Annotated Bibliography
malld614
ORIGINAL PAPER
Examining Systematic Crime Reporting Bias Across Three Immigrant Generations: Prevalence, Trends, and Divergence in Self-Reported and Official Reported Arrests
Bianca E. Bersani1 • Alex R. Piquero2
Published online: 16 July 2016 � Springer Science+Business Media New York 2016
Abstract Objective Mounting evidence reveals that foreign-born, first generation immigrants have significantly lower levels of criminal involvement compared to their US-born, second and
third-plus generation peers. This study investigates whether this finding is influenced by
differential crime reporting practices by testing for systematic crime reporting bias across
first, second, and third-plus generation immigrants.
Methods This study draws on data from the Pathways to Desistance Study, a longitudinal investigation of the transition from adolescence to young adulthood among a sample of
serious adolescent offenders. Self-reported and official reports of arrest are compared
longitudinally across ten waves of data spanning 7 years from adolescence into young
adulthood for nearly 1300 adjudicated males and females.
Results This study reveals a high degree of correspondence between self-reports of arrest and official reports of arrest when compared within groups distinguished by immigrant
generation. Longitudinal patterns of divergence, disaggregated by under-reporting and
over-reporting, in self- and official-reports of arrest indicated a very high degree of sim-
ilarity regardless of immigrant generation. We found no evidence of systematic crime
reporting bias among foreign-born, first generation immigrants compared to their US-born
peers.
Conclusions First generation immigrants are characterized by lower levels of offending that are not attributable to a differential tendency to under-report their involvement in
crime.
& Bianca E. Bersani [email protected]
Alex R. Piquero [email protected]
1 Department of Sociology, University of Massachusetts Boston, Boston, MA, USA
2 Ashbel Smith Professor of Criminology, Program in Criminology, University of Texas at Dallas, Richardson, TX, USA
123
J Quant Criminol (2017) 33:835–857 DOI 10.1007/s10940-016-9314-9
Keywords Immigration and crime � Crime reporting bias � Self-reported arrests � Official arrests � Longitudinal
Introduction
Estimates calculated by the US Census Bureau using data from the 2013 American
Community Survey (ACS) indicate that immigrants account for 13 % of the total US
population, increasing by 1.3 % alone between 2012 and 2013. The immigrant population
is projected to continue to grow, reaching 19 % of the total US population by 2050 (Passel
and Cohn 2008). The growth of the US immigrant population coincides with a renewed
interest into the extent to which immigration and crime are linked. At the community level,
research finds a negative relationship between immigrant concentration and crime, par-
ticularly violent crime, such that areas characterized by high levels of immigrant con-
centration are associated with less crime compared to areas with low immigrant
concentration levels (see e.g., Ewing et al. 2015; Ferraro 2016; Lee and Ramiro 2009;
Sampson et al. 2005). More recently, longitudinal evidence demonstrates a pattern
whereby communities experiencing increases in immigrant concentration (percent foreign-
born living in an area) over time experience decreases in crime over time (MacDonald
et al. 2013; Ousey and Kubrin 2014; Wadsworth 2010). Research conducted at the indi-
vidual level compliments these results with findings showing significantly lower rates of
involvement in crime among foreign-born, first generation immigrants compared to their
second and third-plus generation peers (Bersani 2014; Bersani et al. 2014; Morenoff and
Astor 2006; Rumbaut et al. 2006; Sampson et al. 2005). This pattern emerges in data drawn
from the general US population, immigrant destination cities, and from offender-based
samples. Though the empirical consensus demonstrating a lack of a relationship between
immigration and crime is, perhaps, surprising given the current tenor of discussion (see
Chaves 2008), 1 it is noteworthy that this same relationship has been documented in
empirical research for nearly a century (see Simes and Waters 2014).
Despite this ‘‘emerging scholarly consensus’’ (Lee and Ramiro 2009; see also Sampson
2008) on the (lack of an) immigration-crime nexus, an important omission from this line of
work is the consideration of potential systematic crime reporting bias, which may dis-
proportionately affect foreign-born immigrants and consequently affect the representation
of immigrant involvement in crime. That is, the lower levels of offending observed among
foreign-born, first generation immigrants may be due to differences in offending or to
1 Interestingly, Trump’s (2015) comments during his presidential candidacy regarding immigrants and the
crime and disease they bring to the United States are remarkably similar to early twentieth century com- ments about the biological inferiority of the immigrant population (the pollution of the purer ‘‘American Blood’’ Orebaugh 1929). Senator Henry Cabot Lodge speaks to the risks of immigration stating ‘‘…there is a limit to the capacity of any race for assimilating and elevating an inferior race; and when you begin to pour in unlimited numbers of people of alien or lower races of less social efficiency and less moral force, you are running the most frightful risk that a people can run. The lowering of a great race means not only its own decline, but that of civilization. … [We] are exposed to but a single danger, and that is by changing the quality of our race and citizenship through the wholesale infusion of races whose traditions and inheritances, whose thoughts and beliefs are wholly alien to ours, and with whom we have never assimilated or even been associated in the past. The danger has begun. … There lies the peril at the portals of our land; there is pressing the tide of unrestricted immigration. The time has certainly come, if not to stop, at least to check, to sift, and to restrict those immigrants’’ (1909: 264–266, The Restriction of Immigration).
836 J Quant Criminol (2017) 33:835–857
123
differences in the reporting of crime. 2 Historically, the limitations of official and self-report
data have long been recognized; however, as we explicate below, these limitations may be
compounded when examining immigrant-related crime. For instance, differential treatment
in the process of the administration of the law (affecting officially reported arrests) and the
complexities arising from social desirability reporting and deportation concerns (affecting
self-reported arrests) may over- or under-count, respectively, immigrant involvement in
crime.
This research investigates this important issue by assessing the relationship between
immigrant generational status and systematic crime reporting bias in three key ways. Using
data from a sample of high-risk, adjudicated youth, we first examine the prevalence of self-
reported offending, self-reported arrests and officially reported arrests and compare the
extent to which these measures of criminal behavior converge and diverge within and
across subsamples of first, second, and third-plus generation immigrants. Second, we
examine trends in reports of arrest over a 7-year period across immigrant subsamples
looking at patterns of convergence and divergence in self- and officially-reported arrests
longitudinally. Third, we test whether immigrant generational status is a predictor of crime
reporting bias (over-, under-, and divergence). Collectively, this research contributes to the
small number of studies that compare offending patterns in both official and self-report
records but more importantly fills a neglected gap in the body of research aimed at
advancing understanding of the immigrant-crime nexus.
Prior Literature
Official and Self-Reported Offending Comparisons
In assessing individual criminal activity, researchers have relied on two types of reports,
those from official data bases, such as police contacts, arrests, and convictions, and those
from self-report surveys. Each of these approaches has strengths and weaknesses and one
measure is not inherently better or more preferred than the other. As several authoritative
reviews of this literature exist elsewhere, we do not delve into details here (see Thornberry
and Krohn 2000, 2003). In short, findings from this line of work suggest a high degree of
overlap in measuring offending across both measurement approaches; that is, persons who
appear in official databases tend more often than not to self-report offending and vice versa
(see Thornberry and Krohn 2000; Piquero and Brame 2008). At the same time, when data
are disaggregated across key demographic groups (e.g., race and gender) important dif-
ferences emerge in the extent to which self-report measures map onto officially reported
arrests. In general, the consensus on reporting concordance appears to be mixed and
dependent upon data, crime type, and method of comparison with some studies indicating a
range of differences (Hindelang et al. 1981; Krohn et al. 2013) whereas others document
important similarities (Piquero and Brame 2008; Farrington et al. 1996; Joliffe et al. 2003;
Piquero et al. 2014). For example, some cross-sectional studies utilizing non-offender
samples reveal a moderate to strong concordance between official and self-report records
2 The arguments surrounding differential reporting versus differential involvement have long formed the
foundation of the disproportionate minority contact debate. That is, is the over-representation of minorities in the criminal justice system due to disparities in the involvement in crime across race or to disparities in the differential processing of minorities through the criminal justice system; or a combination of both (Piquero 2008)?
J Quant Criminol (2017) 33:835–857 837
123
but also reveal differential validity across race and/or gender (Hindelang et al. 1981; Kirk
2006; Lab and Allen 1984; Maxfield et al. 2000; Tracy 1987). In some instances, there was
under-reporting, other instances reveal over-reporting, and still other instances yield strong
concordance with no differential validity (see Farrington et al. 1996; Brame et al. 2004).
Two recent studies are worth a more expanded review because each of them, separately,
deal with approaches we combine in the current study.
Krohn et al. (2013) used longitudinal data from the Rochester Youth Development
Study to examine both over and under-reporting in self-reported and official arrests. Their
analyses failed to detect any gender differences for either over- or under-reporting, but they
did find that Blacks under-reported arrests while Whites over-reported arrests. As well,
these authors found that any race or gender differences in self-reports tended to become
insignificant when they controlled for the number of prior arrests. Using data from the
Pathways to Desistance Study, Piquero et al. (2014a, b) investigate the overlap between
official records (using a national search) and self-reports of arrests across race and gender.
