the offender.
Although contemporary versions of this theory have appeared only relatively recently
when compared to other mainstream criminological theories, the depiction of choices in
terms of benefits and costs has a long tradition in economic as well as sociological and
criminological thought. Within criminology, the description of offenders as rational decision makers weighing their alternatives for their pursuit of self-interest dates back to
classical writers of the 18th century, such as Beccaria (1764) and Bentham (1789). Today, several different versions of rational choice theory exist, and rational choice theory
has become one of the main theoretical concepts that is addressed in nearly every
criminological textbook (Hechter & Kanazawa, 1997).
The criminological rational choice model that has attracted by far the most scientific attention, the so-called rational choice perspective, was developed by Cornish and
Clarke in the 1980s (Clarke & Cornish, 1985). Since the theoretical underpinnings of
this dissertation follow the rational choice perspective as outlined by Cornish and
Clarke, a detailed description of this perspective is provided in the following section.

48

The Six Basic Propositions of the Rational Choice Perspective
Essentially, Cornish and Clarke’s rational choice perspective model rests on six
basic propositions, which are stated as below (Clarke & Cornish, 2001, p. 24):
1. Crimes are purposive and deliberate acts, committed with the intention of
benefitting the offender.
2. In seeking to benefit themselves, offenders do not always succeed in making
the best decisions because of the risks and uncertainty involved.
3. Offender decision making varies considerably with the nature of the crime.
4. Decisions about becoming involved in particular kinds of crime (involvement
decisions) are quite different from those relating to the commission of a specific criminal act (event decisions).
5. Involvement decisions can be divided into three stages – becoming involved
for the first time (initiation), continued involvement (habituation), and ceasing
to offend (desistance) – that must be separately studied because they are influenced by quite different sets of variables.
6. Event decisions include a sequence of choices made at each stage of the
criminal act (e.g., preparation, target selection, commission of the act, escape, and aftermath).
These six elements comprise the core of the rational choice perspective. Cornish
and Clarke claim that the combination of these propositions provides a framework applicable to all different types of crimes and therefore of the largest possible criminological
scope (Clarke & Cornish, 2001). The wide scope of applicability is due in part to the
49

concise, yet abstract and all encompassing nature of the statements, which also renders the theory decidedly parsimonious and internally logically consistent. While all six
propositions are essentially hypotheses themselves, only the second one lends itself to
more explicit and better testable hypotheses when applied to the context of hackers.
Thus, the hypotheses of this study are derived from the second proposition. The others,
on the other hand, are statements of such general and abstract nature that they cannot
be operationalized in the same way. Nevertheless, propositions four to six are important
guidelines for the theoretical framework structuring the assessment of hackers, their
hacking careers, motivations, and various decision-making processes in the present
study. The six propositions, as well as the hypotheses and the models derived from
them, are addressed in greater detail in the following section.

The Rationality of Criminal Acts
The starting point of the theory essentially states that crimes are never senseless
or random acts. Rather, they are seen as purposive acts intended to benefit the offender. Beneficial outcomes are most obvious in the case of money and material goods,
but benefits can also include excitement, fun, prestige, sexual gratification, defiance or
domination (Clarke & Cornish, 2001). This first assumption has sparked considerable
criticism from numerous theoretical camps, who concede that certain crimes, particularly pathological crimes or crimes committed by mentally ill persons cannot be adequately conceptualized as rational acts. The dispute over whether all crimes are rational
or not and whether the scientific reconstruction of all crimes as rational acts renders the
concept of rationality universal, tautological, and thus deprived of its explanatory capa50

bilities shall only be mentioned here (Akers, 1990; De Hann & Vos, 2003). However, it is
not addressed in greater detail because, for the reasons listed earlier, it is reasonable to
assume that hacking attacks (as well as probably all other forms of cybercrimes) are deliberately premeditated acts that involve substantial amounts of rational considerations.
Consequently, these types of criminal acts fall well within the scope of the rational
choice perspective and should be at least potentially explainable with this theoretical
concept.

The Limitations of Rationality in Decision Making Processes
While hacking attacks are criminal acts that do require a considerable degree of
premeditation, this circumstance does not automatically imply that all hackers always
make the best decisions. This second proposition is one that separates the different
models and versions of rational choice theory. All rational choice models have in common that they predicate crime as resulting from choices, but they vary considerably in
what they define as a rational choice. Proposing a typology of rational choice models,
Opp (1997) suggests distinguishing rational choice models along a continuum ranging
from wide to narrow concepts of rationality. Wide concepts of rationality, such as the
one employed by the rational choice perspective, conceptualize individual decision making processes as being inevitably bounded or restricted. These models take into consideration that decision-making processes are always limited by situational and cognitive constraints. They are consistent with psychological theories depicting humans as
‘cognitive misers’ who strive to conserve energy and reduce cognitive load (Fiske &
Taylor, 1984). They also take into account that decision-making processes include un51

certainties because they are hardly ever based on knowledge of all relevant facts and
risks. Likewise, it is often impossible to gauge both the exact utility of action alternatives
and the probability that this utility will indeed be the outcome of that action alternative.
Such constraints of cognitive information-processing abilities lead human actors
to employ heuristics and biases in the decision-making process. For example, instead of
considering every single relevant detail, offenders often rely on general tactics and
methods that they have had positive experiences with in the past. Once they decide to
commit the crime, they tend to reinforce their decision by focusing on the potential rewards rather than the risks. This focus on rewards rather than risks prevents offenders
from accumulating the cognitive strain that would result from focusing on the dissonant
cognitions of engaging in criminal activity while potentially having to suffer the consequences of punishment (Festinger, 1957). In cases where offenders take risk considerations into account, they most often focus on the short term costs of getting caught as
opposed to long term consequences, such as getting punished for their actions.
‘Wide’ rational choice conceptualizations further acknowledge time constraints
that are necessarily involved in decision-making situations. Time itself is a valuable resource, and a disproportionate time investment into considerations over low stakes decisions would not be purposively rational. Since wide rational choice models recognize
cognitive and situational imperfections, they can be considered more psychologically
and sociologically plausible. They assume that offenders, like the rest of us, use “satisficing rather than optimizing” (Clarke & Cornish, 2001: 25) decision-making efforts and
strategies. Sometimes limitations of time, cognitive resources, and available information
can lead to decisions that appear irrational.
52

Moreover, as Cacioppo and his colleagues have shown, individuals differ in their
personal need for cognition (Cacioppo & Petty, 1982). Whereas some persons prefer
analytic and rational approaches when processing information and making decisions,
others tend to rely more on their faith in intuition (Cacioppo, Petty, Kao, & Rodriguez,
1986). Individuals who rely on their faith in intuition are generally greater cognitive misers than persons with a strong need for cognition. They oftentimes use heuristics, base
their decisions on similar prior experiences, tend to abbreviate decision-making processes, and terminate them earlier. Rational choice theorists, however, contend that, regardless of the actual degree of rationality involved, to the offenders themselves the selected action alternative appears as the most promising one at the moment it is selected. Thus, while limited, the decision should nevertheless be considered rational.
Several testable hypotheses can be derived from this second proposition for this
study. Hacking is a complex and difficult undertaking, and successful hacking requires
not only extensive knowledge and expertise, but also considerable cognitive efforts.
Thus, compared to hackers with a preference for more heuristic, intuition-based thinking
styles, hackers who prefer to think more analytically and rationally are expected to be
more successful in their hacks because they weigh more alternatives, consider more
factors, and operate more cautiously. It has to be noted that, in the context of hacking,
success refers to the accomplishment of gaining access to the target, and the rating of a
hack as successful is ultimately a subjective assessment by the hacker. The more deliberate preparation of hackers with a preference for rational thinking styles should further result in longer preparation times for hacking attacks, and in a generally higher risk
estimate, in a more risk-oriented focus. The same patterns can be expected for hackers
53

with a lower risk propensity because of their more careful consideration of the involved
risks. Higher risk propensity, on the other hand, should lead hackers to engage in more
hacking attacks because they are less concerned about the involved risks. Precisely
stated, the following testable hypotheses can be formulated for the current study:
H1: Hackers with a higher preference for analytic-rational information processing
are more successful in their hacking activities.
H2: Hackers with a lower risk propensity are more successful in their hacking activities.
H3: Hackers with a higher preference for analytic-rational information processing
estimate the risks involved in hacking to be higher.
H4: Hackers with a lower risk propensity estimate the risks involved in hacking to
be higher.
H5: Hackers with a higher preference for analytic-rational information processing
take longer times to prepare for their hacking attacks.
H6: Hackers with a lower risk propensity take longer times to prepare for their
hacking attacks.
H7: Hackers with a lower risk propensity engage in fewer hacks.
All seven hypotheses are examined in the present study. The preference for rational thinking and the risk propensity of hackers are assessed and their influences on
the overall hacking success, estimated risks, average preparation times, and the total
amount of hacks are measured.

54

The Importance of Situational Adjustments
Different offenses provide different benefits to offenders and are committed with
specific motivations. Some hackers engage in hacking merely to satisfy their curiosity
about how different computer systems and setups work and to find out about their vulnerabilities, while others hack for financial gain, peer recognition, or personal revenge.
Still others deface websites to voice their political opinions or to express their general
antipathy with the host or the content of the site. The variables weighed by offenders
and the factors considered in their decision making vary accordingly. Decision-making
processes are fundamentally influenced by varying situational variables and the situational factors involved in different crimes vary greatly. For example, the situational factors involved in the intrusion into a computer system are substantially different from
those in a bank robbery or the snatching of a purse (De Hann & Vos, 2003).
The above mentioned circumstance demands that situational choice decision
models and flowcharts are not studied in abstract, general terms, but rather are developed specifically for the respective crimes that they are supposed to explain. This notion
disqualifies wide legal categories such as cyber-trespass or cyber-theft as an appropriate basis for event models because they encompass too much variation with regard to
types of offenses, offenders, methods, and goals. Consequently, such categories have
to be broken down into more specific categories that permit models with greater precision and validity.
The increased validity of more specific event models, however, inescapably confines the generalizability of such models. Creating separate models for hacks that are
55

committed with different network mapping tools, for example, is probably not a fruitful
approach, unless the researcher wants to study exactly those differences. Narrow models based on situational assumptions so specific that their applicability becomes restricted to only a very few situations are of limited scientific value. The challenge in the
creation of event models then becomes finding the right balance in the pay-off between
validity and generalizability. Clarke and Cornish suggest guiding the specificity of models along practical considerations, namely whether drawing finer distinctions results in
large enough collections of offenses to justify separate interventions. If the answer to
this question is yes, the distinction should be made. If it is no, they recommend finding
solutions for addressing the variety of crimes subsumed under the broader category
(Clarke & Cornish, 2001, p. 26).
While this proposition does translate directly into testable hypotheses, it presents
an important aspect to consider in the creation of theoretical models and flowcharts. In
order to be able to better understand and predict behavior, researchers have to gain an
understanding of the underlying decision-making processes, the motivations driving
these processes, and the influential factors considered in them. They have to identify
relevant motivations, desired outcomes, and situational variables. Once the influential
variables are identified, researchers can then begin to examine their relative importance
and their specific effects on the outcome of that decision-making process. Eventually,
this line of research will produce path-dependency models that will help to explain and
predict offender behavior and, in some cases, may even be able to prevent criminal behavior by proposing changes in the “choice-structuring properties” (Cornish & Clarke,

56

1987, 1989) of hackers that increase risks and efforts, reduce rewards, and remove excuses (Clarke, 1997; Piquero & Tibbets, 2002).

The Differences between Involvement and Event Decisions
The decisions to generally become involved in criminal activity and to commit a
particular crime can and must be distinguished from each other. The decisions involved
in the commission of a particular crime, referred to as event decisions, include considerations such as the selection of specific targets and of methods to reduce the risk of
apprehension. In order to avoid detection, hackers might, for example, launch their attack from public Internet cafes, or through encrypted networks such as the I2P or the
TOR networks.
Involvement decisions, on the other hand, refer to the general decision of becoming involved in criminal activity. The highly complex nature of such decision-making
processes requires researchers to further subdivide the broad concept of involvement.
Typically, rational choice theorists distinguish three different stages of involvement: (1)
initiation, i.e. the onset of a criminal career when offenders begin to view criminal activity as a viable option to achieve their goals; (2) habituation, i.e. the procedural continuation of a criminal career and; (3) desistance, i.e. the point when offenders decide to discontinue their criminal career and to desist from future offending (Clarke & Cornish,
2001).
While the requirement to specify event decisions models for different types of
crime seems to be obvious, the creation of separate involvement decision models for
different stages and different types of crime is just as important. Involvement decisions
57

at different stages and for different crimes include varying social background factors,
current situational circumstances, available (il-)legitimate opportunities, and differing estimates of costs and risks. Not only are the skills and contacts required for certain types
of crimes different, but it is also reasonable to assume that offenders distinguish the severity and immorality of different crimes. The fact that a person attempts to gain unauthorized access to a computer system, for example, does not imply that this person has
no moral objections against stealing or destroying information that is stored on it.
Clearly, one model size does not fit them all. Instead, models need to be tailored to the
type of crime and the stages of involvement in order to provide a more precise picture of
relevant factors.

The Three Different Stages of Involvement
As was mentioned before, the broad concept of involvement requires further distinction between different phases because the stages of initiation, habituation, and desistance are based on different variables and factors. According to the rational choice
perspective, these variables and factors belong to three different groups: (1) background factors, e.g. bio-psychological factors, socialization, and learned experiences;
(2) current life circumstances, e.g. routines and lifestyles; and (3) situational variables,
e.g. momentary needs and motives and (il-)legitimate opportunities.
The rational choice perspective hypothesizes these three groups of factors to be
disparately influential for the three phases of involvement. Background factors are
thought to exert the greatest influence in the initiation phase because they predetermine
the person’s skills, experiences, morality, and friendships, as well as their socioeco58

nomic statuses. During the habituation phase, the criminal activity and the gains and
risks associated with it gain gradually higher relevance for the current life circumstances, and the immediate life circumstances, in turn, become the most influential factor for decision-making processes. As involvement in criminal activity continues and becomes more and more habitual, mounting negative experiences and increasing realization of the potential costs involved, in combination with certain changes in the offender’s
life circumstances, could eventually lead them to desist from offending. Regardless of
the stage of involvement, however, the rational choice perspective puts great emphasis
on the importance of contextual variables that are present in a specific decision situation. Situational variables such as opportunities and inducements or needs and motives
ultimately tip the scales toward the commission of or abstention from a particular crime.
In proposing different models of the factors likely to be involved in hackingrelated decision-making processes, Figures 1 – 3 show the three different involvement
stages of a hacking career, while Figure 4 presents a model of the concrete event decision to launch a particular hacking attack. All four models are close modifications of the
models proposed by Cornish and Clarke. They adjust Cornish and Clarke’s models by
translating them into the context of hackers and by introducing some factors specific to
hacking, but they assume the same general developments within and between the different phases as the original rational choice perspective models. In addition to criminogenically relevant factors, the models also list computer proficiency related factors because, unlike other types of cybercriminal activity, successful hacking requires the offender to have extensive computer knowledge and skills.

59

Figure 1 displays the variables influencing the initial decision to become involved
in this particular activity. The starting points of the model are the various background
factors commonly identified by criminological theories as being influential to the decision
to engage in criminal behavior. These factors shape the values, attitudes, and behaviors
which predispose people to consider committing crimes. It is important to bear in mind,
however, that some of these background variables have to be treated with caution because hackers (and cyber-criminals in general) exhibit some characteristics that are
substantially different from the attributes of the majority of other criminals, as was stated
in the previous chapter (Wall, 2001). While demanding caution, this problem is at least
partially alleviated by the circumstance that, within the rational choice perspective, these
factors are considered to be only indirectly influential. They exert their influence solely
through their influences on judgments and perceptions in the decision-making process.
The second box includes the various learning experiences the potential offender
has had in his or her life with regard to the decision at hand. Accumulated prior learning
experience is very important because the action alternatives evaluated in the decisionmaking process are interpreted against the background of these experiences, and some
alternatives are enabled only through specific prior learning experiences. Successful
computer hacking, for example, requires substantial knowledge of computers, networking structures, methods of attacks and attack tools, potential exploits, and programming
skills. Without these prerequisites, hacking is probably not a fruitful course of action and
will result in the mere execution of basic prewritten scripts, if it is considered as an action alternative at all.

60

1: BACKGROUND FACTORS
Biopsychological: intelligence; temperament; thrill seeking
Socialization: education; developed level of self-control
Social: social exclusion; computer proficiency of peers

2: EXPERIENCE & LEARNING

3: CURRENT CIRCUMSTANCES,
ROUTINES & LIFESTYLES

1: Computer and programming skills
2: Experience with attack methods
3: Prior contact with law enforcement
4: Foresight and planning
5: Conscience and moral attitude
6: Self-perception

Importance of computer and internet in professional and leisure pursuits; marital status; employment; friends

4: NEEDS AND MOTIVES

5: OPPORTUNITIES & INDUCEMENTS

Greed or need for money; need for excitement, thrill; need for power, status
and prestige; need for media attention;
self-concept boost; revenge; curiosity,
experimentation; political ideology

Legitimate: job offer
Illegitimate: access to illegal opportunity
networks; access to hacking tools & exploits & to insecure (company) networks

6: PERCEIVED SOLUTIONS
Legitimate: work position that allows application of skills
and provides good income
Illegitimate: hacking into a network; other cybercrimes;
commission of other crimes

7: SOLUTIONS EVALUATED
1. Degree of effort; technical difficulties and challenges
2. Amount and immediacy of reward / reputation
3. Risk of trace, capture, and apprehension
4. Likelihood and severity of punishment
5. Moral Crisis

Modification of Clarke &
Cornish (1985)
p. 168, Fig. 1

8: READINESS
to commit a hack

Figure 1: Model for the Initiation of Hacking
61

Box three lists the current life circumstances of the person. These circumstances
are relevant to the decision-making process because they structure the needs and motives displayed in box four as well as the perceived opportunities and inducements to
fulfill the needs and attain the desired goals shown in box five. The perceived solutions
to reach the desired goals listed in box six differentiate between legitimate and illegitimate alternatives. Each of the alternatives under consideration is then evaluated in
terms of the valued utility they provide and the risks and types of potential costs they
include. The actor further estimates the probability that either the benefits or the costs
will be the actual outcome.
Some versions of rational choice theory reconstruct this decision-making process
in great detail and with mathematical equations. They assign expected utility values to
benefits and costs and multiply these values with the probability assigned to them. The
predominant and most attractive alternative, the one that yields the highest expected
utility, is the one that is hypothesized to be invariably selected—otherwise the actor
would not act in accordance with his or her own subjective rationality, and the primary
proposition of rational choice theory would be violated. Cornish and Clarke’s rational
choice perspective does not attempt to reconstruct the decision-making process with
mathematical equations, but it shares the principal assumption of the more elaborate
models that show that the most beneficial alternative is selected once the decisionmaking process is concluded and the actor has selected the one alternative he or she
deems most beneficial.

62

Increasing frequency of computer hacking (due to success) until personal optimum is reached

st

1
hack

nd

2

rd

th

3

4

th

5

CHANGES IN PEER GROUP
Becomes friendly with other hackers; internalizes and adopts self-conception as a hacker; changes friends

INCREASED PROFESSIONALISM
Pride in improved hacking skills and knowledge; successfully reduces risk and increases efficiency through
careful selection of targets and methods; commits higher profile attacks

CHANGES IN LIFE STYLES AND VALUES
Recognition of dependency on hacking; chooses work to facilitate hacking; enjoys life in the fast lane;
devalues legitimate work; justifies criminality

Modification of Clarke & Cornish (1985) p. 168, Fig. 3

Figure 2: Model for the Habituation of Hacking

63

N

th

th

N + 1 HACK
Applies moral
principals; has
increased fear;
loses intellectual
curiosity or kick

CHANGES IN LIFE
CIRCUMSTANCES

REEVALUATION OF READINESS

Reaches maturity;
gets married; gets
job offer

REJECTED ALTERNATIVES

th

N + 2 HACK
Gets (close to
being) traced or
tracked; gets apprehended

Legitimate: no opportunity to apply computer
skills in job; pay too low
Illegitimate: other
crimes ruled out for
various reasons

CHANGES IN LIFE
CIRCUMSTANCES
Gets job offer at
cyber-security
company allowing
legal application of
skills; ultimatum
from loved ones

FURTHER REEVALUATION OF
READINESS

DISTANCE FROM HACKING
Young man’s game; too risky; too much
effort; income too irregular

OTHER CRIME

LEGITIMATE ALTERNATIVES

Rents out botnets or
spam servers; credit
card fraud;
computer espionage

Accepts job offer;
starts own business
Modification of Clarke & Cornish (1985) p. 172, Fig. 4

Figure 3: Model for the Desistance from Hacking
64

In the process of habituation, shown in Figure 2, the central influence on decision
making gradually shifts from background variables to the rewards of successful engagement in the criminal activity and the influences this engagement has on the current
life circumstances. The offender often changes peer groups and becomes closely involved with other persons sharing their interests, because these persons can provide
him or her with valuable information and other resources. The self-perception as a
hacker then becomes internalized and the person takes pride in his or her improved
skills and knowledge. Throughout the successful habituation process, the methods of
attack become increasingly sophisticated, which also reduces the perceived risk of apprehension. Offsetting the risk reduction of increased professionalism, on the hand, is
the probable selection of higher profile targets by more experienced hackers. Higher
profile targets are more attractive to seasoned hackers because they promise more interesting information, more reputation, higher financial gains, or just a greater challenge
and thrill. However, higher profile targets also bear a higher risk and make activities
riskier, not safer. Whether Cornish and Clarke’s theoretical assumption of a decrease in
the subjectively estimated risk potential over the course of a criminal career is valid
when applied to hackers is presently unclear and will be examined in this study.
Cornish and Clarke’s model further predicts that over the course of a criminal career, offenders justify their increasingly criminal and deviant behavior by employing
“techniques of neutralization” (Sykes & Matza, 1957). Applied to hacking, such techniques of neutralization could, for example, be the presentation of the behavior as an
expression of the hacker’s commitment to the free flow of information, or as resistance
to political authoritarianism and corporate domination.
65

Towards the end of the habituation phase, background factors play only a remote
and unimportant role in the decision-making process. As shown in Figure 3, the offender
begins to consider desistance once negative experiences have been made, life circumstances have changed, and/or new legitimate and promising opportunities have opened
up.

The Reconstruction of Crime Events as a Decision Sequence

DECISION
To commit hack
(see the initial involvement model)

SELECTED SERVER / NETWORK

REJECTED SERVER / NETWORK

Easily accessible; has interesting or profitable information; suitable for reputation
gain; antipathy with ideas / opinions of
host

Unfamiliar architecture; no interesting or
profitable information; sympathy with
ideas / opinions of host

COMMITS HACK
REFRAINS FROM HACK
Interesting data, profitable information;
incompetent or negligent administrator;
missing security (wlan encryption, firewalls, virus scanner, updates, etc.);
common vulnerabilities available

Suspicious of ‘honey-pot’; security
measures enacted and up-to-date

Modification of Clarke & Cornish (1985) p. 196, Fig. 2

Figure 4: One Stage of the Crime Event: The Selection of a Target
66

Many criminological theories reduce the actual crime event to the selection of a
suitable target. Rational choice theories, on the other hand, view the crime event itself
as a series of decision-making processes that merely begin with the selection of a target. This seems to be the more accurate perception because oftentimes, crime events
unfold in unforeseen ways and the offender has to adjust accordingly.
Figure 4 displays the hacker’s selection of a suitable target for the attack. However, this selection is only the first in a complex process towards the completion of a
successful attack. In 1994, Cornish borrowed the concept of crime scripts (Cornish,
1994) from cognitive psychology and introduced it into criminology. The crime scripts
approach attempts to reconstruct the detailed procedures used by offenders in the
commission of their crime step-by-step. It is a well-suited theoretical approach for the
reconstruction of many different types of criminal behavior (Cornish & Smith, 2003). Its
application to hacking, however, is a difficult undertaking because, unlike many crimes
that are rather simple acts and similar to other mundane everyday activities (Gottfredson & Hirschi, 1990), hacking is an oftentimes highly complex and non-linear activity
and a single hacking event can involve many different procedures and routes of action.
Thus, the detailed reconstruction of the scripts involved in hacking presents not only a
difficult theoretical challenge, but an even greater challenge for operationalizing and
testing. Tests of such models can probably be best assessed through findings from
crime science studies or with cognitive interviews of individual hackers. Cognitive interviewing is based on a technique called protocol analysis (Ericsson & Simon, 1980,
1984). During a protocol analysis, subjects are asked to verbalize their thoughts as they
perform a particular action or solve a problem. These verbalizations are then recorded
67

and evaluated by the researcher. Repeated cognitive interviews during various types of
hacking attacks will eventually produce theoretical models that accurately reflect the
complexity involved in this activity.
The primary goal that the present study seeks to achieve is not to conduct individual case studies of hacking events, but to generate quantifiable insights into the
hacking community and to assess the general importance of the various factors that,
according to the rational choice perspective, are related to the involvement in hacking
activities. Hence, the analysis focuses on testing the seven hypotheses that were stated
earlier in this chapter. In addition to the influence of rationality and risk propensity on the
conduct and outcomes of hacking activities, the analysis examines the relevance of
several factors that the three involvement models consider to be important. It will be up
to future crime science or protocol analyses to complement the insights that are generated by this study with more specific reconstructions of the crime scripts in hacking
events.
To accomplish the goal of producing generalizable results, the study utilized a
survey as measurement instrument. A survey format was selected because surveys are
the social research method that is best suited to produce quantifiable findings. This feature renders them the most appropriate method for obtaining information about the distributions of sociodemographic characteristics and motivations of hackers, and for examining the pertinence of main factors of the three models shown above. The details of
the research design and process as well as the individual items that were included in
the questionnaire are specified in the following chapters.

