The Gender Gap in Sex Offender Punishment

Abstract

Objectives
This paper tests theoretical arguments that suggest court actors hold gendered views of sex offenders that result in a gender gap in sex offender punishment, where women who commit sexual offenses are treated more leniently than their male counterparts.

Methods
We test this argument with precision matching analyses using 15 years of data on all felony sex offenders sentenced in a single state.

Results
Results indicate that gender disparities in sex offender sentencing exist and are pervasive across sex offense types. Specifically, male sex offenders are more likely to be sentenced to prison, and given longer terms, than female sex offenders. Findings are similar across sex offense severity and whether the offense involved a minor victim.

Conclusions
These findings suggest that female sex offenders are treated more leniently than their matched male counterparts, even in instances of more serious sex offenses and those involving minor victims. Findings support theoretical arguments that contend that court decision-making is influenced by legally-irrelevant characteristics and raise questions about the source of gendered views of sex offenders and their effects on punishment approaches. Findings also raise questions about the virtue of get-tough sentencing policies that provide leeway for such dramatic variation across different groups of people.

Full text

Vol.:(0123456789)
Journal of Quantitative Criminology
https://doi.org/10.1007/s10940-019-09416-x
1 3
ORIGINAL PAPER
The Gender Gap inSex Oender Punishment
RyanT.Shields1· JoshuaC.Cochran2
© Springer Science+Business Media, LLC, part of Springer Nature 2019
Abstract
Objectives This paper tests theoretical arguments that suggest court actors hold gendered
views of sex offenders that result in a gender gap in sex offender punishment, where women
who commit sexual offenses are treated more leniently than their male counterparts.
Methods We test this argument with precision matching analyses using 15years of data on
all felony sex offenders sentenced in a single state.
Results Results indicate that gender disparities in sex offender sentencing exist and are
pervasive across sex offense types. Specifically, male sex offenders are more likely to be
sentenced to prison, and given longer terms, than female sex offenders. Findings are simi-
lar across sex offense severity and whether the offense involved a minor victim.
Conclusions These findings suggest that female sex offenders are treated more leniently
than their matched male counterparts, even in instances of more serious sex offenses and
those involving minor victims. Findings support theoretical arguments that contend that
court decision-making is influenced by legally-irrelevant characteristics and raise ques-
tions about the source of gendered views of sex offenders and their effects on punishment
approaches. Findings also raise questions about the virtue of get-tough sentencing policies
that provide leeway for such dramatic variation across different groups of people.
Keywords Gender· Punishment· Sentencing· Sex offender
Introduction
Scholarship suggests that community members perceive sex offenders to be a special
class of criminal offenders (Jenkins 1998; Meloy etal. 2013; Quinn et al. 2004). Pub-
lic opinion studies, for example, indicate that sex offenders are perceived as especially
dangerous and unlikely to be rehabilitated (Katz-Schiavone etal. 2008; Levenson etal.
2007; Mancini and Pickett 2016; Pickett et al. 2013; Socia and Harris 2016). Crimi-
nal justice policies since the 1990s have largely reflected these perceptions by way of
* Ryan T. Shields
ryan_shields@uml.edu
1 School ofCriminology andJustice Studies, University ofMassachusetts Lowell, 113 Wilder Street,
Room 445, Lowell, MA01854, USA
2 School ofCriminal Justice, University ofCincinnati, Cincinnati, OH, USA

Journal of Quantitative Criminology
1 3
increasing dramatically the punitiveness and control exerted over sex offenders (Farkas
and Stichman 2002; Jenkins 1998; Petrunik 2002; Terry and Ackerman 2009).
The heightened efforts to punish, control, and incapacitate sex offenders—espe-
cially those who victimize children (Simon 2000)—have spurred on a diverse range of
studies evaluating policy impacts. This literature has, however, paid far less attention
to understanding variation in how sex offender punishments are applied. For example,
although prior research documents substantial heterogeneity among felony-convicted
sex offenders and the nature of their offenses (Ackerman etal. 2011), we know little
about how criminal justice actors make sanctioning decisions and how characteristics of
sex offenders and their crimes influence those decisions.
In particular, theoretical arguments about citizens’ perceptions of sex offenders sug-
gest that gender may exert an especially salient influence on court punishment decisions
for sex offenders (e.g., Denov 2001; Mellor and Deering 2010). The sex offender litera-
ture suggests that the typical social construction of a sex offender or sex predator is a
man (Denov 2001; Gakhal and Brown 2011). This aligns in part with actual offending.
For example, sexual abuse perpetration is disproportionately male (Smith etal. 2017).
And, when women do perpetrate sex crimes, they and their offenses are perceived as
less threatening or harmful compared to those committed by men (Denov 2003; Denov
and Cortoni 2006; Mellor and Deering 2010). Studies of public perceptions suggest that
these views about gender and culpability apply across sex offense types despite the fact
that research indicates that, like male-perpetrated crimes, female sex offenses regularly
involve child victims (Williams and Bierie 2015) and exert similar mental, physical, and
behavioral health consequences on victims (for review of consequences of child sexual
abuse, see Institute of Medicine 2013; Kaufman 2010).
Research examining society’s perceptions of, and responses to, sexual violence
through a gendered lens is needed (see, for example, Cortoni 2015). Such work is criti-
cal for advancing theory, research, and policy related to sex crimes, gender, and social
control. It can, for example, test the extent to which hypotheses about gender gaps in
sex offender sentencing hold true and extend prior work that has explored, generally, the
role of gender in criminal sentencing (e.g., Blackwell etal. 2008; Doerner and Demuth
2014). More than that, if there is a consistent gender gap across sex offense types, it
would suggest support for theoretical arguments about the gendered nature of percep-
tions of sex crimes and how those views manifest in criminal justice decisions. It would,
at the same time, underscore critical but unanswered research and policy questions
about the potential consequences of differential treatment of sex offenders by gender
and other demographic characteristics.
To that end, the goal of this study is to advance theory and research on sex offender
punishment and, more broadly, on gender gaps in criminal sentencing by systematically
assessing whether a gender gap exists in sex offender punishments and the robustness of
that gender gap across sex offense types. Our analyses utilize 15years of felony court
sentencing data from a single state. Towards this goal, the article proceeds as follows.
First, we discuss the prior sex offending and sex offender punishment scholarship that
provides context in which to situate the study’s analyses. Second, we review the litera-
ture on female sex offenders and identify gaps in knowledge around this subgroup and
their experiences with sanctions. Third, we present theoretical arguments that address
whether and how gender influences courtroom decision-making. Fourth, we present
the current study, describing the data, analytic approach, and findings. Fifth, we close
with a discussion of results, implications for sentencing theory, and directions for future
scholarship.