Their findings showed moderate agreement between the two measurement approaches in
general as well as across race and gender. Further, their analysis failed to reveal any race
differences in official arrests but did detect gender differences in the prevalence of official
arrests above and beyond the effect of self-reported arrests.
In short, the evidence is rather mixed on the concordance between self-reports and
official records and even less studied with respect to over- and under-reporting. This is an
important issue because key conclusions about the nature and patterning of offending could
be influenced by these reporting differences. It is important to note that crime reporting
differences only matter to the extent that they systematically affect one group more (or
less) than another group. Related to the focus of this research, if self-reports of arrest are
disproportionately lower among first generation immigrants, then conclusions based on
self-report data demonstrating their lower rates of offending are called into question. If
however, all youth undercount their involvement in crime in the same way, then com-
parisons across groups are not biased.
Immigrant Offending: Generational Differences and Crime Reporting Methodology
A recent report summarizing the state of the literature on the immigrant-crime nexus
concludes ‘‘…the overwhelming majority of immigrants are not ‘‘criminals’’ by any commonly accepted definition of the term’’ (Ewing et al. 2015:1). This finding is not
necessarily a novel discovery; research dating back more than a century documents a
pattern whereby the foreign-born are involved in crime at significantly lower rates than
their peers (see e.g., Bowler 1931; Immigration Commission 1911; Simes and Waters
2014; Van Vechten 1941) with concerns of criminal behavior directed at successive
immigrant generations (Bowler 1931; Wickersham Commission 1931; Sutherland 1924/
1934). Much of this research utilized self-reports of criminal behavior where respondents
are asked to detail their involvement, and often frequency of involvement, in behaviors
including a range of risky/delinquent/criminal acts, substance use, and arrest histories.
Regardless of the operationalization of criminal behavior, results show a persistent pattern
of lower problem behavior among first generation immigrants with higher rates among
second and third generation, US-born peers (see e.g., Bersani 2014; Bersani et al. 2014;
Harris 1999; Morenoff and Astor 2006; Powell et al. 2010; Rumbaut et al. 2006; Sampson
et al. 2005; Vaughn et al. 2014).
838 J Quant Criminol (2017) 33:835–857
123
Using data from the PHDCN, Sampson et al. (2005) examined racial/ethnic gaps in the
perpetration of violence among nearly 3000 young adults. Findings suggest that much of
the Hispanic-white gap in offending is explained by immigrant generation and immigrant
neighborhood composition. That is, the odds of violence among first generation immigrants
are nearly half those of the third generation with second generation immigrants’ odds of
violence falling between the first and third generations. While the odds of violence are
lower among first generation immigrants compared to their native-born white peers, the
direction of this relationship changes with successive immigrant generations; by the third
generation the odds of violence outpace their white, native-born peers. Scrutinizing the
complexities of immigrant status further with these data, Morenoff and Astor (2006) find
that markers of increased assimilation (earlier age of migration, higher acculturation) were
associated with a higher likelihood of violence. In short, those in the US for longer periods
of time and more assimilated into the US mainstream have a higher likelihood of
offending.
Studies assessing immigrant generational differences in general delinquent behavior
also consistently demonstrate lower levels of risky behavior among foreign-born, first
generation immigrants compared to their native-born peers (Bersani 2014; Harris 1999;
Powell et al. 2010; Vaught et al. 2014). Powell et al. (2010) used three waves of the
National Longitudinal Survey of Adolescent Health (AddHealth) data, a nationally rep-
resentative, school-based sample of US adolescents in grades 7–12 during the 1994–1995
academic year, to plot trajectories of delinquency spanning adolescence into adulthood and
found consistently lower rates of non-violent delinquency among first generation immi-
grants. More recently, using data from a general population sample, Bersani (2014)
mapped trajectories of offending across immigration generation from early adolescence
into young adulthood for nearly 8000 individuals. Findings again reveal that the prevalence
of offending is significantly lower among foreign-born, first generation immigrants com-
pared to second and third-plus generation, native-born individuals. Moreover, the pattern
of offending observed for first generation immigrants is characterized by a low-rate of
offending lasting into young adulthood while the presence of a chronic, high-rate offender
is virtually non-existent among first generation immigrants.
Research utilizing official reports of criminal behavior reaches similar conclusions.
Foreign-born, first generation immigrants have lower rates of incarceration (Rumbaut et al.
2006) and arrest (Bersani et al. 2014; Jennings et al. 2013) compared to their successive
generation immigrant peers. Rates of incarceration for foreign-born males ages 18–39 in
2000 were dwarfed by rates of incarceration for their native-born peers across racial and
ethnic divisions leading Rumbaut et al. (2006) to report that ‘‘[t]he advantage for immi-
grants vis-à-vis natives applies to every ethnic group without exception.’’ Concentrating on
arrest histories among a sample of Hispanic inmates, Jennings et al. (2013) model tra-
jectories of arrest from 18 to 50 years of age. Using immigrant generation to predict groups
distinguished by their rates of arrest over the life course, the authors find that foreign-born
Hispanics were significantly more likely to be characterized by the low-rate offending
trajectory. More recently, in an effort to isolate and observe a sample of immigrants
arguably most at risk of serious offending, Bersani et al. (2014) examine longitudinal
trajectories of offending among a sample of high-risk, adjudicated adolescents and com-
pare rates of arrest across subsamples defined by immigrant generation. While rates of
official reports of arrest are similar across groups in adolescence, first generation immi-
grants show a pattern of desistance from crime that begins significantly earlier than their
second and third-plus generation black and white peers.
J Quant Criminol (2017) 33:835–857 839
123
Factors Influencing the Differential Reporting of Offending
Despite an extensive body of work demonstrating a persistent pattern of lower crime
among the foreign-born, first generation immigrant population—robust to variation in
measures of crime/offending, sample population (general or offender-based), and modeling
strategies—these findings may be contaminated by immigrant-specific crime reporting
biases which may artificially deflate (or inflate) immigrant involvement in crime. Here, we
highlight several factors that may systematically bias the self- or official-reporting of crime
in ways that differentially affect immigrants. While not an exhaustive list, these issues
draw attention to the unique situation of immigrants that may hold salience when quan-
tifying their involvement in and reporting of crime.
Perhaps the most recurrent theme in the literature concerns the relationship between
immigrants and the police. To the extent that immigrants hold negative perceptions of the
police, their cooperation with and enthusiasm for working with the police and reporting
crime is diminished. Immigrants’ perceptions of crime and the police are shaped by a
variety of experiences; in a broad sense, they may possess a general distrust of authority as
an outgrowth of previous experiences (direct or vicarious) or belief systems. For instance,
research illustrates that the imprint of life experiences in one’s country of origin is resilient
to migratory change (Davis 1985; Menjivar and Bejarano 2004; Pogrebin and Poole 1990).
Some countries of origin have histories of corrupt law enforcement agencies, personnel and
practices. 3 In this sense, negative experiences with the police in one’s home country can
taint perceptions of US criminal justice agents (see also Kirk et al. 2012; Tyler et al. 2010).
Furthermore, immigrants who have had contact with US immigration officials or those who
are unable to distinguish between US immigration officials and local police authorities may
harbor a fear of reprisal and deportation efforts (aimed at themselves, family members, or
friends) from local criminal justice agents (see e.g., Government Accounting Office 2009).
These concerns echo loudly in an era marked by increased attention to the strengthening of
immigration laws. Provine et al. (2012) document that the US has increasingly adopted a
devolution of immigration control whereby the responsibility for immigration enforcement
is shouldered by local police (e.g., authorized to inquire about one’s legal status, access to
deportation status in criminal databases); responsibilities traditionally reserved for federal
level agents (see also Decker et al. 2009). As Gottschalk (2015: 215) states, the ‘‘line
between immigration enforcement and law enforcement is rapidly disappearing in the
United States’’ resulting in the coining of terms such as ‘‘crimmigration’’ and ‘‘imm-
carceration’’. The devolution of responsibilities may have unintended consequences and
increase crime as a result of the alienation of immigrants and severing of their ties to local
law enforcement agents (Kirk et al. 2012; Treyger et al. 2014).
Combined, the negative perceptions of and fear of reprisal from the police likely limit
the degree to which foreign-born, first generation immigrants cooperate with the police. To
the extent that crime is intra-racial/ethnic, differential rates in the reporting of crime and
victimization may undercount immigrant involvement in crime (Kubrin 2014; Theodore
2013). This is complicated further by research finding that immigrants are particularly
vulnerable to victimization and less likely to report crimes (experienced or observed)
(Provine et al. 2012; Zatz and Smith 2012), though a conclusion that immigrants are at
3 As an exemplar of this phenomenon, see documentation emerging from Mexico—where the majority of
US immigrants originate from and whom comprise the largest portion of the foreign-born sample in these data—noting the pervasiveness of corruption in the criminal justice system (see Corchado 2013).