68

CHAPTER SIX:
RESEARCH DESIGN
The goal of the present study is to test all of the hypotheses stated in the previous chapter and, given the exploratory nature of this project, to also generally examine
the sociodemographic characteristics of the hacker community. To achieve these goals,
the research project was designed to produce quantifiable results that are more representative and can be generalized to a wider target population than previous qualitative
case studies of hackers (Jordan & Taylor, 1998; Taylor, 1999). A survey was developed
and used for data collection (Boudreau, Gefen, & Straub, 2001), because surveys are
the one data-collection method particularly suited to produce quantitative results that
can be generalized to other members of the population of interest and oftentimes even
to other similar populations (Newsted, Chin, Ngwenyama, & Lee, 1996).
Even though network administrators (D'Arcy, 2007), software developers (Lakhani & Wolf, 2003), and victims of cybercrimes (L. A. Gordon et al., 2005, 2006), for example, have all been the subject of several survey-based research projects in the past,
no such survey-based study has been conducted on hackers so far. No study exists in
the current cyber-criminological literature that allows scholars to draw conclusions about
any larger population of hackers. Accordingly, the specific details and distributions of
the hacking community continue to remain largely unknown. The aim of the present
study is to undertake this task and to begin filling the remaining gap in the criminological
literature on hackers and the hacking community by providing the first quantifiable insights into the hacking underground.
69

The Difficulty of Sampling Hackers
Although important to achieve, the realization of this study is challenged by a variety of difficult methodological obstacles. Most importantly, the quantitative nature of
this research project demands that the problem of obtaining a sample that is reasonably
representative of any larger hacker population is resolved. Two main characteristics of
the hacking community, in particular, make the process of gathering a representative
sample of hackers a difficult undertaking. First, partly due to their oftentimes illegal activity, hackers form a hard-to-reach and hard-to-identify, sealed off and mostly hidden
underground community. Second, they present a rather rare population. Both of these
circumstances present unique challenges for the sampling process. They require that
researchers find a method for establishing contact with a representative sample of target subjects who are relatively small in number, decentralized, concerned about protecting their real identity, and who interact primarily through specialized online communication channels.
One potential solution to this problem is the utilization of some form of snowball
sampling technique, in which a few active hackers who are known to the researcher recruit other subjects from their acquaintances for the study. The drawback of utilizing
snowball samples is that they are confronted with the crucial problem of study subjects
not being selected from any larger sampling frame. Dependent on the underlying network structures of the participating respondents, this circumstance subjects snowball
samples to several uncontrollable biases (Groves et al., 2004). While some systematic
snowball sampling techniques, such as respondent-driven sampling, are designed to
70

partially alleviate this problem (Heckathorn, 1997), no snowball sampling technique offers any remedies capable of eliminating coverage error. Consequently, these techniques do not permit accurate estimates of the representativeness of the population of
interest.
Overcoming the representativeness problems associated with snowball samples
is a difficult challenge when studying underground target populations because for these
populations, no lists of eligible subjects are available. Hence, the researcher has to select a sampling method that allows the construction of a sampling frame which simultaneously minimizes the systematic undercoverage of eligible subjects and the inclusion
of ineligible subjects (Groves et al., 2004). The difficulty of this particular methodological
challenge is probably part of the reason why no quantitative study of hackers has been
conducted so far.
In the case of hackers, two principle solutions are conceivable for this problem.
One possible solution is to access the communication channels typically utilized by
hackers. Such a study could attempt to advertise links to the survey on public hacker
message boards and then draw its participants from members of the boards. Again, the
problematic aspects of this method are evident. First, it would suffer from several shortcomings generally associated with Internet-based surveys. Unlike traditional methods of
establishing contact with respondents, such as random-digit dialing in telephone surveys or sending surveys via mail, researchers have no direct access to the population
within an online environment. Consequently, the sampling frame of Internet surveys remains a principal issue (Couper, 2000). The vagueness of the sampling frame entails
other problems. For example, it severely limits the possibility to estimate nonresponse
71

rates (Couper, 2000), which, in the case of Internet-based surveys, are typically higher
than those achieved with traditional recruiting methods (Matsuo, McIntyre, Taomazic, &
Katz, 2004). Despite these problems, such an approach should not be generally discarded solely because of its sampling-frame related issues. As Gosling and colleagues
point out, their Internet study, although perhaps not exactly representative of the population in general, nevertheless compared favorably to other publications with regard to
all major sociodemographic characteristics (Gosling, Vazier, Srivastava, & John, 2004).
An additional set of problems confronting this approach is related to the environment in which respondents fill out the survey. In the case of online surveys, this environment is completely out of the researcher’s control or influence. Problems resulting
from self-selection, multiple submissions, non-serious responses, and dropouts are
more severe because Internet surveys are completely self-administered and do not
permit any ad hoc interaction between participant and interviewer. Hence, it is more difficult for the interviewer to recognize and correct problems of comprehension or intentional mischief (Porter & Whitcomb, 2003; Reips, 2002). This circumstance heightens
the possibility of individual participants negatively impacting the quality of the results in
online surveys (Nosek, Banaji, & Greenwald, 2002).
Particularly in the case of studies on hackers, the principal problems of Internetsurveys are likely to be further exacerbated. Many hackers are generally suspicious
about the authenticity and trustworthiness of information they retrieve online. For the
researcher, this implies that the difficult task of establishing the necessary trust without
personal contact becomes a crucial challenge. Failure to convince hackers to trust the
research project will cause many potential participants to refuse taking the survey or to
72

intentionally provide false information. Combined, all of these problematic aspects severely question the validity of the results yielded by this sampling approach.
An additional set of sampling-related problems of Internet-based surveys arises
from the circumstance that their samples include only members of particular message
boards, not hackers in general. Two aspects of this limitation are particularly problematic. Many message boards serve as discussion forums for specialized topics and are interesting only to fractions of the hacking community (e.g. http://sla.ckers.org/forum for
web application security). Equally important is the circumstance that many of the more
popular boards such as “The Phoenix Project” or “Hacker’s Den88” are typically consulted by lesser skilled beginners who use them for basic advice and guidance. Both of
these aspects are likely to introduce systematic sampling biases because they lead to
an underrepresentation of higher skilled hacking experts.
An alternative approach is to field the survey during an occasion when larger
numbers of hackers meet in person. Each year, several hacker conventions take place
in the United States (e.g. BlackHat, ShmooCon, DefCon, ToorCon, or PhreakNIC) and
around the world (e.g. Chaos Communication Camp in Germany, CanSecWest in Canada, PacSec, SysScAN, and Hack in the Box in Asia, or even PakCon in Pakistan).
Fielding a survey at such a convention presents the researcher with the opportunity to
contact more seasoned experts and hackers who are involved enough to undergo the
efforts and costs involved in attending a convention. While eliminating many of the problems associated with Internet-based surveys, researchers following this methodological
approach have to keep in mind that all main hacker conventions have a distinct profile.
Like Internet-based message board samples, they also attract only parts of the hacking
73

community. Consequently, studies based on samples of attendees of hacking conventions are also subjected to limitations regarding the generalizability of their results.
This circumstance holds two implications for research projects following this approach. First, researchers have to carefully consider several sampling-related criteria
when selecting a convention. The world’s largest annual convention DefCon, for example, has a reputation of drawing a large crowd of predominantly young and generally
less experienced hackers who mainly attend the convention to have a good time with
like-minded people. In contrast, other conventions draw more experienced security experts who are interested in the latest security-exploit developments, while others, like
the BlackHat convention, even charge several thousands of dollars for attending advanced seminars in what the BlackHat organizers term “offensive security.” In order for
a convention to draw a sample of attendees who are most representative for the hacker
community, it has to address a broad spectrum of topics and attract a wide variety of
hackers with different security-related interests.
Second, the interpretation of the results has to take this sampling-related limitation into consideration. The fact that hacker conventions offer the most precise sampling
frame for researchers does not imply that this convention-based sampling frame is
equivalent to the target population of hackers in general. Strictly formulated, the findings
of such a study can only be generalized to the greater population of hackers who would
potentially attend the particular conference.
Despite the inevitable restriction regarding the generalizability of the results, this
methodological approach is nonetheless better suited than snowball or Internet samples
because of the strengths it offers in managing the task of sampling hackers. Most im74

portantly, it provides researchers with a delimited sampling frame. Surveys fielded at a
particular convention can also serve as a point of reference to which future surveys that
are fielded at different conventions can compare and contrast their results to. Eventually, such a comparison of results from different conventions will provide estimates that
closely reflect the distributions within the diverse hacking community. For the present
study, this approach presented a viable opportunity to gain the first general insights into
an underground community that remains largely unknown. Because it is the least problematic sampling method and the one most likely to generate generalizable results, this
methodological approach was chosen.

The ShmooCon Convention
The survey for the current research project was fielded during the 2008
ShmooCon convention in Washington D.C. The ShmooCon convention was selected
because its profile closely matches the selection criteria outlined above. Since its first
convening in 2004, ShmooCon has developed into one of the largest annual conventions worldwide. Today, it ranks among the most popular conventions, and it is attended
by both U.S. and international hackers and security experts. In addition, it has one of
the most diverse programs that is attractive to a wide variety of hackers (Grecs, 2008).
The convention is commonly announced as an “annual East Coast hacker convention
hell-bent on offering an interesting and new atmosphere for demonstrating technology
exploitation, inventive software and hardware solutions, as well as open discussion of
critical information security issues” (HackWire, 2004).

75

The 2008 convention was held over the weekend from Friday, February 15 to
Sunday, February 17 in the Marriott Wardman Park Hotel in Washington D.C. After the
official opening remarks on Friday afternoon, various speakers started the convention
with presentations on a number of different topics ranging from “Intercepting Mobile
Phone/GSM Traffic” to “Forensic Image Analysis for Password Recovery” and “Hacking
the Samurai Spirit” to “New Countermeasures to the Bump Key Attack.” The main event
on Friday was a keynote address given by Dr. Edward W. Felten, a highly renowned
professor of Computer Science and Public Affairs at Princeton University and the founding Director of Princeton’s Center for Information Technology Policy.
The main program on Saturday and Sunday was divided into the three general
sections “Build It!,” “Break It!,” and “Bring It On!.” In each of the three sections, expert
speakers gave presentations on a broad variety of topics. The primary target audience
of the “Build It!” section consisted of network administrators and other security experts.
The section covered several topics of general interest to this audience, among them
“They’re Hacking Our Clients! Why are We Focusing Only on the Servers?,” “Practical
Hacker Crypto,” “Using Aspect Oriented Programming to Prevent Application Attacks,”
“Hacking Windows Vista Security,” or “Path X: Explosive Security Testing Tools using
XPath.”
The “Break It!” section, on the other hand, included presentations covering a variety of aspects related to the act of breaking into computer systems and networks. The
section was geared towards black hat hackers and penetration testers alike and contained presentations entitled “Virtual Worlds – Real Exploits,” “Smarter Password Crack-

76

ing,” “VoIP Penetration Testing: Lessons Learned,” and “Malware Software Armoring
Circumvention.”
The third section “Bring It On!” covered topics generally related to the lives of
hackers. The presentations in this section included titles such as “When Lawyers Attack! Dealing with the New Rules of Electronic Discovery,” “The Geek and the Gumshoe
or Can Mathematics and Crimes?,” “You Must Be This Tall to Ride the Security Ride,”
“Legal Issues for Bot-net Researchers and Mitigators,” or “How do I Pwn Thee? Let Me
Count the Ways.” Before the final closing remarks on Sunday afternoon, the main
agenda of the convention ended with a large panel discussion “On the Social Responsibility of Hackers” that involved convention organizers and speakers from all three sections. Combined, the three main sections and the surrounding presentations and discussions addressed a broad variety of hacking and security-related topics that ranged
from purely technical topics to sessions providing hackers with practical advice for coping with the problems and challenges in a hacker’s life.
In addition to the expert speaker presentations in the three sections, the convention offered several other activities and competitions. Participants in the various hacking
contests, for example, had to go on a virtual treasure hunt and gather clues and hints
that were hidden all over the convention. Even the red laser engraved acrylic punch
card badges that attendees received upon registration displayed different dot patterns of
PDP-8 machine code that, if combined, led to other clues as part of one contest. The
progress of the contestants in the hacking arcades was made public in real time by a
projection of their advancement on one of the hallway walls. The winner of each contest
was rewarded with prizes in the form of gaming consoles and, most importantly, the
77

recognition from peers. Prizes were also offered for the winners of the lock picking
competition in the “Lock Picking Village” and the three first places in the “Hack or Halo”
ego-shooter video gaming tournament. Other events during the convention involved
several live-streamed online video feeds and interviews with prominent hackers as well
as a Saturday night party in a nearby dance club.
The unconventional name of the convention was explained during the opening
remarks. The ShmooCon conference received its name from the concept of so-called
“shmooballs.” As part of the registration package, each attendee received a soft rubber
stress ball the size of a tennis ball. All participants were explicitly instructed to throw this
ball at the speaker when, during a presentation, they felt that the speaker made a mistake or did not know the material of the presentation well enough. The act of throwing
the shmooball then entitled the thrower to voice their objections or to correct the speaker. In making this concept a central aspect of the convention, the ShmooCon organizers
attempted to break down the hierarchies between speaker and audience and to solicit
input from all experts in the audience. By doing so, they also demonstrated their adherence to the democratic ideas of the original hacker community. In a humorous demonstration of how the shmooball concept is intended to work, some of the staff members
assaulted the opening speaker during his explanation of the shmooball concept with a
self-made shmooball canon that fired a large magazine of balls across the auditorium at
him.
Aside from the shmooball concept, several other organizational details implemented the creeds and convictions of the traditional hacking community. The application process, for example, was designed so as to give all interested persons the same
78

chance of obtaining an admission ticket. The tickets were sold exclusively online. The
sales events took place on three different occasions on a first come, first serve basis.
The sale of tickets was restricted to only three occasions because the number of admissions was limited to 800 attendees and, according to the convention organizers, the
demand for tickets vastly exceeded their available number. The structure of the organizational team was itself organized in a strictly non-hierarchic, democratic fashion. All
suggestions were due for approval and implemented only after they obtained a majority
of the vote. Furthermore, the organizers underwent great effort to solicit detailed feedback from the attendees and, thereby, include them in the improvement process for future conventions.
Summarizing, it can be recapitulated that the conceptual layout of the 2008
ShmooCon convention made it an ideal candidate for the present research project. Its
main program was designed to be inclusive and to address the interests of a wide variety of hackers. Its agenda combined a mixture of serious discussions of latest security
issues and several events to entertain attendees, thereby making the conference interesting for hackers of all different backgrounds and skill levels. Before the survey was
fielded during the conference, however, several steps were taken to ensure its appropriateness and its success as a scientific measuring instrument.

The Pretest and IRB Approval
Boudreau, Gefan, and Straub (2001) emphasize the need for every survey instrument to be pretested as a preliminary step that prevents unanticipated encounters
during the fielding of the final version of the survey. Therefore, a pretest of the initial
79

draft of the survey instrument was conducted with a convenience sample comprised of
six self-proclaimed hackers known to the researcher. Each of these testers volunteered
to participate in the pretest. They received a copy of the initial draft along with some
general instructions and information about the purpose of the study. The general instructions asked them to provide feedback about the design of the survey with regard to
three aspects: (1) the suitability and appropriateness of the individual items and answer
categories in general, (2) their estimate of the suitability of individual items to address
the larger research questions, (3) the appropriateness of the content and structure of
the instrument for the intended target population and measuring situation.
The long geographical distances between the pretest participants, who are residing in different countries, did not permit the researcher to conduct personal debriefing
sessions. Instead, the panel members were asked to provide detailed written feedback
after their completion of the survey and to return their comments via email. The feedback received from this pretest focused primarily on revisions of the wording and was
aimed at eliminating potential ambiguities in some of the hacking-related questions. It
also included some suggestions for minor changes in the provided standard answer
categories. While there was general agreement among the reviewers on the general
suitability and appropriateness of the items included in the survey and on the exhaustiveness of the provided standard answer categories, one particular item received negative feedback and was therefore excluded from the final survey. All six participants reported that answering the item “What does the term hacking mean for you?” was disproportionally time-consuming since it consumed almost as much time as the remainder
of the survey.
80

In a subsequent step, the revised version of the survey draft was reviewed by
two experienced survey researchers on the sociology faculty at University of Central
Florida. Aside from providing a second scrutiny of the appropriateness of the survey tool
and the unambiguousness of the individual items, this expert assessment was also
used to ensure the appropriateness of the survey as a scientific measurement instrument and to examine the content validity of the items, many of which were developed
specifically for the present study and had not yet been validated. Based on the recommendations of these experts, some modifications and refinements were implemented in
the final version of the questionnaire. In particular, the wording of a few individual items
was revised and some items were rearranged. There was agreement among the reviewers on the importance of all main sections of the questionnaire, on the appropriate
length of the measurement tool, and on the suitability of the included items to address
the intended dimensions of the underlying concepts. Thus, the results of these pretests
suggest that the questionnaire was well designed for the research project and possessed adequate validity.
Following the pretest of the questionnaire, the research proposal was submitted
to the University of Central Florida Institutional Review Board (IRB) for approval. The
submitted proposal materials consisted of the study questionnaire, an informed consent
document, and an online application that detailed the various aspects of the study to the
IRB reviewer. The informed consent form (see Appendix B) served as an identification
of the researcher. It also informed prospective participants about the details of the
project they are going to participate in and granted confidentiality and anonymity. To
ensure the highest possible degree of anonymity, the documentation of consent was
81

waived on the consent form. The study was exempt from a full board review because it
recruited no vulnerable populations and posed only minimal risk for participants since it
did not ask respondents to reveal any specific self-incriminating information. Only a few
formal stipulations were requested by the IRB reviewer in the initial review, and the
study received IRB approval after the requested changes had been implemented.

The Data Collection Process
Once IRB approval was obtained, a formal request for permission to field the
questionnaire during the convention was submitted to the organizers of ShmooCon. Although this request was received positively, the organizers expressed their main concern as brokers for the community of protecting their attendees’ privacy, and they asked
for the submission of additional identification materials. Among the requested identification was the curriculum vita of the principal investigator, the contact information of a
study supervisor, a detailed research proposal, the documentation of IRB approval, and
a statement that the survey was for research purposes only and involved no commercial
or governmental institutions or organizations.
Once all these materials had been submitted and reviewed, the study was approved and the organizers of the convention even offered to announce the survey
project as an integral part of the convention in the opening remarks to build trust among
the attendees. The announcement of the study in the opening remarks included a PowerPoint slide that informed attendees about the research project and its IRB approval
as well as a personal introduction of the principal investigator. This introduction by the
organizers proved to be of invaluable help for the acceptance of the study. Many res82

pondents immediately recognized the researcher from the opening remarks when they
were asked to participate in the study.
Furthermore, the organizers of the convention provided a table in the central
hallway between conference rooms and several chairs for participants to take the survey. The survey table was in an ideal location next to the sponsor booths and between
presentation rooms. Whenever attendees walked from one presentation section into
another, they passed by the study table. Placed on the survey table were copies of the
questionnaire, consent forms, large University of Central Florida printouts, poster-sized
prints inviting attendees to “be part of the first survey on hackers,” business cards of the
researcher, and Snickers candy bars as incentives to participate. Convention attendees
passing by were verbally approached by the researcher and invited to participate in the
study. They were told that the survey referred to hacking as the unauthorized intrusion
into computer systems, networks, or website servers and they were asked to participate
only if they had ever committed such an intrusion. Attendees who indicated that they
worked as penetration testers were asked to only participate if they had ever invaded a
computer system outside of a contractual agreement and instructed to only refer to
these intrusions in their answers. Penetration testers and other attendees who reported
to have never committed such a hack were told that the survey did not pertain to them
and were dissuaded from taking the survey.
Approximately one-third of the approached attendees were rejected because
they had never committed an unauthorized computer intrusion. Most of the rejected attendees were penetration testers who reported to have always followed a strict ethical
code. A smaller fraction of the rejected attendees had simply never attempted a com83

puter intrusion, either because they had just recently become interested in hacking or
because they attended the conference with their significant other who is actively hacking. A total of 164 questionnaires were distributed among qualified attendees. Most of
the persons who agreed to participate in the study filled out the questionnaire on site.
Some, however, asked to take it with them and fill it out at a more convenient situation.
Of the 164 distributed surveys, 129 were returned to the researcher, 124 of which were
filled out completely and included in the analysis of the study. Thus, the response rate
of completed and returned surveys was 75 percent and an estimated 25 percent of all
eligible attendees were included in the study. The large fraction of sampled units and
the high response rate can be attributed to the personal interest many hackers displayed in the survey project and to the trust the official announcement of the study had
helped to build.

84

CHAPTER SEVEN:
THE SURVEY INSTRUMENT
The survey instrument of the current study was designed to test all of the hypotheses stated in the third chapter. In particular, it assessed the main phases in the careers of hackers as proposed by the rational choice perspective and examined the extent of rationality in their involvement and event decision making. In addition, it also
measured the general distributions of sociodemographic characteristics within the hacking community.
At the beginning of the questionnaire, a small introductory section welcomed the
respondents, thanked them again for their willingness to participate in the study, and
provided a brief introduction of the researcher and the purpose of the study. It also
reassured respondents that the survey is strictly noncommercial and for dissertation research purposes only, that there are no institutions or organizations involved or affiliated, and that their participation in this project is voluntary, anonymous, and confidential.
The measurement instrument itself consisted of a total of 72 items in three main
sections. The first section gathered detailed information about the various phases of the
respondents’ hacking careers. It embodied items pertaining to the initiation of the hacking activity, its habituation, and the eventual desistance from hacking. It further assessed several other details of the respondent’s hacking activity, including a variety of
involved decisions and motivations. Given the exploratory nature of this research
project, many items in this first section offered open-ended “Other” answer categories in
85

addition to the answer options provided. The answers recorded in these open “Other”
categories were included as string variables in the dataset. Several of the items also
permitted respondents to give multiple answers. All answer options listed for these
items were coded as separate dichotomous variables.
The second section of the survey instrument consisted of questions measuring
the degree of risk propensity, rationality, and faith in intuition in the respondents’ decision-making processes. The main reason for the assessment of these three personality
traits was to allow the survey instrument to test all of the hypotheses that were derived
from the rational choice perspective in chapter three. The operationalization of theoretical hypotheses concerning the influence and degree of rationality in decision-making
processes presented a principally difficult methodological challenge. Typically, such assumptions are measured with either fictional scenarios of nearly real-life decisionmaking situations (Clarke & Cornish, 2001; Finch, 1987; Harrington, 1996; Kerlinger,
1986) or with social psychological scales (Clarke & Cornish, 2001; Kerlinger, 1986). The
decision to operationalize the three personality traits with social psychological scales in
the present study was made because this assessment format better fitted the setting in
which the survey was fielded. It was less demanding to merely ask respondents to indicate their agreement to some general statements than to have them read through several fictional scenarios before answering questions. Moreover, this measurement technique also reduced the overall length of the survey and the time needed for its completion. Further lengthening the survey with scenarios might have decreased the attendees’ willingness to participate, given that the survey was already eight pages long and
took about 15 minutes of the time they spent at the convention.
86

All items in this second section were taken from well-established scales that were
abbreviated to keep the overall length of the survey within reasonable limits. The decisions about which items to include were based on statistical and conceptual considerations. Items were selected according to their item-to-total correlations and their factor
loads on the respective underlying dimension. To maintain construct validity despite the
shortening of the scales, items were also selected based on their ability to measure different aspects of the underlying concept. All items were anchored on appropriately labeled seven-point Likert-type scales. Seven-point scales were chosen over five-point
scales because they allow for finer distinctions in the measurement of the variables
(Sommer & Sommer, 2002). Another advantage of seven-point scales is that increasing
the scale end points from five to seven also tends to increase the ability to reach the
upper limits of reliability (Krosnick & Fabrigar, 1997; Nunnally, 1978).
The third section concluded the survey instrument with measures of basic sociodemographic information. The items included in the three parts of the survey instrument, which can be found in Appendix A, are presented in greater detail in the following
sections.

The Measurements of General Hacking Activity
The first item in the survey asked respondents to indicate whether they are currently actively hacking. Two dichotomous answer categories, “Yes” and “No,” were provided and respondents who answered with “No” were instructed to think back to when
they were still hacking when answering the subsequent questions. The rationale behind
the placement of this item at the beginning of the questionnaire was to differentiate be87

tween active hackers and respondents who had been hacking at one point in their lives
but had desisted from hacking in the meantime or had given up any illegal hacking activities to become penetration testers.
The items measuring the various aspects of the respondent’s hacking career
were organized in chronological order. The first set of questions examined the onset of
the hacking activity. Respondents were asked to indicate how old they were when they
first became interested in hacking and to also provide the motivations that initially got
them interested in this activity. The motivations listed as standard answer categories for
this question were derived from the existing literature on hackers (Taylor, 2000, 2004;
Thomas, 2002; Yar, 2005a) and arranged according to their assumed importance (Yar,
2005a). The listed motivations were “Intellectual Curiosity,” “Excitement, thrill, fun,” “Experimentation,” “Status and prestige,” “Financial gain,” “Peer recognition,” “Political ideology,” “Media attention,” “Protest against corporations,” “Self-concept boost,” “Feeling
of power,” “Personal revenge,” and an open-answer category “Other (specify).”

A

second item then asked participants to indicate which of these motivations had been
their primary one.
Further examining the onset phase, the following item then asked participants to
indicate the length of the time frame between their initial interest in hacking and their
first actual hack. The purpose of this item was to provide a frequency measure that
permitted comparisons to later stages of the respondent’s hacking career and, thereby,
allowed additional measures of whether the activity has indeed become more intensified. The answer categories for this item gave respondents the choice to enter their an-

88

swers in days, weeks, months, and years. All answers were later recoded into days to
provide a unified scale for the corresponding variable in the dataset.
After the examination of the initial interest phase, the survey shifted to questions
about the details of the first actual hack the respondents had ever committed. Again,
participants were asked about their specific motivations to commit this first hack. The
answer categories listed for this question were the same as for the previous question
about the motivations that had sparked the initial interest in hacking. The motivations for
the first hack were measured separately from the motivations that initially got respondents interested in hacking because this separate measurement provided an indication
of the consistency between the two. Thereby, it also allowed an assessment of the relative importance of situational factors in the decision to commit the first hack. For example, a person could become interested in hacking out of intellectual curiosity but then
decide to launch a hacking attack in an attempt to take personal revenge.
The next question then asked respondents to specify the type of target they selected for their first hacking attack. The set of answer categories for this item distinguished between single hosts, networks, and websites and between private, corporate,
non-profit, and governmental entities. Again, the item also included an “Other” category
that allowed respondents to specify any other type of target. Further detailing the aspects of the target, the subsequent item asked participants to report the criteria their
target selection was based on. The answer options given for this item were “Easy
access,” “Interesting information,” “Profitable information,” “Reputation gain,” “Antipathy,” and again “Other.”

