Professional Documents
Culture Documents
Pekkas-Suggestions 05302023
Pekkas-Suggestions 05302023
Parti
Foriest
Dearden
Hawdon
Räsänen
etc
1
Abstract
This paper addresses the gap in the literature by empirically assessing a cross-national
comparison of institutional anomie theory and self-reported cyber offending. We used survey
data from Finland, Hungary, and the United States to examine the relationship between
institutional anomie and cybercrime. Each country represents different types of societies, which
makes them interesting cases for cross-national comparison. With Hungary representing a
conservative welfare regime, Finland a social democratic welfare regime, and the US a liberal
welfare regime, we predict that the more political regimes intervene in economic relations and
support non-economic institutions, the less applicable institutional anomie theory’s variables will
models suggest that IAT works best in the United States, where individualism, achievement,
fetishism of money increased the indices of cybercrime offending. IAT variables were less
useful in predicting cybercrime offending in the other nations. Yet, also in line with our
prediction, family has a strong external control function in Finland, protecting participants from
where a work-based welfare system suppresses family related values. Education has a mixed
effect on cybercrime offending. Thus, IAT is valuable for understanding cybercrime offending;
however, it should be examined in more countries with social institutions and individual value
2
Institutional Anomie Theory and Cybercrime: A Cross-country Comparison Analysis of
Technology has affected almost all aspects of our lives. From working online to interacting with
strangers or friends, many of our offline behaviors have moved, at least partially, online. Crime
is no different. Crimes are now occurring online or through some form of computer facilitation.
These crimes include online fraud, identity theft, spam, hacking, and cyberbullying (Ngo and
Jaishankar, 2017). For example, aAn American Gallup poll recently reported that 23% of
households were victimized by some type of digital financial crime, including unauthorized
credit card transactions, which is higher than many offline forms of theft such as vandalism
motivations behind computer-facilitated crimes. For example, routine activities and lifestyles
routine-activities theory are often used to explain the convergence in digital space between a
motivated offender, suitable target, and lack of a capable guardian (Choi and Lee, 2017;
Dearden, 2018; Hawdon et al., 2019; Holt and Bossler, 2009; Reyns et al., 2019; Weulen
Kranenbarg et al., 2019). Other common theories utilized in cybercriminology research include
self-control (Dearden and Parti, 2021; Donner et al., 2014; Holt et al., 2018; Nodeland, 2020;
Reisig et al., 2009) and social learning (Hawdon et al., 2019; Holt et al., 2010; Weulen
Kranenbarg et al., 2021). While some work also focuses on cybercrime and strain (e.g., Hay and
Ray, 2020), little work has been done relating to anomie and cybercrime (for an exception, see
3
Dearden et al., 2021). This paper attempts to address this shortfall in the literature by empirically
Literature Review
Merton (1938) defined anomie as the result of the differential access to legitimate, institutionally
defined means to achieve cultural defined goals. Merton suggested that these goals were
generally focused on economic success. This initial premise has been revised by a series of
theories. For example, general strain theory was an individual-level adaptation from Merton’s
original ideas, expanding the range of types of strain one can experience in society (Agnew,
1992). Of specific attention for this project, institutional anomie theory (IAT) was developed by
IAT is situated in anomie theories generally but draws from both Merton and Durkheim
to more comprehensively address the structural strains created in modern industrial societies,
especially the United States. One key focus of IAT is the pursuit of the American Dream.
According to IAT, crime rates are, at least in part, a function of the cultural pressures created
from the American Dream (Messner and Rosenfeld, 2012). Within the American Dream, four
primary values are defined. First, individuals should focus on achievement at any cost. This
partially leads to the second value of intense individualism. That is, the achievement at any cost
is not defined by some group success, only by individual success. Third, everyone should have
the same universal goal. Historically, other societies could have had varying goals, including
non-economic goals. Moreover, some societies could even have different opportunities to obtain
4
success (e.g., non-anomic society). However, in our modern economic system, this is no longer
true. Success is only defined by economic success. Finally, the fourth value identified by IAT is
the fetishism of money. While money has represented an intermediary for trade, helping to
establish systems whereby goods can be traded beyond in-kind trades, today money itself holds
some intrinsic value, at least according to IAT. Given this, money becomes the end goal of
individuals. Money becomes a means to show of success, with increasing money representing
increasing success. This fetishism of money leads to the massive holdings of individuals, without
any intention or even ability to realistically spend their vast amounts of money (Messner and
Rosenfeld, 2012).
IAT was not designed to explain crime directly, but rather through institutional pressures.