840 J Quant Criminol (2017) 33:835–857
123
disproportionate risk for victimization is inconclusive (see Kubrin and Desmond 2015;
MacDonald and Saunders 2012).
While a hearty literature suggests that immigrants may hold concerns that influence their
collaboration with US local police authorities, recent empirical findings challenge this
assumption (see Davis and Henderson 2003) and emphasize important complexities in
understanding variation in perceptions of the law (see e.g., Correia 2010; Kirk et al. 2012;
Piquero et al. 2014; Rengifo and Fratello 2015). Neighborhoods characterized by a higher
concentration of immigrants hold on average less cynical views of the law and are more
cooperative with legal authorities (Kirk et al. 2012). At the individual level, foreign-born, first
generation immigrants hold more positive perceptions of the law, less cynical attitudes
toward the legal system, and perceive more social costs associated with punishment com-
pared to their second generation immigrant and native-born peers (Piquero et al. 2014;
Rengifo and Fratello 2015). In addition to positive perceptions of law enforcement eroding
over time (across immigrant generation), Rengifo and Fratello (2015) find that positive views
of the police are weakened by increases in the frequency of involuntary police contact
(reported police stops). Interestingly, Davis and Henderson (2007) report that while the
foreign-born hold more positive perceptions of the police, they are also significantly less
likely toseekout assistance fromthepolice followinga crime. Ifthis isthe case, then they may
hold positive views of US criminal justice agents while at the same time fail to report crime.
There are also reasons to expect that the reported levels of immigrant crime may be inflated
stemming from differential treatment in the process of the administration of the law (Hagan
and Palloni 1998; see also Sellin 1938; Sutherland 1924/1934). Empirical evidence points to
disparities throughout the criminal justice system from inequalities in arrest, the disposition
of cases, and sentencing outcomes related to immigrant status. Despite their lower levels of
offending, Davies and Fagan (2012) found that immigrant communities experience dispro-
portionately higher levels of police enforcement (arrests and street stops). Similar to
minorities (Roberts 2011), immigrants experience disproportionate police contact for mis-
demeanor offenses that tangle them in the web of the criminal justice system. This exposure to
the criminal justice system leaves an indelible mark, resulting in lasting consequences the
most substantial of which may be the potential for transfer to deportation facilities (Gott-
schalk 2015). A greater likelihood of arrest, all else being equal, means that immigrants will
show up at disproportionate rates in official reports. Once in the criminal justice system,
immigrants may incur additional penalties. First, some are filtered out of the criminal system
to the civil system often losing the right to an attorney and other legal safeguards (for an
exhaustive review see Gottschalk 2015). Second, Hispanics in general receive the harshest
sentencing outcomes compared to their white and black peers and may be the result of
concerns regarding their legal status (Demuth 2003; Schlesinger 2005). Recent research
suggests that the Hispanic penalty observed in prior research masks the noncitizen penalty in
sentencing decisions (Light et al. 2014; Wolfe et al. 2011). Notably, the extent to which
immigrants (non-citizens) specifically incur differential criminal justice processing is
unsettled (c.f. Light et al. 2014; Orrick and Piquero 2015) and may differ across jurisdictional
level, i.e., federal, state, or local.
Current Research
The current research is aimed at investigating systematic crime reporting bias across
subsamples distinguished by immigrant generation. Despite a long history of empirical
tests of reporting biases across race, gender, and to a lesser extent, ethnicity, the question of
J Quant Criminol (2017) 33:835–857 841
123
whether immigrants differentially report arrests has not been addressed. Moreover, the vast
majority of research testing for disparities in offending across self- and official-reports has
relied upon cross-sectional data. We advance knowledge on disparities in arrest reports by
documenting the extent of divergence across self- and official-reports of arrest across first
generation immigrants, second generation and third-plus generation youth. We then use
immigrant generational status to predict reporting divergence (under- and over-reporting).
Data, Measures, and Analytic Strategy
Data
This paper uses data from ten survey waves spanning 7 years from the Pathways to
Desistance study, a longitudinal investigation of the transition from adolescence to young
adulthood among a sample of serious adolescent offenders. Study participants were ado-
lescents who were overwhelming found guilty of a serious offense (almost entirely felony
offenses) in the juvenile or adult court systems in Maricopa County, AZ or Philadelphia
County, PA. Eligible crimes included all felony offenses with the exception of less serious
property crime, as well as misdemeanor weapons offenses and misdemeanor sexual assault.
About 40 % of the offenses that brought the youth into the study were for a ‘person crime’.
These youth were ages 14–17 at the time of enrollment into the study (M = 16.5). A total
of 1354 adolescents were enrolled, representing approximately one in three adolescents
adjudicated on the enumerated charges in each locale during the recruitment period
(November, 2000 through January, 2003). The study sample is comprised mainly of non-
White (41 % Black, 34 % Hispanic) males (86 %), who were, on average, 16 years of age
at baseline with an average age of 14.9 years at their first petition. Study participants
completed a baseline interview at the time of enrollment followed by interviews every
6 months for the first 3 years of the study and annually thereafter through 7 years.
Information regarding the rationale and overall design of the study can be found elsewhere
(Mulvey et al. 2004; Schubert et al. 2004).
Measures
Arrest
The self-report offending data contains a binary indicator of an arrest based on an interview
question which asks the participant if they were ‘‘arrested and charged’’ since the date of
the last interview. Official records (based on petitions found in juvenile records in each
jurisdiction prior to age 18 and arrests appearing in FBI records thereafter) are assigned to
a recall period based on the date of the petition/arrest as mapped onto the dates covered
within a recall period. For both self- and official-reports, we used a binary indicator of a
petition/arrest by wave, where, as noted earlier, the recall periods from waves one through
six correspond to 6-month recalls, while waves seven through ten correspond to 12-month
recalls. 4 The seriousness of the adjudication charge for the initial referring petition (used as
the basis for study eligibility) was coded using a severity ranking system (Gottfredson and
4 For the first six waves, official records were based on petitions and the self-report records were based on
arrests, while for remaining waves both self-reports and official records are based on arrests.
842 J Quant Criminol (2017) 33:835–857
123
Baron 1993). The rank order for the ‘referral seriousness rate’ variable ranges from status
(coded 1; .7 %), misdemeanor (coded 2; 37 %), through to felonious assault, felony with a
weapon (coded 8; 16.5 %), and murder, rape, arson (coded 9; 1.9 %).
Offending
Self-reported involvement of offending across a variety of 22 offenses was captured
annually. This measure is a revised version of a common self-reported delinquency
measure (Huizinga et al. 1991) of the number of crimes committed including: (1) destroyed
or damaged property, (2) set fire to house, building, etc., (3) rape, (4) murder, (5) shot at
someone, (6) beaten up someone badly, (7) been in a fight, (8) beaten up, threatened or
attacked someone as part of a gang, (9) entered/broken in a building to steal, (10) used
checks/credit cards illegally, (11) stolen a car/motorcycle to keep or sell, (12) prostitution,
(13) stolen something from a store, (14) stolen a car/motorcycle to keep or sell, (15)
carjacked someone, (16) taken something from another by force (with a weapon), (17)
taken something from another by force (without a weapon), (18) bought/received/sold
stolen property, (19) sold marijuana, (20) sold other illegal drugs, (21) drove drunk or high,
and (22) carried a gun. These items tap into involvement in a range of aggressive (e.g.,
been in a fight; shot at someone) and income generating (e.g., bought/received/sold stolen
property) offenses. For the current analysis, we use the total offending variety proportion.
In this measure, the numerator is the number of different types of acts endorsed and the
denominator is the number of items for which the subject gives either a ‘‘yes’’ or ‘‘no’’
answer. All items which the subject refused to answer, replied ‘‘don’t know’’, or was not
asked are removed from the denominator. The closer this figure is to ‘‘1’’, the greater the
variety of offenses the youth has committed.
Immigrant Generational Status
We determine immigrant status using information on the country of birth of the youth and
his/her parents (Pew Research Center 2013; Rumbaut and Ewing 2007). First generation
immigrants are those respondents born outside the US with foreign-born parents (n = 83;
6.4 % of the total sample). Second generation immigrants include respondents born in the
US with at least one foreign-born parent (n = 210; 15.5 % of the total sample). Because
the vast majority of immigrant youth in these data are of Latino ancestry the results may
not generalize to immigrants of other ancestries. The largest group of first generation
immigrants in the sample is of Mexican ancestry (61.8 % of the first generation subsample)
and this subset resides exclusively in Phoenix. Clearly Mexican immigrants dwarf other
immigrant ethnicities in these data; however, this disproportionality corresponds with
general population trends whereby immigrants of Mexican ancestry comprise the largest
proportion of immigrants in the US (Fortuny and Chaudry 2011). The second largest
subset, those with Puerto Rican ancestry (26.6 %), is located exclusively in Philadelphia.