89

Two additional items were included in the questionnaire to explicitly assess the
role financial motivations and considerations played during the onset of the hacking career. The first asked participants whether they were employed at the time they committed their first hack. The answer options for this item were “Yes, full-time,” “Yes, parttime,” and “No.” The second item directly asked respondents whether “economic profits
[were] a motivation at all.” The answer categories for this item were “Yes, an important
one,” “Yes, but not important,” and “No.”
To provide a more detailed assessment of the first hack ever committed by the
participants, the survey asked them to indicate the methods they used in their first technical intrusion. The item referred to the first hack as the first technical intrusion because
it was intended to measure only the technical methods employed in the first hack. The
term “technical intrusion” basically denotes all attempts to subvert exploitable system or
network defaults, holes, bugs, or passwords. It excludes other methods that can be
used in a hack, such as social engineering methods. Specifically, respondents were instructed to mark all technical methods they had used in their first hack from a list of 24
of the most common technical intrusion methods (Kanellis, Kiountouzis, Kolokotronis, &
Martakos, 2006). The methods listed as answer options for this item were arranged in a
sequential order from reconnaissance tactics over different ways to gain access to a
computer system to methods for covering up any potential traces. The reconnaissance
tactics included in the list were “Footprinting,” “Ping sweeping,” “DNS zone transfer,”
“Whois,” “Network mapping,” “Port scanning,” “Versatile scanning tools,” and “Vulnerability scanning.” The methods of gaining access to computer systems contained were
“RPC port/end-point dump,” “Packet sniffing,” “Session hijacking,” “Grinding pass90

words,” “Password cracking,” “Password theft,” “Directory traversal/climbing,” “Buffer
overflows,” “Format strings,” “Resource mismatches,” “CGI,” “Root shell/kits,” and “Keyloggers.” The list ended with three common methods for covering up any potential traces: “Spoofing,” “Bouncing,” and “Source routing.” Even though this list basically covered
the methods most commonly used by hackers (Kanellis et al., 2006), the number of
possible attack tactics is far greater. Thus, the answer categories were again complemented by an open-ended “Other” category. Since it cannot be assumed that all readers
are familiar with all different hacking techniques, the skill-level they require, and the type
of target they are usually used against, the different methods are briefly introduced in
the Glossary.
The assessment of individual hacking methods was included for a twofold purpose. The primary reason was to provide a measurement of the prevalence of several
known hacking tactics and methods. Currently, no quantitative estimate of the popularity
of different hacking methods exists in the literature, because the victimization surveys
conducted on the issue thus far can assess the methods that are being used in hacking
attacks only indirectly and incompletely. Some methods simply leave no traces on the
target machine. Others can be rendered undetectable to victims or even cyber-forensics
by several potential circumstances. For example, many reconnaissance methods leave
no traces on the target computer and, oftentimes, efforts to cover up traces succeed in
rendering the detection of intrusion methods impossible.
The second purpose of the inclusion of item measuring the employed hacking
methods was to also provide a potential assessment of the level of technical skills the
respondent had when they launched their first hacking attack. Some hacking methods
91

require a much more advanced skill level, whereas others can be performed by merely
executing common software tools. Thus, the item also provided an indirect proxy measure of the skill-level during the onset of the hacking activity. The decision to include an
indirect measurement of the respondents’ skill level at the beginning of their hacking careers rather than a direct measurement was made in an attempt to avoid the potential
self-perception and social-desirability biases introduced by a direct measurement. While
suited to avoid such biases, conclusions about the skill-level of the attacker have to be
drawn cautiously from this indirect assessment because the selection of a particular method is not solely dependent on the skill-level of the attacker. The decision to use a particular method is also influenced by the type of target and the specific vulnerabilities of a
particular target. Directory traversals, for instance, are attack methods that are used
primarily for website attacks, but not for attacks on computer systems or networks. Furthermore, conclusions about the skill-level of the attacker have to be drawn with caution,
because some of the attack methods denote categories that can subsume different attacks of varying difficulty.
The next set of questions on the questionnaire asked hackers about their hacking
career in general. The questions in this part were broken down into three different hacking activities that were measured separately. The questionnaire distinguished between
technical intrusions, so-called “Wetware” social methods, and the distribution of malicious code. Wetware is a term commonly used among hackers and computer programmers. It is an analogy to software, but refers to the act of programming a brain, as
opposed to computer code. Six common wetware methods were listed as answer op-

92

tions: “Social Engineering,” “Deception,” “Bribery,” “Shoulder surfing,” “Impersonation,”
and “Dumpster diving.” The six methods are briefly introduced below.
Social engineering basically denotes all attempts of hackers to establish and
subvert trust relationships with victims or to predict their behavior. Once a trust relationship is established, the attacker tricks the victim into revealing information or performing
an action, such as a password reset, for example, that can then be used in the attack. A
subcategory of social engineering methods are so-called reversed social engineering
methods, in which attackers use their expertise to induce their victims to reveal sensitive
information when seeking their help and advice. Deception is a method similar to impersonation, but demands a higher commitment than the latter. Whereas an impersonation
can be as simple as sending an email pretending to be somebody else, deception typically refers to attempts to gain inside access to a network in order to be able to launch
an attack from behind its firewall. Shoulder surfing and dumpster diving are two other
popular social methods that directly exploit the oftentimes careless behavior of targets.
Shoulder surfing describes an attack method in which the hacker simply monitors the
moment the victim types passwords or other sensitive information on their keyboard.
Once the login information is known, the hacker can exploit the user account without
having to worry about any counter measures such as firewalls or intrusion detection
systems because they can no longer distinguish between the legitimate user and the
attacker. Similarly, dumpster diving, a slang term for searching through the target’s garbage, is a simple and oftentimes legal method to gain valuable information because
many people dispose of sensitive information without destroying it first. Bribery is probably the easiest way to hack wetware, but it is not as common in developed western na93

tions as it is in developing countries. More profound bribery attempts than simple direct
cash payments usually align the financial interests of the target with those of the hacker.
Whereas technical intrusions and social engineering methods are the two attack
methods typically associated with hackers, a third category of attacks, the distribution of
malicious code, was also included in the questionnaire. This method of attacking computer systems is not necessarily limited to only hackers, but is oftentimes also used by
other cyber-criminals such as email phishers or spammers. Nevertheless, it was included in the questionnaire to measure hacker attempts of gaining control over target
computers by distributing malicious code. The answer categories given for the question
“Have you ever distributed?” were “Trojan Horses,” “Adware/Spyware,” “Viruses/Worms,” “Spam/Phishing mails,” “None of the above,” and “Other.” Readers unfamiliar with these types of malicious codes can find brief descriptions in the glossary accompanying this text.
The engagement in each of the three types of hacking methods was measured
with three items asking respondents to: (1) estimate how often they had attempted
these types of hacking methods in their life, (2) indicate the frequency with which they
engaged in those types of methods, and (3) provide an approximation of their success
with these methods. The answer categories provided for the frequency measurement
were attempts per weeks, months, or years. All answers were later recoded into days to
obtain a unified interval-scaled variable. The estimated success rates of technical intrusions and social methods were measured on scales of ten percent increments ranging
from zero to 100 percent. The success rate of malware distributions was not measurable this same way because the success of malicious code distributions primarily de94

pends on how many computers become infected with that code. Thus, the success with
malware distributions was measured with a seven-point Likert-type scale ranging from
1”Not at all” to 4 “Somewhat” to 7 “Very” successful.
Further examining the continuation of the respondents’ hacking careers as proposed by the rational choice perspective, the next set of items asked respondents: (1)
whether they changed friends and peers since their first hack to include other hackers;
(2) whether they believe they improved their skills since their first hacking attempts; (3)
whether their hacking activity has become more frequent since their first hacking attempts; and (4) whether their motivations had changed since their first hacking attempts. The three answer options given for these items were “Yes, very much,” “Yes,
somewhat,” and “No.” The item measuring the frequency of hacking attempts further
distinguished between the two answer options “No, it’s the same” and “No, it’s less frequent.” A last item in this set asked respondents to specify their currently predominant
motivation for hacking in an open format.
The respondents’ selection decisions of targets for hacking attacks were measured with five items. The first two items asked whether respondents had changed targets since their first attacks and whether their targets had become higher profile targets.
Again, the answer options “Yes, very much,” “Yes, somewhat,” and “No” were given.
The third item then detailed the type of currently preferred targets. Like the answer options given for the question about the target of the first hacking attempt, the answer categories provided for this item distinguished between single hosts, networks, and websites, and between private, corporate, non-profit, and governmental entities. Again, an

95

open answer category was provided to allow respondents to specify any other types of
targets.
The last two target-related items asked about the criteria influencing the selection
or rejection of a particular target. The criteria provided for the question “What are your
current selection criteria for targets?” were the same as for the item measuring the selection of the target in the first hacking attack: “Easy access,” “Interesting information,”
“Profitable information,” “Reputation gain,” “Antipathy,” and “Other (specify).” The criteria listed for the question “What might be a reason for you to reject a potential target?”
were “Unfamiliarity with architecture,” “No interesting information,” “No profitable information,” “Sympathy with host,” “None of the above,” and again “Other (specify).”
The set of questions following the target-related items interviewed participants
about various aspects of their hacking attacks. Respondents were asked whether they
had changed their methods and tactics since their first hack and what their currently
preferred methods and tools are for the different stages of a hack. For the item measuring the currently preferred methods, the six main stages of a hack “Reconnaissance,”
“Gaining access,” “Persisting,” “Propagating,” “Paralyzing,” and “Covering up” were
listed and respondents were instructed to enter their preferred methods and tools for
each of the main phases. They were then asked whether they follow a persistent pattern
in all their hacks or vary their methods and software tools between hacks. The answers
to both questions about the variability in methods and tools were recorded on sevenpoint Likert-type scales ranging from 1 “Very persistent pattern” and “Always same
tools” to 7 “Vary a lot.”

96

The section detailing the aspects of the respondents’ current hacking activity
then ended with an item measuring the methods they employed to remain anonymous.
The answer options provided for this question were “Spoofing,” “Bouncing,” “Source
routing,” “Proxies,” “TOR, I2P, etc.,” “Tunnels, Covert channels,” “War driving,” “Public
access points (e.g. Internet cafes),” and “Other (specify).” All eight anonymizing methods utilize different techniques to hide the offender’s true identity. Spoofing is a tactic
hackers commonly use to pretend they are somebody else. The disguise is accomplished by entering the wrong source IP addresses in the TCP/IP headers of data packets. Bouncing is a method that is similar to spoofing. In the case of bouncing, the hacker
uses proxy servers or computers to relay, or bounce, their requests to the attacked machines. Thereby, the impression is evoked that the attack originated from the proxy,
when in fact it was only channeled through it. In the context of computer networks, proxies denote gateways that relay one Internet session to another. Proxies cannot only be
used for bouncing techniques, but some proxy servers are even specifically designed
for anonymizing purposes. The use of such proxies makes the real source of data
packet untraceable. Source routing denotes a technique that allows a sender of data
packets to specify the route these packets take through a network. Source routing can
be used in hacking attacks to reach otherwise unreachable targets through intermediate
computers that are connected to the target. TOR (The Onion Router) and I2P (Invisible
Internet Project) are both open source projects that are similar in that they enable their
users to communicate anonymously on the Internet. Tunnels and covert channels, on
the other hand, are methods used by hackers to circumvent firewall protections and
evade intrusion detection systems. Lastly, war driving and public access points are two
97

methods to access the Internet anonymously through other persons’ private or public
networks. In the case of war driving, the hacker drives through an area in a moving vehicle with a WiFi-equipped computer and searches for open or weakly secured wireless
networks that are then used as Internet access points.
The last set of questions in this first part of the questionnaire examined possible
doubts the respondents had about their hacking activity and their readiness to desist
and end their hacking careers. The first item in this last set of questions asked respondents whether hacking is a source of income for them. The answer options given for this
item were “Yes, the main,” “Yes, but not the main,” and “No.” In addition to providing an
explicit measure of financial considerations, this item also enabled the researcher to
compare the relevance of financial considerations during the onset and the later stages
of the hacking activity. The item was followed by the direct question about whether the
respondent had ever thought about quitting hacking. The three answer options for this
item were "Yes, often,” “Yes, a few times,” and “Never.” An additional item was included
to measure the reasons for thoughts about quitting. The reasons listed for this item were
“Involved risk,” “Involved effort,” “Income too irregular,” “Fear of detection,” “Fear of apprehension,” “Fear of prosecution,” “Legal alternatives,” and “Other (specify).”
The first part of the questionnaire ended with seven items measuring the respondents’ estimation of risks involved in their hacking activity. The first risk-related item
asked them whether they are more or less afraid of being traced, tracked, or apprehended now than when they began to hack. The four answer options for this item were
“More afraid,” “Less afraid,” “Unchanged,” and “I was never afraid.” Respondents were
then asked to estimate the time they usually invested in preparation, planning, consid98

eration of routes of action, and the selection of techniques and tools. The answers were
recorded in hours, days, weeks, and months and were entered in the dataset in hours
with one day equating to eight hours. An additional measure of the development of
carefulness was provided by the item “Do you spend more or less time for planning today than when you started to hack?” The three answer options for this item were “More
time today,” “About the same time,” and “Less time today.”
The next risk-related items then asked respondents to estimate the risk they run
of being detected and of being apprehended during or after a hack. Both items were
recorded on seven-point Likert-type scales ranging from 1 “No risk at all” to 4 “Some
risk,” to 7 “Very high risk.” The second to last item in this section measured the percentage of attacks the respondents had aborted because they were afraid of being detected. The percentage of aborted attacks was again measured on a scale of ten percent increments between zero and 100 percent. The first section of the questionnaire
concluded with asking respondents whether they were more focused on potential rewards or potential risks during their hacks. The risk versus rewards focus question was
measured on a seven-point Likert scale ranging from 1 “Only on rewards” to 4 “Balance
both” to 7 “Only on risks.”

The Measurements of Risk Propensity and Rationality
The second part of the questionnaire consisted of items from different scales designed to assess the influence of three personality factors. In particular, the three scales
measured the respondents’ risk propensity as well as their preference for rational versus heuristic or experiential-based thinking styles. In an attempt to limit the overall
99

length of the survey, each of the three scales was abbreviated to five items. All items in
this second section were worded as general statements, and respondents were instructed to indicate how much they agree or disagree with these statements on sevenpoint scales ranging from 1 “Strongly disagree” to 7 “Strongly agree.”
The decision to measure general risk propensity with a scale instead of a single
item was based on findings by MacCrimmon and Wehrung (1990), who reported that
the concept of risk propensity is too broad to be accurately captured with a single item.
The five items measuring risk propensity were taken from different scales and slightly
modified for best thematic fit. The first item “I always try to avoid situations involving a
risk of getting into trouble” was modified from a scale developed by Dahlback (1990).
The second item, “I always play it safe even when it means occasionally losing out on a
good opportunity,” was adapted from Gomez-Mejia and Balkin’s (1989) “willingness to
take risks” scale, which is an advancement of the original scale developed by Slovic
(1972) and the modifications introduced by Gupta and Govindarajan (1984). The remaining three items were taken from Dulebohn (2002), who developed them to measure
general risk propensity and who reported a Cronbach alpha of .73 for this three-item
scale. The second item “I am rather bold and fearless in my actions” was reversed to
prevent biases introduced by “acquiescence” response strategies of participants who
give superficial answers because they want to get through questions quickly (Krosnick &
Fabrigar, 1997).
Two other scales were included to assess the degree to which respondents generally rely on their rationality versus their intuition when making decisions. All items in
the rationality and the faith in intuition scales were taken from the latest version of the
100

Rational-Experiential Inventory (REI) scale (Pacini & Epstein, 1999). The REI is a well
established and supported measurement instrument for rational versus heuristic thinking styles. Today, considerable evidence exists in the social psychological literature for
its construct validity (Epstein, 2003; Epstein, Pacini, Denes-Raj, & Heier, 1996; Handley, Newstead, & Wright, 2000; Pacini & Epstein, 1999).
The full version of the REI consists of 40 items in two main scales measuring the
preference for analytical-rational or intuitive-experiential information processing. Each of
the main scales is further divided into subscales of self-assessed effectiveness and of
engagement in both thinking styles. More precisely, rational effectiveness refers to the
confidence persons have in their logical reasoning, whereas rational frequency or engagement refers to the pleasure derived from rational thinking (Handley et al., 2000).
Conversely, experiential ability measures the confidence in relying on personal intuitions
and experiential engagement measures the enjoyment of using intuition as the basis of
one’s decision making. The internal consistency reliabilities are reported with .87-.90 for
the two REI scales and .79-.84 for the four subscales (Epstein, 2003). The full version of
the REI scale was abbreviated in the survey. The questionnaire contained five items
from each of the two REI scales. Three of the five items in each scale were taken from
the ability subscales and two from the engagement subscale. The selection of items followed the criteria that were outlined earlier in this chapter.

The Measurements of Sociodemographics
The questionnaire concluded with some measurements of basic sociodemographic characteristics in the third section. Respondents were first asked to indicate
101

their sex and to enter their year of birth. They were then asked to identify the last grade
they completed. The answer categories provided for this question were “None, or
grades 1-8,” “High school incomplete (grades 9-11),” “High school graduate (grade 12
or GED certificate,” “Business, technical, or vocational school AFTER high school,”
“Some college, no 4-year degree,” “College graduate (B.S., B.A., or other 4-year degree) ,” and “Post-graduate training/professional school after college (Master’s degree
or Ph.D.).”
The racial and ethnic background of the participants was assessed with two
items. The first item asked whether the respondents are “of Hispanic or Latino origin or
descent, such as Mexican, Puerto Rican, Cuban, or some other Spanish background.”
The second item then asked them to indicate their race. The answer categories provided for this item were “White,” “Black,” “Asian,” and “Other mixed race: (specify).” The
marital status of respondents was measured with the question “Are you married, living
as married, divorced, separated, widowed, or never been married.” The last two questions on the survey asked “Are you now employed full-time, part-time, retired, or not
employed for pay?” and “Are you also a full- or part-time student.” The survey concluded with thanking the respondents for their participation and wishing them a nice day
at the conference.

102

CHAPTER EIGHT:
METHODS AND RESULTS
The analysis conducted in this research project addressed all the study goals
that were outlined in the previous chapters. Aside from examining the distributions of
sociodemographic characteristics within the hacking community, it also specified the rational choice perspective models for the continued involvement in hacking by measuring
the shifting importance of various motives as well as several other hacking-related aspects. Furthermore, it tested all hypotheses raised in Chapter 5 by investigating the effects of varying degrees of rationality and risk propensity on hacking-related decisions.
Given the exploratory nature of the research project, the first section describes
the distributions of sociodemographic characteristics in a univariate analysis. Descriptive statistics are also used to identify the relative importance of several hacking-related
factors. In the second section, the validity and reliability of the three personality scales
are briefly examined, and the scales are transformed into indices. These indices are
then used to test the hypotheses from Chapter 5 in several additive multivariate linear
regressions.

The Sociodemographic Composition of the Sample
The sociodemographic characteristics displayed in Table 1 show a vastly skewed
gender distribution among hackers. Only seven of the 124 participants (5.6 percent)
were females. The wide gender gap revealed in this study confirms other reports that
describe hacker communities as being predominantly male (Adam, 2004; Taylor, 1999).
103

Table 1: Sociodemographic Characteristics of Sample Respondents
Variable

N1

%2

Sex
Male
Female

117
7

94.4
5.6

Age3

120

30.6 (6.7)

Education
None, or grades 1-8
High school incomplete
High school graduate
Vocational school
Some college
College graduate
Post-graduate Master’s or Ph.D.

0
4
7
2
30
47
34

0.0
3.2
5.6
1.6
24.2
37.9
27.4

Race
Hispanic descent
White
Black
Asian
Other

3
116
2
5
1

2.4
93.5
1.6
4.0
0.8

Marriage status
Never married
Living as married
Married
Divorced

63
17
43
1

50.8
13.7
34.7
0.8

Employment
Full-time
Part-time
Unemployed

92
22
10

74.2
17.7
8.1

Student status
Yes, full-time
Yes, part-time
Not a student

14
31
79

11.3
25.0
63.7

Actively hacking
Yes
No

97
27

78.2
21.8

4

1

The total sample size is n=124.
Percentages may not add up due to rounding.
3
Measured in years, means reported (std. dev. in parentheses).
4
Some categories were excluded because they yielded no results.
2

104

The general underrepresentation of women in all areas related to computing and
information technology—except in office or administrative positions—has already received considerable scrutiny in the literature (Webster, 1996). Against this background,
the vast domination of males in the hacking community is not surprising. However, the
gender difference in this study exceeded even the discrepancies found in other areas of
computing and information technology, in which women are estimated to account for 10
to 30 percent of participants (Zarrett & Malanchuk, 2005).
Taylor traces the absence of women in the hacking community, which he finds to
be an “unexplained statistic,” (Taylor, 1999, p. 32) back to what he sees as the fundamentally masculine nature of hacking. He describes the hacking culture as young, male,
technology-oriented, and laden with factors that discourage women from joining. Among
the factors listed by Taylor are social stereotyping, a masculine “locker room” environment, and a gender biased computing language (Taylor, 1999, p. 36). Adam goes one
step further by describing the hacker culture as one that, despite the explicit egalitarianism expressed in the Hacker Ethic, is nevertheless characterized by a “frontier masculinity,” a “Wild West brand of masculinity,” and a deeply rooted misogyny displayed by
men who hide behind the anonymity of the Internet and associate “technology with desire, eroticism and artificial creation” (Adam, 2004, p. 6). The data collected in the
present study confirmed the existence of a substantial gender gap, but it did not include
any additional attitude measurements with regard to gender. Hence, it is not possible to
confirm or reject any of the explanations mentioned above. Nevertheless, some of the
elements identified by Adam and Taylor were clearly visible during the convention. For

105

example, some attendees wore T-Shirts with labels such as “Penetration Expert” or other print designs that connected technology with expressions of male sexuality.
Aside from the large gender gap, the data also displays a heavily skewed race
distribution. Over 93 percent of the hackers in the sample were White, a fraction that
vastly exceeded the percentage of Whites in the U.S. population. Another noteworthy
finding in the race distribution is the fact that Asians were the largest minority in the
sample. While the low cell count of all minorities in the present study did not permit accurate generalizations of this finding, this result again reflects the racial distributions in
most IT professions (Zarrett & Malanchuk, 2005). A common explanation for this finding
is the particular prevalence of positive attitudes toward math, science, and computerrelated occupations among Whites and many Asian cultures (Bement, Ward, Carlson,
Frase, & Fecso, 2004).
The age distribution of the convention attendees shows a much higher mean value than the one suggested by the common notion of the prototypical hacker as a juvenile delinquent teenager (Yar, 2005a). It is reasonable to assume that the higher average age in this study of ShmooCon convention attendees was caused by the sampling
frame of this particular research project. The profile of the ShmooCon convention was
geared more toward security experts and computer professionals than to teenagers who
pursue their hacking interests merely as a leisure-time hobby. Thus, while the distribution in this particular sample is certainly not enough evidence to falsify any claims that
the majority of hackers are teenagers, nevertheless, it indicates that the hacking community is by no means limited to only teenagers. To the contrary, it involves many mature security experts and many seasoned hackers who pursue their hacking activity in a
106

professional manner. The data clearly show that hacking is not just a ‘young man’s
game.’ The oldest active hacker in the sample was 52 years old and reported to have
been hacking for close to three decades.
The professionalism of most respondents was also reflected in their educational
attainments. Ninety percent of the hackers in the present sample had at least some college education, and about one-fourth of them reported to have obtained a Master’s or
Ph.D. degree. Moreover, about one-third of all respondents were enrolled either as fulltime or part-time students. An examination of the four cases with an incomplete high
school education revealed that most of them were young participants (between 18 and
19 years old) who also reported to be full-time students. These four cases were most
likely high school students who had not yet graduated.
The high fraction of students in the survey sample is particularly surprising when
considering the fact that over 90 percent of all respondents were also employed. About
three-fourth reported to be employed full-time and an additional 18 percent were parttime employees. The high employment rate was probably part of the reason why more
than double as many respondents indicated to be part-time students than full-time students. When asked about their marital status, about half of all respondents said that
they were never married. A significantly smaller fraction, only about one-third of all participants, reported to be married.
In short, the sociodemographic characteristics that were sampled in this study
paint the picture of a hacking community that is predominantly male, White, and comprised of highly educated members. They also work regular jobs and are oftentimes also studying but appear to be hesitant to engage in serious relationship commitments.
107

The Descriptive Statistics of Continued Involvement
The control item at the beginning of the questionnaire asked participants whether
they “are currently actively hacking.” Almost 80 percent of the respondents answered
this question with “yes.” A brief examination of the remaining 20 percent revealed that
most of them were very active former hackers who had given up their hacking activities
in the meantime. The result obtained from this control item allows the conclusion that
the goal of the present study, which was to sample persons with actual hacking experiences and not just persons who are merely interested in computer-security related issues, was indeed accomplished.

The Initiation Phase
Some of the most interesting questions about the initiation phase of hacking are
(1) what exactly sparked the initial interest in hacking, (2) what led hackers to commit
their first actual hacking attempt, and (3) at what age did they do so. All three questions
relate to main components of the initiation model as proposed by the rational choice
perspective, and all three questions were addressed in the present study. The results
show that many hackers became interested in hacking even before their early teenage
years. One person reported that he was only nine years old when he first became interested in hacking. While this respondent was the youngest, he was no exception. Table
2 shows that the first peak in the initial interest distribution was as early as 12 years.
Twenty percent of all respondents reported to have already been interested in hacking
by that age. The median of the distribution was at 15 years, and the mean at 16 years.
108

Table 2: Motivations for Interest in Hacking and First Hack
N1

Variable

%2

Age interested in hacking3

124

16.0 (4.3)

Motive for initial interest4
Intellectual curiosity
Experimentation
Excitement, thrill, fun
Feeling of power
Peer recognition
Self-concept boost
Status and prestige
Personal revenge
Other
Political ideology
Protest against corporations
Financial gain
Media attention

118
105
82
26
23
22
19
12
7
6
4
3
2

95.2
84.7
66.1
21.0
18.5
17.7
15.3
9.7
5.6
4.8
3.2
2.4
1.6

Primary motive for interest4
Intellectual curiosity
Experimentation
Excitement, thrill, fun
Feeling of power
Other
Self-concept boost
Political ideology
Peer recognition
Personal revenge

74
21
15
4
4
2
2
1
1

59.7
16.9
12.1
3.2
3.2
1.6
1.6
0.8
0.8

Motive for first hack4
Intellectual curiosity
Experimentation
Excitement, thrill, fun
Feeling of power
Peer recognition
Self-concept boost
Status and prestige
Personal revenge
Other
Protest against corporations
Financial gain

91
84
56
13
10
10
4
6
3
2
2

73.4
67.7
45.2
10.5
8.1
8.1
3.2
4.8
2.4
1.6
1.6

1

The total sample size is n=124.
Percentages may not add up due to rounding.
3
Measured in years, means reported (std. dev. in parentheses).
4
For better readability, the motives are rank ordered by importance.
2

109

The self-reported motives for the initial interest in hacking show that the majority
of participants became interested in hacking because of “intellectual curiosity” (95 percent), “experimentation” (85 percent), and “excitement, thrill, or fun” (66 percent). A
second set of motives revolving around self-expression and peer-recognition turned out
to be of significantly lesser importance. Among these motives were “feeling of power”
(21 percent), “peer recognition” (19 percent), “self-concept boost” (18 percent), “status
and prestige” (15 percent), and “personal revenge” (10 percent).
Some of the motives that are oftentimes associated with hackers in media reports (Alexander, 2005) as well as scientific (Grabosky & Smith, 1998; Kilger, Arkin, &
Stutzman, 2004) and governmental publications (Krone, 2005), on the other hand,
played only a marginal role as initial interests. Among these motives were “political ideology” (5 percent), “protest against corporations” (3 percent), “financial gain” (2 percent),
and “media attention” (2 percent). The results clearly demonstrate that motives associated with youth, boredom, frivolity, mischief, or curiosity are the main reasons for
young persons to become initially interested in hacking. In contrast, only a very few respondents became interested in hacking because of political or financial considerations,
or other motives with a stronger criminal intent.
A similar pattern emerged from the question about the single most important motive for the initial interest. Here, roughly four times more respondents (60 percent) answered with “intellectual curiosity” than with the next popular answer option “experimentation” (17 percent). “Media attention”, “financial gain”, “protest against corporations”,
and “status and prestige,” on the other hand, were not mentioned at all and were excluded from Table 2.
110

Only five “Other” reasons were specified. Out of those, the desire to spy on a
girlfriend, who the respondent believed to be cheating, was named twice. The other
reasons were independence, learning of security, and playing pranks on friends. Overall, the few individual reasons that were given in addition to the list of standard answer
options suggest that the list was comprehensive. The only item that should probably be
included in the theoretical model and future measurements was spying.
The separate measure of the motives for the first actual hack produced roughly
the same results as the item measuring the motives for the initial interest. The main difference between the two items was that the reasons for the first actual hack were more
specific than for the initial interest. Accordingly, most respondents marked fewer motives, which resulted in lower percentages for all motives. The patterns between the different motives, however, were very similar to the ones emerging from the question
about initial interests. Two noteworthy findings in the distribution of motives for the first
hack were that “political ideology” was not mentioned by any respondent and that “financial gain” was a motive for only two respondents.
Aside from the motivations for the initial interest and the first actual hack, the survey also measured the length of the time span between these two events. The time
measure was recorded in days in the dataset but is presented in more meaningful categories in Table 3. Interestingly, about one-third of all respondents committed their first
hacking attempt already within the first week of becoming interested in hacking. Also, an
additional 20 percent committed their first hack within the first month of becoming interested. This finding suggests that the initial interest of many hackers is not an abstract,
intellectual enterprise, but a preparation for their first hack. It indicates that the initial in111

terest is guided by the intent to actually launch hacking attacks. Less than 50 percent
were interested in hacking longer than a month before they actually attempted a hack.
The longest reported time span in the sample was clearly an outlier (10 years). Table 3
displays the recorded time spans between initial interests and first hacks.