As mentioned above, the economic pursuits and pressures create an environment with poor
opportunity for other means of success. Historically, and likely in other systems or economies,
market institutions did not dominate other institutions. Institutions such as family, religion, and
politics all served a vital role in balancing opportunities and goals for individuals. Success in
such societies can be defined in other ways, including political aspirations, familial relations, or
religious devotion. In the American system, these institutions have been weakened, leading to
The non-dominant institutions struggle to maintain some degree of relevance, while also
acknowledging they are not as important as the dominant economic institution. According to
Messner and Rosenfeld (2012), this is done in three ways. All non-economic institutions are
devalued, serving only to perpetuate the values of the dominant institutions. Education provides
a good example, where the values of education are often only seen in relation to economic
principles, such as the return-on-investment of education, a focus on marketable majors, and the
5
general commodification of education (Lawrence and Sharma, 2002). Non-economic institutions
also must accommodate the economic institution whenever there is a conflict. This may be a
physical accommodation, such as the transition in business from generally not operating on
religious days. Religious institutions are then forced to respond and accommodate those who
work during typical service times through recorded sermons, televised opportunities, or even
sermons via a podcast. Finally, penetration occurs, where even non-economic institutions
attempt to “look like” the economic institution to showcase their relevance. For example,
politicians are often forced to adopt a singular focus on the “bottom line” or even shift economic
priorities away from the values of politics and more towards the values of the economic
As a structural theory, IAT has little to say about individual behavior. However, this has
not discouraged scholars from offering hypotheses and even creating individual-level
assessments based on IAT. Given that both original authors of IAT have discussed individual
assessment in follow-up articles (e.g., Messner et al., 2019), it appears that IAT has at least some
IAT can and has been assessed cross-nationally to consider the different institutional pressures
associated with both the system and age of the government. Anomie generally, and IAT
specifically, have a focus on later-staged market economies. Empirical work cross-nationally has
6
most often been tested using homicide data (Hughes et al., 2015; Kim and Pridemore, 2005;
Messner and Rosenfeld, 1997; Rogers and Pridemore, 2022). Other crimes have been considered
as well, although these studies generally focused on cross-county or cross-state variation using
data within the United States. Examples of these studies include applications of IAT to violent
crime (Piquero and Piquero, 1998), property crime (Chamlin and Cochran, 1995), white-collar
crime (Schoepfer and Piquero, 2006), and cybercrime (Dearden et al., 2021). Overall, these
accurate data and conceptual measurement of variables, complicate the discussion of the theory’s
Of specific interest to our focus, individual-level tests have often found support for IAT
(e.g., Dearden et al., 2021; Muftić, 2006; Rosenberger, 2016; Tuliao and Chen, 2019). For
example, using a sample of college students, Muftić (2006) found that individual measures
other non-economic institutions, finding that individuals who were less committed to family and
politics were more likely to cheat. Relating to cybercrime, using a survey of US adults, Dearden
and colleagues (2021) found a positive correlation between IAT and self-admitted cybercrime
relationship between IAT-identified beliefs and cybercrime. Individuals with the lowest and
highest levels of individual IAT beliefs were the most likely to have admitted to committing a
cybercrime.
For example, Hirtenlehner and colleagues (2013) used multi-level modelling to test the effects of
7
cultural forces and social institutions on morally dubious behaviors. These scholars found little
support for the “cultural imperative of the ‘American Dream’” across Europe. Consequently,
while support for IAT at the individual level remains fairly impressive, tests including
possible. Moreover, whenever possible, cross-national tests of IAT should include predictions
about the cross-national applicability of IAT for the nations included in the specific sample used
to conduct the test. Logically, IAT should work best in the United States since the theory is
based on the concept of the “American Dream.” While American culture has certainly
penetrated the world system, the extent to which the “American Dream” is held in other
countries undoubtedly varies. Similarly, the extent to which the economic system has come to
economic success that leads to criminal behavior according to IAT; rather, it is doing so at the
expense of all other means of success, believing that success must be achieved individually, and
valuing money intrinsically rather than as a tool for other means of success. As with other strain
theories, it is when these values are deeply held and success does not occur that crime ensues.
Thus, logically, IAT should apply in some countries more than it does in others.
Here, we purposefully include three nations that should provide variability in terms of the
extent to which IAT applies. We include the United States as the “baseline” nation since it is the
basis of the theory. We also include Finland and Hungary as these nations provide stark contrasts
to the United States in terms of their economic systems and the relationship of the state to the
citizenry vis-à-vis the welfare system. Consideration of these factors allow us to derive specific
8
hypotheses concerning the ability of IAT-informed variables to account for individual-level
We anticipate that the welfare regime type a nation adopts will influence the applicability
ofr IAT. Since IAT argues that crime is the product of “cultural pressures for the unrestrained
pursuit of monetary success and weak social control deriving from an imbalanced institutional
structure” (Messner and Rosenfeld 2012: 99), we predict that the more welfare regimes intervene
in economic relations and support for non-economic institutions, the less applicable IAT-related
variables will be. While it is true likely true that surrounding cultural circumstances contribute
here. This is why we relied Relying on Esping-Andersen’s (1990) widely used typology on
welfare state regimes. The , the three nations included in our study exemplify his three original
regimes, and social democratic regimes. These institutional differences have several implications
The American welfare model is an example of a liberal regime. Liberal regimes tend to
provide modest, means-tested assistance typically to the poor and working classes, and they
typically rely on market solutions to problems. The American system is based on a principle of
“minimal state interference” (Hass, 2006: 69), and the state becomes involved in welfare
provision only when the market, voluntary organizations, and the family fail to provide services
(see Gilbert, 2002). As an “enabling state” (Gilbert, 2002), the US state mandates private social
9
indirectly through tax incentives (Gilbert, 2002; Katz, 2002; Hacker, 2002). As a result, the US
social safety net is comparatively limited. In addition, there has always been a strong anti-statism
in the United States, and state intervention into economic affairs is generally unpopular
Next, Hungary is an example of a conservative welfare regime (see Aspalter et al., 2009).