Although Puerto Ricans are US citizens by birth, we follow precedent set in previous
research which excludes them from the native-born US sample as they often experience
many of the obstacles to incorporation that other immigrant groups face (see e.g.,
Hirschman 2001). Third-plus generation youth refer to respondents born in the US with
US-born parents (n = 1061). The majority of native-born youth are Black (n = 549; 41 %
of the total sample), followed by White (n = 253; 19 % of the total sample) and Hispanic
(n = 208; 15 % of the total sample), with a small group designated as ‘‘other’’ race/
ethnicity (n = 51; 4 % of the total sample). Examined across immigrant generational
J Quant Criminol (2017) 33:835–857 843
123
subsample, we find no significant differences in the proportion male or mean age at
baseline (see Table 1). Looking at the race/ethnic breakdown across immigrant status, we
find a significantly higher proportion of Black respondents and a significantly lower pro-
portion of Hispanic respondents among third-plus generation youth compared to first and
second generation immigrants.
Controls
We include two factors known to influence offending and the modeling of trajectories of
offending overtime: early onset of problem behavior and time spent off the street (e.g.,
incarcerated). A series of five questions are used to establish the early onset of problem
behavior (problems before the age of 11). These questions were selected from the Psy-
chopathy Checklist—Youth Version (Forth et al. 2003). Early problem behavior is oper-
ationalized as a summary of total items endorsed including: cheating, disturbing class,
being drunk/stoned, stealing, and fighting. Time secure is a time-varying covariate cap-
turing the proportion of time spent in a secure facility. Prior research has highlighted the
importance of controlling for street time when measuring longitudinal patterns of
offending (see Eggleston et al. 2004; Piquero et al. 2001).
Analytic Strategy
The analytic framework for the current research moves from description to prediction of
systematic crime reporting bias. In agreement with Laub (2010: 423) noting the lost art of
‘‘descriptive quantitative criminology’’ we see description as a key tool to understanding
the fundamental elements of behavior. Exploiting the detailed self- and official reported
measures of arrest, we begin by examining the prevalence and longitudinal arrest trends
within subsamples of immigrant generations (level of correspondence of self-reported and
officially-reported arrests among first, second, and third-plus immigrant generations) and
between subsamples of immigrant generations (extent of under-reporting, over-reporting,
and divergence in reporting patterns). Following this descriptive discussion, we next
perform a series of robustness analyses to examine the sensitivity of our findings to
alternative modeling approaches including analyses conducted on the subsample of
Table 1 Sample descriptive statistics by immigrant generation, pathways to desistance study
First generation Second generation Third-plus generation
N 83 210 994
% of sample 6.40 16.30 77.20
Mean age at baseline 16.13 16.11 16.02
Proportion male 0.86 0.88 0.86
Proportion black 0.01 0.05 0.52
Proportion hispanic 0.88 0.82 0.20
Early onset 1.08 1.58 1.54
Pre-baseline mean official arrest rate 2.81 3.08 2.90
Arrest referral seriousness rate 3.88 4.23 4.51
Cumulative mean self-report arrest rate 1.52 2.03 2.05
Cumulative mean official-report arrest rate 1.53 2.15 2.08
844 J Quant Criminol (2017) 33:835–857
123
Hispanic youth, permutation analyses (Good 2013), and an analysis of trajectories of
offending from adolescence through to young adulthood using a series of hierarchical
linear models (HLM version 6.04, Raudenbush and Bryk 2002).
Results
Arrest Trends
All youth were selected into the sample because they were adjudicated of a crime. While
this means that all respondents have a history of an official criminal justice sanction, there
is heterogeneity in the extent to which respondents had arrest experience prior to the
baseline interview in this study. To understand the extent to which criminal involvement
pre-baseline varies by immigrant generational status, we assessed mean differences in
official arrest rates prior to baseline. Results are plotted in the first series in Fig. 1. The
average arrest rate pre-baseline does not significantly differ across first (mean = 2.81),
second (mean = 3.08), and third-plus (mean = 2.90) generation youth. In other words,
regardless of immigrant generational status, the youth in this sample have similar levels of
involvement in crime prior to the baseline interview. While offending prior to baseline is
statistically similar across immigrant groups, we test for differences in the seriousness of
the arresting charge to investigate whether, for example, immigrant youth are more likely
to be referred for less serious crimes. We find no significant differences in the seriousness
of the arresting charge across immigrant generations (second series, Fig. 1).
Next, we compared re-arrest patterns following the baseline interview by aggregating
the total number of arrests across all ten waves. Results are presented in the third and
fourth series in Fig. 1. Using both self-reported and officially-reported measures of arrest
we find that first generation immigrants have significantly lower mean arrest rates com-
pared to their second and third-plus generation immigrant peers. Taken together, the arrest
trend findings suggest a pattern whereby youth, distinguished by immigrant generation, are
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
Pre-Baseline Mean Official Arrest Rate
Arrest Referral Seriousness Rate
Cumulative Mean Self-Report Arrest Rate
Cumulative Mean Official- Report Arrest Rate
M ea
n Ra
te
First Genera�on Immigrants Second Genera�on Immigrants Third-plus Genera�on Immigrants
Fig. 1 Mean rate of arrest by immigrant generation, pathways to desistance study
J Quant Criminol (2017) 33:835–857 845
123
statistically similar at baseline, but evidence a pattern of divergence in their offending
behavior over time. 5
Finally, we compared the level of offending over time and plot the prevalence rates for
each group by wave in Fig. 2 for self-reported offending (Fig. 2a), self-reported arrest
(Fig. 2b) and officially recorded arrest (Fig. 2c). In the first few waves following the
adjudication arrest (baseline), rates of offending across all three measures are fairly similar.
In the final waves of the data, involvement in crime and arrest among first generation
immigrants diverges from their second and third-plus generation peers with first generation
immigrants displaying the lowest self-reported and officially recorded involvement in
crime/arrest. Rates of offending, particularly in the final waves, are very similar comparing
second and third-plus generation immigrants.
Within Immigrant Generation Crime Reporting Analysis
Next, we exploit the wave specific crime data and compare self-reported offending, self-
reported arrests and officially reported arrests to test for divergence in reporting practices
across immigrant generations. We begin by assessing the correlation between self-reported
offending and our two indicators of arrest (see Table 2). For each immigrant generation,
rates of self-reported offending decline over time. Comparisons between self-reported
offending and self-reported arrest are moderate for each immigrant generation with a
similar average correlation across all waves (.362 first generation immigrants; .352 second
generation immigrants; .347 third-plus generation immigrants). The strength of association
between self-reported offending and officially recorded arrests is weaker (average strength
of association across waves: .140 first generation immigrants; .241 second generation
immigrants; .248 third-plus generation immigrants).
Next, we compare self-reported arrests and official reported arrests within each immi-
grant generation. In general, the correlation between self and official reports of arrest is
moderate to high across waves for each immigrant generation with a similar average
correlation across all waves (.557 first generation immigrants; .533 second generation
immigrants; .582 third-plus generation immigrants) with few differences in terms of the
variation in wave specific reporting convergence across immigrant generations. We note
that the periods of weaker associations across self-reported and official reported arrests for
first and second generation immigrants, compared to their third-plus generation peers, may
be influenced by differences in sample sizes for these groups. While we cannot rule out that
these statistical differences are not ‘‘real’’ differences, it is notable that in these instances
the trend is representative of higher self-reported arrests compared to official reported
arrests. In other words, first and second generation immigrants appear to over-report their
arrests.
While the correlation values are fairly strong comparing self-report and official reported
arrests, these measures of association combine data across respondents and compare total
numbers of arrests across measurement type. To assess whether individuals who self-report
an arrest have a corresponding officially-recorded arrest, we conducted a series of Chi-
5 An anonymous reviewer observed that a particular strength of this finding, using a sample of adolescents
with an official juvenile record, is that it identifies a similar pattern of lower offending among the first generation immigrant group mimicking the results obtained in research using data from general population samples (see i.e., Bersani 2014; Sampson et al. 2005). This provides an additional level of confidence in the results and adds to the generalizability of the immigrant-specific finding of lower offending across different sampling frames.
846 J Quant Criminol (2017) 33:835–857
123
square tests and compare cell convergence and divergence across immigrant generations;
results reported in the final columns in Table 2. These analyses tell us that significant
variation exists in the data. More specifically, regardless of immigrant generation, sig-
nificant variation in individual patterns of reported arrest by measurement type is evident in
each wave.