Journal of Quantitative Criminology
1 3
Background
Sex Oending andSex Oender Punishment
The United States experienced a dramatic decline in rates of sexual violence over the past
two decades. According to the National Crime Victimization Survey, rates of sexual vic-
timization of teens dropped by approximately 40–60% between 1993 and 2004 (Finkelhor
and Jones 2006). Reports from other national surveys, official criminal and juvenile justice
data, and child welfare agencies suggest similar downward trends in rates of sexual vio-
lence (Finkelhor and Jones 2006, 2012; Planty etal. 2013).
During the same period of time that child sexual victimization was declining, media and
public attention to sexual violence and victimization increased substantially (Leon 2011;
Sample and Bray 2003). Scholars argue that this increased attention to “sexual predators”
amplified the public’s fear about the threat of sex crimes, especially those involving chil-
dren, and led to greater support for less treatment and more punitive approaches towards
them despite actual downward trends in actual victimization (Levenson etal. 2014; Man-
cini 2014; Meloy etal. 2013). For example, in a nationally representative telephone sur-
vey, the vast majority of respondents (approximately 97%) supported the view that any sex
offenses that involved minor victims warranted incarceration over probation, community
treatment, or fines (Mears etal. 2008). Further, nearly half of all respondents (48%) in that
survey reported that they would be unwilling to pay additional taxes to fund sex offender
treatment. Other public opinion surveys identified similar results for other sex crime poli-
cies (Levenson etal. 2007; Levenson etal. 2014; Mancini etal. 2010).
Policymakers responded quickly to public concern about sex offenders. Beginning in
the 1990s, several policies were directed squarely at the incapacitation, management, and
supervision of sex offenders (Mancini et al. 2013; Sample 2011). For example, scholar-
ship suggests that sex offenders receive especially lengthy terms of incarceration (Budd
and Desmond 2014; Cohen and Jeglic 2007; Greenfeld 1997). Apart from incarceration,
sex offenders are subjected to a range of unique post-incarceration civil sanctions that
include civil commitment evaluation (Harris 2009), placement on the sex offender regis-
try (Levenson 2009), community notification requirements (Zevitz 2006), GPS monitoring
(Armstrong and Freeman 2011), residency restrictions (Levenson 2009), castration (Scott
and del Busto 2009), and other types of restrictions on community participation (see, e.g.,
Chaffin etal. 2009).
Female Sex Oenders
The literature on sexual offending and its attendant policy responses has been largely
informed by samples of male offenders. The omission of females from empirical analyses
of this research may be due to substantially small sample sizes of female offenders in avail-
able datasets (e.g., Amirault and Beauregard 2014). It may also stem from the fact that typ-
ical representations of sexual crimes involve a male offender sexually assaulting a woman
or child (Vandiver and Walker 2002). As Embry and Lyons (2012, p. 152) described, the
social construction of sex offender is that of a male predator, and women “are seen as inac-
tive participants or merely a bystander to sexual assault.” Thus, sex offenders are inherently
considered to be male, and male sex offenders are inherently perceived to be more threat-
ening to the community and as exerting more agency over their offending behaviors.