Conservative regimes tend to encourage family-based assistance where the state intervenes only
after the family is unable to provide assistance. Conservative welfare state regimes are
distinguished by their “status differentiating” programs where benefits are earnings‐related and
geared toward maintaining existing social relations. The family and other institutions play an
important role in the provision of benefits, and the redistributive impact of transfer payments is
minimal (Esping-Andersen, 1990). Prior to the imposition of the Soviet system after WWII,
Bismarckian policies provided sickness and disability insurance as well as old age pensions
(Szikra, 2018). Then, under Soviet-style state socialism, social protection was extended to almost
the entire population and included near full-employment guarantees, cash-transfers, and price
subsidies. The Bismarkckian traditions returned in the post-Soviet era, and these policies largely
protected Hungary’s citizens from the harshest effects of the turbulent transition to a capitalist
democracy (Aspalter et al., 2009; Bohle and Greskovits, 2012; Szikra, 2018). This was indeed
the case in Hungary following the Soviet Union’s collapse, as the country increasingly adopted
the policies of a conservative welfare regime (Aspalter et al., 2009). However, these protections
started to erode as early as the 2010s when the old-age pension was partially privatized. In
addition, a radical departure from the European social model occurred during Prime Minister
Orbán’s second term when he argued that instead of a Western type of welfare state that is not
10
transformation resulted in changes to pension policies in 2010-2012 and family policy in 2010-
2017. These policy changes dramatically weakened social provisions (Szikra, 2018), but
Finally, Finland is a classic Nordic welfare state or social democratic regime (see
Esping-Andersen, 1990; Räsänen, 2006). Social democratic regimes attempt to avoid market-
based access to welfare services by adopting a universalistic system that offers high levels of
social protection in the form of universal coverage of old-age pensions, sickness insurance, child
allowance and parental leave, and occupational injury insurance (Greve, 2007; Lin, 2004).
Nordic welfare systems have extensive fiscal interventions in the labor market, and the marginal
tax rates are very high (Alesina and Glaeser, 2006). In addition, there is historical support for an
interventionist state and a tradition of “statism” in Finland (Lin, 2004; Midttun and Witoszek,
2010). In the language of IAT, the economic system is far less dominant in Finland than it is in
the United States. Other Finnish institutions are not as devalued, forced to accommodate the
Considering the welfare regimes in the three nations, we would predict that IAT would be
most applicable in the United States. American citizens are relatively unprotected from market
forces and other institutions are not widely supported by the welfare state. For example, parental
leave policies are far more limited in the US than in social democratic regimes (see Gilbert,
2004). Conversely, IAT should be far less applicable in Finland. Finnish citizens enjoy universal
welfare provisions that de-commodify welfare (Esping-Andersen, 1990), which is often reflected
11
countries they have family policies that provide generous support for child allowances and
Finally, we predict that IAT would apply moderately well in Hungary. While the polices
of conservative regimes provide less protection against market forces than those of social
democratic regimes, two factors can potentially counteract the effect of these policies on crime.
First, the focus on maintaining existing status relations should reduce the hyper-competitiveness
found in nations like the United States. Second, the clear recognition of the importance of non-
economic institutions such as the family implies that there has been less devaluation of these
institutions than seen in liberal regimes. Therefore, we predict that, if IAT applies in Hungary at
all, its applicability would be somewhat tempered, at least compared to the United States.
Methods
Sample
Samples of adults from Hungary, Finland, and the US were collected in March 2022.
The samples were selected from Dynata’s (formerly SSI) panels of respondents. Dynata, a
professional sampling and survey fielding company, uses random digit dialing, banner ads, and
other permission-based techniques to recruit participants to create their panel databases. Dynata
offers a small fee or reward to users who qualify for and participate in a survey. Online
such as sex, race, and ethnicity, have been found to yield similar results as random probability-
based samples (Simmons and Bobo, 2015; Weinberg et al., 2014; authors, forthcoming; although
MacInnis et al., 2018, offers a differing perspective). Demographically balanced panels ensure
one-time participation through IP address matching and eliminate speeders through attention
12
checks and mean-time comparisons with the overall sample (Evans and Manthur, 2005;
Wansink, 2001). Moreover, pre-panel interviews and participation incentives increase participant
interest and the overall response validity of the data (see Wansink, 2001).
Dynata provided the online samples to the researchers, and the surveys were designed
and hosted by a research team from universities in each of the three nations. The surveys were
designed by the research team in English and translated to Finnish and Hungarian by native
All samples are representative of their respective nation’s population in terms of age and
sex. Although the US sample is representative of the population for race and ethnicity as well,
we did not collect data on racial and ethnic origins in Europe due to privacy protection and data
regulatory rules (the European Data Protection Regulation does not allow collecting data about
race and ethnicity as these are considered sensitive data in European Union member states). 1
For consistency, we decided to omit racial and ethnic data from the current analysis entirely.
Individuals who failed to complete or sped through the survey were excluded from the
analysis. In total, 4,763 respondents started the survey (1,585 in Finland, 1,553 in Hungary, and
1,625 in the US). We dropped 486 speeders (167 in Finland, 110 in Hungary, and 209 in the US)
and 72 people (28 in Finland, 22 in Hungary, and 22 in the US) who did not complete the survey
from the sample. Another seven cases were dropped from the analysis because they were missing
a country code. In total, 4,205 participants remained in the final sample, out of which 1,390
respondents were from Finland, 1,421 were from Hungary, and 1,394 were from the US. In the
total sample 903 participants admitted to at least one cybercrime (in Finland 303, in Hungary 321
and in the US 279). Table 1 shows the demographic characteristics of the sample.
13
Measures
Dependent variable
behaviors during the 12 months preceding the survey. Counts and percentages of each offending
We combined the binary responses of each offending behavior to create a general index
of cyber offending for each country. For this general cyber offending variable, the minimum
value was 0, indicating no offending behavior within the 12 months prior to completing the
survey, and the maximum value was 10, indicating self-reporting all offending behaviors within
the 12 months before participating in the survey. Of all participants who reported cyber
offending (n=903), only 12.4% (n=112) reported only one offending behavior. All others
Independent variables
Fetishism of Money, Family, and Education, as our primary independent variable (see Table 3
for measures).