Table 3: Details of the First Hacking Attempt
Variable

N1

%2

45
23
31
25

36.3
18.5
25.0
20.2

Time span between interest and hack3
Up to 1 week
Up to 1 month
Up to 1 year
2 to 10 years

1st target owner / type
Private
Corporate
Non-profit
Government

N

(%)

N

Single host
50 (40.3)
5 (4.0)
0
1 (0.8)

(%)

Network
29 (23.4)
7 (5.6)
4 (3.2)
1 (0.8)

N

Website
4 (3.2)
3 (2.4)
1 (0.8)
0

1st target selection criteria
Easy access
Interesting information
Profitable information
Reputation gain
Antipathy
Other

70
36
0
0
7
11

56.5
29.0

28
28
68

22.6
22.6
54.8

0
5
119

4.0
96.0

5.6
8.9

Employed when 1st hacked
Yes, full-time
Yes, part-time
No

Economic profit a motive at all
Yes, an important one
Yes, but not very important
No
1

The total sample size is n=124.
Percentages may not add up due to rounding.
3
Categories are not cumulative.
2

112

(%)

Table 3 further shows that the most popular targets of the first hack were single
private computer hosts (40 percent) and private networks (23 percent). Corporate computers and networks were the second-most popular targets (4 percent and 6 percent).
With regard to corporate targets, the relationship between single hosts and networks
was reversed. More corporate networks were attacked than single computers. This difference is probably due to accessibility reasons. While many single private hosts can be
located in unprotected wireless networks or public networks, an attack on corporate
computers typically requires a preceding attack on the network in which the computer is
located. Only one hacker selected a government host and network as the targets for his
first attack. For all others, these targets were probably too risky and too high profile to
be considered as a first target.
Most hackers selected their first target based on practical considerations. The
majority of participants (57 percent) reported that the ease of gaining access was their
primary selection criteria. About half as many chose a particular target because it offered interesting information (29 percent). Revenge or antipathy with the host played
only a minor role as selection criteria. Only seven respondents attacked their targets
because of personal dislike (6 percent). Some specifications of answers in the “Other”
category (9 percent) revealed that some respondents counted attempts to hack their
own computer system or network as their first hacking attempt. Future survey designs
will need to be more explicit to rule out this interpretation of the question.
The answers to the selection criteria question again confirmed the irrelevance of
commonly assumed motives in the initiation phase. None of the respondents attacked
their targets in search for profitable information or because they were particularly suited
113

for gaining a reputation as a hacker. The finding that financial interests played hardly
any role during the onset of hacking activity was also confirmed by the answers to the
explicit question asking whether economic profits were a motive at all. Only five respondents (4 percent) said it played a minor role. The vast majority (94 percent), on the other
hand, indicated that economic considerations or potential financial gains had nothing to
do with their decision to start hacking.
The finding that a majority of respondents (55 percent) were unemployed when
they first hacked is not surprising given the young age of most respondents when they
started to hack. During their first hacks, most of them were still dependent teenagers
with little or no income of their own. Despite little or no income, however, it is important
to note that economic interests hardly played any role in the decision to engage in hacking activities.

The Popularity of Different Hacking Methods
Table 4 shows the popularity of the 24 different hacking attacks that were listed
in the questionnaire. When interpreting the results, it is important to bear in mind that
the corresponding item asked respondents only about attacks they had used in their first
hack. Hence, the frequencies in Table 4 relate to only one single hacking event, not all
hacking attempts over the course of a career. Despite the fact that the question referred
to merely one single event, some of the attack methods were used by almost half of all
respondents. The most popular methods were packet sniffing (49 percent), port scanning (47 percent), buffer overflows (47 percent), and network mapping (44 percent). Of
the 24 answer options, bouncing was the only method that had not been applied by an114

ybody in the sample. The high frequency turnout of most items in the list, however, was
a strong indication that the listing of methods in this study was appropriate.
The fact that only a few specialized methods were specified as answers in the
“Other” category further suggests that the list was close to being exhaustive. Most of the
methods specified in the “Other” category were antique methods that were popular in
the 1980s and 1990s but have since become obsolete. The implication that the provision of standard answer options for this question had the disadvantage of systematically
excluding outdated methods that older hackers used in their first attacks was reaffirmed
by the verbal feedback from some of the older hackers. Nevertheless, the provision of
standard answer categories for this question produced much better results than the
relatively unsuccessful open-ended measurement technique that was used for the question about currently preferred methods and tools. Most respondents were hesitant to reveal their methods or tools and either skipped the question, or gave nondescript answers such as “custom tools.” By way of comparison, they were much more willing to
mark already provided answer categories. The differential success of the two measurement techniques suggests that future surveys should operate with standard answer categories for questions about preferred hacking methods because this question format
minimizes the reluctance of many respondents to reveal this information.
The results shown in Table 4 further suggest that the description of most hacks
as complex events with different phases is correct. Many hackers combined different
reconnaissance methods with different intrusion and cover-up techniques. Two of the
most popular reconnaissance methods were network mapping and port scanning. Oftentimes, these two methods were used in combination. After hackers had created a
115

map of the network and had located the target computer within that network, they then
searched for vulnerable open ports as entry points to that computer system. Of the different methods to gain access to a system, the various techniques to obtain passwords
were the most frequently used. These results suggest that the classic exploitation of
password weaknesses remains popular among many hackers.

Table 4: Methods Used in First Hacking Attack
N1

%2

34
29
7
36
54
58
29
45
2
61
18
14
38
36
14
58
5
2
4
9
29
14
0
4
26

27.6
23.2
5.8
29.0
43.5
46.5
23.2
36.3
1.5
49.4
14.5
11.6
30.5
29.0
11.6
46.5
4.4
1.5
2.9
7.3
23.2
11.6

Variable
Methods used in 1st technical intrusion
Footprinting
Ping sweeping
DNS zone transfer
Whois
Network mapping
Port scanning
Versatile scanning tools
Vulnerability scanning
RPC port/end-point dump
Packet sniffing
Session hijacking
Grinding Passwords
Password cracking
Password theft
Directory traversal/climbing
Buffer overflow
Format string
Resource mismatch
CGI
Root shell/kits
Keylogger
Spoofing
Bouncing
Source routing
Other

Average number of methods (std. dev.)

5.0 (19.5)

1

The total sample size is n=124.
Percentages may not add up due to rounding.
3
Categories are not cumulative.
2

116

2.9
21.0

The Prevalence and Success Rates of Different Kinds of Attack Methods
Table 5, 6, and 7 list the prevalence and success rates for three different types of
hacking attacks. While many persons think of hacking attacks as performed solely
through technical means and exploits, they are in fact more diverse and oftentimes involve a combination of technical methods, social methods, and circulations of different
kinds of malicious code (J. Erickson, 2008). To gain a clearer picture of the prevalence
of each of the three types of attacks and to obtain a better understanding of the composition of typical hacking attacks, all three types of attacks were assessed independently.

Table 5: Measurements of Hacking Activity – Technical Intrusions
N1

%2

27
16
20
11
25
25

21.8
12.9
16.1
8.9
20.2
20.2

26
30
12
17
22
17

21.0
24.2
9.7
13.7
17.7
13.7

124

48.1 (2.5)

Variable
Total number of technical intrusions
Up to 5
6-10
11-50
51-100
101-500
501-20,000

Frequency of technical intrusions (per year)
1
2-6
7-12
13-36
37-156
157-2,500

Success rate in % (st. dev.)
1
2

The total sample size is n=124.
Percentages may not add up due to rounding.

Table 5 shows a bimodal distribution for the number of technical intrusions. Most
hackers were either very engaged or hardly active at all. The four most active hackers in
117

the sample each reported to have committed an estimated total of 20,000 hacking attacks over the course of their careers. A total of 20 percent of all respondents in the
sample indicated having hacked between 500 and 20,000 times and another 20 percent
said they had hacked between 100 and 500 times. The results in Table 5 clearly show
that the convention was attended by many veteran hackers.
The frequency measure of technical intrusions per year further confirmed that
many of the interviewed hackers were regularly hacking at the time the survey was conducted. A total of 30 percent of all respondents reported having attempted from three
hacks per month to an astonishing seven hacks per day. Obviously, a number of seven
hacking attempts per day can only be achieved when the single hacking attempts are
committed through highly automated attack routines and without any sophisticated
preparation routines. An investigation into the cases with extremely high frequencies
and numbers of total hacks revealed that they did indeed report very short preparation
times and many of them also reported low to moderate success rates. Overall, the success rate reported by all respondents showed that they considered about half (48 percent) of all technical intrusions to have been successful.
A 50 percent success rate of all technical intrusions was already very high, but,
as Table 6 shows, the success rate of social methods was even higher (62 percent).
The very high success rate of social methods was one of the most surprising findings in
this study. It demonstrates that the popular image of hackers as social hermits who
launch their hacking attacks solely through remote computer and network technology, or
even do so mainly to compensate for social deficits, has to be revised. The opposite

118

seems to be the case. Hackers seem to be very socially intelligent persons who know
how to successfully manipulate and trick other persons.
Table 6: Measurements of Hacking Activity – Social Methods
N1

%2

94
81
51
46
36
18

75.8
65.3
41.1
37.1
29.0
14.5

17
24
20
14
11
21
17

13.7
19.4
16.1
11.3
8.9
16.9
13.7

18
31
23
7
9
25
11

14.5
25.0
18.5
5.6
7.3
20.2
8.9

107

71.8(2.4)

Variable
Types of social methods
Social engineering
Shoulder surfing
Dumpster diving
Impersonation
Deception
Bribery

Total number of social methods
0
1-5
6-10
11-50
51-100
101-500
501-5,000

Frequency of social methods (per year)
0
1
2-6
7-12
13-36
37-156
157-365

Success rate in % (st. dev.)
1
2

The total sample size is n=124.
Percentages may not add up due to rounding.

More detailed case examinations of the hackers who reported the highest success rates with their social hacks revealed that they were the same hackers who also
reported to have the most success with their technical intrusions. Moreover, most of
these very successful hackers gave almost identical numbers and frequencies for the
119

two methods. Based on these observations, one can conclude that the most successful
hackers are the ones who combine social and technical approaches in their hacks.
All six of the listed social attack methods shown in Table 6 were used by an unexpectedly high number of hackers. The most popular type of the social attack methods
was social engineering. Three-fourth of all respondents reported to have used social
engineering methods in their hacks. Surprisingly, shoulder surfing was the second-most,
and dumpster diving the third-most popular method. The large popularity of these two
simple methods (65 and 41 percent) shows that hacking attacks do not always have to
be sophisticated attacks and require a lot of knowledge, preparation, and expertise.
They can be as simple as watching other computer users type in their login passwords
or searching their garbage for sensitive information that can then be exploited. The two
more sophisticated social methods, impersonation and deception, were used by fewer
hackers than the more simple methods, but they were still considerably popular. Thirtyseven percent of all respondents reported to have used impersonations, and 29 percent
indicated that they have engaged in deceptions to trick persons for hacking purposes.
Bribery was the least common technique of all social methods. Still, a total of 15 percent
of all respondents had employed this method for their hacks.
The unexpected popularity of social methods was corroborated by the two items
measuring the total number of social hacks and the frequency with which they were
employed. Even though the highest total number of social hacks was considerably lower
than that for technical intrusions, 30 percent of all respondents still reported to have
conducted more than 100 such attacks over the course of their careers. Only 14 percent
of all participants said that they had never engaged in any such methods. Another sur120

prising finding was that, even though the highest frequency with which social methods
were applied (1 per day) was again lower than the highest frequency of technical intrusions (7 per day), the percentage of hackers who reported to engage in three social attacks per month or more was roughly the same as for technical intrusions (29 versus 30
percent).

Table 7: Measurements of Hacking Activity – Malware Distribution
N1

%2

20
0
11
9
15
87

16.1
8.9
7.3
12.1
70.2

87
22
4
2
5
4

70.2
19.4
3.2
1.6
4.0
3.2

87
24
3
8
0
2

70.2
18.5
2.4
6.4

37

=3.6 (0.9)

Variable
Types of distributed malware
Trojan Horses
Adware/Spyware
Viruses/Worms
Spam/Phishing mails
Other
None

Total number of distributed malware
0
1-5
6-10
11-50
51-100
101-1,000

Frequency of distributed malware (per year)
0
1
2-6
7-12
13-36
37-200

Success rate scale 1-7 (std. dev.)
1
2

The total sample size is n=124.
Percentages may not add up due to rounding.

121

1.6

In addition to technical intrusions and social methods, the survey also asked respondents about types of malicious code that they had ever distributed. Table 7 shows
the descriptive statistics for the distribution of malicious code. Different from the large
popularity of the first two methods, distributions of malicious code were relatively uncommon, and the type of malicious codes that had been distributed was highly diversified. Seventy percent of all hackers in the sample reported never having distributed any
type of malicious codes. The types of malware that were most often used were Trojan
horses (16 percent) and viruses (11 percent). The 15 persons who marked “Other”
forms of malware most often specified that they designed their own scripts and were
circulating custom-made program codes or other forms of highly specialized code such
as SQL injections.
A noteworthy finding was the unpopularity of spam or phishing emails (7 percent)
and adware or spyware (0 percent). The results in Table 7 suggest that hackers and
phishers or spammers are two separate types of cyber-criminals. This conjecture was
reaffirmed by additional case analyses of active distributors of malware. The respondents who most often circulated malicious code rarely engaged in technical or social
hacking methods. Thus, while classic hackers oftentimes combined technical intrusions
and social engineering methods, or at least engaged frequently in both of them, they
were distinctively different from the persons who circulated the most malicious code.
The latter appeared to have a hacking profile that is distinctively different from the former two.
Since the success rate of malware distributions is not simply determined by any
definite success versus failure ratio but by the total numbers of infected computers, the
122

success rate of malware distributions was measured on a seven-point Likert type scale.
The results of this measure suggest that hackers who distributed malware were estimated to have a medium level of success with their efforts ( =3.6).

The Developments During the Habituation of Hacking
The lengths of hacking careers in Table 8 reaffirmed the considerable experience
of most hackers in the present sample. The normal-shaped distribution of hacking experiences ranged from less than a year to 28 years and averaged at ten years. The length
of most hacking careers in the present sample was a clear indication that the majority of
respondents were not beginners, but had already habitualized their hacking activities.
As was predicted by the rational choice perspective, most respondents befriended other hackers during their time as active hackers. Seventy-five percent of all
respondents said they had changed their social networks to include other hackers, and
23 percent did so “very much.” Alongside the changes in their networks, most hackers
also reported changes in their motives, their engagement in hacking, and their skills.
Only one respondent said that he had not improved his hacking skills since he began
hacking. As expected, this particular hacker was one of the least experienced in the
sample. He had less than one year of hacking experience and had committed only one
single hack. All other respondents claimed to have improved their skills over the course
of their careers, and 72 percent said they did so “very much.”

123

Table 8: Developments during Hacking Career
N1

%2

8
30
47
20
10
9

6.5
24.2
37.9
16.1
8.1
7.3

28
66
30

22.6
53.2
24.2

89
34
1

71.8
27.4
0.8

37
39
18
30

29.8
31.5
14.5
24.2

38
37
49

30.6
29.8
39.5

Variable
Time hacking (in years)
Up to 1
2-5
6-10
10-15
16-20
20-28

Change friends (more hackers)
Yes, very much
Yes, somewhat
No

Improved skills
Yes, very much
Yes, somewhat
No

Hacking more frequent
Yes, very much
Yes, somewhat
No, it’s the same
No, it’s less frequent

Motives changed
Yes, very much
Yes, somewhat
No

Current primary motive (initial interest)
Intellectual curiosity
Financial gain
Experimentation
Other
Excitement, thrill, fun
Self-concept boost
Feeling of power
Political ideology
Peer recognition
Personal revenge
1
2

37
28
22
21
14
2
0
0
0
0

The total sample size is n=124.
Percentages may not add up due to rounding.

124

(74)
(0)
(21)
(4)
(15)
(2)
(4)
(2)
(1)
(1)

29.8 (59.7)
22.6
17.7 (16.9)
16.9 (3.2)
11.3 (12.1)
1.6 (1.6)
(3.2)
(1.6)
(0.8)
(0.8)

A majority of respondents reported that their hacking activities had intensified
over the course of their careers. Of all hackers in the sample, 30 percent said they are
hacking much more frequently now than when they started, and 32 percent reported
that their hacking activities have somewhat increased. Only 30 respondents (24 percent) said their hacking activities had become less frequent. Of those, 27 also said that
they are no longer actively hacking. Thus, only 3 active hackers had decreased their
hacking frequency and 60 percent of the still active hackers had increased it. The data
shows an apparent trend toward an intensification of hacking activities over time.
Sixty percent of respondents further indicated that their motives had changed
since their initial interest in hacking. Indeed, the comparison of initial motives with current ones revealed three dramatic changes that had occurred between the two measures. First, the importance of intellectual curiosity as the primary motive had decreased
by 50 percent over time (from 60 to 30 percent). Second, financial gain, a motive of no
importance for the initial interest, had become the second-most important motive for
hacking. Twenty-three percent of all subjects said that their main motives for continuing
to hack were financial gains. The sharp increase of financial gains as motives for hacking is an intriguing finding. It means that, while most hackers set out to become hackers
because they are curious about the technology and keen to experiment with it, along the
way some of them realize the financial possibilities that are achievable through their engagement in hacking. The third main difference between the two measures is the reduction of motives. While the list of initial motives included ten different motives, this list
was reduced to only six persistent motives. Feelings of power, political ideology, peer
recognition, and personal revenge no longer played a role for the continued engage125

ment in hacking. The changes in motives demonstrate that for many, hacking efforts
evolved into a professional business. This trend was also reflected in the “Other” category. Most entries in this category pertained to the gathering of sensitive and securityrelated information.
The changes in motives were mirrored in the changes that occurred in the preferences for certain targets. As Table 9 illustrates, 71 percent of all respondents reported
to have changed their targets over the course of their careers. Also, 50 percent said
they are now attacking higher profile targets and 86 percent reported to have changed
their methods and tools to attack the different kinds of targets.
The increased preference of many hackers for higher profile targets was visibly
reflected in their preferred types of targets. Both corporate and governmental targets
were attacked much more frequently. The preference for corporate computers quadrupled (from 4 to 17 percent), the preference for corporate networks septupled (from 6
to 40 percent), and the preference for corporate websites increased twelvefold (from 2.4
to 28 percent). Similarly, governmental targets, which were virtually not targeted during
the onset of the hacking activity, were much more popular among experienced hackers.
Fifteen percent reported having attacked governmental hosts, 25 percent attacked governmental networks, and 20 percent targeted governmental websites.

126

Table 9: Target Preferences
N1

%2

52
36
36

41.9
29.0
29.0

29
34
61

23.4
27.4
49.2

Variable
Targets changed since 1st hack
Yes, very much
Yes, somewhat
No

Higher profile targets
Yes, very much
Yes, somewhat
No

Current target owner / typeN (%)3

N (%)3

N (%)3

Single host
49 (39.5)
21 (16.9)
4 (3.2)
18 (14.5)

Network
56 (45.2)
49 (39.5)
4 (3.2)
31 (25.0)

Website
23 (18.5)
35 (28.2)
7 (5.6)
25 (20.2)

Private
Corporate
Non-profit
Government

Current target selection criteria (initial criteria)
Easy access
Interesting information
Profitable information
Reputation gain
Antipathy
Other

58
87
31
2
2
11

(70)
(36)
(0)
(0)
(7)
(11)

46.8 (56.5)
70.2 (29.0)
25.0
1.6
1.6 (5.6)
8.9 (8.9)

Rejection reasons
No interesting information
Unfamiliarity with architecture
Sympathy with host
No profitable information
Other
None of the above

60
48
23
19
9
30

48.4
38.7
18.5
15.3
7.3
24.2

51
37
14

50.0
36.3
13.7

Variability of methods (scale 1-7)

123

4.7 (1.7)

Variability of tools (scale 1-7)

123

3.9 (1.7)

Change in methods and tactics
Yes, very much
Yes, somewhat
No

1

The total sample size is n=124.
Percentages may not add up due to rounding.
3
Multiple answers were possible. % values refer to complete sample.
2

127

The selection criteria for targets changed in accordance with the motives and the
targets. The prospect of obtaining profitable information, initially irrelevant during the
onset of hacking activities, had become the third-most important criterion. Twenty-five
percent of all respondents said this criterion was relevant for their selection of targets.
The significantly increased importance of profitable information further confirmed the
trend toward a professionalization of illegal activities. Easy access remained the most
important criterion, but its significance was notably reduced (from 57 to 47 percent). Following an opposite trend, the prospect of interesting information had vastly gained importance. More than double as many hackers listed interesting information as one of
their selection criteria (from 29 to 70 percent).
Among rejection criteria, the absence of interesting information was the most frequent one. Almost half of all participants (48 percent) listed it as a reason to refrain from
an attack. Unfamiliarity with the architecture of a computer system or network was the
second-most common reason for a rejection (39 percent), followed by sympathy with
the host of that system or network (19 percent). Analogous to its importance as a selection criterion, 15 percent of all hackers in the sample marked the absence of profitable
information as a reason for rejecting a particular target. This result underlines the profound change many hackers undergo over the course of their hacking careers. Hackers
apparently become more professional and many of them begin to see hacking not only
as an intellectual challenge, but as a potential source of income.
Table 10 illustrates that the question asking respondents whether hacking for
them is a source of income was positively confirmed by 41 percent. Fifteen percent
even identified hacking as their main source of income. An additional case analysis
128

found that these 15 percent were mostly the same hackers that also reported to reject a
target if it has no profitable information to offer.

The Risk-Avoidance Efforts and Desistance Considerations
The last section of hacking-activity related questions in the survey asked respondents about their efforts to minimize risks, their risk estimates, and their readiness to
desist from hacking. The results of these items are displayed in Tables 10 and 11. The
results in Table 10 confirm the appropriateness of the anonymizing tactics that were
provided as answer categories. Only 5 percent said they used methods other than the
ones provided for this item. In contrast, the least popular method on the list (source
routing) was used by 15 percent of the respondents in the sample. More than 50 percent of all participants used the three most common methods. Proxies (60 percent), war
driving (57 percent), and spoofing (54 percent) all share in common that they are relatively simple and effortless methods. As Table 10 shows, an inverse correlation seemed
to exist between the technical difficulty of anonymizing methods and their popularity.
The explanation for this inverse correlation is simple. The multitude of relatively simple
methods that already grant a high degree of anonymity in cyberspace made using more
strenuous tactics unnecessary.
When asked whether they had ever considered desisting from hacking, a total of
45 percent of all respondents answered “yes.” However, only six hackers (5 percent)
said they had done so often. An examination of these six cases showed that all of them
also reported to no longer be actively hacking. Unfortunately, the small case number in
the present sample renders this dataset unsuitable for further investigations of signifi129

cant differences between those cases and the ones that had not considered quitting or
those who considered quitting a few times.

Table 10: Anonymizing Methods and Desistance Considerations
N1

%2

74
71
67
55
54
49
33
19
6

59.7
57.3
54.0
44.4
43.5
39.5
26.6
15.3
4.8

18
33
73

14.5
26.6
58.9

6
50
68

4.8
40.3
54.8

27
21
15
11
9
8
6
5

21.8
16.9
12.1
8.9
7.3
6.5
4.8
4.0

Variable
Anonymizing methods3
Proxies
War driving
Spoofing
Public access points
Tunnels, Covert channels
TOR, I2P, etc.
Bouncing
Source routing
Other

Hacking source of income
Yes, the main
Yes, but not the main
No

Considering quitting
Yes, often
Yes, a few times
Never

Reason for considering quitting3
Involved risk
Fear of prosecution
Fear of apprehension
Involved effort
Fear of detection
Other
Income too irregular
Legal alternatives
1

The total sample size is n=124.
Percentages may not add up due to rounding.
3
For better readability, categories are rank ordered by importance.
2

130

The hackers who had considered quitting reported to have done so mainly because they were worried about the involved risk in general (22 percent). Seventeen percent considered quitting because they were afraid of being prosecuted, 12 percent did
so because they were afraid of being apprehended, and 7 percent thought about quitting because they were afraid of being detected. The lesser importance of the fear of
detection and apprehension might be initially counter intuitive, given that a prosecution
implies a prior detection and apprehension. However, the greater fear of prosecution
simply illustrates the fact that the prospect of being prosecuted is a greater deterrent
than the prospects of being detected or apprehended.
Compared to the fears of getting caught, a relatively small number of hackers (18
percent) also reported having considered quitting because of various practical reasons.
For 9 percent, the involved effort was a reason to contemplate giving up hacking.
Another 5 percent said that the income yielded by hacking was just too irregular, and 4
percent considered switching to legal alternatives.
Table 11 shows that, when asked whether they were more or less afraid of being
tracked and apprehended now than when they started to hack, slightly more than onethird of all respondents reported to be more afraid today (37 percent). Thirty-four percent said their level of being afraid had not changed, and another 30 percent said they
were less afraid today (15 percent) or had never been afraid (15 percent). The slightly
higher percentage of hackers who report to be more afraid today suggests that the increased efforts of law enforcement agencies in combating cyber-criminals has not gone
unrecognized in the hacking community. However, they also indicate that many hackers
are well aware of the continuing slim chances of becoming tracked and apprehended.
131

Many hackers further rated the risk of becoming apprehended significantly lower than
that of being detected. Thus, they demonstrated awareness of the fact that, due to the
various difficulties involved in the prosecution of cybercrimes, detection does not necessarily imply apprehension. Nevertheless, the popularity of the various anonymizing methods and the high fraction of aborted hacking attacks (about 20 percent) were clear indicators of the efforts many hackers undergo to avoid detection.
Table 11: Risk Estimates and Carefulness
N1

Variable

%2 (sd)

Afraid of being tracked and apprehended
More afraid
Unchanged
Less afraid
I was never afraid

46
42
18
18

37.1
33.9
14.5
14.5

10
20
30
24
24
13
3

8.1
16.1
24.2
19.4
19.4
10.5
2.4

63
44
17

50.8
35.5
13.7

124

3.6 (1.1)

Risk of apprehension estimate (scale 1-7) 124

2.5 (1.4)

Focus on rewards (1) vs. risks (7)

124

3.7 (1.2)

Rational (1) vs. emotional (7) during hacks 124

2.4 (1.4)

% of aborted attacks

20.4 (2.0)

Preparation and planning time
Less than 2 hrs
2 hrs
3 hrs – 1 day
2 days
3 – 5 days
2 – 3 weeks
Over 3 weeks

Change in preparation time
More time today
About the same time
Less time today

Risk of detection estimate (scale 1-7)

1
2

124

The total sample size is n=124.
Percentages may not add up due to rounding.

132

In concordance with the changes found in motives, targets, and methods, the increasing professionalism of many hackers was also manifested in their more careful
preparations. About half of all respondents reported to invest more time in their preparations today than when they started, a significantly higher number than the 14 percent
who prepare less today. On average, hackers dedicated between three to eight hours to
preparing their hacks. The length of preparation distribution, however, shows great variability and ranges from less than two hours to several weeks in some extreme cases.