0.000
0.020
0.040
0.060
0.080
0.100
0.120 (a) Propor�on variety self-reported offending
(b) Prevalence of self-reported arrest
(c) Prevalence of officially recorded arrest
n n
0 5
10 15 20 25 30 35
n n n
0 5
10 15 20 25 30 35
n n
Fig. 2 Average level of offending over time by immigrant generation, pathways to desistance study
J Quant Criminol (2017) 33:835–857 847
123
Table 2 Proportion reporting arrest by wave and immigrant generation, pathways to desistance study
SR crime
SR arrest
OR arrest
Correlation: SR crime by SR arrest
Correlation: SR crime by OR arrest
Correlation: SR arrest by OR arrest
Group difference
p value Sig p value Sig p value Sig Chi-sq Sig
First generation
0–6 months 0.069 20.5 19.3 0.383 *** 0.113 0.638 *** 31.778 ***
7–12 months 0.061 16.7 9.6 0.289 ** 0.132 0.643 *** 32.203 ***
13–18 months 0.062 20.5 19.3 0.436 *** 0.349 ** 0.638 *** 31.778 ***
19–24 months 0.040 13.3 16.9 0.318 ** 0.180 0.546 *** 22.336 ***
25–30 months 0.027 17.3 15.7 0.364 * 0.023 0.281 * 5.903 *
31–36 months 0.028 15.7 14.5 0.325 ** 0.080 0.569 *** 22.629 ***
37–48 months 0.027 17.8 16.9 0.127 0.051 0.501 *** 18.311 ***
49–60 months 0.029 18.8 19.3 0.477 *** 0.053 0.310 ** 6.625 **
61–72 months 0.023 13.0 12.0 0.552 *** 0.170 0.489 *** 16.492 ***
73–84 months 0.021 17.5 9.6 0.348 ** 0.247 * 0.704 *** 31.192 ***
Average 0.362 0.140 0.557
Second generation
0-6 months 0.106 25.3 18.1 0.415 *** 0.214 ** 0.443 *** 37.992 ***
7–12 months 0.097 20.8 19.5 0.397 *** 0.293 *** 0.608 *** 70.944 ***
13–18 months 0.069 19.7 20.0 0.295 *** 0.306 *** 0.495 *** 45.974 ***
19–24 months 0.074 25.4 22.4 0.253 *** 0.136 0.546 *** 57.523 ***
25–30 months 0.049 15.7 14.8 0.273 *** 0.260 *** 0.635 *** 76.944 ***
31–36 months 0.057 19.0 14.8 0.303 *** 0.258 *** 0.616 *** 71.711 ***
37–48 months 0.069 22.1 18.6 0.422 *** 0.111 0.235 *** 10.470 ***
49–60 months 0.066 28.7 31.4 0.355 *** 0.280 *** 0.609 *** 69.804 ***
61–72 months 0.063 24.6 23.3 0.438 *** 0.350 *** 0.576 *** 60.665 ***
73–84 months 0.049 23.7 31.9 0.369 *** 0.198 ** 0.567 *** 54.336 ***
Average 0.352 0.241 0.533
Third-plus generation
0-6 months 0.081 15.8 15.2 0.344 *** 0.209 *** 0.550 *** 298.960 ***
7––12 months 0.069 17.8 16.6 0.345 *** 0.265 *** 0.616 *** 376.135 ***
13–18 months 0.065 17.4 17.4 0.311 *** 0.272 *** 0.637 *** 391.173 ***
19–24 months 0.060 20.5 19.9 0.356 *** 0.265 *** 0.634 *** 386.517 ***
25–30 months 0.050 19.2 19.0 0.395 *** 0.237 *** 0.602 *** 351.198 ***
31–36 months 0.050 23.9 18.4 0.310 *** 0.247 *** 0.592 *** 341.019 ***
37–48 months 0.067 30.9 17.5 0.395 *** 0.253 *** 0.427 *** 173.355 ***
49–60 months 0.059 28.5 30.4 0.336 *** 0.260 *** 0.605 *** 347.484 ***
61–72 months 0.052 27.7 19.1 0.341 *** 0.257 *** 0.645 *** 385.078 ***
73–84 months 0.046 22.6 24.5 0.341 *** 0.210 *** 0.515 *** 238.679 ***
Average 0.347 0.248 0.582
* p \ .05
** p \ .01
*** p \ .001
848 J Quant Criminol (2017) 33:835–857
123
Between Immigrant Generation Crime Reporting Analysis
The results of the within immigrant generation crime reporting analyses are suggestive of
important differences in reporting practices across self-reported and official reported
measures of offending. The extent to which these differences influence findings comparing
patterns of involvement in crime across subsamples distinguished by immigrant generation
rests on whether we observe systematic crime reporting differences by immigrant gener-
ation. That is, if all youth similarly under- or over-report their involvement in crime, then
comparisons across groups are not affected by this difference. We investigate systematic
crime reporting bias by comparing wave-specific rates of convergence (0 arrests reported
in both self-reports and official reports; 1 arrest reported in both self-reports and official
reports) and divergence (a self-reported arrest but no official recorded arrest; and vice
versus) across arrest reporting measures. As shown in Fig. 3, we find a high degree of
convergence (lines in the top portion of the figure) across arrest report measures regardless
of immigrant generation or wave. The average rate of convergence across waves is 86.83
for first generation immigrants, 83.88 for second generation immigrants, and 85.84 for
third-plus generation immigrants. As the figure shows, the trend in convergence across
arrest reporting measure is fairly linear for each immigrant group.
While convergence (and divergence) rates are similar for all immigrant generational
groups, there may be important patterns of divergence (over- or under-reporting) in arrest
reporting measures. We disaggregate divergence in reporting practices and compare pat-
terns of over-reporting (self-reporting an arrest in a wave with no officially recorded arrest)
and under-reporting (officially recorded arrest in a wave with no self-reported arrest) across
immigrant generation. Because much of the criticism of research on immigrant involve-
ment in crime is levied at potential under-reporting of crime among first generation
immigrants, we show patterns of under-reporting in Fig. 3 (lines in the bottom portion of
the figure; results for over-reporting are available upon request). Again, while we observe
0
10
20
30
40
50
60
70
80
90
100
Wave 1 Wave 2 Wave 3 Wave 4 Wave 5 Wave 6 Wave 7 Wave 8 Wave 9 Wave 10
First Genera�on Second Genera�on Third-Plus Genera�on
Fig. 3 Convergence in self-report and official-report arrests (top portion) and under-reporting of arrests (bottom portion) by Wave and immigrant generation, pathways to desistance study
J Quant Criminol (2017) 33:835–857 849
123
minor fluctuation in reporting over time, the general trend of under-reporting over time is
characterized by similarity across immigrant generational groups.
Robustness Analyses
We perform a series of supplemental analyses to test the robustness of our findings presented
above. First, to ensure that our findings are not confounded by differences in the proportions
of ethnic and racial groups found in the three immigrant generational subsamples, we select
out Hispanic youth and reanalyze the data. The findings using the Hispanic only sample are
remarkably similar to those found using the full sample reported above.
Second, we perform a series of permutation analyses (Good 2013) to test whether facets
of the data may have distorted the results. Prior analyses include assumptions about the
normality of the distribution, homogeneity of variance, and independence. Permutation
tests are a type of randomization method that offer an improvement over standard para-
metric approaches in that they constitute a distribution free test. In these tests, real data are
‘‘shuffled’’ to obtain all possible arrangements of the data values. Information on the self-
reported arrest variable is resorted 1000 times and then the relationship between officially
reported arrests and immigrant generational status conditional on the shuffled self-reported
arrest variable are compared. Because the total number of possible permutations is large,
we utilize a Monte Carlo approach and randomly estimate 1000 permutations of the data
(StataCorp 2013). The obtained value is the probability of producing values equal to or
more extreme than the observed value given the number of permutations calculated. In our
example, if we shuffle data on the self-reported arrest variable, these analyses test whether
our observed relationships between immigrant generation and crime maintain. Results
from the permutation analyses (available upon request) confirm the pattern of findings
regarding the relationships between immigration and crime reporting and provide further
confidence in the results detailed above.
Finally, we perform a stronger test of systematic crime reporting bias by conducting a
series of hierarchical longitudinal models using immigrant generation as a predictor of
divergence in arrest reporting measures (analytic details for the longitudinal hierarchical
models are provided in the ‘‘Appendix’’). Controlling for the influence of age, gender,
race/ethnicity, criminal history, and time spent incarcerated, we examine whether immi-
grant status predicts arrest reporting divergence. The inclusion of the controls means the
reference group for the intercept is white, third-plus generation, women. To situate this
research with extant literature on the immigration-crime nexus, we begin by predicting
general involvement in crime (results available upon request) using three different
dependent variables: proportion of involvement in a variety of self-reported offenses,
prevalence of self-reported arrest, and prevalence of officially recorded arrest. Similar to
previous research, first generation immigrant status is associated with a reduced level of
offending when assessing self-reports of offending generally and the prevalence of self-
reported arrests; first generation immigrant status was not significant in the model assessing
the prevalence of an officially-reported arrest. No significant effects were observed for the
second generation immigrant control for any outcome.