As stipulated by IAT, four cultural values constitute the foundation of the American
Rosenfeld, 2001). Respondents were asked questions on a 5-point Likert scale on separate
elements of the four cultural value systems. For instance, respondents were asked how much they
14
agreed with the following statements (5 representing strongly agree and 1 representing strongly
disagree): “I will sacrifice a lot of other things to have a lot of money,” and “I intend to do
Next, IAT proposes that social institutions regulate human conduct (Messner and
Rosenfeld, 2001). The institutions discussed in IAT are the economy, the political system
(polity), the family, and educational institutions. Respondents were asked 5-point Likert scale
question about their associations with two of these four institutions: family and education. For
estimate on a 5-point Likert scale (1 representing strongly agree and 5 representing strongly
disagree) how strongly they agree with statements such as “College education is important so I
can be a better person” and “College education is important so I can be a better citizen.” Factor
analyses were conducted for each subset of institution categories, producing a composite
measure of the two institutional categories. Separate scales were then created to represent family
and education. Higher scores indicate higher participation in or commitment to the institution. At
the same time, higher scores represent weaker commitment to the belief of the American Dream.
Cronbach’s alphas were calculated to determine internal consistency for each scale. Table 3
reports the 23-item IAT scale breakdown for the whole sample as well as the three participating
countries.
Control variables
15
We also include common variables used in analyses of cybercrime, including measures of
routine activities. To measure time spent on the darkweb, participants were asked, “How many
hours per week do you spend on the dark web,” and responses were recorded as a continuous
variable. In addition, we included a control for sex (male=1, other=0), and education level.
Education level was measured on a Guttman-like scale ranging from “Less than a high school
diploma,” “High school degree,” “Some college,” through “A college degree,” “A master's or
Analysis.
Several models investigated the effect of individual levels of IAT on general cyber
offending. As cyber offending, the dependent variable, was a count variable (i.e., an index of
binary responses to specific cyber offending in the past 12 months) and because the data were
regression is not restricted to the assumption that the variance is equal to the mean. Otherwise,
tThese models are similar to regression in that they allow correlations to be established, but they
Results
In the first negative binomial regression model, cyber offending was regressed on the
overall IAT index to examine the independent relationship between IAT and cyber offending in
all three countries (AIC=6245.51). Of the six IAT measures, four were significantly related to
16
Individualism and FoM increased the probability of cybercrime offending while Universalism
The second binominal regression model included all covariates with the general index of
IAT. This model was created to understand the multivariate relationship between IAT and
cyberoffending by using the above-mentioned control variables. Compared to the first model, the
model fit improved (AIC=5109.003). The results of this analysis and subsequent analyses are
reported in Table 4. First, a unit increase on the individualism scale resulted in an approximately
7% increase in cybercrime (IRR=1.073, p<0.001), and a unit increase on the FoM scale resulted
1-unit increase on the Universalism scale decreased the risk of cybercriminal activity by 10%
(IRR=.903, p<0.001). Achievement, Family, and Education were not significantly related to the
cyberoffending index.
The risk of committing a cybercrime increased by approximately 37% with every hour per week
increase of time spent on the darkweb (IRR=1.374, p<0.001); men committed 54% more
cybercrimes than did women (IRR=1.537, p<0.001); and people with the lowest (less than high
school) and highest education levels (PhD or professional degrees) committed online crimes
Next, we ran binominal regression on country samples separately. As Table 4 shows, the
model’s fit improved in all countries compared to the full model. In the Finnish sample, the
model was well fitted (AIC=1668.025), but the only IAT variable that was a statistically
significant predictor of cyberoffending was Family. A unit increase in the scale reduced the
17
probability of committing a cybercrime by 5% (IRR=.954, p<0.05), demonstrating the external
control function of family as an institution in Finland. Among the control variables, the results
were similar to the whole model. A one-hour-per-week increase of time spent on the darkweb
increased the risk of cyber offending by 28% (IRR=1.283, p<0.05), and men were more than
education, the highly educated and those with very low levels of education were also more likely
The model for the Hungarian sample also had a better fit compared to the previous model
(AIC= 1659.861). Here, however, none of the IAT indexes achieved statistical significance.
Among the control variables, education was not significant, but darkweb activity and sex were.
Similar to the full model, a one-hour-increase per week in time spent on the darkweb increased
the risk of committing cyberoffenses by approximately 21% (IRR=1.208, p<0.001). Men were
62% more likely to report cybercrime activities than were women (IRR=1.624, p<0.001). Table
Finally, compared to the full model, the model fits best in US sample (AIC= 1726.318).
Among the IAT indices, Family and Education were not statistically significant, however, a one-
(IRR=1.105, p<0.001), and a one-unit increase in Achievement increased the risk of committing
a one-unit increase of FoM was related to an 8% higher risk of offending (IRR=1.082, p<0.05).
Based on these results, it appears the institutions of family and education are so weak that they
fail to control individuals who intend to engage in cybercrimes. As for the controls, neither sex,
18
nor education level were significant predictors of offending; however, a one-hour increase per
week of time spent on the darkweb increased the probability of engaging in a cybercrime by
approximately 37% (IRR=1.367, p<0.001). For the IAT scales distribution per country sample,
see Table 5.