The Assessment of Scale Validity and Reliability
The regression models used for the testing of the research hypotheses operated
with two of the personality constructs (risk propensity and rationality) as independent
variables. To ensure the appropriate operationalization of all personality variables in the
regression models, the validity and reliability of the personality constructs was assessed
prior to the calculation of the regression models. When estimating the validity of a theoretical construct, two aspects are of particular importance: discriminant and convergent
validity (Schnell, Hill, & Esser, 1999; Trochim, 2002). Discriminant validity is achieved
when a construct is empirically distinguishable from other constructs (Straub, 1989),
and convergent validity is confirmed when all items comprising the construct are measuring only aspects of the intended construct. At the same time, a construct also has to
be reliable, i.e. internally consistent. Internal consistency is achieved when all items
comprising the construct are highly inter-correlated. Since the scales used to measure
the personality constructs were abbreviated and partially modified, the validity and reliability of these scales was analyzed in an exploratory validation phase.
133

The Exploratory Factor Analysis
According to Thompson (2004), an exploratory factor analysis (EFA) should be
conducted when the relationships between individual items and underlying factors are
not exactly known. Since the risk propensity scale consisted of items from different
scales and the Rational-Experiential Inventory scale was heavily abbreviated, EFA was
used for a confirmation of the convergent and discriminant validity of the measures. The
particular type of EFA was a principal component analysis with promax rotation and
Kaiser normalization (calculated with SPSS 16.0). As Hair and his colleagues suggested, the selection of an orthogonal or oblique rotation should be made according to
the specific demands of a particular research problem (Hair, Anderson, Tatham, &
Black, 1998). According to Hair et al., orthogonal rotation methods are most appropriate
when the research goal is to reduce the number of items in a construct, regardless of
how meaningful the resulting underlying factors are. On the other hand, if the intent is to
create or verify theoretically meaningful constructs, oblique rotation methods are better
suited. Since the purpose of this factor analysis was to reveal the appropriateness of the
scales used in this study, promax rotation, an oblique rotation method, was chosen over
orthogonal rotation. All 15 items were entered into the EFA and three factors were extracted. Table 12 presents the EFA results for all three personality variables.

134

Table 12: Personality Scales, Item, Factor, and Index Analysis
Item to
total
correlation

Items

Factors1
1
2
3

Risk propensity scale (α=.83)
I always try to avoid situations involving
a risk of getting into trouble.
I always play it safe even when it means
occasionally losing out on a good opportunity.
I am a cautious person who generally avoids risks.
I am rather bold and fearless in my actions.2
I am generally cautious when trying something new.

.65

.81

.69

.88

.71
.52
.53

.83
.63
.65

Rationality items (α=.75)
I usually have clear, explainable reasons
for my decisions.
I don’t reason well under pressure.2
Thinking hard and for a long time about
something gives me little satisfaction.2
I prefer complex to simple problems.
I enjoy solving problems that require hard thinking.

.62

.79

.55
.44

.81
.57

.42
.63

.57
.82

Intuition-experience items (α=.86)
Using my gut-feelings usually works well
for me in working out problems in my life.
I trust my initial feelings about situations.
I like to rely on my intuitive impressions.
I often go by my instincts when deciding on a
course of action.
I don’t think it is a good idea to rely on one’s
intuition for important decisions.2

Eigenvalue
Variance explained (%)
Cumulative variance (%)

.57

.73

.66
.79
.79

.82
.87
.86

.61

.86
2.71
18.1
18.1

Indices
Summative risk propensity index
Summative rationality index
Summative intuition index
1

α

N Range

.83
.75
.86

124 5-35
124 11-35
124 10-35

2.21
14.8
32.9

4.31
28.7
61.6

(sd)
22.1
27.2
23.6

6.1
5.0
5.3

Principal Component Analysis with Promax Rotation Method and Kaiser Normalization.
Loadings less than .4 not shown.
2
Items were reversed.

135

Table 12 shows that the EFA produced three factors with eigenvalues greater
than 2.0, a level that confirms the independence of the concepts. The high eigenvalues
of all three factors also indicated that the factors explained large fractions of the variance within their respective set of variables. The three factor solution accounted for
63.4 percent of the total variance, a value above the generally accepted 60 percent level in social research (Hair et al., 1998; Thompson, 2004). To assess the factor loadings
in the individual item analysis, guidelines from Kim and Mueller (1978) were used. According to these guidelines, loadings of 0.4 to 0.54 are considered fair; 0.55 to 0.62 are
considered good; 0.63 to 0.70 are considered very good; and over 0.71 are considered
excellent.
As Table 12 shows, all of the 15 items loaded higher than 0.55 on their respective factors, and none of the items loaded higher than 0.4 on any other factors. Thus, all
three constructs were extracted “cleanly” as factors. The fact that none of the items
loaded on multiple factors was a strong sign that all three personality constructs had
high levels of discriminant validity. Similarly, the high to excellent loadings of all individual items on their respective factors further suggested that all three constructs also had
high levels of convergent validity. Based on the positive EFA results, all of the 15 items
were retained in the analysis
All 15 items also correlated highly with their respective scales. The lowest item to
total correlation of any item was 0.42, which shows that all items contributed in a meaningful way to the scale scores. The high internal consistency of all three scales is further reflected in their high Cronbach’s alpha values. The risk propensity scale reached
an alpha level of 0.83, the rationality scale a level of 0.75, and the experience scale a
136

level of 0.86. All three values were within 0.70 and 0.90, the range that is typically considered to be ideal for internal consistency measures (Hair et al., 1998).
Overall, the loading patterns of the REI items in this factor analysis compared favorably to the factor analysis findings for the complete scales that were reported by
Handley and colleagues (Handley et al., 2000). The similarity between the patterns of
both factor analyses confirms the appropriateness of the item selections that were used
to create the abbreviated scales. The comparison to Handley’s results further reveals an
interesting finding. When compared to the general public sampled in Handley’s study,
the sample of hackers yielded a considerably higher average rationality value (5.4 compared to 3.4 in Handley’s analysis). Hackers also reported to have a higher confidence
in their experience-based decision making (4.7 compared to 3.4), even though this difference was not as large as the one found between the two rationality measures. These
comparisons suggest two important differences between hackers and the general public: (1) hackers prefer a more analytical and rational thinking style than the average person, and they (2) display a generally higher confidence in their ability to make decisions,
regardless of whether these decisions are based on rational considerations or on intuition and experience.

The Test of the Study Hypotheses
All of the hypotheses tested in this study were assessed using additive multivariate linear regression models. The personality scales were recoded into summative indices and included as independent variables in the regression models. For the creation
of the indices, all negative items were reversed so that higher values in all variables re137

flected more pronounced personality characteristics. The calculation of the summative
indices was unproblematic because none of the scale items had any missing values and
none of the indices contained any extreme outliers. Table 12 provides detailed statistical
descriptions of all three indices.
In addition to the two variables testing the predictions derived from RCP, seven
main sociodemographic variables were included in the saturated models to control for
potential effects of external factors. The seven variables included in the analysis were
age, gender, race, educational attainment, employment status and student status. The
age variable was calculated by subtracting the respondents’ year of birth from 2008,
and included as a continuous variable. Given that the conference took place in February, this calculation method yielded the correct age of most participants. Sex (1 indicating females), race (1 indicating Non-Whites), marital status (reference category never
been married), employment status (1 indicating unemployed), and student status (1 indicating students) were coded as indicator variables. The decision to collapse some of
the subcategories in these variables was made in order to avoid problems that might
result from the low cell counts of some of these variables. Black, Asian, Hispanic, and
Other respondents, which together had a cell count of 11 cases, were collapsed into
one single category Non-Whites. The reference category “never been married” was
chosen for marital status because most respondents fell in this group. The one respondent who reported to be divorced was included in the next closest category “married.”
The variable measuring educational attainment was included in the models in its original
coding. The inclusion of ordinal scaled variables in linear regression models, even
though debated by some researchers since the early 1970s (Hawkes, 1971; Morris,
138

1970), has become a common practice and has repeatedly been proven to produce viable results (Allan, 1976; Kim, 1978; O'Brian, 1979).
Before the regression models were run, the conformity of the present data with
several assumptions of these models was verified to avoid any potential violations. First,
all variables were screened for outliers. Given the small number of cases in some of the
subgroups (7 females, 11 Non-Whites, 10 unemployed), this step was particularly important. As a result of this screening process, three cases were excluded from the third
regression analysis because their values on the dependent variable were four to ten
times higher than the next highest values. If included in the regression model, these
cases would have exerted a disproportional influence on the results. Next, all variables
were screened for multicollinearity. With the tolerance levels of all independent variables over 0.9, no linear inter-correlation was detected among them. The results of the
various regression models are displayed in Tables 13, 14, 15, and 16.

The Test of Hypotheses H1 and H2
Table 13 shows the influence of the two personality characteristics, risk propensity and rationality, on the overall success of hacking activities. To reflect the overall success of all hacking activities most accurately, the success rates of the three different attack methods (technical intrusions, social methods, and malicious code distributions)
were weighed with the proportion of total hacking attempts that was accounted for by
the respective attack method, and all three products were then summarized into the total success rate for all methods. For example, if a hacker reported having undertaken a
total of 100 hacking attempts, out of which 70 were technical intrusions, 20 were social
139

engineering attacks, and 10 were distributions of malicious code, the total success rate
for this hacker was calculated as the sum of the success rate of technical intrusions
multiplied by 0.7, the success rate of social methods multiplied by 0.2, and the success
rate of malicious code distributions multiplied by 0.1.
The regression results presented in Table 13 clearly support the predictions that
were derived from the propositions of the rational choice perspective. Despite the low
number of cases in the models (n=124), a circumstance that usually causes high ingroup variances, both models were highly significant (Model 1 and 2 p<.001).

Table 13: OLS Regression Coefficients for Estimated Effects of Rationality and Risk
Propensity on Hacking Success
Variable

B

Model 1
β

Model 2
B
β

Hypothesized characteristics
Rationality index
Risk propensity index

.14 **
-.09 **

.29 (.04)
-.23 (.04)

.14 **
-.09 *

.30 (.04)
-.23 (.04)

.00
.13
-1.66 *
-.12

.00 (.04)
.01 (.97)
-.19 (.77)
-.06 (.22)

-.41
.20
-1.30
-1.03 *

-.06 (.64)
.04 (.57)
-.15 (.97)
-.20 (.97)

(1.27) 5.03 **
.21

(2.00)

Sociodemographic controls
Age
Female
Non-White
Education
Marital status
Living as married
Married
Unemployed
Student

Constant
R-squared

4.09 **
.11

*p<.05. **p<.01. ***p<.001

140

In the first model, the two personality characteristics alone explained 11 percent
of the variance in the success of hacking attacks. In this model, both variables exerted a
highly significant influence on the dependent variable (p<.01). Moreover, rationality
turned out to be the most influential variable in both models. As was predicted in the hypotheses, the effect of rationality on the success of hacking attacks was positive and the
effect of risk propensity negative. The results confirm the predictions made in hypotheses H1 and H2 in Chapter 5: The higher the preference for an analytic-rational approach
to thinking and the lower the risk propensity of a hacker, the more successful this hacker is in his or her hacking activities.
The inclusion of the sociodemographic control variables in the second, saturated
model had only a slight impact on the effect of both personality variables. While the
standardized coefficients for both variables remained roughly the same, the inclusion of
the control variables reduced the effect of risk propensity to a p<.05 significance level.
Overall, the inclusion of the additional variables raised the amount of explained variance
to 21 percent in the second model. Nevertheless, the impact of the individual sociodemographic variables was surprisingly small. Only two of the variables reached a significant level. The variables age and sex had virtually no impact on the dependent variable.
Particularly for the age variable, this finding was surprising because it implies that, despite the fact that virtually all hackers in the sample reported to have increased their
hacking skills over the course of their hacking careers, hackers of all ages report roughly the same success rates. One possible explanation for this finding is that more seasoned hackers also report to seek out higher profile targets and that this change in tar-

141

get preference countervails the improvements of hacking skills. This explanation, however, remains speculative and requires further testing to be confirmed.
Different from the surprising finding with regard to age, the complete absence of
a gender effect in the data is unfortunately not very meaningful because only seven of
the respondents were females. Since seven respondents are not enough cases to allow
a confident generalization of the results, future studies with more female hackers are
needed to confirm this finding. The only two sociodemographic variables in the second
model to reach a significant level were race and student status. Students report significantly lower success rates than persons who are not or no longer studying. Also, when
compared to White hackers, hackers belonging to minority groups report a significantly
lower success rate (I<.05). However, this finding also has to be interpreted with caution,
given the small number of minority hackers in the present sample.

The Test of Hypotheses H3 and H4
The predictions of hypotheses H3 and H4—that hackers with a higher preference
for analytic-rational thinking and a lower risk propensity will also perceive their hacking
activity as riskier—were tested with a second regression analysis. For the assessment
of the overall risk estimation in the dependent variable, the two items measuring the estimated risks of (1) being detected and (2) being apprehended during or after a hack
were combined into one summative index. As did the other indices, this index had no
missing values and showed no significant deviation from a normal distribution.

142

Table 14: OLS Regression Coefficients for Estimated Effects of Rationality and Risk
Propensity on Perceived Risk Involved in Hacking

Variable

Model 1
B
β

Model 2
B
β

Hypothesized characteristics
Rationality index
Risk propensity index

.11 **
-.10 **

.25 (.04)
-.27 (.03)

.13 *** .30 (.04)
-.11 *** -.30 (.03)

Sociodemographic controls
Age
Female
Non-White
Education
Marital status
Living as married
Married
Unemployed
Student

Constant
R-squared

-.11 **
1.67 *
-1.99 **
-.09

-.32 (.04)
.17 (.86)
-.25 (.67)
-.05 (.19)

.16
.02 (.57)
.22
.05 (.50)
-1.27
-.16 (.86)
-1.43 *** -.30 (.44)
5.32 ***
.11

(1.17) 9.37 ***
.29

(1.76)

*p<.05. **p<.01. ***p<.001

The results of the regression models corroborated the predictions stated in the
theoretical hypotheses. Both models were highly significant (Model 1 and 2 p<.001),
and the two personality characteristics exerted highly significant effects on the dependent variable in Model 1 (p<.01) and Model 2 (p<.001). Thus, hackers with a preference
for an analytic-rational thinking style estimate the risks involved in their hacking activities to be significantly higher, as do hackers with a lower risk propensity. While this result is little surprising for the risk propensity variable, the finding that the standard coefficient of the rationality variable is almost as high as that of the risk propensity variable
implies that hackers with higher rationality scores weigh the potentially involved risks
considerably more in their decision-making processes. In combination with the results of
143

the first regression analysis, this finding suggests that it is the most successful hackers
who are also most susceptible to changes in the estimated risks. Together, these two
findings hold important policy implications because they indicate that successful manipulations of risk estimations might be a suited tactic to deter especially higher skilled
hackers.
The circumstance that the significance level of both personality characteristics
increased with the inclusion of the control variables in the saturated model hints at a
slight suppression effect. The one sociodemographic variable that was identified to exert this suppression effect was student status, a variable that had a highly significant effect itself (p<.001). Aside from the student status variable, age (p<.01), gender (p<.05),
and race (p<.01) also exerted a significant effect on the risk perception. The findings
that student hackers have significantly lower risk estimates, whereas female hackers
estimate risks to be significantly higher are consistent with many other studies in the literature (Bouffard, 2007; Earnest, 2003; Paternoster, 1989; Wright, Caspi, Moffit, & Paternoster, 2004). Contrariwise, the findings that older hackers have lower risk perceptions, while minority hackers have higher ones, appear to be counterintuitive at first. The
effect measured with regard to age, however, is consistent with the predictions of RCP.
Rational choice perspective assumes that the awareness of involved risks decreases
during the habituation of the activity. During this phase, the criminal activity becomes a
routine exercise and the absence of any negative consequences diminishes the awareness of involved risks over the years.
The finding that minority hackers report higher risk estimates, while interesting,
has to be interpreted with caution because the number of minority hackers included in
144

the sample was too small to allow meaningful generalizations. Future studies with larger
samples of minority hackers are again required to confirm this finding.

The Test of Hypotheses H5 and H6
As was mentioned earlier, three cases were excluded from the third regression
analysis due to their extreme values on the dependent variable. The third analysis regressed the time that hackers reported to use for the preparation of their attacks on personality traits and sociodemographic characteristics. Three of the respondents reported
preparation times that were four to ten times longer than the next longest time intervals.
Two of the cases reported to prepare their attacks for three months, and one respondent indicated it takes him a whole year to prepare a single hack. An examination of the
three cases revealed that all three cases engaged in hardly any hacking activity and
probably counted the time intervals between hacks as preparation time. Based on the
results of this examination, all three cases were excluded. After the exclusion of the
three outliers, the dependent variable was no longer significantly skewed. The results of
the third regression analysis are shown in Table 15.
The results shown in Table 15 corroborate hypothesis H5, but they lend only partial support to hypothesis H6. As was the case in the other regression analyses, both
models in the third regression analysis were significant (p<.01). The rationality variable
was again the strongest predictor variable in both models (p<.01 in Model 1 and p<.05
in Model 2). The effect of the risk propensity variable pointed in the predicted direction
in both models, but the effect reached a significant level only in the second, saturated
model (p<.05). Again, the student variable was identified as responsible for the slight
145

suppression effect (Cramer, 2003). Despite the fact that none of the sociodemographic
variables reached a significant level, the proportion of explained variance in the dependent variable increased from 11 percent in the first model to 21 percent in the second
model.

Table 15: OLS Regression Coefficients for estimated Effects of Rationality and Risk
Propensity on Preparation Time

Variable

Model 11
B
β

Model 21
B
β

Hypothesized characteristics
Rationality index
Risk propensity index

1.3 **
-.76

.25 (.50)
-.17 (.41)

1.2 *
-.87 *

.22 (.53)
-.20 (.44)

.20
-1.40
-1.57
1.97

.05 (.51)
-.11(13.0)
-.16 (9.1)
.09 (2.6)

-3.04
-6.40
6.18
1.12

-.04 (7.6)
-.11 (7.0)
.06(11.7)
.02 (5.9)

(14.6) -4.54
.21

(23.8)

Sociodemographic controls
Age
Female
Non-White
Education
Marital status
Living as married
Married
Unemployed
Student

Constant
R-squared

2.61
.11

*p<.05. **p<.01. ***p<.001
1
n=121. 3 outliers were excluded.

The analysis demonstrates that hackers with a higher preference for a rational
thinking style prepare for their hacking attacks longer. The longer time period used for
preparation is an indication for an overall more thorough preparation process. The longer preparation phase reduces the involved risks and it probably also contributes to the
146

overall higher success rate of more rational-acting hackers and hackers with a lower
risk propensity.

The Test of Hypotheses H7
The last hypotheses H7 stated that risk propensity also exerts an influence on
the total number of hacking attempts. The dependent variable total number of hacking
attempts was calculated as a summative index of the total number of technical intrusions, social methods, and malware distributions a person had attempted. The wide
range of the index (it ranged from 1 to 23,000) and the rounded estimates many respondents gave to the questions about the total number of attacks caused the dependent
variable to have a platykurtic shape with a multimodal, rounded peak and wide shoulders. Despite the significant deviation from the mesokurtic shape of a normal distribution, the distribution of the dependent variable was not significantly skewed, and was
therefore included in the regression.
Different from the other three regression analyses, the effect of the risk propensity variable (p<.001 in both models) was stronger in this analysis than the effect of the
rationality variable. Nevertheless, an unpredicted significant effect was also found for
the rationality variable (p<.05 in Model 1 and p<.01 in Model 2). The effects of both variables are shown in Table 16.
The risk propensity of respondents influenced the number of total hacking attempts as predicted in hypothesis H7. Persons with a higher risk propensity engaged in
significantly more hacking attempts. Surprisingly, the level of rationality also exerted a
significant influence on the number of total hacks. Hackers with a preference for analyt147

ic-rational thinking styles also committed significantly more attacks. One possible explanation for this finding is that, despite the higher risk awareness and the longer preparation times, hackers with a preference for analytic-rational thinking styles engage in more
hacks simply because they are more successful in their hacks.

Table 16: OLS Regression Coefficients for estimated Effects of Rationality and Risk
Propensity on Total Amount of Hacking Attacks
Variable

Model 1
β

B

B

Model 2
β

Hypothesized characteristics
Rationality index
Risk propensity index

174.60 *
228.44 ***

.21 (74.72) 192.51 **
.33 (61.96) 243.44 ***

.23 (74.23)
.35 (61.26)

Sociodemographic controls
Age
Female
Non-White
Education
Marital status
Living as married
Married
Unemployed
Student

Constant
R-squared

1981.59
.12

-19.42
16.69
-17.10
-539.06

-.03 (68.87)
.00 (1613.7)
-.00 (1279.4)
-.16 (368.80)

1940.18
510.49
4110.45 **
-2226.61 **

.16 (1073.5)
.06 (956.86)
.27 (1623.2)
-.25 (829.90)

(2215.30) 5438.62
.29

(3345.22)

Note. Standard errors are listed in parenthesis.
*p<.05. **p<.01. ***p<.001.

Two of the sociodemographic control variables entered in the saturated model
also exerted a significant effect on the number of hacks. Unemployed hackers reported
a significantly higher number of hacking attacks than hackers who were employed
(p<.01). One possible explanation for this finding could be related to the circumstance
148

that hacking is a time-consuming activity. Unemployed hackers simply have more time
at their hands that they can dedicate to hacking. Time considerations could also be the
reason why student hackers report to commit significantly fewer attacks (p<.01), since
most of the students in the sample were part-time students who also had full-time jobs.
Overall, the regression analyses corroborated the hypotheses that were derived
from the propositions of the rational choice perspective. With the exception of hypothesis H6, all predicted effects were found to be strongly significant and in the predicted
direction in all models. The implications of these findings are discussed in the next
chapter.

149

CHAPTER NINE:
DISCUSSION, IMPLICATIONS, AND LIMITATIONS
In the first section of the final chapter of this study, some of the main findings
about hackers and the hacking community are summarized and briefly discussed. The
chapter continues with a succinct discussion of the implications that the findings of this
study hold for the application of the rational choice perspective to hackers. Next, some
of the implications of this study for policy efforts to combat cybercrimes are discussed.
The chapter concludes with a list of suggestions for future cyber-criminological research
studies.

The (In)Accuracy of the Common Hacker Stereotype
The present study showed that the common hacker stereotype as a clever, lonesome deviant male adolescent whose computer proficiency compensates social shortcomings is barely beginning to tell the whole story of who hackers are. That is not to say
that this stereotypical portrayal of hackers is completely mistaken. Several aspects of
this characterization were indeed confirmed by the study results as well as the researcher’s personal observations during the conference. First, the participants in this
study were indeed highly educated, intelligent persons who had their inquiring minds set
on technological developments. Many of these technophiles also seemed to be equally
inventive, creative, and determined.
Second, the convention attendees were also predominantly males, and minority
hackers were rare exceptions. The near uniformity with regard to the sex and race dis150

tributions, however, stood in sharp contrast to the strong emphasis of many attendees
on an individualistic appearance. Many hackers conveyed their individualistic nature in
conversations with the researcher as well as through their physical appearance. The
physical expressions of individualism ranged from extravagant haircuts and hair colors,
to unusual clothing styles, to large tattoos on various body parts, sometimes even on
faces.
The two most important inadequacies of the hacker stereotype seem to be the
notions that hackers are invariably young and that they are socially inept. The study
found that hacking is by no means only a young man’s game as Yar suggested (Yar,
2005a). It remains to be seen what fraction of hackers is actually comprised by teenagers, but the findings of this study clearly showed that persons of various age groups engage in hacking activities. More importantly, the data also revealed that hackers undergo a maturation process over the course of their hacking careers and that the more experienced and seasoned hackers tend to be the most dangerous ones. They are more
likely to attack higher profile targets and some of them even engage in their illegal hacking activities with the stronger criminal intent of making financial profits.
Young and inexperienced hackers can certainly cause damage with their mischief, but the study showed that these hackers attack primarily private targets and do
so out of intellectual curiosity, love for knowledge, experimentation, boredom, or youthful tomfoolery. Many hackers first became interested in hacking very early in their lives,
and, typically, they were not driven by a pronounced initial criminal intent. As their hacking activities continued to become habitualized, however, many of them developed into
more professional and ambitious hackers. Over the course of their hacking careers,
151

many intensified their hacking activities and began to also attack higher profile targets
such as governmental and corporate information systems. Some hackers even reported
having turned their once merely deviant juvenile behavior into a criminal business activity. A total of 15 percent of all respondents said that hacking had become their main
source of income and that they would reject a target unless it was profitable. Undoubtedly, these experienced veteran hackers are the ones with whom law enforcement
agencies should be most concerned about and to whom they should direct their attention.
Although the comparatively high fraction of unmarried hackers showed that many
of them are indeed hesitant to engage in serious relationships and commitments, the
vast popularity of social hacking methods and their high success rates also indicated
that the commonly presumed social incompetence of hackers is wrong and misleading.
The falseness of this assumption was further reaffirmed by some of the observations
the researcher made during the convention. Most attendees appeared to be outgoing
and sociable. Many attended the convention together with their friends, and most of the
attendees seemed to share a distinct sense of humor and mingled quickly. Certainly,
the informal observations during the convention and the findings that hackers are skilled
in manipulating and ‘programming’ other persons and oftentimes manage to exploit the
trust or carelessness of other computer users for their hacking purposes are not enough
evidence for a strong rebuttal of the notion that hackers are social hermits. It might be
the case that the sociability of hackers is limited to interactions with other, like-minded
technophiles and that, although many appear to be skilled manipulators, genuine and
affectionate social relations are of lesser importance to them. Additional examinations of
152

the social networks of hackers, their amount, frequency, and quality of interactions with
close contacts, the types of contacts they engage in (face-to-face or online), and the
importance they attribute to these social contacts are needed before a conclusion about
the appropriateness of the assumption that hackers are recluses can be reached.
The debate about the sociability of hackers aside, one of the most important findings of the present study was the significant role of social hacking methods. The common perception of hacking attacks as being executed solely through technical means
and the perception of hackers as socially incompetent are probably part of the reason
why the danger posed by social engineering attacks is oftentimes underestimated. Unless these perceptions are revised and the awareness of social hacks is raised, social
engineering methods will continue to be very successful and will pose a serious threat
for many organizations.
The separate analysis of the three main hacking techniques in this study showed
that many hackers combine social and technical methods and launch attacks that are
comprised of both tactics. The examination of respondents who reported using both
techniques in roughly equal amounts further revealed that they were the most successful hackers. Different from social and technical means and strategies, which appear to
be commonly combined, the surveyed hackers largely refrained from distributing and
spreading malicious code. Thereby, they demonstrated having a strong preference for
directed attacks on selected targets over widely dispersed and randomly distributed attacks without specific targets. It appears that phishers, spammers, and virus coders are
a group of cybercriminals that is distinctively different from ‘traditional’ hackers. Unfortunately, the limited amount of persons who distributed large amounts of malicious code
153

in the present study did not permit the analysis of whether these persons also exhibit
other distinctive differences. More studies with larger subgroups of malware distributors
are needed to draw such comparisons.