Journal of Quantitative Criminology
1 3
Women contribute a non-trivial amount of sex offender perpetration and they engage in
sex crimes of similar severity as men (Denov 2003; Williams and Bierie 2015). Estimat-
ing precisely the prevalence of female sex offending is a challenge, however, for at least
two reasons. First, scholars believe that female-perpetrated sexual offending is underre-
ported (Center for Sex Offender Management [CSOM] 2007; Vandiver and Kercher 2004).
Underreporting of female-perpetrated sexual abuse may stem in part from societal per-
ceptions of female sexual passivity or incapability of committing sexual harm (Anderson
and Swainson 2001; Denov 2003; Denov and Cortoni 2006). Second, a limited number of
studies suggest that female sex offenders are more likely to be diverted away from official
punishments (Vandiver and Walker 2002) and thus be omitted from official conviction and
punishment statistics. In this latter scenario, and as we discuss below, scholars suggest that
females receive more lenient treatment from the courts because of their perceived lack of
dangerousness (CSOM 2007).
Despite these challenges, best estimates indicate that females account for anywhere from
1 to 20% of known sex offenses depending on the type of offense. For example, a recent
meta-analysis of studies using police and victimization data from 12 countries including
the U.S. estimated that while females account for approximately 2% of officially reported
sex crimes, victimization data reported female perpetrators in 12% of cases (Cortoni etal.
2017). CSOM (2007) estimated that women account for only 1% of all rape arrests, but
they account for 6% of all other sex offense-related arrests. McLeod (2015) examined sub-
stantiated child sexual abuse reports in the United States in 2010 and found that 20% of
cases involved a primary female offender. Finkelhor etal. (2009) focused on youth crimes
and estimate that 7% of all sexual abuse perpetrated by youths involve a female offender.
Indeed, the prevalence of female offenders may be highest among child sexual abuse
cases, where estimates indicate that 15–20% of cases involve female perpetrators (see, e.g.,
McLeod 2015).
In light of the overwhelming focus on male sex offending patterns, scholars have argued
that a gendered approach to understanding female sex offending is warranted (Cortoni and
Gannon 2016; Gannon et al. 2014; ten Bensel etal. 2016; Williams etal. 2019). Though
there are similarities between male sex offenders and female sex offenders, important dif-
ferences exist. For example, compared to male sex offenders, female sex offenders are more
likely to offend against children in their families or under their care (Wijkman etal. 2010),
demonstrate more fluid gender preferences (Freeman and Sandler 2008; West etal. 2011;
Williams and Bierie 2015), and report significant abuse and trauma histories (Elliott etal.
2010; Gannon etal. 2008; West etal. 2011; Strickland 2008).
Notably, compared to male sex offenders, female sex offenders are also more likely to
co-offend, usually with another male (Wijkman etal. 2010; Williams and Bierie 2015; Wil-
liams etal. 2019), and there are differences between females who sexually offend alone
and those who offend with a counterpart. For example, in a recent study of female solo-
and co- sex offenders and male sex offenders, Williams etal. (2019) found that female sex
offenders demonstrated lower self-esteem, assertiveness, and higher levels of loneliness
than male offenders. Female solo offenders also reported more mental health problems
than female co-offenders or male sex offenders.
Researchers have developed different typologies to classify female sex offending (e.g.,
Matthews etal. 1991; Vandiver and Kercher 2004). More recently, Gannon etal. (2008)
expanded this line of work and developed the Descriptive Model of Female Sex Offending,
theorizing three primary patterns of female offending: Explicit Approach, where women
planned their offending behaviors, experienced little coercion from a counterpart, and had
experienced positive affect as a result of their offending; Directed-Avoidant, where women

Journal of Quantitative Criminology
1 3
did not plan their offenses, experienced high levels of coercion, and experienced negative
affect from their offending; and Implicit-Disorganized, where women did not plan their
offenses, experienced either positive or negative affect, and whose offending appeared to
be highly impulsive. A more recent test of this theory found preliminary support for these
pathways (Gannon etal. 2014).
Public opinion about sex offenders has been largely informed by sensationalized popu-
lar media accounts of sexual abuse (Christensen 2018; Malinen etal. 2014) and the result-
ant fear of sex offenders has been a driving force behind legislation aimed at punishment
and management of these offenders. Although public views of sex offenders are well-doc-
umented, a smaller literature has focused on how the public views female sex offenders
specifically. Scholars have noted that female sexual violence tends to be either ignored,
minimized, or viewed more positively than their male counterparts (King and Roberts
2017; Rogers and Davies 2007), even by professionals who work in the field of sexual
abuse (Clements etal. 2014; Gakhal and Brown 2011). In one study of Nebraska residents
(Cain etal. 2017), respondents were asked to respond to the prompt that female sex crimes
are less serious than crimes committed by men. Although 64% of respondents reported
that they disagreed or disagreed strongly with the statement, nearly 12% either agreed
or strongly agreed that female sex crimes are less serious than those committed by men.
Given the harms associated with female perpetrated sexual abuse, this finding is striking. If
the general public demonstrates less fear over female sexual offending, then it is logical to
expect that punishment decision-making will reflect a similar value system.
Is There aGender Gap inSex Oender Punishment?
Extant scholarship shows that gender structures both sexual offending patterns and how
sex offenders are perceived—even when sex offenses are serious and involve children—
which suggests that there may be gender disparities in how sex offenders are punished. It
is unclear whether, in the context of the punitive turn for sex offenders, punishments for
sex offenders are uniformly assigned or if they are disproportionately more severe for men.
As we argue below, there are strong theoretical reasons to anticipate that gendered social
perceptions of sex crimes play a role in how courts respond to sex offenders, that gender
impacts sanctioning regardless of the type of sex crime, and that male sex offenders, all
else equal, receive harsher treatment from state courts than do their female sex offender
counterparts.
The primary basis of our hypotheses is formed by two key arguments. The first argu-
ment stems from prior literature on attributions theory, which suggests a pathway through
which extralegal factors like gender can influence criminal sentencing (Doerner and
Demuth 2014; Spohn 2013; Spohn and Beichner 2000; Steffensmeier and Demuth 2006).
Attribution-based theoretical frameworks (Albonetti 1991; Steffensmeier etal. 1998) sug-
gest that the social constructions of offense seriousness and offender dangerousness among
society members likely influence court decision making because judges and prosecutors
rely on perceptual shorthands about individuals to account for limitations in knowledge
about the offender in a context where rapid case decisions are required (Albonetti 1991).
Use of these shorthands allows for biases to manifest. In the context of sex offenses, to
the extent that gender structures how court actors perceive the nature and seriousness of
sex offenses and that it does so regardless of the type of sex crime or the type of victim
involved, we should expect a gender gap in sex offender sentencing generally, across the
various types of sex offenses (see also Daly 1987).