Discussion
Using survey data from Finland, Hungary, and the United States, the current study
provides a cross-national, micro-level test of the relationship between institutional anomie theory
and self-reported cyber offending. As expected, results from our full sample model indicate an
overall significant, positive association between particular IAT cultural values (Individualism
and Fetishism of Money) and self-reported cyber offending. Consistent with our predictions,
cross-nationally, IAT is most applicable in the United States—a liberal regime where the
economic social safety net is limited. In fact, the positive correlations between institutional
anomie theory’s cultural variables and cyber offending are only significant in the United States,
as results from our split-sample, within-group models, do not find a statistically significant
cyber offending in our conservative regime (Hungary) and social democratic regime (Finland).
concerning the relative ability of social institutions to control crime. According to IAT, the
criminogenic strain of economic institutions is not buffered, nor mediated by the otherwise
protective influence of noneconomic social institutions. As expected, our analyses indicate that,
in the United States, where we find institutional anomie theory to be most pronounced, the
family and educational institutions do not exert any independent influence on cybercrime. We do
19
find that involvement in and commitment to the family does discourage cyber offending in
Finland.
Taken together, the results presented here are generally in support of institutional anomie
theory, at least in the United States, but our study is consistent with other comparative analyses
that find the criminogenic strain of the American Dream to be uniquely American. This is
unsurprising as the United States is the original impetus from whence institutional anomie theory
is derived. Although we improve upon the extant literature by enhancing the generalizability of
offending, our study raises an important question: Why is Universalism, a core tenet of
institutional anomie theory that is expected to increase the probability of offending significantly
and negatively associated with crime? To account for this inconsistency, we propose Merton's
deviance typology (Murphy and Robinson, 2008), which offers various individual-level
adaptations to strain, including, but not limited to criminal behavior. Of these, it is Merton’s
“Conformity” adaptation that best explains why institutional anomie theory can sometimes
protect against offending. Merton’s Conformists are individuals who accept anomic cultural
goals but will only rely on conventional means to attain said goals (Murphy and Robinson,
2008). Consistent with this view, Merton’s Conformists can recognize the ubiquitousness of the
American Dream without legitimizing the anomic goals imposed upon them by society. Further,
whereas cultural values associated with institutional anomie theory embody egoism, goal
orientation, and materialism, Universalism does not assume the manifestation of IAT-related
values conducive to crime, per se. Instead, Merton’s universalist Conformists may select into the
American Dream due to an organic affinity for accountability, personal responsibility, and
altruistic achievement. For Conformists, who may be as invested in the conventional means as
20
they are inspired by the intense pressure to succeed in achieving conventional goals, the
American Dream might function as a source of informal social control. Thus, our discovery of a
significant inverse association between institutional anomie theory and cybercrime warrants
future directional tests of institutional anomie theory and prescribes an extension of its
Our study is not without limitations. Although the sample represented the three
participating nation states with respect to age and sex, the survey’s representativity is vulnerable
because survey participants in paid panels apply self-selection and tend to be more educated and
disproportionately from higher income brackets (Singer and Kulka, 2002). Second, although the
countries in this study represent liberal, conservative, and social democratic economic systems,
IAT must be further examined involving more countries with social institutions, cultural factors
and individual value systems different from that of the United StatesS’s. Third, more nuanced
scales must be applied to examine the effect of structural and individual factors that predict
engagement in cyber offenses. Although we found IAT being supported best in the US, the
nation where the concept of the American Dream was invented, we need to study how different
social, economic, and political systems and institutions induce cyber deviant activities. Despite
these limitations, our analyses suggest that IAT is a valuable theory for explaining cyber
Conclusion
theory. Over the years, it has been adopted to explain individual-level offending and applied to a
variety of crimes. Our analysis continues this tradition by using common individual-level
21
measures of the various cultural values and institutional variables to predict involvement in
cybercrime. Yet, we extend existing work by considering the theory’s applicability cross-
appropriate to consider how cross-national variation in the adoption of the American Dream can
account for cross-national variation in criminal behavior. We contribute to the limited body of
While we find support for IAT in the United States, we predicted and found that it would be less
useful in Hungary and Finland because of the welfare regimes found in these nations. Additional
theorizing and testing is needed to determine its utility, and we argue that considering the role
played by the welfare state in mitigating the dominance of the economic institution could be a
fruitful. Ultimately, a much larger sample of nations that varying on the extent to which they
adopt the American Dream is needed to truly assess IAT’s value in explaining cross-national
22
Notes
23
References
Agnew R (1992) Foundation for a general strain theory of crime and delinquency. Criminology 30(1):
47–88.
Alesina A and Glaeser EL (2006) Why are welfare states in the US and Europe so different?. Horizons
Aspalter C, Jinsoo K and Sojeung P (2009) Analysing the welfare state in Poland, the Czech
43(2):170–185.
Bohle D and Greskovits B (2012) The political economy of protest and patience in East-Central
Chamlin MB and Cochran JK (1995) Assessing Messner and Rosenfeld’s institutional anomie theory:
Choi K-S and Lee JR (2017) Theoretical analysis of cyber-interpersonal violence victimization and
offending using cyber-routine activities theory. Computers in Human Behavior 73: 394–402.
Dearden TE (2018) The conjunction fallacy in profiles of victims of homicide. Journal of Investigative
Dearden TE and Parti K (2021) Cybercrime, differential association, and self-control: knowledge
transmission through online social learning. American Journal of Criminal Justice 46(6): 935–
955.
Dearden TE, Parti K and Hawdon J (2021) Institutional anomie theory and cybercrime –cybercrime
and the American Dream, now available online. Journal of Contemporary Criminal Justice
37(3): 311–332.