The Suitability of the Rational Choice Perspective
The discussion of the various theoretical difficulties facing the newly developing
field of cyber-criminology in the fourth chapter expounded the main problems of transposing traditional, ‘terrestrial’ criminological theories to the online environment. The discussion established that the irrelevance of central assumptions of many traditional theories precludes their meaningful application to cybercrimes and cyber-criminals, and the
violation of some traditional assumptions about offenders or other criminogenically relevant environmental aspects even renders some of them inapplicable to cybercrimes. As
a result, the selection of potential theories is severely limited. The rational choice perspective was selected for the present study because it does not rest on any environmentspecific assumptions and its main offender-related propositions are applicable to cybercriminals. The results of this study suggest that the rational choice perspective is indeed
a viable theory for the explanation and description of their behavior.
With only one exception, the hypotheses that were derived from the rational
choice perspective were corroborated in the present study. The only hypothesis that received only partial support was the assumed effect of risk propensity on the length of
preparations for hacking attacks. Nonetheless, the effect was found in the predicted direction and also reached a significant level in the second, saturated model. The analysis
demonstrated the importance of the personality traits, rationality and risk propensity, for
154

various hacking-related decision-making processes and outcomes. Both traits were
found to be influential factors on the overall success as a hacker, the length hackers
take to prepare for their hacking attacks, the risk assessments of hackers, and the
overall number of hacking attacks. The importance of rationality as a factor was further
underlined by the finding that it was the most important factor in almost all regression
models. The study established both factors as essential dimensions of cybercrime offender typologies.
The three involvement decision models of the rational choice perspective that
were modified and transposed to fit the involvement in hacking proved to be suitable
guidelines for the assessment of hacking-related decisions. While the study was not
able to test all of the many factors thought to be influential for involvement in hacking,
most of the examined factors were indeed found to be accurate. Some factors, however, deviated from the way they were proposed in the models. Most importantly, financial
considerations were found to be irrelevant for the initiation phase. The model that was
suggested in Chapter 3 emphasized financial needs, motives, opportunities, and inducements and it also considered employment to be relevant to the decision to engage
in hacking activities. The results of this study, on the other hand, proved this assumption
to be false. Different from other criminal activities, hackers seem to not be driven by financial considerations when they decide to begin to hack. Instead, the data documented that the importance of financial motivations as a driving motive for the continued
involvement in hacking develops over time. Therefore, it should be included in the
second model as one aspect of increased professionalism. With regard to considerations to desist from hacking, the study found risk-related deliberations to be most in155

fluential. Practical reasons were found to be of generally lower importance. The lower
importance of practical reasons suggests that even if hackers find their hacking activities unsuited to yield a sustained income, they seem to continue to hack on a hobby basis. The decision to completely give up hacking, on the other hand, is most often made
because hackers are afraid that their hacking activities could get them into trouble and
jeopardize their careers in legitimate jobs and their lives.
Despite the general support for the theory that was found in this study, some of
its drawbacks also became apparent. First, the generality of the principal propositions of
the rational choice perspective rendered the deduction of precise hypotheses difficult.
Only one of the six main propositions lent itself to meaningful and testable hypotheses.
Furthermore, the testing of these hypotheses itself was found to be principally complicated because it had to resort either to psychological personality traits or to fictional
scenarios. Second, the models that can be derived from propositions four to six, while
evidently helpful in clarifying and structuring influential factors of the continued involvement in criminal activities, rest on many different causal assumptions that are exceedingly difficult to operationalize. Consequently, in order to account for the empirical specification of these models for the different types of crimes and offenders, the operationalization had to resort to univariate descriptions to determine the relative importance of
the theoretically proposed factors and developments.
Rational choice perspective emphasizes the importance of two particular personality traits—the ability of offenders to deliberately weigh the outcomes of alternative
actions and their willingness to take risks. Both factors, while corroborated as influential
in the present study, do not lend exclusive support to the rational choice perspective.
156

The general theory of crime (Gottfredson & Hirschi, 1990) considers both personality
traits as two of the six components that comprise low self-control. Future studies should
examine all six dimensions of low self-control and investigate the influence of this personality construct on hacking activities. Furthermore, Jaishankar recently proposed the
so-called space transition theory, the first criminological theory that was explicitly designed for the application to crimes committed in cyberspace (Jaishankar, 2008). Space
transition theory provides an explanation for why otherwise law-abiding persons, who do
not commit crimes in the terrestrial world, engage in cyber-criminal activities. It argues
that people behave differently when they move from one space to another. Future cyber-criminological studies should devote special attention to this first exclusively cybercrime-related theory and test whether it is indeed better suited for the explanation of cybercrimes than traditional criminological theories.

The Policy Implications of the Study Findings
The conclusions that can be derived from this study are not limited to contributions to the scientific discourse about cybercrime offenders. They also hold some important implications for the efforts to combat cybercrimes. Experts agree that present efforts to combat cybercrimes are facing a multitude of challenges that have to be addressed. Aside from the resource shortages and other practical difficulties that were outlined in Chapter 3, law enforcement efforts to combat cyber-criminals are also hampered by a shortage of substantive and reliable information that can be used for the
creation of cybercrime-offender profiles. Detailed profiles of the different types of cybercriminals, their skill levels, and their motivations are critical because they provide helpful
157

guidance for the investigation of cybercrimes and, thereby, increase the effectiveness of
current prosecution efforts. A more effective response by the criminal justice system is
urgently needed—not only because it would increase the number of convicted cybercriminals but, most importantly, because it would also have a preventive deterrence effect on the larger hacking community.
The findings of this study suggest that the creation of a deterrence effect has to
become an essential component of efforts to combat cybercrimes. Unfortunately,
present efforts to curb cybercrimes are hardly suited to exert a pronounced deterrence
effect. Despite the annually increasing number of cybercrimes, only a relatively few high
profile cybercrime cases are presently successfully tried, and many of them do not lead
to swift or severe punishments (Brenner, 2006). The continuing unlikeliness of punishments is particularly problematic because it severely undermines any efforts to deter
criminal behavior in cyberspace. Indeed, the findings of the present study demonstrate
that many hackers are aware of the slim chances of being detected and punished. The
current improbability of becoming prosecuted even led some hackers to report that they
have never been afraid of being apprehended or prosecuted. Furthermore, the risk
awareness of most hackers seems to decrease over time as they repeatedly learn that
their actions have no negative consequences for them.
Nevertheless, several measures in this study also signify that deterrence can be
a successful strategy to prevent cybercrimes. The study showed that many hackers
have a nuanced risk awareness. For example, the majority of hackers report having become more concerned about risks in recent years, a finding that suggests that increased efforts to combat cybercrimes do not go unnoticed in the hacking community.
158

Furthermore, many hackers evidently distinguish between the chances of becoming detected and apprehended and the consequences of these two events. Most importantly,
the data also indicate that the most successful hackers are the ones that also have the
highest risk awareness. Thus, these hackers seem to be the ones that are most susceptible to changes in risk estimates.
Deterrence undoubtedly is an indispensable component in the control of all different types of criminal behaviors, but is seems to be particularly suited to prevent cybercrimes. Unlike other, less deliberately acting types of criminals, hackers plan their
hacking attacks, and they oftentimes do so in an explicitly rational manner. Consequently, they should be more easily dissuaded than criminals who commit their crimes spontaneously whenever opportunities arise.
Taken together, the findings of this study suggest that a more pronounced deterrence perspective needs to become a central addition to the existing technical approaches to cybercrime prevention. However, merely adding deterrence as one separate component will not suffice. To be effective, a deterrence perspective has to be integrated into all currently existing national policy efforts to prevent computer crimes, ranging from legislation and regulation to law enforcement strategies. Moreover, it would also have to be implemented outside of national or state policies. One promising approach to establish deterrence policies in the private sector could be directed at businesses and organizations. The present study showed that most hackers pursue legal
careers in legitimate jobs and companies. Organizations and companies that offer IT
security services or are otherwise attractive to hackers should be encouraged to promote awareness of the potential consequences of committing cybercrimes. For exam159

ple, they could distribute information about punishments that have been given to convicted computer criminals as well as other informational materials that directly highlight
what constitutes a crime under legal as well as ethical perspectives. Other informal control mechanisms, such as extra-legal social stigmata or the systematic introduction of
negative effects on job opportunities might also be strong incentives to prevent particularly young middle-class computer experts from becoming involved in computer crime
activities.
Unquestionably, the establishment of effective deterrence efforts as an integral
part of cybercrime prevention strategies will not be an easy undertaking. The vast range
of cybercrime activities and the multitude of different offenders considerably complicate
the selections of the most appropriate deterrence policies. Strategies that are most effective for leisure-time juvenile hackers will most likely be unfit to deter destructive computer-security experts or other seasoned hackers from attacking computer systems for
monetary gains. Nonetheless, deterrence should be pursued as a mitigation strategy,
because even limited accomplishments can prevent some crime incidents and provide
some protection from an increasingly serious problem.
Companies in branches that typically employ hackers can certainly be particularly
helpful in deterring computer crimes, but the results of this study also indicate that all
companies and organizations need to do more to actively prevent victimization, regardless of their branch. The analysis of the different hacking methods showed that, of all
three main types of attack methods, social engineering attacks are the most successful
ones. It also revealed that the various methods to obtain user passwords, be it the systematic guessing of weak or standard passwords or the theft of user logins remain the
160

most common ways hackers gain access to their targets. Moreover, the most successful
hackers seem to launch combined hacking attacks that utilize both technical intrusion
methods and information they obtained through social inquiries and deceptions of their
victims. Thus, it seems that the weakest points of companies and organizations are their
employees and members. Corporations have to educate their employees about social
hacking methods. They need to raise awareness of the seriousness and frequency of
the problem, educate their staff about the wetware tactics commonly used by hackers,
and give them instructions of how to avoid becoming victimized.
The education of employers, while definitely an important protective measure, is
not the only contribution that will be required from organizations. They also need to start
reporting all their victimization incidents to the authorities. The current situation, in which
individual organizations refrain from reporting incidents to protect their own interests
and thereby harm the interest of all businesses, needs to be changed because, unless
more incidents are reported, computer crimes are unlikely to become controllable. The
benefits and detriments of a mandatory reporting system are debatable, but a reporting
requirement would certainly benefit efforts to manage cybercrimes. It would put law enforcement agents in the position to decide which cases to devote their attention to rather
than be dependent on the willingness of organizations to submit their cases in order to
press charges.

The Limitations of the Present Study and Suggestions for Future Research
Even though it produced valuable insights into the sociodemographic composition of the hacking underground and the various developments hackers undergo over
161

the course of their hacking careers, the present study was limited in certain ways. One
set of potential shortcomings relates to the sampling frame and the sample size of the
study. The study analyzed only data from one particular convention, a circumstance that
constricts the confidence with which the present findings can be generalized to larger
populations. Although the ShmooCon convention attracted a diverse clientele, it remains unclear how general the profile of this particular convention really was. It also
remains uncertain whether there are significant differences between the attendees of
different conventions. More datasets from different conventions are needed to enable
researchers to draw comparisons between them and to assess the reliability and validity
of the present data. Once multiple studies from different conventions exist, meta-studies
will eventually be able to compare the results of these studies and extract highly reliable
and valid findings.
Although repeated studies from different conventions will eventually be able to
generate valid and generalizable results, these results will, to a certain degree, be generalizable only to the subset of hackers who consider attending hacker conventions or,
more narrowly, have already attended them. It remains to be seen whether there are
systematic and consistent differences between hackers who potentially attend conventions and those who do not. The average age of respondents in this study, for example,
was considerably higher than the typical age of hackers that other authors have suggested (Yar, 2005a). This finding indicates that studies operating with conventions as
their sampling frames are indeed suffering from some systematic selection biases. An
assessment of the exact areas in which such systematic differences exist and the degree to which they render the results of convention studies distinctively different from
162

other studies with different sampling frames can only be achieved by comparative studies. Until other sampling frames, such as message boards, have been utilized and until
their results have been compared with the ones produced by convention studies, researchers have to remain cautious when generalizing convention-based study findings
to all hackers.
Second, studies with larger sample sizes are needed to confirm some of the findings in the present study. The overall small sample size of this survey reduced the case
numbers in some subgroups beneath commonly accepted margins of statistical generalizability. The regression results with regard to female hackers, minority hackers, and
unemployed hackers, for example, have to be interpreted with the utmost caution and
their validity should be reassessed with larger samples to verify their accuracy.
One important sample-size related aspect that has to be considered in this context is that, while larger sample sizes are certainly desirable, their creation bears practical problems. Despite the fact that the ShmooCon conference is one of the largest international hacker conventions, it was attended by ‘only’ about 800 persons, many of
whom were not eligible for participation in the study. Accordingly, even though it
achieved a relatively high response rate among eligible attendees, the present study
yielded less than 130 cases. Two possible solutions for this problem come to mind.
First, researchers could solve this problem by collecting data from the world’s
largest hacking convention DefCon. DefCon is attended by over 7,000 persons and it
also has a reputation for attracting many black hat hackers. The large size of this convention makes it the ideal candidate for studies that seek to obtain larger sample sizes.
Researchers attempting to utilize the DefCon convention for their research purposes,
163

however, are most likely facing a different kind of challenge. DefCon has a reputation of
being a less professional convention and a convention that is attended by many hackers
mainly to enjoy a fun weekend in Las Vegas with like-minded people. The resulting
challenge for potential research projects is to conduct the study before attendees become intoxicated and no longer take the research project seriously.
An alternative solution for the sample-size problem would be to combine the datasets from different conventions. The results from different studies of various conventions could be merged into one larger dataset. Although this approach promises to provide larger case numbers and will likely yield generalizable results, it is not without disadvantages. The individual surveys would have to repeatedly ask the same item subsets to be comparable, thus hindering and delaying the assessment of different hackingrelated aspects and the development of more advanced survey instruments.
Aside from potential biases resulting from the sampling frame and the problems
associated with the small sample size, it is reasonable to assume that the present research project was also confronted with the problem of social-desirability biases. Socialdesirability biases are introduced through the propensity of respondents to give socially
desirable responses. They are a common problem of studies that rely on indirect, subjective information provided by respondents rather than objective or direct measures, or
a combination of the two (Fisher, 1993). In the case of cyber-criminals, socialdesirability biases are extremely difficult to overcome because objective measures of
cybercriminal activities are difficult to obtain. One possible assessment of socialdesirability biases could be achieved by conducting a research study that combines a
survey section with a direct measurement of hacking skills and expertise. For example,
164

a honeypot could be used as one possibility to obtain a direct measurement. In computer terminology, the term ‘honeypot’ refers to a forensic software trap which can, among
other things, be used to log and analyze hacking activities. This honeypot environment
could be utilized to verify the skill levels respondents claim to have. The inclusion of
such a direct measurement, however, complicates the study and the obtainment of IRB
approval and significantly increases the effort for respondents. For that reason, the conduction of such a combined study during a convention is highly unfeasible.
The present study was a first attempt to generate quantifiable information about
the hacking underground, and, as such, it was also naturally limited with regard to how
many aspects of this community were assessable. As does every extension of our
knowledge, the current study provided some answers but also raised many more questions. Future studies need to include other measurements of attitudes, social networks,
personal background information, and many other aspects to refine and extend our understanding of hackers. Such studies could specify and detail many additional characteristics in a more precise way. The large fraction of college-educated hackers in this
study, for example, rendered the educational achievement variable close to a constant.
To better assess the impact of varying educational backgrounds, future studies could
ask respondents what their study subject is or was and what type of college they attend(ed). The same is true for the measures of employment, to name but a second aspect. It would be interesting to know the exact profession of respondents and whether
and how their occupations are related to their hacking activities.
Although the paper-pencil format was, for several practical reasons, the only viable design of the present study, this format was not ideally suited for the assessment of
165

the great diversity of different types of hackers. Electronic computer-based surveys
should be designed for future studies because they allow more complex path dependencies and, thus, are better able to capture the wide range of different types of hackers. For example, such surveys could begin by asking respondents whether they identify themselves as white or as black hat hackers and then ask different sets of questions
for each type of hacker. Studies operating with such measurement techniques would be
geared more specifically to the different types of hackers and would yield more detailed
and more encompassing results without systematically excluding any type of hackers.
Thereby, they could provide answers to pressing questions such as what exactly the
ratio between white hat and black hat hackers is, and what fraction of the hacking community is comprised by law-abiding white hat hackers and penetration testers. They
could also assess tensions within the hacking community by revealing the attitudes and
opinion white hat and black hat hackers have about each other. Our understanding of
the hacking underground is just beginning to evolve and such studies could add much
needed details.
While the survey format of the present study was appropriate to address the
main components and factors of the involvement decisions as proposed by rational
choice perspective, its ability to capture the intricacies of event decisions was limited.
Thus, aside from the various types of surveys suggested above, cybercrime researchers should also conduct more event specific studies or analyze forensic crime science
protocols to better capture and reconstruct the intricate decision-making processes that
are involved in such complex events.

166

Parallel to analyzing the various personality traits that influence the behavior of
hackers, cybercrime researchers should further begin to construct typologies of different
hacker profiles. The multitude of motives and skills that were confirmed by this study
suggests that a variety of different types of hackers exist. Researchers should attempt
to isolate prototypical types of hackers, construct typologies of the various types, collect
empirical evidence to ensure the included types of hackers are exhaustive and mutually
exclusive, and examine how exactly the various types of hackers differ from and relate
to each other.
The long list of future studies that were suggested in this last section calls to
mind that cyber-criminology is only just beginning to develop and that our knowledge
about cybercrime offenders remains fragmentary at best. The present study yielded
some important insights into the composition of the hacking underground and it shed
some light on the motivations and maturation processes of hackers. Nevertheless, it
was but one step toward the establishment of cyber-criminology as a distinct subfield of
criminological research. A long and difficult road is still ahead for this young field of criminological research.

167

APPENDIX A:
THE SURVEY QUESTIONNAIRE

168

Hacking Questionnaire 2008
Hello and thank you for your willingness to participate in this survey!
My name is Michael Bachmann and I am a student at University of Central Florida doing a study
about hackers and their hacking activities.
First, let me assure you that:
- The survey is strictly noncommercial and for my private dissertation research only.
- There are no institutions or organizations involved in or affiliated with this project.
- All information is completely anonymous, confidential and your participation is voluntary.
- Completion of this survey takes about 15 minutes.
- It is for research purposes only.
OK, let’s begin:
First, let me ask you some questions about your hacking activity in general.
01 Are you currently actively hacking?

□ Yes □ No

If you answered this question with no, please think back to when you were still hacking
when answering the questions below.

02 At approximately what age did you begin to become interested in hacking? ______
03 What were your initial motivations to become interested in hacking? (Mark all that apply)

□ Intellectual curiosity
□ Excitement, thrill, fun □ Experimentation
□ Status and prestige
□ Financial gain
□ Peer recognition
□ Political Ideology
□ Media attention
□ Protest against corporations
□ Self-concept boost
□ Feeling of power
□ Personal revenge
□ Other (specify) ___________________________________________________________

04 What would you say was your primary motivation? (select one answer from question 3)
______________________________________________
05 How long was the timeframe between your initial interest and your first actual hack?
______ Days

______ Weeks

______ Months

169

______ Years

06 What were the initial motivations for your first hack? (Please mark all that apply)

□ Intellectual curiosity
□ Excitement, thrill, fun □ Experimentation
□ Status and prestige
□ Financial gain
□ Peer recognition
□ Political Ideology
□ Media attention
□ Protest against corporations
□ Self-concept boost
□ Feeling of power
□ Personal revenge
□ Other (specify) ___________________________________________________________

07 What was the type of target of your first hack?

□ Private single host
□ Private network
□ Private website
□ Corporate single host
□ Corporate network
□ Corporate website
□ Non-profit single host
□ Non-profit network
□ Non-profit website
□ Government single host □ Government network □ Government website
□ Other (specify) __________________________________________________________

08 Based on which criteria did you select your first target?

□ Easy access
□ Reputation gain

□ Interesting information □ Profitable information
□ Antipathy
□ Other (specify) _____________

09 Were you employed when you first started to hack?

□ Yes, full-time

□ Yes, part-time

10 Was economic profit a motivation at all?

□ Yes, an important one

□ No

□ Yes, but not very important

□ No

11 What methods did you use in your first technical intrusion? (Please mark all that apply)

□ Footprinting
□ Ping sweeping
□ DNS zone transfer
□ Whois
□ Network mapping
□ Port scanning
□ Versatile scanning tools □ Vulnerability scanning □ RPC port/end-point dump
□ Packet sniffing
□ Session hijacking
□ Grinding Passwords
□ Password cracking
□ Password theft
□ Directory traversal/climbing
□ Buffer overflow
□ Format string
□ Resource mismatch
□ CGI
□ Root shell/kits
□ Keylogger
□ Spoofing
□ Bouncing
□ Source routing
□ Other (specify) ___________________________________________________________

12 Approximately how often have you attempted technical intrusions in your life? ______ times

170

13 How frequently do you attempt technical intrusions?
______ Per Week or

______ Per Month or

______ Per Year

14 Approximately how many of them are successful?

□ 0% □ 10% □ 20% □ 30% □ 40% □ 50% □ 60% □ 70% □ 80% □ 90% □ 100%

15 Have you ever used “WetWare” social methods? (Please mark all that apply)

□ Social Engineering
□ Shoulder Surfing

□ Deception
□ Impersonation

□ Bribery
□ Dumpster diving

16 Approximately how often have you attempted social methods in your life? ______ times
17 How frequently do you attempt social methods?
______ Per Week or

______ Per Month or

______ Per Year

18 Approximately how many of them are successful?

□ 0% □ 10% □ 20% □ 30% □ 40% □ 50% □ 60% □ 70% □ 80% □ 90% □ 100%

19 Have you ever distributed? (Please mark all that apply)

□ Trojan Horses
□ Viruses/Worms
□ None of the above

□ Adware/Spyware
□ Spam/Phishing mails
□ Other (specify) __________________________

20 Approximately how often have you distributed malware in your life? ______ times
21 How frequently do you distribute malware?
______ Per Week or

______ Per Month or

______ Per Year

22 How successful would you say you were with your distributions?

□

□

Not at all

□

□

□

□

Somewhat

□
Very

23 Approximately how long have you been hacking now? ______ Months

______ Years

24 Did you change friends and peers since your first hack to include more other hackers?

□ Yes, very much

□ Yes, somewhat

171

□ No

25 Would you say you have improved your skills since your first hacking attempts?

□ Yes, very much

□ Yes, somewhat

□ No

26 Has your hacking activity become more frequent since your first hacking attempts?

□ Yes, very much
□ No, it’s the same

□ Yes, somewhat
□ No, it’s less frequent

27 Have your motivations changed since your first hacking attempts?

□ Yes, very much

□ Yes, somewhat

□ No

28 Currently, what is your main motivation for hacking? (specify): _______________________
29 Have your targets changed since your first hacks?

□ Yes, very much

□ Yes, somewhat

□ No

30 Would you say your targets have become higher profile targets?

□ Yes, very much

□ Yes, somewhat

□ No

31 What is the type of target you currently prefer?

□ Private single host
□ Private network
□ Private website
□ Corporate single host
□ Corporate network
□ Corporate website
□ Non-profit single host
□ Non-profit network
□ Non-profit website
□ Government single host □ Government network □ Government website
□ Other (specify) __________________________________________________________

32 What are your current selection criteria for targets?

□ Easy access
□ Reputation

□ Interesting information □ Profitable information
□ Antipathy
□ Other (specify) _____________

33 What might be a reason for you to reject a potential target? (Please mark all that apply)

□ Unfamiliarity with architecture
□ No profitable information
□ None of the above

□ No interesting information
□ Sympathy with host
□ Other (specify) __________________________

34 Have your methods and tactics changed since your first hack?

□ Yes, very much

□ Yes, somewhat

172

□ No

35 What is your currently preferred method and tool for the different stages of a hack?
(Please write in what you prefer to use for each of the following stages)
Reconnaissance: __________________________________________________________
Gaining access:

__________________________________________________________

Persisting:

__________________________________________________________

Propagating:

__________________________________________________________

Paralyzing:

__________________________________________________________

Covering up:

__________________________________________________________

36 Do you follow a persistent pattern in all your hacks or do you vary or change your methods?

□

□

□

□

□

□

Very persistent pattern

□

Vary a lot

37 Do you have preferred tools you use in all your hacks or do you vary your software tools?

□

□

□

□

□

□

Always same tools

□

Vary tools a lot

38 Which methods do you employ to remain anonymous? (Please mark all that apply)

□ Spoofing
□ Bouncing
□ Source routing
□ Proxies
□ TOR, I2P, etc.
□ Tunnels, Covert channels
□ War driving
□ Public access points (e.g. Internet cafes)
□ Other (specify) __________________________________________________________

39 Is hacking a source of any income for you?

□ Yes, the main

□ Yes, but not the main

40 Have you ever thought about quitting hacking?

□ Yes, often

□ Yes, a few times

□ No
□ Never

41 If yes, what caused you to question continuing to hack? (Please mark all that apply)

□ Involved risk
□ Fear of detection
□ Legal alternatives

□ Involved effort
□ Income too irregular
□ Fear of apprehension □ Fear of prosecution
□ Other (specify) __________________________________

42 Are you more/less afraid of being traced, tracked, or apprehended then when you started?

□ More afraid

□ Less afraid

□ Unchanged
173

□ I was never afraid

43 How much time do you usually invest in preparation, planning, considering routes of action,
selection of techniques and tools?
______ Hours

______ Days ______ Weeks

______ Months

44 Do you spend more or less time for planning today than when you started to hack?

□ More time today

□ About the same time

□ Less time today

45 Please estimate the risk you run of being detected during/after a hack.

□

□

□

No risk at all

□

□

Some risk

□

46 Please estimate the risk you run of being apprehended during/after a hack.

□

□

□

No risk at all

□

Very high risk

□

□

Some risk

□

□

Very high risk

47 Approximately what percentage of attacks have you aborted because you were afraid of detection?

□ 0% □ 10% □ 20% □ 30% □ 40% □ 50% □ 60% □ 70% □ 80% □ 90% □ 100%

48 During the hacks, do you tend to focus more on the potential rewards or the potential risks?

□

□

□

Only on rewards

□

□

Balance both

□

□

Only on risks

49 Would you say you act rational and deliberate during your hacks or do you sometimes get
caught-up in the moment and act emotionally?

□

□

□

□

□

Very rational

□

□

Very emotional

Now, please indicate how much you agree or disagree to the following statements
(you are almost finished):
50 I always try to avoid situations involving a risk of getting into trouble.

□

□

□

□

□

Strongly disagree

□

□

Strongly agree

51 I always play it safe even when it means occasionally losing out on a good opportunity.

□

□

□

□

□

Strongly disagree
52 I am a cautious person who generally avoids risks.

□

□

□

□

Strongly agree

□

□

Strongly disagree

□

□

□

Strongly agree

174

53 I am rather bold and fearless in my actions.

□

□

□

□

□

Strongly disagree
54 I am generally cautious when trying something new.