Journal of Quantitative Criminology
1 3
By extension, the second argument is based on public perceptions research described
earlier which suggests that gender does indeed structure perceptions of sex offender dan-
gerousness and will do so even when sex offenses are violent, involve minor victims, or
would otherwise be considered serious. Gender appears to be uniquely salient in the nar-
ratives and rhetoric that surround sex offenses, as offenders are consistently portrayed as
being highly dangerous, predatorial, and male (Quinn etal. 2004). In line with attribu-
tions perspectives, the small proportion of sex crimes committed by females paired with
the relative lack of public concern over female perpetration of sexual assault suggest the
possibility that court actors may perceive female sex offenders as posing less of a threat to
community members and, in turn, uniformly treating female sex offenders more leniently
than male sex offenders.
Our logic aligns with prior applications of attributions perspectives developed to under-
stand how sentencing disparities emerge. For example, one of the most consistently cited
attribution-based theoretical frameworks, focal concerns (Steffensmeier et al. 1998),
argues that judges rely on three dimensions for sentencing decisions: offender danger-
ousness, blameworthiness, and practical constraints of the court. In the context of gender
and sex offenders, female sex offenders may be treated less harshly simply because males
are consistently portrayed as the more dangerous and culpable group (Embry and Lyons
2012). Studies find, for example, that court actors may be more likely to perceive victims
of female perpetrators, minor victims included, as more compliant in the crime than with
male offenders (Hassett-Walker etal. 2014) and to be more generally inclined to diminish
the seriousness of male sexual victimization by a female (e.g., Hayes and Carpenter 2013).
In contrast, the gender gap in sex offender sentences could be conditional on the charac-
teristics of the sex crime. This idea is supported by a set of corollary arguments of attribu-
tions frameworks that suggest that female sex offenders may receive more serious punish-
ments than male sex offenders in some instances. For example, the “selective chivalry”
or “evil woman” thesis argues that paternalistic judges might generically perceive female
offenders as weak and requiring protection and so seek to punish female offenders more
leniently except when their behavior contradicts typical gender roles attributed to them
(Bernstein etal. 1977; Moore and Padavic 2010). In these instances, and according to this
argument, we would anticipate a sex offender gender gap to diminish, if not go away alto-
gether. The expectation is that females who engage in crimes that deviate from normative
attributions and their perceived social status (e.g., that of a caretaker or mother) may in
fact be punished twice: once for the violation of the law and again for the violation of an
implicit, and gendered, social code. This perspective is especially important for consider-
ing court treatment of females who commit sex offenses, and especially those that involve
a child victim, because judges may view those crimes as gender atypical or antithetical
to “being female.” In these cases, it may be that female sex offenders are punished three
times: once for violating law, another for violating gender norms, and once again for violat-
ing sexual norms. (Our empirical analyses will shed some light on this possibility.)
Prior Empirical Studies ofGender andSex Oender Punishment
Four prior empirical studies of court decisions suggest that male and female sex offenders
are processed and punished differently. In one exploratory study of case processing of male
(n = 122) and female (n = 61) youth who sexually offended, Vandiver and Teske (2006)
found that boys received longer sentences than girls. Specifically, 27% of boys in the sam-
ple and 5% of girls received sentences of 10 or more years. The remaining three studies

Journal of Quantitative Criminology
1 3
focused on adult sex offenders. The results of these analyses indicate that women are typi-
cally sanctioned more leniently than men under certain conditions. For example, Sandler
and Freeman (2011) used arrest and sentencing data from New York covering a period of
20years (1986–2005) and found that although there were no significant sex differences in
convictions, female sex offenders were less likely to be incarcerated upon conviction than
their male counterparts. Embry and Lyons (2012) examined National Corrections Report-
ing Program data from 1994 to 2004 and found that men were sentenced to longer prison
terms for rape, child sexual assault and forcible sodomy, but not for sexual assault, the
very offense type where the male–female perpetrator ratio is most even. In Hassett-Walker
etal.’s (2014) analysis, only those cases involving noncustodial adolescent victims demon-
strated differences in female sentencing.
Although these studies have provided important insight into female sex offender sen-
tencing, they are limited in number and in methodological rigor, and overall indicate mixed
results. For example, among these studies, analyses are restricted by small samples, and
when sample sizes are larger (e.g., national datasets), the empirical analyses have been
unable to account for important characteristics of the offenders and the crimes, such as the
severity of a given offense and whether an offense involved a minor victim. These factors
may vary substantially across perpetrator gender, and not accounting for these differences
is likely to influence results.
Current Study
The analyses presented below address the research gaps described above and test the
hypothesis that female sex offenders receive overall more lenient sentencing outcomes.
Analyses utilize a large, multi-year sample of all convicted felony sex offenders in a sin-
gle state (Florida). Our approach employs a precision (exact) matching methodology that
matches female perpetrators to male perpetrators using a range of theoretically relevant
dimensions. We describe the data and matching dimensions in more detail below, followed
by a description of our study’s results and the conclusions that stem from them.
Data
The data for this study come from the Florida Department of Corrections’ Sentencing
Guidelines database. Sentencing guidelines were enacted in Florida in 1983 in response
to concerns about the lack of uniformity in sanctions (Griswold 1989). Revised guidelines
appeared in 1995, when the guidelines established felony severity scores, and again in
1998, which lowered mandatory prison thresholds and allowed judges greater discretion
in departing from the guideline’s recommended sentences. Sentencing guidelines assigned
points to case/offender characteristics and make recommendations for type of sanction
(incarceration or not) and length of sanction. Scoresheets were completed by court staff
and submitted to the Florida Department of Corrections.
This dataset includes information on all convicted felony sex offenders sanctioned in
the state of Florida between 1995 and 2010. Females represent a relatively small but not
negligible proportion of convicted sex offenders and this large observation window allows
us to capture a substantial proportion of female sex offenders convicted of a variety of sex
offenses. The dataset used here includes only unique sex offenders—in instances when a