Next, we shift our focus to measures of arrest reporting divergence (overall divergence,
under-reporting, and over-reporting) and examine the influence of immigrant generation
controlling for age, gender, race/ethnicity, criminal history and time spent incarcerated;
results reported in Table 3. We include a new time-varying covariate in these models, pro-
portion of involvement in a variety of self-reported offenses, to control for individual dif-
ferences in the severity of offending (i.e., heightened levels of criminal behavior) on crime
850 J Quant Criminol (2017) 33:835–857
123
reporting practices under the assumption that more serious offenders may be less accurate
sources of information. Notably, results are robust and maintain in models excluding this
control. We find no evidence that first generation immigrants exhibit a significantly differ-
ential pattern of reporting compared to their native-born peers. The only significant effect
observed for immigrant generational status is found when analyzing the likelihood of under-
reporting (officially-recorded arrest without a corresponding self-reported arrest) where
under-reporting is slightly higher among second generation immigrants.
Discussion
Research has repeatedly shown a persistent pattern of lower rates of offending among
foreign-born, first generation immigrants compared to their US born peers. Despite the
consistency of this finding, the extent to which systematic crime reporting biases differ-
entially affect first generation immigrants introduces an important point of contestation to
this body of research. While the challenge of crime reporting biases affect all demographic
groups, experiences with and facets of the criminal justice system, both in the US and
supplanted from one’s country of origin, uniquely affect the foreign-born and create the
potential for heightened rates of reporting biases among first generation immigrants. The
purpose of this research was to attend to this gap in knowledge by first, measuring the
prevalence of crime reporting disparities across subsamples of immigrants distinguished by
generational status, and second, using immigrant generational status to predict divergence
(under- and over-reporting) in arrest reports across self- and official measures of arrest.
Our analyses revealed a series of notable findings. First, while we find statistical similarity
in average arrest rates and arrest referral seriousness in the period leading up to adjudication
(and entry into the study) across groups distinguished by immigrant generation, this similarity
evaporated when we assess patterns of arrest overtime. Specifically, cumulative counts of
Table 3 Hierarchical, multivariate analysis of divergence in arrest reports by immigrant generation, pathways to desistance study
Divergence in arrest reports
Under-reporting arrests
Over-reporting arrests
Coeff SE Coeff SE Coeff SE
Intercept -2.276*** 0.069 -3.108*** 0.071 -2.748*** 0.079
First generation -0.039 0.101 0.046 0.099 -0.101 0.125
Second generation 0.097 0.066 0.192** 0.068 -0.060 0.074
Male 0.601*** 0.063 0.731*** 0.063 0.386*** 0.073
Black 0.027 0.056 0.060 0.058 -0.017 0.065
Hispanic 0.057 0.062 -0.040 0.063 0.125 0.070
Early onset 0.028 0.019 0.021 0.021 0.026 0.023
Time-varying covariates
Time secure -0.540*** 0.054 -0.530*** 0.066 -0.368*** 0.060
Variety self-reported offending 1.610*** 0.199 -1.370*** 0.212 3.163*** 0.220
* p \ .05
** p \ .01
*** p \ .001
J Quant Criminol (2017) 33:835–857 851
123
involvement in crime, captured with self-report and official reports of arrest, diverge with
first-generation immigrant youth displaying a lower average rate of self-reported and offi-
cially recorded arrest compared to their second and third-plus generation peers. This findingis
particularly notable given that the sample was selected based on having an official juvenile
record. That is, similar to studies focusing on samples culled from the general population (see
e.g., Bersani 2014; Sampson et al. 2005), we find that within a sample of juveniles with an
official record, first generation immigrants display a pattern of lower levels of offending
compared to their second and third-plus generation peers.
Second, when we compared the correspondence between self-reported offending, self-
reports of arrest and official reports of arrest within group distinguished by immigrant gen-
eration, we found a high degree of similarity. In general, while the strength of association was
weaker when comparing self-reported offending with both measures of arrest for all groups,
the correspondence across arrest measures was equally strong for first, second, and third-plus
generation immigrants. The few instances where differences were apparent reflect patterns
where first and second generation immigrants show a tendency to over-report arrests (more
self-reported arrests than officially reported arrests). Moreover, when we assessed longitu-
dinal patterns of convergence (and divergence) in self- and official-reports of arrest, we found
a very high degree of convergence regardless of immigrant generation with average rates of
convergence in the mid-80 %. While convergence across crime reporting measures in these
data are not perfect, they are consistently high regardless of immigrant generational status
reaching convergence levels similar to those found in studies using general population
samples (see review in Piquero et al. 2014).
Third, turning to between immigrant generation comparisons, when we considered
differences in the likelihood of offending with respect to the frequency of self-reported
crime, prevalence of self-reported arrest, and the prevalence of officially-recorded arrest by
immigrant status, consistent with previous literature first-generation immigrant youth
display the lowest likelihood of offending for two of our three offending outcomes (fre-
quency of self-reported crime; prevalence of self-reported arrest) compared to their peers.
When we focused on investigating divergence in arrest reports, including a focused
attention on the under- and over-reporting of arrests, we find no evidence of systematic
crime reporting biases across groups distinguished by immigrant generation. That is, first-
generation youth did not exhibit a significantly different pattern of arrest reporting com-
pared to their US-born peers. In one instance, results reveal that second generation
immigrants were significantly more likely to under-report arrests. Collectively, these
analyses bear little evidence to suggest that immigrants, particularly foreign-born, first
generation immigrants, exhibit systematic crime reporting bias.
The findings from this research contribute to the body of knowledge on the immigrant-
crime nexus generally, and with respect to potential crime reporting differences across
groups in particular. An oft-repeated finding in the literature has been that the foreign-born
do not pose a unique threat in terms of their criminal proclivities; in fact, research con-
sistently shows that first generation immigrants are characterized by significantly lower
rates of offending compared to their similarly situated peers, a finding that shows up in
research utilizing different data sources and different sampling frames. This research
augments this work by demonstrating that this persistent pattern is not the result of
divergence in crime reporting practices. While measures of arrest across self- and official-
reports are not perfectly aligned, a limiting feature of the research in this area more
generally, divergence in reporting practices does not differentially affect first generation
immigrants. Stated simply, we find no evidence that foreign-born, first generation
852 J Quant Criminol (2017) 33:835–857
123
immigrants under-report their arrest history. In fact, when evidence of divergence exists it
is in the direction of immigrants over-reporting arrests.
We envision several directions for subsequent work. First, as our sample sizes were
small, we encourage replication of this research using larger samples as well as the use of
data covering a longer time period; to date, we are unware of data meeting these char-
acteristics that also contain official and self-reports of offending for immigrants and their
peers. Second, while this research examines offending behavior among mostly Mexican,
and to a lesser extent, Puerto Rican immigrants, who have been the focus of immigrant-
crime concerns, heterogeneity in immigrant experiences based on ethnicity and country of
origin were limited in this study due to sampling strategy; different groups of immigrants,
especially Asian, European, and Middle-Eastern immigrants should be examined to test for
the generalizability of these findings. The latter group in particular, which is ill-studied in
criminology, would be of special interest considering perceptions that exist about Middle-
Eastern immigrants. Third, a more general area of work would consider how individuals
interpret the experience of an arrest or police contact. That is, do some youth, especially
immigrant youth whom do not fully understand the American criminal justice system,
perceive police stops, searches, or arrests in the same manner?
The current sociopolitical context is rife with disputed concerns levied at the criminal
proclivities of the immigrant population; the results of this research inform these debates
by providing empirically grounded evidence that the foreign-born pose no unique criminal
threat. Rather, foreign-born first generation immigrants are characterized by significantly
lower levels of offending which, as we show here, is not attributable to a differential
tendency to under-report their involvement in crime.
Appendix
The longitudinal analysis of individuals is complicated by a number of issues. First,
observations gathered over time are nested within each individual violating the indepen-
dence assumption of ordinary least squares regression which affects the estimation of
standard errors and increases the likelihood of obtaining a type 1 error. Second, because
individuals display at least a modest level of behavioral stability over time, individuals are
more similar to themselves across observations than they are to other individuals (Osgood
2009). Third, the variable number and spacing of observations across respondents results in
an unbalanced model. HLM is flexible and can accommodate these factors (Raudenbush
and Bryk 2002; Snijders and Bosker 1999).
Equations are specified using either Poisson (count dependent variables) or Bernoulli
(dichotomous dependent variables) extensions of HLM. In the interest of space, we detail
here the equations using a Poisson model; modeling decisions are consistently applied. We
allow for overdispersion in each model as the addition of the overdispersion parameter has
been shown to result in more accurate significance tests (Osgood 2000). To examine
offending trajectories over time we use a two-level hierarchical model where offending is
measured at level 1 and includes repeated measures of self or official reports of arrest for
individuals across 10 waves. The level 1, within individual equation is:
git ¼ log kitð Þ
git ¼ p0i þ p1i ageð Þit þ p2i age 2
� � it þ p3i age3
� � it þ p4i timesecureð Þit
J Quant Criminol (2017) 33:835–857 853
123
where gitis the log of the average arrest rate for individual i at age t. To capture the potential of non-linear arrest (or divergence) trajectories, the equation is specified to follow
a cubic function of age. The age terms were grand mean centered to allow for more
stable estimation due to collinearlity between the age terms. Grand mean centering yielded
an interpretation of p1i as the average rate of growth across the entire observation period. We observe standard growth parameter models prior to the addition of predictors to assess
whether a cubic model is necessary to capture accurately the longitudinal trend in our
outcome of interest excluding higher order polynomials when not significant. The pro-
portion of time spent in a secure facility is included as a time-varying covariate and is
grand-mean centered. The subscript i attached to variables at level 1 indicates that these
variables can take on different values for each individual. The subscript t indicates that the
variables can take on different values over time.