24
Donner CM, Marcum CD, Jennings WG, et al. (2014) Low self-control and cybercrime: Exploring the
utility of the general theory of crime beyond digital piracy. Computers in Human Behavior 34:
165–172. https://doi.org/10.1016/j.chb.2014.01.040
Durkheim E (1969[1898]) Individualism and the intellectuals. In: Lukes S (trans) and Hamilton P (ed)
Esping-Andersen G (1990) The three worlds of welfare capitalism. Princeton, New Jersey: Princeton
University Press.
Evans J and Mathur A (2005) The value of online surveys. Internet Research 15(2): 195–219.
https://doi.org/10.1108/10662240510590360
Gilbert N (2002) Transformation of the welfare state: The silent surrender of public responsibility.
Greve B (2007) What characterise the Nordic welfare state model. Journal of social sciences 3(2): 43–
51.
Hacker JS (2002) The divided welfare state: The battle over public and private social benefits in the
Hawdon J, Bernatzky C and Costello M (2019) Cyber-routines, political attitudes, and exposure to
https://doi.org/10.1093/sf/soy115
Hay C and Ray K (2020) General strain theory and cybercrime. In: The Palgrave Handbook of
https://doi.org/10.1007/978-3-319-78440-3_21
25
Hirtenlehner H, Farrall S and Bacher J (2013) Culture, institutions, and morally dubious behaviors:
Testing some core propositions of the institutional-anomie theory. Deviant Behavior 34(4):
291–320.
Holt TJ and Bossler AM (2009) Examining the applicability of lifestyle-routine activities theory for
https://doi.org/10.1080/01639620701876577
Holt TJ, Burruss GW and Bossler AM (2010) Social learning and cyber-deviance: Examining the
importance of a full social learning model in the virtual world. Journal of Crime and Justice
Holt TJ, Van Wilsem J, Van de Weijer S, et al. (2018) Testing an integrated self-control and routine
Hughes LA, Schaible LM and Gibbs BR (2015) Economic dominance, the “American Dream,” and
homicide: A cross‐national test of institutional anomie theory. Sociological Inquiry 85(1): 100–
128.
Katz MB (2002) The price of citizenship: Redefining the American welfare state. Macmillan.
Kim S and Pridemore WA (2005) Poverty, socioeconomic change, institutional anomie, and homicide.
Lawrence S and Sharma U (2002) Commodification of education and academic labour – using the
677.
Lin K (2004) Sectors, agents and rationale: A study of the Scandinavian welfare states with
26
Long JS and Freese J (2006) Regression models for categorical dependent variables using Stata. 2nd
MacInnis B, Krosnick JA, Ho AS, et al. (2018) The accuracy of measurements with probability and
nonprobability survey samples: Replication and extension. Public Opinion Quarterly 82(4):
707–744. https://doi.org/10.1093/poq/nfy038
Merton RK (1938) Social structure and anomie. American Sociological Review 3(5): 672–682.
Messner SF and Rosenfeld R (1997) Political restraint of the market and levels of criminal homicide:
Messner SF and Rosenfeld R (2012) Crime and the American dream. Cengage Learning.
Messner SF and Rosenfeld R (2001) Crime and the American Dream (3rd ed.). Belmont, CA:
Wadsworth.
Messner SF, Rosenfeld R and Hövermann A (2019) Institutional anomie theory: An evolving research
program. In: Krohn MD, Hendrix N, Hall GP, et al. (eds) Handbook on Crime and Deviance,
20779-3_9
Midttun A and Witoszek N (2010) Introduction: The Nordic model – How sustainable or exportable is
it? In: Midttun A and Witoszek N (eds) The Nordic model: Is it sustainable and exportable?,
Muftić LR (2006) Advancing institutional anomie theory. International Journal of Offender Therapy
Murphy DS and Robinson MB (2008) The Maximizer: Clarifying Merton's theories of anomie and
27
Ngo F and Jaishankar K (2017) Commemorating a decade in existence of the international journal of
cyber criminology: A research agenda to advance the scholarship on cyber crime. International
Nodeland B (2020) The effects of self-control on the cyber victim-offender overlap. International
https://doi.org/10.52306/03020220onxt9834
Parti K, Dearden T and Hawdon J (Forthcoming) Perspectives of paid panel survey research in
cybercrime victimization and offending: Validity of global online market research sampling
and data collection. In: Graham R, Humer SG, Lee CS, et al. (eds) International Handbook of
Piquero A and Piquero NL (1998) On testing institutional anomie with varying specifications. Studies
Quadagno J and Street D (2005) Ideology and public policy: Antistatism in American welfare state
Reinhart RJ (2018) One in four Americans have experiences cybercrime. Gallup Politics.
Reisig MD, Pratt TC and Holtfreter K (2009) Perceived risk of internet theft victimization: Examining
the effects of social vulnerability and financial impulsivity. Criminal Justice and Behavior
Reyns BW, Fisher BS, Bossler AM, et al. (2019) Opportunity and self-control: Do they predict
multiple forms of online victimization? American Journal of Criminal Justice 44(1): 63–82.
https://doi.org/10.1007/s12103-018-9447-5
28
Rogers ML and Pridemore WA (2022) Not just another test of institutional anomie theory: Assessing
https://doi.org/10.1080/07418825.2022.2102535
Rosenberger JS (2016) Television consumption and institutional anomie theory. Sociological Focus
Räsänen P (2006) Consumption disparities in information society: Comparing the traditional and
digital divides in Finland. International Journal of Sociology and Social Policy 26(1–2): 48–62.
https://doi.org/10.1108/01443330610644425
Schoepfer A and Piquero NL (2006) Exploring white-collar crime and the American dream: A partial
https://doi.org/10.1016/j.jcrimjus.2006.03.008
Simmons AD and Bobo LD (2015) Can non-full-probability internet surveys yield useful data? A
comparison with full-probability face-to-face surveys in the domain of race and social
https://doi.org/10.1177/0081175015570096
Singer E and Kulka RA (2002) Paying respondents for survey participation. In: Ploeg MV, Moffitt RA
and Citro CF (eds) Studies of welfare populations: Data collections and research issues, pp.