Journal of Quantitative Criminology
1 3
given individual experienced multiple felony sex convictions during this time period, only
the most recent conviction event is included in these analyses.
Florida is advantageous as a study site for analyzing patterns of the use of formal social
control for sex offenders for at least two reasons. First, Florida’s (registered) sex offender
resident rate is one of the highest in the country (National Center for Missing and Exploited
Children 2015). Second, Florida’s sentencing policies have undergone significant changes
over the past two decades, at the center of which are a number of increasingly punitive
responses to sex offenders. These changes are overall reflective of the national shift in
sanction policies for sex offenders (Budd and Desmond2014; Cohen and Jeglic 2007).
Analytic Strategy
Our analyses employ a precision (i.e., perfect, exact, variable-by-variable) matching tech-
nique. Precision matching, like other matching methodologies, is useful for estimating
“treatment” effects on an outcome—here, gender effects on sentencing decisions for sex
offenders—by creating equivalent groups and approximating random assignment (Guo and
Fraser 2010). In this case, we match “treated” cases (females) to a control group (males)
using a range of observed covariates. Although gender status is not a traditional treatment
that could be randomly assigned, the goal here is to create samples of convicted sex offend-
ers that are identical along all potential confounders, except for gender. This procedure
provides a more rigorous estimate of the impact of gender on punishments issued to sex
offenders by the court than what might be accomplished by a typical control methodology
in a regression framework (Nagin etal. 2009; Apel and Sweeten 2010).
The precision matching methodology is the most rigorous matching design for control-
ling for potential confounding variables (Nagin etal. 2009; Bales and Piquero 2012). It is also
uniquely suited to serve the purposes of this study, more so than more common methodologies
like propensity score matching, for two reasons. First, because of the relatively large size of our
sample, and the especially large number of control (male) cases compared to treated (female)
cases, we have the capacity to conduct precision matching. Precision matching, especially when
it includes a large number of matching dimensions and continuous matching variables, can be
difficult. Researchers have referred to this as a dimensionality curse—fine-grained, precision
matching is rigorous, but it can make identifying a matched sample difficult. The dimensional-
ity curse is most problematic when samples are too small and when there are a large number
of treated cases (Blackwell etal. 2008; Savolainen etal. 2013). For this analysis, our pool of
treated cases (female sex offenders) is small and we have a substantially large pool of potential
male matches, which makes our sample ideal for conducting a precision matching analysis.
Second, the precision matching approach does not require an estimate of an individual’s
propensity to receive treatment. In our case, we would be estimating a propensity for an
offender to be female, which is an implausible conceptual design. Instead we match on a
variable-by-variable basis. This is an intuitive procedure that creates a group of female sex
offenders identical to that of the matched male offenders on all observed dimensions.
For the precision matching analyses described below, we employed a 1:k (“one to many”)
matching design. After matching, the matched cohort is then used to estimate multinomial
logistic regression models of sentencing outcomes on gender, along with sentencing year and
judicial circuit dummy variables, to test whether gender is associated with the likelihood that
a sex offender receives tougher or more lenient sanctions. The multinomial logistic regression
analyses include weights to account for the unbalanced nature of the matched cohort due to

Journal of Quantitative Criminology
1 3
the use of the 1:k matching specification. Ancillary analyses using a 1:1 precision match-
ing analysis without replacement led to a substantially smaller matched sample, but findings
using the smaller sample were substantively the same as those presented in the paper.
To test the robustness of gender’s influence on sex offender sentencing, we conducted
two additional sets of analyses, which are also presented below. First, we conducted a
series of analyses focused on subgroups, based on the type of offense for which an individ-
ual was convicted. These analyses are used to differentiate between two competing argu-
ments discussed above—one that suggests that there should be a general gender gap robust
across offense seriousness or victim status and another that suggests that for more serious
sex offenses the gender gap should diminish. For these analyses, we examine multinomial
logistic regression models on subsamples, based on the type of sex crime for which an
individual was convicted and whether the offense included a minor victim.
Second, we assessed whether gender influences the length of a given sanction. We con-
ducted three separate precision matching analyses—for offenders who received probation,
intensive probation, and prison—and then used negative binomial regression to assess
whether gender influences the amount of time included in a sentence. Similar to above, these
models assess a gender impact and also the possibility that as offenses become more severe,
such as if they involve a sexual assault or a minor victim, gender disparities will diminish.
Dependent Variables
The dependent variable for the multinomial logistic regression analyses is a four-category
sanction outcome variable. Sex offenders in this sample were assigned to one of four sanctions:
probation, intensive probation, county jail, or state prison. For the negative binomial regres-
sion analyses examining length of sentence, we used a count measure of court-designated sen-
tence length for probation, intensive probation, and state prison, measured in monthly units.
Gender andMatching Covariates
We performed precision matching using gender status as treatment (male = 1, female = 0) and
using a range of matching variables that tap into key demographic, prior record, court, and
offense characteristics. Specifically, cases were matched on a continuous measure of age, and
three mutually exclusive dummy variable indicators of racial and ethnic status (non-Latino
white, non-Latino black, and Latino). We included a dummy variable indicator of whether a
given case went to trial along with court-designated sentencing score components based on the
Florida sentencing guidelines database. The guidelines database provides a continuous prior
record score that accounts for the number of prior convictions and seriousness of prior convic-
tions for any offense committed as a juvenile or adult (see Florida Department of Corrections
2015). We also used a measure of offense severity, based on the guidelines database, which
rates the relative severity of a given offense from 1 to 10, with higher values indicating greater
offense severity. Offense severity levels are predetermined and comparable across different
offense types. The levels are assigned based on a given statute violation and theoretically aim
to be “…commensurate with the harm or potential for harm to the community that is caused
by the offense, as determined by statute” (Florida Department of Corrections 2015, p. 16). In
addition, we included a “scored to prison” dummy variable measure which indicates whether
the combined sentencing score, after accounting for prior record and severity level together,
scored an individual high enough to justify a prison sentence (according to Florida’s sentenc-
ing guidelines). Not least, we included a dummy variable indicator of whether a given offense