□

□

□

□

□

Strongly disagree

□

□

□

□

Strongly disagree

□

□

□

□

□

Strongly disagree

□
□

□

□

□

□

Strongly disagree

□

□

□

□

Strongly disagree

□

□

□

□

□

Strongly agree

59 I enjoy solving problems that require hard thinking.

□

□

Strongly agree

58 I prefer complex to simple problems.

□

□

Strongly agree

57 Thinking hard and for a long time about something gives me little satisfaction.

□

□

Strongly agree

56 I don’t reason well under pressure.

□

□

Strongly agree

55 I usually have clear, explainable reasons for my decisions.

□

□

Strongly agree

□

□

□

Strongly disagree

□

□

Strongly agree

60 Using my gut-feelings usually works well for me in working out problems in my life.

□

□

□

□

□

Strongly disagree

□

□

□

□

Strongly disagree

□

□

□

□

Strongly disagree

□

□

□

□

Strongly agree

63 I often go by my instincts when deciding on a course of action.

□

□

Strongly agree

62 I like to rely on my intuitive impressions.

□

□

Strongly agree

61 I trust my initial feelings about situations.

□

□

□

□

□

Strongly disagree

□

□

Strongly agree

175

64 I don’t think it is a good idea to rely on one’s intuition for important decisions.

□

□

□

□

□

□

Strongly disagree

□

Strongly agree

Finally, let me ask you some general questions about yourself.
65 What is your gender?

□ Male

□ Female

66 What year were you born? 19__
67 What is the last grade or class you completed in school?

□ None, or grades 1-8
□ High school incomplete (grades 9-11)
□ High school graduate (grade 12 or GED certificate)
□ Business, Technical, or vocational school AFTER high school
□ Some college, no 4-year degree
□ College graduate (B.S., B.A., or other 4-year degree)
□ Post-graduate training/professional school after college (Master's degree or Ph.D.)

68 Are you, yourself, of Hispanic or Latino origin or descent, such as Mexican, Puerto Rican,
Cuban, or some other Spanish background?

□ Yes

□ No

69 What is your race?

□ White □ Black

□ Asian

□ Other mixed race: (specify) ___________________

70 Are you married, living as married, divorced, separated, widowed, or never been married?

□ Never been married
□ Divorced

□ Living as married
□ Separated

□ Married
□ Widowed

71 Are you now employed full-time, part-time, retired, or not employed for pay?

□ Employed full-time □ Employed part-time
□ Not employed for pay □ Retired

72 Are you also a full- or part-time student?

□ Yes, full-time

□ Yes, part-time

□ No

Thank you very much for your time! Have a nice day and enjoy the rest of the conference.

176

APPENDIX B:
THE SURVEY CONSENT FORM

177

What Makes Them Click? Hacking Experiences and the
Decisions Involved in Hacking
Michael Bachmann, Principal Investigator
Informed Consent Form
I am a sociology Ph.D. student at the University of Central Florida (UCF). This
survey is part of my dissertation research which explores general hacking experiences
and the involved decision-making processes. Your participation will address these issues and involves you and other consenting attendees of this conference.
You will be asked questions about your hacking experiences, your motivations,
methods, and decision-making processes. Altogether, the survey takes only around fifteen minutes.
The primary purpose of this study is to further our understanding and knowledge
of hackers and the hacking community, a group that has so far been largely ignored by
sociological researchers. The study does not include any potential risk for participants
because it contains no identifiers, is completely anonymous, and asks only general
questions about hacking activity. It does not collect any specific illegal or selfincriminating information.
You are free not to answer any question you choose not to answer. You are free
to end the survey at any time. Deciding not to answer questions or to end the survey
altogether will not affect your status as an attendant of this conference in any way. Your
participation in this research is completely voluntary.
This research is governed by a privacy certification that prohibits the researcher
from disclosing any individual information to any outside party, including other attendees
of this conference. All information you provide in the survey is completely anonymous
and will be kept in the strictest confidence by the researcher. The researcher cannot divulge any information and no information can be linked specifically to you.
Questions
Please feel free to ask any questions you want before you fill out the survey or at any
time during the survey. If you feel that I have not adequately answered all your questions, contact my faculty supervisor Dr. Harold J. Corzine (407) 823-2202. You may call
that number but be sure to identify yourself as a participant in the hacking study.

178

Your consent
I understand the basic procedure of this study and am aware that I may discontinue participation at any time. I hereby consent to participate in a survey of about fifteen minutes
that focuses on general hacking experiences.
I understand that this study is anonymous, confidential, strictly non-commercial, and for
dissertation research purposes only. Only the researchers will have access to the questionnaire. No other person or organization, including other attendees at this conference
will have access to my questionnaire. I understand that I must be at least 18 years old
to participate in this research project. I understand that my participation or nonparticipation in this study has no bearing one way or the other on my status as an attendee of this conference. I understand that no compensation can be provided for my participation in this study.
If you believe you have been injured during participation in this research project, you
may file a claim against the State of Florida by filing a claim with the University of Central Florida’s Insurance Coordinator, Purchasing Department, 4000 Central Florida Boulevard, Suite 360, Orlando, FL 32816, (407) 823-2661. University of Central Florida is
an agency of the State of Florida and the university’s and the state’s liability for personal
injury or property damage is extremely limited under Florida law. Accordingly, the university’s and the state’s ability to compensate you for any personal injury or property
damage suffered during this research project is very limited.
Information regarding your rights as a research volunteer may be obtained from:
Barbara Ward, CIM
Institutional Review Board (IRB) Coordinator
University of Central Florida (UCF)
Office of Research & Commercialization
12201 Research Parkway, Suite 501
Orlando, Florida 32826-3246
Telephone: (407) 823-2901

You do not need to sign this consent form. Submission of a completed questionnaire
constitutes your consent to participate in this research project.

179

APPENDIX C:
THE IRB APPROVAL LETTER

180

181

GLOSSARY
ARPANET
A network created in 1969 by the US Defense Department's Advanced Projects
Research Agency (ARPA) to develop a system of data communications for scientific and military operations. ARPANET adopted the TCP/IP communications
standard, which defines data transfer on the Internet today.
Black Hats
Hackers (more correctly called crackers) who engage in destructive computer
exploits, motivated by greed, revenge, sabotage, blackmail, that can result in
harm to property and/or to people.
Bouncing
Bouncing is a method similar to spoofing. The attacker uses proxy servers or
computers to relay, or bounce, their requests to the attacked machines and to
evoke the impression the attack originated from the address it was channeled
through.
Buffer
A region of memory set aside to temporarily store output or input data waiting to
be directed to a device.
Buffer Overflow
In computer security and programming, a buffer overflow, or buffer overrun, is a
programming error which may result in a memory access exception and program
termination, or in the event of a malicious attack, a breach of system security.
182

CGI
CGI (Common Gateway Interface) denotes an interface program that enables an
Internet server to run external programs to perform specific functions, such as
creating “dynamic” content. Several broad-spectrum scanners exist that detect
known vulnerabilities in CGI programs. CGI attacks are among the most popular
attack methods on Internet servers. Most website defacements exploit the CG Interface.
Cracking
Gaining unauthorized access to computer systems to do harm or commit a crime.
Cracker
Cracker is a label for a malicious black hat hacker who breaks into remote computer systems without authorization in order to commit crimes.
Cybercrime
A crime related to technology, computers, and the Internet, resulting in harm to
property and/or persons.
Cyberspace
Composed of millions of interconnected computers, servers, routers, switches,
and fiber optic cables, cyberspace allows critical infrastructures to work. Today, it
is the nervous system of the global economy.
Directory Traversal/Climbing
Directory traversal or directory climbing, sometimes also referred to as backtracking, denotes one of the most common exploits on the Internet. By altering the

183

path of an URL, the attacker is reading files from a website that are not intended
to be shared over the Internet.
DNS
DNS stands for Domain Name System, a computer program running on web
servers that translates URL domain names into IP addresses in order to direct
easy-to-remember domain name requests (e.g. http://www.google.com) to the
respective binary IP address (216.239.39.99) of the domain host.
DNS Zone Transfer
DNS, or Domain Name System, is a program that is used by DNS servers to
translate user-friendly Uniform Resource Locator (URL) domain names (e.g.
http://www.google.com) into the respective binary IP address (in the case of
http://www.google.com the number 216.239.39.99) of the domain host. When receiving such a request, DNS servers look for the domain name in their database
and direct the request to the IP address associated with that domain name. Most
of the queries that are received by Domain Name Systems request translations
of URL domain names into IP addresses. The functionality of the Domain Name
System, however, is not limited to such translations. One additional feature of
DNS servers is the so-called “Zone Transfer”. The Zone Transfer feature allows
downloading the entire table of names and IP addresses contained within a network from the DNS database. Downloading this table enables the hacker to effectively map a network. Typically, the information contained in the database reveals the IP addresses of hosts that are likely to be active as well as the location
of important servers and sometimes even user names. DNS Zone Transfers are
184

typically used to spy out lower profile and lesser-secured target networks because many professional network administrators disable the Zone Transfer functions of their DNS servers or ensure that no compromising information is sent out
through such transfers.
Dumpster Diving
Dumpster diving is a slang term for searching through somebody’s garbage. For
attackers, this search can be a legal and powerful source of information because
many people do not destroy sensitive information before their disposal due to social taboos.
Footprinting
Sometimes also referred to as fingerprinting, is a term that can have two meanings. Sometimes footprinting is equated with reconnaissance tactics. In this understanding, all different kinds of activities aimed at gathering information about
computer systems, networks, or websites, and the entities they belong to are
subsumed under this term. When equated with the reconnaissance phase, footprinting commonly involves a range of technical methods, such as ping sweeps,
port scans, operation system identifications, network enumerations, registrar
queries, and many others. The second, more common understanding of the term
is of a much more narrow scope and denotes one specific activity, namely, the
identification of a particular operation system through the interpretation of specific
responses of a computer system that reveal the identity of that particular system.
In order to trigger the system to reveal identifying responses, the attacker submits unexpected combinations of code to the target. While most systems respond
185

in the same way to correct response requests, they rarely respond the same way
to requests that the programmers could not anticipate. A Windows Vista operation system, for instance, responds to faulty data requests in a way that is distinctively different from a Mac OS, a UNIX OS, a prior Windows OS, or any other
kind or operation system. Thus, footprinting techniques allow the hacker to identify the exact host computer system or platform he or she is targeting and allows
him or her to gear the attack towards the specific vulnerabilities of that particular
system. Footprinting techniques allow the hacker to identify the exact host computer system or platform he or she is targeting and allows him or her to gear the
attack towards the specific vulnerabilities of that particular system. The method is
used for attacks on individual hosts and networks as well as websites, because it
also allows the hacker to identify weaknesses and points-of-entry in the operation
system of a server that hosts the targeted website. Footprinting is used by hackers of different skill levels. Whereas lesser skilled attackers will resort to readily
available footprinting tools, attackers having the necessary programming knowledge typically prefer to use their own, custom-made footprinting requests.
Format String
Format strings denote an attack method similar to buffer overflows. By programming input with special formatting codes, attackers can override memory areas
on a target host and use this memory for crashing the program, get access to information from different locations in the computer’s memory, or even for breaking
into the host system.

186

Grinding Passwords
The act of remotely guessing login passwords is oftentimes referred to as grinding passwords. The attacker continually logs on remote machines and guesses a
different password each time. Alternatively, the hacker can attempt to log onto
many different machines with the same commonly used password.
Hacker
A person who enjoys learning the details of computer systems and how to stretch
their capabilities. In the public discourse, hacker is often used synonymously with
the term cracker.
Honeypot
In computer terminology, a honeypot is a trap set to detect, deflect, or in some
way counteract attempts of unauthorized intrusions.
I2P
I2P, originally also termed the Invisible Internet Project, is an anonymous network that can be used for anonymous surfing and to transfer files anonymously.
ICMP
ICMP stands for “Internet Control Message Protocol”, an integral part of the Internet Protocol that handles error and control messages. Routers and hosts use
ICMP to send reports of problems about data packets back to the originating
source.
IP address
Internet Protocol, or IP, addresses uniquely identify a computer accessible over a
TCP/IP-based network or the internet in a four-part numerical format.
187

Keylogger
Once an attacker has successfully gained access to a target system, he typically
installs packet sniffers to record network communications and keyloggers to
record all keystrokes users of that system enter on the keyboard. Several versions of keyloggers exist for special purposes such as, for example, recording
only passwords.
LAN
Local Area Network: A user-owned and operated data transmission network that
typically connects a number of communicating devices such as computers, routers, gateways, terminals, or printers located within a small geographical area.
LANs are typically distinguished from WANs, Wide Area Networks such as the
Internet, which span very large geographical areas or even the entire world.
Network Mapping
While ping sweeps provide the attacker with the addresses of active remote
hosts, they do not provide any information about the networks these hosts are located in. In order to explore and document the layout of a network, its topology,
traffic paths, structures, services, and its performance between nodes, attackers
typically use additional tools. Among the oldest and simplest of these tools is traceroute, a utility that displays the times and locations of intermediate stops along
the route data packets travel to reach their destination. Modern network mappers
oftentimes combine a vast array of powerful and versatile utilities that allow users
to identify a multitude of network details, among them active hosts and open
ports, running applications and services, the rules of firewalls within the network,
188

and even the contents of data packets sent within the network. While network
scanning is often done with such versatile scanning tools, it is not limited to only
such utilities. Typically network mapping tools are used in attacks against networks or specific hosts located within those networks. The basic skill level required to perform a network scan is not high, but network scans of more experienced hackers can be sophisticated and performed with advanced and customized tools. Three main subcategories of network scans exist that were asked
separately in the questionnaire: port scanners, versatile scanning tools, and vulnerability scanners.
Packet
A piece of data of fixed or variable size that is sent through a communication
network like the Internet. A message is typically broken up into packets before it
is sent over a network.
Packet Sniffing
Packet sniffing is the computer equivalent of wire-tapping a telephone communication. Sniffing programs disable the filter of Ethernet cards, thus allowing them
to monitor all communication within a network.
Password Cracking
Password cracking is the act of decrypting and recovering the passwords stored
on a computer system. Every system has to store passwords to be able to authenticate users. Password crack utilities are used to decipher the encryption and
reveal the account passwords to the attacker.

189

Password Theft
Password theft is one of the simplest and oldest hacking methods. Once the legitimate username and password are known, the attacker can exploit the user account without having to worry about any countermeasures such as firewalls or intrusion detection systems because they cannot distinguish between the legitimate user and the attacker.
Phreaking
Phreaking is a slang term coined to describe the activity of a subculture of people
who study, experiment with, or explore telephone systems, the equipment of telephone companies, and systems connected to public telephone networks. Oftentimes, phreaking denotes the act of breaking into a telephone network for the
purpose of making free calls or to charge calls to another person's account.
Ping
Ping is the short version for Packet Internet Groper, a utility that sends out data
packets to check whether a remote host is currently connected to a LAN network
or the Internet and to test the quality of the connection to that remote host. The
utility sends out ICMP “echo request” packets to the target host and listens for
ICMP “echo response” replies. The tool then estimates the round-trip time between sending the request and receiving the response and records any packet
loss in a statistical summary.
Ping Sweeping
The term “ping sweeping,” sometimes also referred to as Internet Control Message Protocol (ICMP) sweep, denotes a basic network scanning technique used
190

to determine the IP addresses of active, or live, hosts within a network. Whereas
single ping commands are used to determine the network location and availability
of one particular remote computer host, a ping sweep consists of ICMP “echo request” packets that are send to multiple hosts within one IP address range or entire network. All currently available hosts within the scanned range will reveal
their availability through returned “echo replies”, which are typically summarized
by the ping sweeping tool in a list of all active IP addresses within the specified
address range. Ping sweeps are typically used in attacks against networks, but
they can also be used to identify one specific target host within a network. They
do not require a high level of sophistication, are among the dated and slower
network scanning methods, and suffer from two main drawbacks. Ping sweeps
are detectable and will alert professional network technicians of an imminent attack. Also, many network administrators sometimes simply block ICMP echo requests being sent from their networks, in which cases the attacker then has to
resort to scanning open TCP ports. Since they bear some danger of compromising the attack and reveal only limited amounts of information, ping sweeps are
commonly used by lesser skilled hackers and in combination with other network
scanning methods. They are contained in many popular network scanners such
as nmap, nessus, and others, and their execution does not require a high level of
sophistication. To the contrary, the execution of a ping sweep is relatively simple
and merely requires the attacker to know the range of the targeted IP addresses.

191

Port
In the context of TCP/IP and UDP based computer networks, the term port refers
to an end point of a logical service connection that is numerically identified between 0 and 65536. For example, port 80 is used for HTTP traffic.
Port Scanning
Just as network scanners, port scanners are commonly used by network administrators to test the security of their networks and by hackers to compromise it. In
order to understand how exactly a port scan works and why many hacking exploits rely on open ports, it is import to know what a port is. Ports are one of the
main two components of the common Internet protocol TCP/IP. Within the
TCP/IP protocol, the destination of a data packet is determined by two components, the destination address of the computer the packet is sent to and the destination port for the particular type of data packet. Even though firewalls close the
majority of ports on most computers, specific ports or ranges of ports have to remain open for services and applications that are running on a host to be able to
receive incoming information. Two vulnerabilities, in particular, are associated
with open ports: security concerns associated with the program responsible for
delivering the service and with the operating system running on the host. Port
scanners allow an attacker to scan a range of listening ports on a target host for
ports that are open and can be used as entry points to this host. Port scanners
are commonly used in a variety of different hacking attacks, and they do not require a high level of sophistication. The successful exploitation of open ports is a
more difficult undertaking than their mere detection in a scan.
192

Proxy
In the context of computer networks, proxies are gateways that relay one Internet
session to another. Many proxies can be used for anonymizing purposes.
Reconnaissance
Reconnaissance in the context of a hacking attack usually consists of several
pre-test phases that are performed prior to the actual attack. The purpose of
these pre-tests is to gather as much information about the target computer host,
network, or website and its owners and administrating technicians as possible.
During the reconnaissance phase, the hacker attempts to gather initial information, determine the network range, identify active hosts and clients, discover active ports and access points, fingerprint the operating systems in use, uncover
the active services on ports, and combine all gathered intelligence into a map of
the network.
Resource Mismatch
Resource mismatches are one of the most common Denial of Service attacks.
They utilize the circumstance that some connections use significant more resources on one side of the connection than on the other. An attacker can exploit
such a connection to effectively shut down the target host.
Root shell/kits
On UNIX systems, only the “root” user, the equivalent to a Windows “Administrator,” has complete control over the machine. Several root shell or root kit exploits
exist that attempt to obtain a shell prompt with root privileges so as to allow the
attacker to enter and execute any command on the target system.
193

RPC
Remote Procedure Calls are application-to-application network requests designed to execute commands on remote host computers.
RPC port/end-point dump
RPC port/end-point dump scans typically supplement port scans by listing all
RPC (Remote Procedure Calls) services running on a system.
Session Hijacking
The term session hijacking refers to the exploitation of a valid computer session
to gain unauthorized access to information or services on a computer system.
The attacker commandeers a TCP session from a legitimate user after this user
has achieved authentication, thereby removing the need for the attacker to authenticate himself.
Shoulder Surfing
A slang word used among hackers to describe an attack method in which the attacker simply monitors the moment the victim enters passwords or other sensitive information on their keyboard.
Social Engineering
The term social engineering encompasses all attempts of attackers to establish
and subvert trust relationships with their victims or to predict the behavior of their
victims. Once such a relationship is established, the attacker tricks the victim into
revealing information or performing an action, such as a password reset, for example, that can then be used in the attack. A subcategory of social engineering
methods are so-called reversed social engineering methods, in which attackers
194

use their expertise to induce their victims to reveal sensitive information when
seeking their help and advice.
Source Routing
Source routing denotes a technique that allows a sender of a data packet to specify the route this packet takes trough a network. Source routing can be used in
hacking attacks to reach otherwise unreachable targets through intermediate
computers that are connected to the target.
Spoofing
Spoofing is a method attackers use to pretend they are somebody else. This is
accomplished by creating TCP/IP packets using somebody else’s IP address as
the source IP address.
TCP/IP
Stands for “Transmission Control Protocol / Internet Protocol” and summarizes a
collection of Internet communication protocols between computers that together
form the basis for most communications on the Internet.
TOR
Tor (The Onion Router) is a free software implementation of so-called secondgeneration onion routing. Tor is a system that is similar to I2P in that it enables its
users to communicate anonymously on the Internet.
Tunnel
Methods used by hackers to circumvent firewalls and evade intrusion detection
systems.

195

UDP
Or “User Datagram Protocol,” denotes a connectionless information transport
protocol that offers a limited amount of services when information is exchanged
over the Internet Protocol (IP). It is a faster alternative to the TCP protocol. It
does not guarantee delivery or provide sequencing of packets, but is favored for
time-sensitive data such as video or audio streams.
URL
URL, or Uniform Resource Locator, is an address that can uniquely specify any
Internet resource or file. It consists of the protocol, the domain, and the name of
the file.
Versatile Scanning Tools
Versatile scanning tools are software tools that combine a variety of different
scanning utilities in one application. A large number of free and commercial
scanning tools are available today. As is the case with many other hacking tools,
many of these applications are also used for legitimate purposes by system administrators to verify the configurations of hosts in their networks. Versatile scanning tools appeal to attackers because they conveniently combine several of the
mapping and scanning functions mentioned earlier. One of the most popular versatile scanning tools is Nmap, a software application that is more than a simple
scanner. Nmap allows scanning for RPC (remote procedure calls) services on a
target machine, sending decoy scans with fake source addresses, sending scans
with different timing options to avoid detection, and identifying a computer’s operating system via fingerprinting, to name but a few functions. Versatile scanning
196

tools are among the essential hacker tools and are used by hackers of all skill levels.
Vulnerability Scanning
Vulnerability scanning denotes the act of searching for and mapping systems for
weaknesses in an application, computer, or network. This search is typically performed by automated vulnerability scanners that combine many functions, such
as active network scanning functions and vulnerability scanning functions, in one
application. Vulnerability scanners evaluate several types of vulnerabilities including general system weaknesses such as faulty operating system code, faulty
application code, or faulty configurations.
War Driving
War driving is the act of driving through an area in a moving vehicle with a WiFiequipped computer to detect wireless networks that permit the attacker to access
the Internet anonymously.
Wetware
A slang word commonly used among hackers and computer experts to signify all
methods of programming a brain, as opposed to software. It denotes all social
engineering methods.
Whois
Another method for finding out general information about networks, hosts, and
their owners are whois requests. Whois is both a searchable public database
system of domain names and their registrars and a query utility that provides detailed ownership information about the domain name holders. The whois data197

base system contains information on how to contact the registrar and the technical administrators as well as information on the domain name servers, the registration and expiration date, and the dates the domain was updated. The system
was originally developed as a method for system and network administrators to
research contact information of other administrators, ISPs, and certified computer
incident response teams, and to determine the availability of domain names. Law
enforcement agencies also commonly utilize whois queries to lookup domain
owners in their combat against online copyright infringements, racial discrimination, hatred, violence, and child pornography. Undoubtedly, the whois system has
many positive uses that contribute to a higher user confidence in the Internet as
a reliable and efficient communication medium, but it can also be abused to
gather intelligence for hacking attacks, especially for attacks on websites. The
whois system is not restricted to network administrators, but allows everybody to
place queries, regardless of their motivations or intentions. In recent years, the
unrestricted access to the whois system has led to increasing abuses by email
spammers, who use the system to acquire email addresses for their spam databases. This development that has caused many web site hosts to “lock” their
domains in order to prevent spammers from accessing their email addresses,
thereby rendering the system less useful than it used to be.

198

LIST OF REFERENCES
Adam, A. E. (2004). Hacking into Hacking: Gender and the Hacker Phenomenon. Computers and Society, 32(7), 3-18
Aguila, N. (2008). The Fifteen Greatest Hacking Exploits: The Birth Of Hacking. Retrieved March 16, 2008, from
http://www.tomshardware.com/2008/03/14/the_fifteen_greatest_hacking_exploits
/index.html
Akers, R. L. (1990). Rational Choice, Deterrence, and Social Learning Theory in Criminology: The Path Not Taken. The Journal of Criminal Law & Criminology, 81(3),
653-676
Akers, R. L., & Sellers, C. S. (2004). Criminological Theories: Introduction, Evaluation,
and Application (4 ed.). Los Angeles: Roxbury.
Alaganandam, H., Mittal, P., Singh, A., & Fleizach, C. (2005). Cybercriminal Activity.
Retrieved February 5, 2008, from http://www.cs.ucsd.edu/~cfleizac/WhiteTeamCyberCrime.pdf
Alexander, S. (2005). Cyber hackers' motives shift to money. Retrieved May 20, 2008,
from http://www.detnews.com/2005/technology/0509/20/A04-320608.htm
Allan, G. J. B. (1976). Ordinal-scaled variables and multivariate analysis: comment on
Hawkes. American Journal of Sociology, 81(6), 1498-15000
Bachmann, M. (2008). Book Review: Clifford, R. D. (2006). Cybercrime: The Investigation, Prosecution, and Defense of a Computer-Related Crime. International Criminal Justice Review, 18(2). DOI 10.1177/1057567708319404
199

Badham, J. (Writer) (1983). War Games. In H. Schneider (Producer). USA: MGM/UA
Entertainment.
Beccaria, C. (1764). Of Crimes and Punishments (D. Young, Trans.). Indianapolis, Ind:
Hackett.
Bednarz, A. (2004). Profiling Cybercriminals: A Promising but Immature Science. Retrieved May 03, 2008, from
http://www.networkworld.com/supp/2004/cybercrime/112904profile.html
Bement, A. L., Ward, W. E., Carlson, L. T., Frase, M. J., & Fecso, R. S. (2004). Women,
Minorities, and Persons with Disabilities in Science and Engineering. Arlington,
VA: National Science Foundation, Division of Science Resources Statistics,
Document Number 04317)
Bentham, J. (1789). An Introduction to the Principles of Morals and Legislation. London:
T. Payne.
Bequai, A. (1999). Cyber-Crime: the US Experience. Computers and Security, 18(1),
16-18
Blake, C. (2000). Casting the Runes. Paper presented at the British Computer Society
Information Security Specialist Group seminar, London.
Boudreau, M. C., Gefen, D., & Straub, D. W. (2001). Validation in Information Systems
Research: A State-of-the-art Assessment. MIS Quarterly, 11(1), 1-16
Bouffard, J. A. (2007). Predicting Differences in the Perceived Relevance of Crime's
Costs and Benefits in a Test of Rational Choice Theory. International Journal of
Offender Therapy and Comparative Criminology, 51(4), 461-485

200

Brenner, S. (2006). Defining Cybercrime: A Review of State and Federal Law. In R. D.
Clifford (Ed.), Cybercrime: The Investigation, Prosecution, and Defense of a
Computer-Related Crime (pp. 13-94). Durham, NC: Carolina Academic Press.
Cacioppo, J. T., & Petty, R. E. (1982). The need for cognition. Journal of Personality
and Social Psychology, 42(1), 116-131
Cacioppo, J. T., Petty, R. E., Kao, C. F., & Rodriguez, R. (1986). Central and peripheral
routes to persuasion: An individual difference perspective. Journal of Personality
and Social Psychology, 51(5), 1032-1043
Cammell, D. (Writer) (1977). Demon Seed. In H. Jaffe (Producer). USA: United Artists.
Capeller, W. (2001). Not Such a Neat Net: Some Comments on Virtual Criminality. Social and Legal Studies, 10, 229-242
Cartwright, T. (1618). A confutation of the Rhemists translation, glosses, and annotations on the New Testament a 1603.
Casey, E. (2004). Digital Evidence and Computer Crime: Forensic Science, Computers
and the Internet (2 ed.). San Diego, CA & London: Academic Press.
Castells, M. (2002). The internet galaxy: Reflections on the internet, business, and society. Oxford: Oxford University Press.
Chirillo, J. (2001). Hack Attacks Revealed: A Complete Reference with Custom Security
Hacking Toolkit. New York, NY: Wiley.
Clarke, R. V. (1992). Introduction. In R. V. Clarke (Ed.), Situational Crime Prevention:
Successful Case Studies. Albany, NY: Narrow and Heston.
Clarke, R. V. (1997). Situational Crime Prevention: Successful Case Studies (2 ed.). Albany, NY: Harrow and Heston.
201

Clarke, R. V. (2004). New Challenges for Research: Technology, Criminology and
Crime Science. In E. U. Savona (Ed.), Crime and Technology: New Frontiers for
Regulation, Law Enforcement and Research (pp. 97-104). Dordrecht, Netherlands: Springer.
Clarke, R. V., & Cornish, D. B. (1985). Modeling offender's decisions: A framework for
research and policy. In T. M. & N. Morris (Eds.), Crime and Justice (Vol. 6). Chicago: University of Chicago Press.
Clarke, R. V., & Cornish, D. B. (2001). Rational Choice. In R. Paternoster & R. Bachman (Eds.), Explaining Criminals and Crime: Essays in Contemporary Criminological Theory. Los Angeles: Roxbury.
Clarke, R. V., & Felson, M. K. (1993). Routine Activity and Rational Choice. Advances in
Criminological Theory (Vol. 5). New Brunswick, NJ: Transaction Press.
Clifford, R. D. (2006). Introduction. In R. D. Clifford (Ed.), Cybercrime: The Investigation,
Prosecution, and Defense of a Computer-Related Crime (2 ed.). Durham, NC:
Carolina Academic Press.
Cohen, L. E., & Felson, M. K. (1979). Social Change and Crime Rate Trends: A Routine
Activity Approach American Sociological Review, 44(4), 588-608
Computer Crime and Intellectual Property Section, C. D. (2002). Searching and seizing
computers and obtaining electronic evidence in criminal investigations. Retrieved. from
http://permanent.access.gpo.gov/websites/usdojgov/www.usdoj.gov/criminal/cyb
ercrime/s&smanual2002.htm#_VA_.