Szikra D (2018) Welfare for the wealthy: The social policy of the Orbán-regime, 2010–
https://library.fes.de/pdf-files/bueros/budapest/14209.pdf
29
Tuliao KV and Chen C (2019) Economy and supervisors’ ethical values: Exploring the mediating role
Wansink B (2001) Editorial: The power of panels. Journal of Database Marketing and Customer
Weinberg JD, Freese J and McElhattan D (2014) Comparing data characteristics and results of an
Weulen Kranenbarg M, Holt TJ and Van Gelder JL (2019) Offending and victimization in the digital
https://doi.org/10.1080/01639625.2017.1411030
Weulen Kranenbarg M, Ruiter S and Van Gelder JL (2021) Do cyber-birds flock together? Comparing
https://doi.org/10.1177/1477370819849677
30
Table 1: Demographics and other characteristics of total sample and countries
Whole sample US sample Finnish sample Hungarian sample
Total Offender Total Offender Total Offenders Total Offenders
N=4205 s N=1394 s N=1390 N=303 N=1421 N=321
(100.00 N=903 (100.00%) N=279 (100.00% (100.00% (100.00% (100.00%)
%) (100.00 (100.00% ) ) )
%) )
Count measures
Sex
Male 2001 497 680 (50%) 154 627 158 688 185 (59%)
(48%) (56%) (56%) (46%) (54%) (49%)
Female 2131 385 690 (50%) 121 732 133 708 131 (41%)
(52%) (44%) (44%) (54%) (46%) (51%)
Non-binary 16 9 (1.0%) 5 (0.4%) 3 (1.0%) 8 (0.6%) 6 (2.0%) 3 (0.2%) 0 (0.0%)
(0.4%)
Race
White n/a n/a - - n/a n/a n/a n/a
Non-white n/a n/a - - n/a n/a n/a n/a
Education
Less than 328 (8%) 84 (9%) 65 (5%) 11 (4%) 164 52 (17%) 99 (7%) 21 (7%)
high school (12%)
High 1649 320 347 (25%) 55 (20%) 542 95 (31%) 760 170 (53%)
school (39%) (35%) (39%) (54%)
Some 631 137 311 (22%) 64 (23%) 198 48 (16%) 120 (9%) 25 (8%)
college (15%) (15%) (14%)
College 1006 222 428 (31%) 88 (32%) 267 63 (21%) 307 71 (22%)
degree (24%) (25%) (19%) (22%)
Master’s or 563 140 232 (17%) 61 (22%) 206 45 (15%) 124 (9%) 34 (11%)
professiona (13%) (16%) (15%)
l degree
Darkweb 576 240 308 (25%) 151 94 (7%) 42 (15%) 174 47 (16%)
(15%) (29%) (63%) (13%)
Continuous measures
Age N=4090/ 884/42.8 1357/45.9/ 274/35.1/ 1341/47.4 293/45.8/ 1385/47.1 317/46.7/1
Mean=46 /15.5/18- 17.5/18-88 12.1/18- /16.4/18- 15.3/19- /16/18-89 5.9/18-82
.8/ 85 85 87 77
SD=16.7/
Min-max
18-89
31
Table 2: Self-reported offending behavior 12 months preceding the survey – Whole sample;
Country samples
Total US Finnish Hungarian
offenders: offenders: offenders: offenders:
n; % of total n; % of US n; % of FI n; % of HU
sample sample sample sample
Types of offending behavior N=903; 22% N=279; 21% N=303; 22% N=321; 23%
Posted hurtful information about someone 355 (9%) 205 (15%) 73 (5%) 77 (6%)
on the internet
Threatened and insulted others through 201 (5%) 133 (10%) 48 (4%) 20 (1%)
email or instant messaging
Excluded someone from an online 592 (14%) 178 (13%) 201 (15%) 213 (15%)
community
Hacked into an unauthorized area of the 169 (4%) 108 (8%) 43 (3%) 18 (1%)
internet
Distributed malicious software 178 (4%) 127 (9%) 35 (3%) 16 (1%)
Illegally downloaded copyrighted files or 339 (8%) 131 (10%) 97 (7%) 111 (8%)
programs
Illegally uploaded copyrighted files or 224 (8%) 133 (10%) 54 (4%) 37 (3%)
programs
Used someone else’s personal information 178 (4%) 129 (9%) 31 (2%) 21 (2%)
on the internet without their permission
Purchased illegal drugs online or bought 172 (4%) 123 (9%) 36 (3%) 13 (1%)
prescriptions without a prescription on
online pharmacies or websites
Posted nude photos of someone else 167 (4%) 117 (9%) 32 (2%) 18 (1%)
without their permission
32
Table 3: Institutional Anomie Scales in the US, Finnish, and Hungarian samples
Total US sample FI sample HU sample
sample
IAT measure M SD M SD M SD M SD
Individualism (⍺=.772)
Being successful is more important than being 2.12 1.08 2.42 1.32 1.94 0.90 2.02 0.92
happy
I intend to do whatever it takes to have some of 2.09 1.14 2.52 1.29 1.97 1.50 1.77 0.91
the really expensive things in life