Journal of Quantitative Criminology
1 3
included a minor victim (under 18), along with six mutually exclusive dummy variables to
measure the type of offense as one of six felony-level statute violation types: (1) lewd and las-
civious molestation, (2) other sexual assault, (3) second degree sexual battery, (4) first degree
sexual battery, (5) forced sexual assault, or (6) capital sexual assault. These offense types are
listed here and in the table in order of lowest to highest seriousness.
Results
Table1 presents descriptive statistics for the full sample along with a descriptive break-
down of the male subgroup and the female subgroup. Inspection of Table1 reveals sev-
eral important findings, including the fact that females, as expected, constitute only a small
Table 1 Descriptive statistics, convicted felony sex offenders in Florida, 1994–2011
Sentence length measures only applicable to those sentenced to prison, jail, intensive probation, or proba-
tion, respectively
Full sample
(N = 27,243)
Males (N = 26,677) Females
(N = 566)
Mean SD Mean SD Mean SD
Treatment variable
Female (1/0) 0.02 0.14 – – – –
Sentencing outcomes
Prison (1/0) 0.57 0.49 0.58 0.49 0.34 0.47
Prison sentence length (mos.) 132.10 162.47 132.58 162.96 93.63 108.85
Jail (1/0) 0.03 0.18 0.03 0.18 0.05 0.21
Jail sentence length (mos.) 10.43 66.81 10.49 67.62 8.01 6.92
Intensive probation (1/0) 0.12 0.32 0.11 0.32 0.20 0.40
Intensive probation sentence length (mos.) 20.70 11.77 20.77 11.90 18.64 6.70
Probation (1/0) 0.28 0.45 0.28 0.45 0.42 0.49
Probation sentence length (mos.) 135.64 812.21 136.74 817.26 91.05 571.21
Matching covariates
Age (cont.) 34.70 13.18 34.78 13.25 31.04 8.82
White (1/0) 0.56 0.50 0.56 0.50 0.72 0.45
Black (1/0) 0.28 0.45 0.28 0.45 0.19 0.39
Latino (1/0) 0.15 0.36 0.15 0.36 0.08 0.27
Trial (1/0) 0.07 0.25 0.07 0.25 0.05 0.21
Offense severity (cont.) 7.34 1.45 7.35 1.44 6.86 1.55
Minor victim (1/0) 0.83 0.38 0.82 0.38 0.92 0.28
Prior record score (cont.) 5.60 15.25 5.70 15.38 1.06 3.86
Scored to prison (1/0) 0.90 0.30 0.90 0.30 0.82 0.39
Lewd and lascivious (1/0) 0.68 0.46 0.68 0.47 0.70 0.46
Other sexual assault (1/0) 0.01 0.07 0.00 0.07 0.02 0.14
2nd deg. sexual battery (1/0) 0.12 0.33 0.12 0.33 0.16 0.37
1st deg. sexual battery (1/0) 0.08 0.27 0.08 0.27 0.07 0.25
Forced sexual assault (1/0) 0.05 0.22 0.05 0.22 0.02 0.14
Capital sexual assault (1/0) 0.06 0.24 0.06 0.24 0.02 0.14

Journal of Quantitative Criminology
1 3
proportion of convicted felony sex offenders. In Florida, 2% of felony sex offenders are
females. Over the 17-year study period, there are 566 female felony sex offenders con-
victed in the state. By contrast, there were 26,977 male felony sex offenders convicted.
The average age of female sex offenders in our sample was 31, and the average age
of men in our sample was 34.8. A greater proportion of female sex offenders were white
(72% compared to 56% for men). In general, men are receiving convictions for overall
more severe statute violations and have more serious prior records—men scored higher
in offense severity (
x=7.35
; SD 1.44; and prior record (
x=5.7
; SD 15.38) than women
[(
x=6.86
; SD 1.55 and (
x=1.06
; SD 3.86), respectively). Across sentencing outcomes
for sex offenders, prison is the most commonly assigned sanction type for the full sample
and for males (57% and 58%, respectively). Females, however, received prison a substan-
tially lower 34% of the time. Probation is the most frequently used sanction for females,
which was assigned 42% of the time, compared to 28% for males. Females also receive
overall higher proportions of jail (5% for females compared to 3% for males) and intensive
probation sentences (20% for females compared to 11% for males) than males.
Juxtaposed against the substantial differences in the proportion of males compared to
females who go to prison, we see only an 8 percentage point difference between the pro-
portion of males (90%) compared to females (82%) actually eligible, according to the sen-
tencing guidelines score totals, for a prison sentence. Yet, the percentage of males receiv-
ing prison is 24 percentage points higher than that of females (58 vs. 35). This discrepancy
between prison eligibility and actual assignment to prison suggests that there is an overall
greater inclination among court actors to imprison male sex offenders than to do so for
females.
Gender Eects onSex Oender Sanction Decisions
This sanction disproportionality may stem from a perception held by court actors that
male sex offenders are more predatorial or pose greater threats to communities than do
female sex offenders. At the same time, these differences may stem from other sources,
such as gender differences in average offense severity or offenders’ prior record or experi-
ences with incarceration. For example, we see that offense severity levels are higher for
males compared to females (7.35 vs. 6.86, respectively) and that males have overall higher
prior record scores (5.70 vs. 1.06, respectively). These and other compositional differences
might explain any disparities we see in sanction assignment. We test this possibility below
using precision matching analyses.
Table2 provides results of the 1:k precision matching analyses and descriptive statistics
of the matched sample divided by control (male) and treatment (female) groups. Gener-
ally, the matching estimation resulted in the identification of at least 1 matched male sex
offender for the vast majority of female offenders. Specifically, 481 out of 566 female
sex offenders could be exactly matched; a success rate of 85%. This treatment group was
matched to 6080 male control cases. The total matched sample includes 6561 cases. The
descriptive statistics and multivariate analyses below include sampling weights to account
for the unbalanced matching procedure (i.e., more control cases than treatment cases).
The top of Table 2 includes the percentage breakdown of sanction assignment for
matched males and females. Here again, even after matching, we see two substantial dis-
proportionalities: (1) a substantially larger percentage of males receive prison than do
females (47% compared to 33%, respectively), and (2) a substantially larger percentage
of females receive probation than do males (42% compared to 34%, respectively). The