Individual level characteristics were entered into the equation at level 2. Coefficient
effects at this level indicate how much variation in the intercept (i.e., initial offending
level) and slopes (i.e., growth in the offending trajectory) is explained by between-indi-
vidual characteristics. The level 2, between-individual equations are:
p0i ¼ b00 þ b01...0k controlsð Þiþr0i
p1i ¼ b10 þ r1i
p2i ¼ b20 þ r2i
p3i ¼ b30 þ r3i
p4i ¼ b40 where variation in the log of the crime rate at the age coded as zero (the mean age for the
sample) is explained by an array of sociodemographic and behavioral controls. The
dichotomous measures were left in their original form (not centered) allowing for the
generation of offending ‘‘profiles’’ for specific demographic groups. We allow for variation
between individuals on the dependent crime/arrest variables (p0i) and age (p1i, p2i, p3i) parameters as indicated by the error terms (r0i), (r1i), (r2i) and (r3i). The inclusion of an
error term on the age slope indicates that the estimated rate of crime/arrest is allowed to
vary across individuals and that the model allows for gradual change in arrest over time
across individuals (see Horney et al. 1995).
References
Bersani BE (2014) An examination of first and second generation immigrant offending trajectories. Justice Q 31:315–343
Bersani BE, Loughran TA, Piquero AR (2014) Comparing patterns and predictors of immigrant offending among a sample of adjudicated youth. J Youth Adolesc 43:1914–1933
Bowler AC (1931) Recent statistics on crime and the foreign born. In: National Commission (ed) On law observance and enforcement: report on crime and the foreign born (part II). Government Printing Office, Washington, DC, pp 79–196
Brame R, Fagan J, Piquero AR, Schubert CA, Steinberg L (2004) Criminal careers of serious delinquents in two cities. Youth Violence Juv Justice 2:256–272
Chaves LR (2008) The Latino threat: constructing immigrants, citizens, and the nation. Stanford University Press, Stanford
Corchado A (2013) Midnight in Mexico: a reporter’s journey through a country’s descent into darkness. Penguin Books, London
854 J Quant Criminol (2017) 33:835–857
123
Correia ME (2010) Determinants of attitudes toward police of Latino immigrants and non-immigrants. J Crim Justice 38:99–107
Davies G, Fagan J (2012) Crime and enforcement in immigrant neighborhoods: evidence from New York City. Ann Am Acad Polit Soc Sci 641:99–124
Davis WE (1985) Language and the justice system: problems and issues. Justice Syst J 10:353–364 Davis RC, Henderson NJ (2003) Willingness to report crimes: the role of ethnic group membership and
community efficacy. Crime Delinq 49:564–580 Davis RC, Henderson NJ (2007) Immigrants and law enforcement: a comparison of native-born and foreign-
born americans’ opinions of the police. Int Rev Victimol 14:81–94 Decker SH, Lewis PG, Provine DM, Varsanyi MW (2009) On the frontier of local law enforcement: local
police and federal immigration law. In: McDonald WF (ed) Sociology of crime, law, and deviance. Emerald/JAI Press, Bingley, pp 261–278
Demuth S (2003) Racial and ethnic differences in pretrial release decisions and outcomes: a comparison of hispanic, black and white felony arrestees. Criminology 41:873–908
Eggleston EP, Laub JH, Sampson RJ (2004) Methodological sensitivities to latent class analysis of long- term criminal trajectories. J Quant Criminol 20:1–26
Ewing WA, Martı́nez DE, Rumbaut RG (2015) The criminalization of immigration in the United States. American Immigration Council Special Report, Washington, DC
Farrington DP, Loeber R, Stouthamer-Loeber M, Van Kammen WB, Schmidt L (1996) Self-reported delinquency and a combined delinquency seriousness scale based on boys, mothers, and teachers: concurrent and predictive validity for African-Americans and Caucasians. Criminology 34:493–517
Ferraro V (2016) Immigration and crime in the new destinations, 2000–2007: a test of the disorganizing effect of migration. J Quant Criminol 32(1):23–45
Forth AE, Kosson DS, Hare RD (2003) The hare psychopathy checklist: youth version. Technical manual. Multi-Health Systems, Inc., New York
Fortuny K, Chaudry A (2011) Children of immigrants: growing national and state diversity. Brief 1. The Urban Institute, Washington, DC
Good PI (2013) Permutation tests: a practical guide to resampling methods for testing hypotheses. Springer, Berlin
Gottfredson DC, Baron WH (1993) Deinstitutionalization of juvenile offenders. Criminology 31(4):591–611 Gottschalk M (2015) Caught: the prison state and the lockdown of American politics. Princeton University
Press, Princeton Hagan J, Palloni A (1998) Immigration and crime in the United States. In: Smith JP, Edmonston B (eds) The
immigration debate: studies on the economic, demographic, and fiscal effects of immigration. National Academy Press, Washington, DC, pp 367–387
Harris KM (1999) The health status and risk behaviors of adolescents in immigrant families. In: Hernandez DJ (ed) Children of immigrants: health, adjustment, and public assistance. National Academy Press, Washington, DC, pp 286–347
Hindelang MJ, Hirschi T, Weis JG (1981) Measuring delinquency. Sage, Beverly Hills Hirschman C (2001) The educational enrollment of immigrant youth: a test of the segmented-assimilation
hypothesis. Demography 38:317–336 Horney J, Wayne Osgood D, Marshall IH (1995) Criminal careers in the short-term: intra-individual
variability in crime and its relation to local life circumstances. Am Sociol Rev 60:655–673 Huizinga D, Esbensen F-A, Weiher AW (1991) Are there multiple paths to delinquency? J Crim Law
Criminol 82:83–118 Immigration Commission (1911) Reports of the immigration commission: immigration and crime.
Government Printing Office, Washington, DC Jennings WG, Zgoba KM, Piquero AR, Reingle JM (2013) Offending Trajectories among native-born and
foreign-born hispanics to late middle age. Sociol Inq 83:622–647 Jolliffe D, Farrington DP, David Hawkins J, Catalano RF, Hill KG, Kosterman R (2003) Predictive,
concurrent, prospective and retrospective validity of self-reported delinquency. Crim Behav Mental Health 13:179–197
Kirk DS (2006) Examining the divergence across self-report and official data sources on inferences about the adolescent life-course of crime. J Quant Criminol 22:107–129
Kirk DS, Papachristos AV, Fagan J, Tyler TR (2012) The paradox of law enforcement in immigrant communities: does tough immigration enforcement undermine public safety? Ann Am Acad Polit Soc Sci 641:79–98
Krohn MD, Lizotte AJ, Phillips MD, Thornberry TP, Bell KA (2013) Explaining Systematic bias in self- reported measures: factors that affect the under- and over-reporting of self-reported arrests. Justice Q 30:501–528
J Quant Criminol (2017) 33:835–857 855
123
Kubrin CE (2014) Secure or insecure communities? Criminol Public Policy 13:323–338 Kubrin C, Desmond SA (2015) The power of place revisited: why immigrant communities have lower levels
of adolescent violence. Youth Violence Juv Justice 13(4):345–366 Lab SP, Allen RB (1984) Self-report and official measures: a further examination of the validity issue.
J Crim Justice 12:445–455 Laub JH (2010) Nurturing the journal of quantitative criminology through late childhood: retrospective
memories (distorted?) from a former editor. J Quant Criminol 26:421–424 Lee MT, Ramiro M Jr (2009) Immigration reduces crime: an emerging scholarly consensus. In: Mcdonald
WF (ed) Immigration, crime and justice, sociology of crime, law and deviance, vol 13. Emerald Group Publishing Limited, Bingley, pp 3–16
Light MT, Massoglia M, King RD (2014) Citizenship and punishment: the salience of National membership in US. Criminal courts. Am Sociol Rev 79:825–847
Lodge HC (1909) Speeches and addresses, 1884–1909. Houghton Mifflin, Boston MacDonald JM, Saunders J (2012) Are Immigrant youth less violent? Specifying the reasons and mecha-
nisms. Ann Am Acad Polit Soc Sci 641:125–147 MacDonald JM, Hipp JR, Gill C (2013) The effects of immigrant concentration on changes in neighborhood
crime rates. J Quant Criminol 29:191–215 Maxfield MG, Weller BL, Widom CS (2000) Comparing self-reports and official records of arrests. J Quant
Criminol 16:87–110 Menjivar C, Bejarano C (2004) Latino immigrants’ perceptions of crime and police authorities in the United
States: a case study from the phoenix metropolitan area. Ethn Racial Stud 27:120–148 Morenoff JD, Astor A (2006) Immigrant assimilation and crime: generational differences in youth violence
in Chicago. In: Martinez R Jr, Valenzuela A Jr (eds) Immigration and crime: race, ethnicity and violence. New York University Press, New York, pp 36–63
Mulvey EP, Steinberg L, Fagan J, Cauffman E, Piquero AR, Chassin L, Knight GP, Brame R, Schubert CA, Hecker T, Losoya SH (2004) Theory and research on desistance from antisocial activity among serious adolescent offenders. Youth Violence Juv Justice 2:213–236
Orebaugh DA (1929) Crime, degeneracy and immigration: their interrelations and interreactions. The Gorham Press, Boston
Orrick EA, Piquero AR (2015) Assessing the impact of Mexican nativity on sentence length. Crim Justice Policy Rev 26:643–664
Osgood DW (2000) Poisson-based regression analysis of aggregate crime rates. J Quant Criminol 16:21–43 Osgood DW (2009) Statistical models for life events and criminal behavior. In: Piquero AR, Weisburd D
(eds) Handbook of quantitative criminology. Springer, New York, pp 375–396 Ousey GC, Kubrin CE (2014) Immigration and the changing nature of homicide in US cities, 1980–2010.