I expect to make as many sacrifices as are 2.40 1.16 2.83 1.30 2.02 1.01 2.35 1.02
necessary in order to advance my work/career
I expect to devote whatever time and energy it 2.48 1.22 2.94 1.31 2.24 1.15 2.25 1.05
takes to move up in my job/career
Being happy is more important than being 2.10 1.09 2.18 1.13 1.93 1.02 2.19 1.09
successful
Achievement (⍺=-0.367)
I will sacrifice a lot of other things to have a lot of 2.29 1.11 2.61 1.26 2.13 1.02 2.14 0.98
money
I don’t need help from others to succeed 2.86 1.09 3.01 1.21 2.77 1.01 2.80 1.01
I am getting/have gotten education because it’s 3.51 1.19 3.13 1.27 3.70 1.14 3.70 1.04
expected by my friends or parents
Success is measured by the amount of money a 2.50 1.19 2.57 1.30 2.44 1.16 2.49 1.11
person makes
I am getting a college education because it’s 3.83 1.21 3.38 1.31 3.99 1.16 4.12 1.01
expected by my parents
Universalism (⍺=0.626)
Anyone that works hard enough can be successful 3.52 1.11 3.76 1.12 3.52 1.09 3.29 1.08
You only have yourself to blame for your failures 2.97 1.11 3.37 1.14 2.67 1.06 2.86 1.02
in life
Fetishism of Money (⍺=0.667)
Having lots of money is one of my major goals in 2.55 1.18 2.95 1.29 2.37 1.13 2.35 1.02
life
Education is only helpful to get a good job 2.86 1.13 2.89 1.26 2.88 1.08 2.80 1.05
Education is useful because it helps you make 3.25 1.02 3.40 1.08 3.34 .94 3.02 .99
more money
Family (⍺=0.841)
How important to you is the opinion of your 2.59 1.28 3.36 1.21 2.09 1.13 2.32 1.11
mother?
How important to you is the opinion of your 2.50 1.29 3.43 1.26 2.00 1.07 2.09 1.00
father?
How important to you is the opinion of your 2.80 1.24 3.03 1.30 2.67 1.23 2.70 1.15
siblings?
How important to you is the opinion of your 2.44 1.19 2.70 1.22 2.34 1.21 2.30 1.12
friends?
Education (⍺=0.803)
A college education is important so I can be a 2.85 1.25 2.65 1.25 3.16 1.30 2.76 1.11
better person
A college education is important so I can be a 2.82 1.31 2.85 1.27 2.99 1.39 2.63 1.24
better spouse/parent
A college education is important so I can learn 2.95 1.34 2.57 1.22 3.31 1.39 2.95 1.30
33
about different cultures
A college education is important so I can be a 3.22 1.26 2.63 1.25 3.53 1.20 3.48 1.12
better citizen
34
Table 4: Regression models
Whole model Finland Hungary US
Goodness of Fit
AIC BIC AIC BIC AIC BIC AIC BIC
5109.00 5186.62 1668.02 1732.89 1659.86 1725.00
3 7 5 9 1 6 1725.784 1784.874
Parameter estimates
Parameter B Exp(B) B Exp(B) B Exp(B) B Exp(B)
Individualis 0.071** 1.073** 0.100** 1.105**
m * * 0.010 1.010 0.037 1.037 * *
Achievemen
t 0.015 1.015 0.025 1.025 -0.023 0.977 0.071* 1.073*
Universalis - -
m 0.102** 0.903** 0.243** 0.784**
* * -0.060 0.942 -0.044 0.957 * *
FoM 0.042* 1.043* -0.027 0.974 0.048 1.049 0.079* 1.082*
Family 0.001 1.001 -0.047* 0.954* 0.015 1.015 0.025 1.026
Education -0.011 0.989 0.000 1.000 0.019 1.019 -0.017 0.983
Darkweb 0.318** 1.374** 0.249** 1.283** 0.189** 1.208** 0.312** 1.367**
* * * * * * * *
Male 1.537** 0.704** 2.022** 0.485** 1.624**
0.43*** * * * * * 0.227 1.255
Less than
high school 0.306* 1.358* 0.482* 1.620* 0.311 1.365 0.282 1.326
High school -
0.322** 0.725** -0.386* 0.680* -0.052 0.949 -0.216 0.805
Some
college -0.097 0.908 -0.144 0.866 -0.422 0.656 0.075 1.078
College
degree -0.105 0.9 -0.003 0.997 0.000 1.000 -0.133 0.876
*p<0.05, **p<0.01, ***p<0.001
35
Table 5: IAT scale items’ distributions by country
Finland Hungary United States
N=1,086 N=1,109 N=696
IAT index IRR b p IRR b p IRR b p
Individualism 1.01 .01 .611 1.037 .037 .074 1.105 .100 <.001
Achievement 1.025 .025 .414 .977 -.023 .443 1.073 .071 <.05
Universalism .942 -.060 .112 .957 -.044 .209 .784 -.243 <.001
Fetishism of .974 -.027 .392 1.049 .048 .088 1.082 .079 <.05
Money
Family .954 -.047 <.05 1.015 .015 .444 1.026 .025 .180
Education .000 1.000 .979 1.019 .019 .288 .983 -.017 .326
Intercept .474 -.747 .206 .140 -1.966 <.01 .101 -2.293 <.001
36