Journal of Quantitative Criminology
1 3
descriptive statistics across all matching measures for matched male and female offenders
are identical, which is indicative of a successful precision matching procedure and which
suggests that sanction disproportionalities are not due to the observed variables included in
the analysis.
We turn now to a series of analyses that employ the matched sample created above to
estimate whether gender exerts a statistically significant effect on sanction assignment
across the four main sanction types after accounting for sentencing year and judicial cir-
cuit. We then test whether this effect is conditional on the nature of a given sex offense.
Table3 presents two multinomial logistic regression analyses. Panel A includes only
a dummy variable measure of gender status in the regression equation and uses robust
standard errors to account for clustering of cases within years and judicial circuits. To
test the robustness of the gender coefficient estimate, panel B presents an alternative,
more rigorous modeling strategy to account for this clustering by adding year and judi-
cial circuit dummy variables to the regression equation. To conserve space and to ease
interpretation of the results, coefficient estimates for year and judicial circuit measures
are not shown, but are available upon request.
Table 2 Post-matching statistics, 1:1 precision matching, with weighting
Males Females
Matching results
All cases 26,677 566
Matched cases 6080 481
Percent matched 23 85
Mean SD Mean SD
Sentencing outcomes
Prison (1/0) 0.47 0.50 0.33 0.47
Jail (1/0) 0.03 0.18 0.04 0.19
Intensive probation (1/0) 0.16 0.36 0.21 0.41
Probation (1/0) 0.34 0.47 0.42 0.49
Matching covariates
Age (cont.) 30.55 8.59 30.55 8.60
White (1/0) 0.76 0.43 0.76 0.43
Black (1/0) 0.16 0.37 0.16 0.37
Latino (1/0) 0.07 0.26 0.07 0.26
Trial (1/0) 0.03 0.16 0.03 0.16
Offense severity (cont.) 7.00 1.28 7.00 1.28
Minor victim (1/0) 0.95 0.21 0.95 0.21
Prior record score (cont.) 0.49 2.03 0.49 2.03
Scored to prison (1/0) 0.86 0.35 0.86 0.35
Lewd and lascivious (1/0) 0.75 0.43 0.75 0.43
Other sexual assault (1/0) 0.01 0.08 0.01 0.08
2nd deg. sexual battery (1/0) 0.16 0.36 0.16 0.36
1st deg. sexual battery (1/0) 0.06 0.23 0.06 0.23
Forced sexual assault (1/0) 0.01 0.12 0.01 0.12
Capital sexual assault (1/0) 0.02 0.14 0.02 0.14

Journal of Quantitative Criminology
1 3
Table 3 Multinomial logistic regression analysis of a 4-category sanction measure on gender using a
matched sample, with weighting (N = 6561)
Probation serves as the reference category. RSE robust standard errors. Panel A does not include judicial
circuit dummy variables or sentencing year dummy variables. Instead, robust standard errors are employed
to account for clustering. Panel B includes the circuit and year dummy variables. Coefficients not shown,
but are available upon request
*p < .05; **p < .01; ***p < .001
Intensive probation Jail Prison
b RSE b RSE b RSE
Panel A
Female 0.061 0.111 − 0.103 0.251 − 0.584*** 0.078
Intercept − 0.769*** 0.158 − 2.266*** 0.180 0.339*** 0.069
Log pseudolikelihood − 7426.166
Pseudo R-squared 0.0025
Intensive probation Jail Prison
b SE b SE b SE
Panel B
Female 0.033 0.134 − 0.199 0.257 − 0.685*** 0.114
Circuit dummies (not shown) – – – – – –
Year dummies (not shown) – – – – – –
Intercept − 1.723** 0.500 − 3.659** 1.110 1.588*** 0.270
Log pseudolikelihood − 6833.887
Pseudo R-squared 0.0821
0.46
0.18
0.02
0.33
0.35
0.13
0.02
0.50
0.0
0.1
0.2
0.3
0.4
0.5
0.6
Probation Intensive
Probation
Jail Prison
Females Males
Fig. 1 Predicted likelihood of sanction type, by gender (panel B estimates)