J Quant Criminol 30:453–483 Passel JS, Cohn D (2008) U.S. population projections: 2005–2050. Pew Research Center, Washington, DC Pew Research Center (2013) Second-generation americans: a portrait of the adult children of immigrants.
Pew Research Center, Washington, DC Piquero AR (2008) Disproportionate minority contact. Future Child 18:59–79 Piquero AR, Brame R (2008) Assessing the race-/ethnicity-crime relationship in a sample of serious ado-
lescent delinquents. Crime Delinq 54:390–422 Piquero AR, Blumstein A, Brame R, Haapanen R, Mulvey EP, Nagin DS (2001) Assessing the impact of
exposure time and incapacitation on longitudinal trajectories of criminal offending. J Adolesc Res 16:54–74
Piquero AR, Bersani BE, Loughran TA, Fagan J (2014a) Longitudinal patterns of legal socialization in first- generation immigrants, second-generation immigrants, and native-born serious youthful offenders. Crime Delinq. doi:10.1177/0011128714545830
Piquero AR, Schubert CA, Brame R (2014b) Comparing official and self-report records of offending across gender and race/ethnicity in a longitudinal study of serious youthful offenders. J Res Crime Delinq 51:525–555
Pogrebin MR, Poole EE (1990) Culture conflict and crime in the Korean–American community. Crim Justice Policy Rev 4:69–78
Powell D, Perreira KM, Harris KM (2010) Trajectories of delinquency from adolescence to adulthood. Youth Soc 41:475–502
Provine DM, Varsanyi M, Lewis PG, Decker SH (2012) Growing tensions between civic membership and enforcement in the devolution of immigration control. In: Kubrin Z, Martinez J (eds) Punishing immigrants: policy, politics and injustice. New York University Press, New York, pp 42–61
Raudenbush SW, Bryk AS (2002) Hierarchical linear models: applications and data analysis methods, 2nd edn. Sage Publications, Thousand Oaks
856 J Quant Criminol (2017) 33:835–857
123
Rengifo AF, Fratello J (2015) Perceptions of the police by immigrant youth: looking at stop-and-frisk and beyond using a New York City sample. Youth Violence Juv Justice 13(4):409–427
Roberts J (2011) Why misdemeanors matter: defining effective advocacy in the lower criminal courts. UC Davis Law Rev 45:277–372
Rumbaut RG, Ewing WA (2007) The myth of immigrant criminality and the paradox of assimilation: incarceration rates among native and foreign-born men. Immigration Policy Center, Washington, DC
Rumbaut RG, Gonzales RG, Komaie G, Morgan CV, Rosaura T-E (2006) Immigration and incarceration: patterns and predictors of imprisonment among first- and second-generation young adults. In: Martinez R Jr, Valenzuela A Jr (eds) Immigration and crime: race, ethnicity, and violence. New York University Press, New York, pp 64–89
Sampson RJ (2008) Rethinking crime and immigration. Contexts 7:28–33 Sampson RJ, Morenoff J, Raudenbush SW (2005) Social anatomy of racial and ethnic disparities in vio-
lence. Am J Public Health 95:224–232 Schlesinger T (2005) Racial and ethnic disparity in pretrial criminal processing. Justice Q 22:170–192 Schubert CA, Mulvey EP, Steinberg L, Cauffman E, Losoya S, Hecker T, Chassin L, Knight GP (2004)
Operational lessons from the pathways to desistance project. Youth Violence Juv Justice 2:237–255 Sellin T (1938) Culture conflict and crime. Social Science Research Council, New York Simes JT, Waters MC (2014) The politics of immigration and crime. In: Bucerius SM, Tonry M (eds) The
oxford handbook of ethnicity, crime, and immigration. Oxford University Press, New York Snijders T, Bosker R (1999) Multilevel analysis: an introduction to basic and advanced multilevel modeling.
Sage Publications, Thousand Oaks StataCorp. (2013) Stata statistical software: release 13. StataCorp LP, College Station Sutherland EH (1924/1934) Criminology. J. B. Lippincott Co, Philadelphia Theodore N (2013) Insecure communities: Latino perceptions of police involvement in immigration
enforcement. Department of Urban Planning and Policy, University of Illinois at Chicago, Chicago Thornberry TP, Krohn MD (2000) The self-report method for measuring delinquency and crime. In: Duffee
D, Crutchfield RD, Mastrofski S, Mazerolle L, McDowall D (eds) Criminal justice 2000: vol. 4: measurement and analysis of crime and justice. National Institute of Justice/Department of Justice, Washington, D.C, pp 33–83
Thornberry TP, Krohn MD (2003) Comparison of self-report and official data for measuring crime. In: Pepper JV, Petrie CV (eds) Measurement problems in criminal justice research: workshop summary. National Academy Press, Washington, DC, pp 43–94
Tracy PE (1987) Race and class differences in official and self-reported delinquency. In: Wolfgang ME, Thornberry TP, Figlio RM (eds) From boy to man, from delinquency to crime. University of Chicago Press, Chicago, pp 87–121
Treyger E, Chalfin A, Loeffler C (2014) Immigration enforcement, policing, and crime: evidence from the secure communities program. Criminol Public Policy 13:285–322
Trump D (2015) Donald trump presidential campaign announcement [Video file]. http://www.c-span.org/ video/?326473-1/donald-trump-presidential-campaign-announcement
Tyler TR, Schulhofer S, Huq A (2010) Legitimacy and deterrence effects in counter-terrorism policing: a study of Muslim Americans. Law Soc Rev 44:365–401
U.S. Census Bureau’s (2013) American community survey (ACS). http://www.migrationpolicy.org/article/ frequently-requested-statistics-immigrants-and-immigration-united-states
U.S. Government Accountability Office (2009) Immigration enforcement: better controls needed over program authorizing state and local enforcement of federal immigration laws. GAO-09-109. Report to congressional requesters. GAO, Washington, DC. http://www.gao.gov/products/GAO-09-109
Van Vechten CC (1941) The criminality of the foreign born. J Crim Law Criminol 32:139–147 Vaughn MG, Salas-Wright CP, DeLisi M, Maynard BR (2014) The immigrant paradox: immigrants are less
antisocial than native-born Americans. Soc Psychiatry Psychiatr Epidemiol 49:1129–1137 Wadsworth T (2010) Is immigration responsible for the crime drop? An assessment of the influence of
immigration on changes in violent crime between 1990 and 2000. Soc Sci Q 91:531–553 Wickersham Report (1931) National commission on law observance and enforcement no. 10. Report on
crime and the foreign born. Government Printing Office, Washington, DC Wolfe SE, Pyrooz DC, Spohn CC (2011) Unraveling the effect of offender citizenship status on federal
sentencing outcomes. Soc Sci Res 40:349–362 Zatz MS, Smith H (2012) Immigration, crime and victimization: rhetoric and reality. Ann Rev Law Soc Sci
8:141–159
J Quant Criminol (2017) 33:835–857 857
123
Reproduced with permission of copyright owner. Further reproduction prohibited without permission.
- Examining Systematic Crime Reporting Bias Across Three Immigrant Generations: Prevalence, Trends, and Divergence in Self-Reported and Official Reported Arrests
- Abstract
- Objective
- Methods
- Results
- Conclusions
- Introduction
- Prior Literature
- Official and Self-Reported Offending Comparisons
- Immigrant Offending: Generational Differences and Crime Reporting Methodology
- Factors Influencing the Differential Reporting of Offending
- Current Research
- Data, Measures, and Analytic Strategy
- Data
- Measures
- Arrest
- Offending
- Immigrant Generational Status
- Controls
- Analytic Strategy
- Results
- Arrest Trends
- Within Immigrant Generation Crime Reporting Analysis
- Between Immigrant Generation Crime Reporting Analysis
- Robustness Analyses
- Discussion
- Appendix
- References