202

Cornish, D. B. (1994). The procedural analysis of offending and its relevance for situational prevention. In R. V. Clarke (Ed.), Crime Prevention Studies (Vol. 3). Monsey, NY: Criminal Justice Press.
Cornish, D. B., & Clarke, R. V. (1987). Understanding Crime Displacement: An Application of Rational Choice Theory. Criminology, 25(4), 933-948
Cornish, D. B., & Clarke, R. V. (1989). Crime specialisation, crime displacement and rational choice theory. In H. Wegener, F. Losel & J. Haisch (Eds.), Criminal behavior and the justice system: Psychological perspective (pp. 103-117). New York,
NY: Springer.
Cornish, D. B., & Clarke, R. V. (Eds.). (1986). The Reasoning Criminal. New York, NY:
Springer.
Cornish, D. B., & Smith, M., J. (2003). Theory for Practice in Situational Crime Prevention (Vol. 16). Monsey, NY: Criminal Justice Press.
Couper, M. P. (2000). Web Surveys: A Review of Issues and Approaches. Public Opinion Quarterly, 64(4), 464-494
Cramer, D. (2003). A cautionary tale of two statistics: Partial correlation and standardized partial regression. Journal of Psychology, 137(5), 507-511
Cronenberg, D. (Writer) (1999). eXistenZ. In D. Cronenberg, A. Hamori & R. Lantos
(Producer). Canada: Alliance Atlantis.
Curran, K., Morrissey, C., Fagan, C., Murphy, C., O'Donnell, B., Firzpatrick, G., et al.
(2005). Monitoring hacker activity with a Honeynet. International Journal of Network Management, 15(2), 123-134

203

D'Arcy, J. P. (2007). The Misuse of Information Systems: The Impact of Security Countermeasures. New York, NY: Lfb Scholarly Pub.
Dahlback, O. (1990). Personality and risk-taking. Personality and Individual Differences,
11(12), 1235-1242
Damsell, K. (2003). Ethical Hackers' Test for Weakness. The Globe and Mail, p. B1,
De Hann, W., & Vos, J. (2003). A crying shame: The over-rationalized conception of
man in the rational choice perspective. Theoretical Criminology, 7(1), 29-54
Dinello, D. (2005). Technophobia!: Science Fiction Visions of Posthuman Technology.
Austin, TX: University of Texas Press.
Dowland, P., Furnell, S., Illingworth, H., & Reynolds, P. (1999). Computer Crime and
Abuse: A Survey of Public Attitudes and Awareness. Computers and Security,
18(8), 715-726
Dulebohn, J. H. (2002). An investigationof the determinants of investment risk behavior
in employer-sponsored retirement plans. Journal of Management 28(1), 3-26
Earnest, T. L. (2003). Deterrence Dynamics and Gender: Extending the Deterrence/Rational Choice Model. Unpublished Dissertation, Mississippi State U, Oxford, MI.
Edgerton, D. (1995). Technophobia then and now. Nature, 376(6542), 653-654
Epstein, S. (2003). Cognitive-experiential self-theory of personality. In T. Millon & M. J.
Lerner (Eds.), Comprehensive Handbook of Psychology, Volume 5: Personality
and Social Psychology (Vol. 5, pp. 159-184). Hoboken, NJ: Wiley.

204

Epstein, S., Pacini, R., Denes-Raj, V., & Heier, H. (1996). Individual differences in intuitive-experiential and analytical-rational thinking styles. Journal of Personality and
Social Psychology, 71(2), 390-405
Erickson, J. (2008). Hacking: The Art of Exploitation (2 ed.). San Francisco, CA: No
Starch Press.
Erickson, K., & Howard, P. N. (2007). A Case of Mistaken Identity? News Accounts of
Hacker, Consumer, and Organizational Responsibility for Compromised Digital
Records. Journal of Computer-Mediated Communication, 12(4), 1229-1247
Ericsson, K., & Simon, H. (1980). Verbal Reports as Data. Psychological Review, 87(3),
215-251
Ericsson, K., & Simon, H. (1984). Protocol Analysis: Verbal Reports as Data. Cambridge, MA: The MIT Press.
Felson, M. K. (2000). The routine activity approach as a general social theory. In S.
Simpson (Ed.), Of crime and criminality: The use of theory in everyday life. Thousand Oaks, CA: Sage.
Festinger, L. (1957). A theory of cognitive dissonance. Evanston, IL: Row & Peterson.
Finch, J. (1987). The Vignette Technique in Survey Research. Sociology, 21(1), 105114
Fisher, R. J. (1993). Social desirability bias and the validity of indirect questioning. Journal of Consumer Research, 20(2), 303-315
Fiske, S. T., & Taylor, S. E. (1984). Social cognition (1 ed.). Reading, MA: AddisonWesley.

205

Freaky. (2004). ShmooCon. Retrieved Mai 01, 2008, from
http://www.hackwire.com/index.php?action=news&catid=11
Furnell, S. (2002). Cybercrime: Vandalizing the information society. London: Addison
Wesley.
Gibson, W. (1988). Mona Lisa Overdrive. Toronto: Bantam Books, Spectra.
Gomez-Mejia, L. R., & Balkin, D. B. (1989). Effectiveness of individual and aggregate
compensation strategies Industrial Relations, 28(3), 431-445
Gordon, L. A., Loeb, M. P., Lucyshyn, W., & Richardson, R. (2005). Computer Crime
and Security Survey: CSI & FBI
Gordon, L. A., Loeb, M. P., Lucyshyn, W., & Richardson, R. (2006). Computer Crime
and Security Survey: CSI & FBI
Gordon, S., & Ford, R. (2006). On the definition and classification of cybercrime. Journal
in Computer Virology, 2(1), 13-20
Gosling, S. D., Vazier, S., Srivastava, S., & John, O. P. (2004). Should we Trust WebBased Studies? A Comparative Analysis of Six Preconceptions About Internet
Questionnaires. American Psychologist, 59(2), 93-104
Gottfredson, M., & Hirschi, T. (1990). A General Theory of Crime. Palo Alto, CA: Stanford University Press.
Grabosky, P. N. (2001). Virtual criminality: Old wine in new bottles? Social and Legal
Studies, 10(2), 243-249
Grabosky, P. N., & Smith, R. (1998). Crime in the digital age. Sydney: Federation Press.
Graham, P. (2004). Hackers and Painters: Big Ideas from the Computer Age. Sebastopol, CA: O'Reilly Media.
206

Grecs. (2008). ShmooCon 2008 Infosec Conference Event. Retrieved April 25, 2008,
from http://www.novainfosecportal.com/2008/02/18/shmoocon-2008-infosecconference-event-saturday/
Groves, R. M., Fowler, F. J., Couper, M. P., Lepkowski, J., M., Singer, E., & Tourangeau, R. (2004). Survey Methodology. Hoboken, NJ: Wiley.
Gupta, A. K., & Govindarajan, V. (1984). Business unit strategy, managerial characteristics, and business unit effectiveness at strategy implementation. Academy of
Management Journal, 27(1), 25-41
Hafner, K., & Lyon, M. (1998). Where Wizards Stay Up Late: The Origins Of The Internet. New York, NY: Simon & Schuster.
Hair, J. F., Anderson, R. E., Tatham, R. L., & Black, W. C. (1998). Multivariate Data
Analysis. Englewood Cliffs, NJ: Prentice Hall.
Handley, S. J., Newstead, S., E., & Wright, H. (2000). Rational and Experiential Thinking: A Study of the REI. In R. Riding & S. G. Rayner (Eds.), International Perspectives on Individual Differences: Cognitive Styles (Vol. 1). Stamford, CT: Ablex
Publishing.
Harrington, S. J. (1996). The Effect of Codes of Ethics. and Personal Denial of Responsibility on Computer Abuse Judgments and Intentions. MIS Quarterly, 20(3), 257278
Harvey, D. (1989). The Condition of Postmodernity. Oxford: Blackwell.
Hawkes, R. K. (1971). The multivariate analysis of ordinal measures. American Journal
of Sociology, 76(5), 908-925

207

Hechter, M., & Kanazawa, S. (1997). Sociological rational choice theory. Annual Review
of Sociology, 23, 191-214
Heckathorn, D. D. (1997). Respondent-Driven Sampling: A New Approach to the Study
of Hidden Populations. Social Problems, 44(2), 174-199
Howell, B. A. (2007). Real-World Problems of Virtual Crime. In J. M. Balkin, J. Grimmelmann, E. Katz, N. Kozlovski, S. Wagman & T. Zarsky (Eds.), Cybercrime:
Digital Cops in a Networked Environment. New York, NY: New York University
Press.
Howitt, P. (Writer) (2001). AntiTrust. In D. Hoberman, A. Amritraj, C. O. Erickson & J.
Chasman (Producer). USA: Metro-Goldwyn-Mayer (MGM).
Jaishankar, K. (2007). Cyber Criminology: Evolving a Novel Discipline with a New Journal. International Journal of Cyber Criminology, 1(1), 1-6
Jaishankar, K. (2008). Space Transition Theory of Cyber Crimes. In F. Schmalleger &
M. Pittaro (Eds.), Crimes of the Internet (pp. 281-283). Upper Saddle River, NJ:
Pearson.
Jewkes, Y. (2006). Comment on the book 'Cyber crime and Society by Majid Yar. Retrieved September 09, 2007, from
http://www.sagepub.co.uk/booksProdDesc.nav?prodId=Book227351
Johnson, B. (2008). Nato says cyber warfare poses as great a threat as a missile attack. Retrieved May 02, 2008, from
http://www.guardian.co.uk/technology/2008/mar/06/hitechcrime.uksecurity
Johnston, P. (2008). Tories want new cybercrime police unit. Retrieved March 07, 2008,
from http://www.crime-research.org/news/06.03.2008/3236/
208

Jordan, T., & Taylor, P. A. (1998). A Sociology of Hackers. The Sociological Review,
46(4), 757-780
Joseph, J. (2003). Cyberstalking: An International Perspective. In Y. Jewkes (Ed.),
Dot.cons: Crime, Deviance and Identity on the Internet. Cullompton: Willan.
Kanellis, P., Kiountouzis, E., Kolokotronis, N., & Martakos, D. (2006). Digital Crime and
Forensic Science in Cyberspace. Hershey, PA: Idea Group.
Kerlinger, F. N. (1986). Foundations of Behavioral Research (3 ed.). New York, NY: Rinehart & Winston.
Kilger, M., Arkin, O., & Stutzman, J. (2004). Profiling. In Honeypotproject (Ed.), Know
your enemy: learning about security threats (2 ed.). Boston: Addison Wesley.
Kim, J.-O. (1978). Multivariate analysis of ordinal variables revisited. American Journal
of Sociology, 84(2), 448-456
Kim, J.-O., & Mueller, C. W. (1978). Introduction to Factor Analysis: What It Is and How
To Do It (Quantitative Applications in the Social Sciences). Newbury Park, CA:
Sage.
Koops, B.-J., & Brenner, S. (2006). Cybercrime Jurisdiction - An Introduction. In B.-J.
Koops & S. Brenner (Eds.), Cybercrime and Jurisdiction (pp. 1-9). The Hague:
TMC Asser Press.
Kovacich, G. (1999). Hackers: Freedom Fighters of the 21st Century. Computers and
Security, 18(7), 573-576
Koziol, J., Litchfield, D., Aitel, D., Anley, C., Eren, S., Neel, M., et al. (2004). The Shellcoder's Handbook: Discovering and Exploiting Security Holes (2 ed.). Indianapolis, IN: Wiley.
209

Kozlovic, A. K. (2003). Technophobic themes in pre-1990 computer films. Science as
Culture, 12(3), 341-373
Krone, T. (2005). Hacking motives. Canberra: Australian Institute of Criminology
Krosnick, J., & Fabrigar, L. (1997). Designing Rating Scales for Effective Measurement
in Surveys. In L. Lyberg, P. Biemer, M. Collins, E. de Leeuw, C. Dippo, N.
Schwarz & D. Trewin (Eds.), Survey Measurement and Process Quality (pp. 141164). New York, NY: Wiley.
Kubrick, S. (Writer) (1969). 2001: A Space Odyssey. In S. Kubrick (Producer). UK/USA:
Metro-Goldwyn-Mayer (MGM).
Kurzweil, R. (2005). The Law of Accelerating Returns. In C. Teuscher & D. Hofstadter
(Eds.), Alan Turing: Life and Legacy of a Great Thinker. Berlin: Springer.
Lakhani, K. R., & Wolf, R. G. (2003). Why Hackers Do What They Do: Understanding
Motivation and Effort in Free/Open Source Software Projects: SSRN.
Landler, M., & Markoff, J. (2007). Digital Fears Emerge After Data Siege in Estonia. The
New York Times. Retrieved August 25, 2007, from
http://www.nytimes.com/2007/05/29/technology/29estonia.html?pagewanted=1&
ei=5070&en=15ee9940d96714da&ex=1188187200
Lemos, R. (2006). Mass defacement puts Turkey on the map. SecurityFocus. Retrieved August 29, 2007, from http://www.securityfocus.com/brief/212
Lenk, K. (1997). The Challenge of Cyberspatial Forms of Human Interaction to Teritorrial Governance and Policing. In B. D. Loader (Ed.), The Governance of Cyberspace: Politics, Technology and Global Restructuring. London: Routledge.

210

Levitt, M. (2007). Worldwide Email Usage 2007–2011 Forecast: Resurgence of Spam
Takes Its Toll. Retrieved August 25, 2007, from
http://www.idc.com/getdoc.jsp?containerId=206038
Levy, S. (1984). Hackers: Heroes of the Computer Revolution. New York: Doubleday.
Leyshon, A., Webb, P., French, S., Thrift, N., & Crewe, L. (2005). On the Reproduction
of the Musical Economy after the Internet. Media, Culture & Society, 27(2), 177209
Li, F. (2006). What is e-business: How the Internet Transforms Organizations. Malden,
MA: Wiley-Blackwell.
Lisberger, S. (Writer) (1982). Tron. In D. Kushner (Producer). USA: Buena Vista Pictures.
Littlewood, A. (2003). Cyberporn and moral panic: an evaluation of press reactions to
pornography on the internet. Library and Information Research, 27(86), 8-18
Longo, R. (Writer) (1995). Johnny Mnemonic. In S. Ahrenberg, D. Carmody, V. Hamburg & R. Lantos (Producer). USA: Sony Pictures.
MacCrimmon, K. R., & Wehrung, D. A. (1990). Characteristics of risk taking executives.
Management Science, 36(4), 422-435
Mann, D., & Sutton, M. (1998). NetCrime. More Change in the Organisation of Thieving.
British Journal of Criminology, 38(2), 210-229
Marshall, L. (2004). The Effects of Piracy upon the Music Industry: A Case Study of
Bootlegging. Media, Culture & Society, 26(2), 163-181
Matsuo, H., McIntyre, K. P., Taomazic, T., & Katz, B. (2004). The Online Survey: Its
Contributions and Potential Problems [Electronic Version]. ASA Secion on Sur211

vey Research Methods, 1, 3998-4000. Retrieved Mai, 01 2008, from
http://www2.bc.edu/~mcintykc/files/Jsm2004OnlinePaper.pdf
McClintock, R. (1999). Educating America for the 21st Century: A Strategic Plan for
Educational Leadership January 2000 through December 2004 Version 2.1
[Electronic Version], 1-25. Retrieved April 01, 2008, from
http://www.ilt.columbia.edu/publications/docs/ILT_Plan_new.pdf
Mitchell, W. (1995). City of Bits. Cambridge, MA: MIT Press.
Mitnick, K. D., & Simon, W. L. (2005). The Art of Intrusion: The Real Stories Behind the
Exploits of Hackers, Intruders & Deceivers. New York, NY: Wiley.
Mitnick, K. D., Simon, W. L., & Wozniak, S. (2002). The Art of Deception: Controlling the
Human Element of Security. New York, NY: Wiley.
Morris, R. (1970). Multiple correlation and ordinally scaled data. Social Forces, 48(3),
299-311
Naughton, J. (2000). A Brief History of the Future: The Origins of the Internet. London:
Phoenix.
NCIRC. (2008). NATO opens new centre of excellence on cyber defense. Retrieved
May 03, 2008, from http://www.nato.int/docu/update/2008/05-may/e0514a.html
Negroponte, N. (1996). Being Digital (First Vintage Books ed.). New York, NY: Random
House.
Newman, G. R., & Clarke, R. V. (2003). Superhighway Robbery: Crime Prevention and
E-commerce Crime. Cullompton: Willan Press.

212

Newsted, P. R., Chin, W., Ngwenyama, O., & Lee, A. (1996). Resolved: Surveys have
Outlived their Usefulness in IS Research. Paper presented at the Seventeenth
International Conference on Information Systems, Cleveland, OH.
Nosek, B. A., Banaji, M. R., & Greenwald, A. G. (2002). E-Research: Ethics, Security,
Design, and Control in Psychological Reseach on the Internet. Journal of Social
Issues, 58(1), 161-176
Nunnally, J. C. (1978). Psychometric Theory (2 ed.). New York, NY: McGraw-Hill.
Nuwere, E., & Chanoff, D. (2003). Hacker Cracker: A Journey from the Mean Streets of
Brooklyn to the Frontiers of Cyberspace. New York: HarperCollins Publishers.
O'Brian. (1979). The use of Pearson's R with ordinal data. American Sociological Review, 44(5), 851-857
Opp, K.-D. (1997). Limited rationality and crime. In G. R. Newman, R. V. Clarke & S. S.
Giora (Eds.), Rational Choice and Situational Crime Prevention: Theoretical
Foundations (pp. 47-63). Dartmouth UK: Ashgate.
Oxford English Dictionary (2008). Oxford: Oxford University Press.
Pacini, R., & Epstein, S. (1999). The relation of rational and experiential processing
styles to personality, basic beliefs, and the ratio-bias phenomenon. Journal of
Personality and Social Psychology, 76(6), 972-987
Painter, C. (2004). Threats to the Net: Trends and Law. In E. U. Savona (Ed.), Crime
and Technology: New Frontiers for Regulation, Law Enforcement and Research.
New York, NY: Springer.

213

Paternoster, R. (1989). Decisions to Participate in and Desist from Four Types of Common Delinquency: Deterrence and the Rational Choice Perspective. Law & Society Review, 23(1), 7-40
Pease, K. (2001). Crime futures and foresight: Challenging criminal behaviour in the information age. In D. S. Wall (Ed.), Crime and the Internet. London: Routledge.
Piquero, A. R., & Tibbets, S. G. (2002). Rational Choice and Criminal Behavior. Recent
Research and Future Challenges. New York, NY & London: Routledge.
Porter, S. R., & Whitcomb, M. E. (2003). The Impact of Contact Type on Web Survey
Response Rates. Public Opinion Quarterly, 67(4), 579-588
Reips, U. D. (2002). Standards for Internet-Based Experimenting. Experimental Psychology, 49(4), 243-256
Richtel, M., & Tedeschi, B. (2007). Online Sales Lose Steam. The New York Times. Retrieved September 09, 2007, from
http://www.nytimes.com/2007/06/17/technology/17ecom.html?ex=1339732800&e
n=60dbe8c0ab8fec6f&ei=5088
Schell, B. H., Dodge, J. L., & Moutsatsos, S. (2002). The Hacking of America: Who's
Doing It, Why, and How. New York, NY: Quorum.
Schell, B. H., & Martin, C. (2004). Cybercrime: A Reference Handbook. Santa Barbara,
CA: ABC-CLIO.
Schlegel, K., & Cohen, C. (2007). The Impact of Technology on Criminality. In J. M.
Byrne & D. J. Rebovich (Eds.), The New Technology of Crime, Law and Social
Control (pp. 23-49). Monsey, NY: Criminal Justice Press.

214

Schnell, R., Hill, P. B., & Esser, E. (1999). Methoden der Empirischen Sozialforschung
(6 ed.). Muenchen: Oldenbourg.
Schrutzki, R. (1989). Die Hacker Ethik. In Das Chaos Computer Buch (pp. 166-180).
Hamburg: Rowohlt.
Shelley, M. W. (1995). Frankenstein. In P. J. Hunter (Ed.), Frankenstein (Norton Critical
Editions ed.). New York, NY, London: W. W. Norton.
Shields, R. (Ed.). (1996). Cultures of the Internet: Virtual Spaces, Real Histories, Living
Bodies. London: Sage.
Shinder, D. L., & Tittel, E. (2002). Scene of the Cybercrime: Computer Forensics Handbook. Rockland, MA: Syngress Publishing.
Simon, L. (2004). Dark Light: Electricity and Anxiety from the Telegraph to the X-Ray.
Orlando, FL: Harcourt.
Slovic, P. (1972). Information processing, situation specificity and the generality of risktaking behavior. Journal of Personality and Social Psychology, 22(2), 128-134
Snyder, F. (2001). Sites of Criminality and Sites of Governance. Social and Legal Studies, 10, 251-256
Softley, I. (Writer) (1995). Hackers: Boot Up or Shut Up! In M. Peyser (Producer). USA:
Metro-Goldwyn-Mayer (MGM).
Sommer, R., & Sommer, B. (2002). A practical guide to behavioral research: Tools and
techniques (5 ed.). New York: Oxford University Press.
Straub, D. W. (1989). Validating instruments in MIS research. MIS Quarterly, 13(2),
255-276

215

Sykes, G., & Matza, D. (1957). Techniques of neutralization: a theory of delinquency.
American Sociological Review, 22, 664-670
Taylor, P. A. (1999). Hackers: Crime in the digital sublime. London and New York, NY:
Routledge.
Taylor, P. A. (2000). Hackers - Cyberpunks or Microserfs. In D. Thomas & B. Loader
(Eds.), Cybercrime: Law Enforcement, Security and Surveillance in the Information Age. London: Routledge.
Taylor, P. A. (2004). Hacktivism - Resistance is Fertile? In C. Sumner (Ed.), The Blackwell Companion to Criminology. Oxford: Blackwell.
Thomas, D. (2002). Hacker Culture. Minneapolis, MN: University of Minnesota Press.
Thomas, D. (2002). Notes from the Underground: Hackers as Watchdogs of Industry.
Online Journalism Review. Retrieved March 2, 2008, from
www.ojr.org/ojr/business/1017969515.php
Thomas, D., & Loader, B. D. (2000). Introduction - Cybercrime: law enforcement, security and surveillance in the information age. In D. Thomas & B. Loader (Eds.), Cybercrime: law enforcement, security and surveillance in the information age. London: Routledge.
Thompson, B. (2004). Exploratory and Confirmatory Factor Analysis: Understanding
Concepts and Applications. Washington, DC: American Psychological Association.
Tourney, C. P. (1992). The Moral Character of Mad Scientists: A Cultural Critique of
Science. Science, Technology, & Human Values, 17(4), 8

216

Trochim, W. M. K. (2002). The research methods knowledge base (2 ed.). Cincinnati,
OH: Atomic Dog Publishing.
Tudor, A. (1989). Monsters and Mad Scientists: A Cultural History of the Horror Film.
Oxford: Blackwell.
Turkle, S. (1995). Life on the Screen: Identity in the Age of the Internet. New York, NY:
Simon and Schuster.
Turner, D., Fossi, M., Johnson, E., Mack, T., Blackbird, J., Entwisle, S., et al. (2008).
Symantec Global Internet Security Threat Report: Trends for July - December 07.
Retrieved May 08, 2008, from
http://eval.symantec.com/mktginfo/enterprise/white_papers/bwhitepaper_internet_security_threat_report_xiii_04-2008.en-us.pdf
Twist, J. (2003). Cracking the hacker underground. Retrieved February 27, 2008, from
http://news.bbc.co.uk/1/low/technology/3246375.stm
Computer Fraud and Abuse Act, 1030 C.F.R. (1986).
Vallee, J. (2003). The Heart of the Internet: An Insider's View of the Origin and Promise
of the On-Line Revolution. Charlottesville, VA: Hampton Roads.
Voiskounsky, A., Babeva, J., & Smyslova, O. (2000). Attitudes towards Computer Hacking in Russia. In D. Thomas & B. Loader (Eds.), Cybercrime: Law Enforcement,
Security and Surveillance in the Information Age. London: Routledge.
Wall, D. S. (2001). Cybercrimes and the Internet. In D. S. Wall (Ed.), Crime and the Internet. London: Routledge.
Warner, J. (2007). U.S. Online Banking Population Grew 9.5% In 2006. Online Media
Daily. Retrieved September 20, 2007, from
217

http://publications.mediapost.com/index.cfm?fuseaction=Articles.showArticle&art
_aid=59070
Webster, J. (1996). Shaping women's work: Gender, employment and information technology. London: Longmans.
Wright, B. R. E., Caspi, A., Moffit, T. E., & Paternoster, R. (2004). Does the Perceived
Risk of Punishment Deter Criminally Prone Individuals? Rational Choice, SelfControl, and Crime. Journal of Research in Crime and Delinquency, 41(2), 180213
Yar, M. (2005a). Computer Hacking: Just Another Case of Juvenile Delinquency? The
Howard Journal, 44(4), 378-399
Yar, M. (2005b). The Novelty of 'Cybercrime': An Assessment in Light of Routine Activity Theory. European Journal of Criminology, 2(4), 407-427
Yar, M. (2006). Cybercrime and Society. London: Sage.
Zarrett, N. R., & Malanchuk, O. (2005). Who's computing? Gender and race differences
in young adults' decisions to pursue an information technology career. New Directions for Child and Adolescent Development, 2005(110), 65-84
Zone-H. (2007). Attackers Top List. Digital Attacks Archive. Retrieved September 09,
2007, from http://www.zone-h.org/component/option,com_topatt/Itemid,48/

218