Journal of Quantitative Criminology
1 3
The models presented in panel A and B tell a consistent story—female sex offend-
ers are significantly less likely to receive a prison sanction than are male sex offend-
ers. Across the three other sanction categories, coefficients indicate that females are less
likely to receive jail than probation and slightly more likely to receive intensive pro-
bation than probation, compared to males. However, coefficient estimates for the non-
prison sanction categories did not reach statistical significance.
To ease the interpretation of these results, Fig.1 presents the predicted likelihoods
of sanction assignment, based on the more rigorous model estimate in panel B of
Table3. Inspection of the figure indicates that females are substantially less likely to
go to prison—their predicted likelihood is 0.33, which is 17 points lower than that of
their precisely matched male sex counterparts (0.50). We see nearly identical and low
predicted likelihoods of a jail sanction for male and female sex offenders, a higher like-
lihood of receiving community control for females than males (0.18 vs. 0.13, respec-
tively), and a substantially higher likelihood of probation for females than males (0.46
vs. 0.35). In short, female sex offenders, all else equal, are most likely to receive a pro-
bation sanction, followed by prison, intensive probation, then jail. For male sex offend-
ers, they are most likely to go to prison, followed by probation, intensive probation, then
jail.
Results to this point support arguments that male sex offenders, all else equal, are per-
ceived to be more dangerous and culpable than their female counterparts and so experience
more punitive and incapacitating sanctions. Is this gender gap robust across different types
of sex crimes and victims? On one hand, theory suggests that all sex offenses, even those
that are more serious and that involve minor victims, are perceived to be more serious and
dangerous when committed by men. Thus, the gender gap should be consistent, regard-
less of offense and victim characteristics. On the other hand, a selective chivalry argument
suggests that more serious sex crimes, especially those against children, violate gender
stereotypes when they are committed by women and so there should be less of a gender
discount. We now turn to analyses that explore these possibilities by estimating the gender
gap across theoretically relevant subgroups.
Table 4 provides estimates from four different models. Here, we split the matched
sample into two broad offense categories: (1) lewd and lascivious offenses and (2) sexual
assault and sexual battery offenses. The latter category includes all sex offense types except
lewd and lascivious offenses. Smaller subsamples—for example, subsamples based on
the degree of sexual assault or sexual battery—create overly small sample sizes with lim-
ited statistical power that could not support the analysis. We ran separate models for these
two subgroups. We also ran one additional model for each subgroup, analyzing only those
offenses of a given type that involved a minor victim—arguably the most egregious version
of either of these offense subgroups. The models in Table4 include robust standard errors
to account for year and circuit clustering. We ran ancillary analyses including year and cir-
cuit dummy variables in the modeling, and findings were substantively the same. However,
in some instances, the samples were too small to allow for inclusion of such a high number
of variable estimates in a single model, so we show the results that employ robust standard
errors instead.
Across panels A—D, findings parallel those of the previous analyses. Specifically, we
see a negative, statistically significant effect of female status on the likelihood of incar-
ceration compared to males and relative to probation in all four models. The coefficient
estimates for the other sanction types do not reach statistical significance. Thus, findings
in Table4 suggest that gender status impacts sex offender-sanctioning decisions generally

Journal of Quantitative Criminology
1 3
and regardless of the seriousness of the offense and regardless of whether a minor victim is
involved.
Gender Eects onSentence Length Decisions
Last, we can also assess whether a gender gap exists within sanction types in the length
of sentences handed down by the court. For these analyses, we again employed precision
matching using the same matching covariates above, but here we used four separate match-
ing analyses, one for each sanction subgroup. That is, we matched, among offenders who
received probation, female sex offenders to a male control pool. We followed the same pro-
cedure, separately, for intensive probation, jail, and prison subgroups. “Appendix” presents
descriptive information about the success of those matching procedures, including the fact
that, except for those receiving jail, we successfully matched the majority of female offend-
ers to males across each sanction type. For jail, we could only identify precise matches for
Table 4 Multinomial logistic regression analysis of a 4-category sanction measure on gender, conditional
on offense type and victim status, with weighting
Probation serves as the reference category. RSE robust standard errors
*p < .05; **p < .01; ***p < .001
Intensive probation Jail Prison
b RSE b RSE b RSE
Panel A: lewd and lascivious
Female 0.066 0.136 − 0.233 0.246 − 0.560*** 0.120
Intercept − 0.805*** 0.145 − 2.271*** 0.199 0.209 0.086
Log pseudolikelihood − 5629.677
Pseudo R-squared 0.002
N = 5829
Panel B: lewd and lascivious, with minor
Female 0.052 0.142 − 0.243 0.243 − 0.570*** 0.132
Intercept − 0.778*** 0.148 − 2.236*** 0.195 0.245** 0.084
Log pseudolikelihood − 5554.457
Pseudo R-squared 0.002
N = 5814
Panel C: sexual assault and sexual battery
Female 0.031 0.305 0.253 0.566 − 0.668** 0.222
Intercept − 0.637** 0.244 − 2.245*** 0.332 0.734*** 0.127
Log pseudolikelihood − 1761.587
Pseudo R-squared 0.004
N = 732
Panel D: sexual assault and sexual battery, with minor
Female − 0.048 0.301 0.377 0.370 − 0.874*** 0.189
Intercept − 0.505* 0.243 − 2.679*** 0.471 0.796*** 0.114
Log pseudolikelihood − 1478.21
Pseudo R-squared 0.006
N = 648

Journal of Quantitative Criminology
1 3
a handful of females, most likely due to the relatively small pool of potential male matched
cases, and so we did not conduct further analysis of the jail subgroup.
Table5 presents the results of negative binomial regression models predicting sentence
length across each sanction type. Panel A estimates the effect of gender on probation sen-
tence length using the probation matched group, panel B focuses on intensive probation,
and panel C focuses on prison. Parallel to our earlier findings, we identify a statistically
significant gender gap in sentence length decision-making. Male sex offenders receive
longer sentences than female sex offenders, and the largest difference emerges for the most
severe sanction, which is imprisonment. Figure2 illustrates this pattern by estimating the
predicted sentence length across the three sanction types. Here we see that males on aver-
age receive prison sentences that are more than 20 months longer than those assigned
to females (102.9 compared to 81.4), intensive probation sentences that are on average
1.5months longer (20 compared to 18.5), and probation sentence lengths that are on aver-
age 5months longer (62.7 compared to 57.).
Table 5 Negative binomial
regression analysis of sentence
length on gender, using 1:k
matched samples within sanction
type, with weighting
Each panel represents a unique matching analysis and subsequent neg-
ative binomial regression analysis. See “Appendix” for specific details
about each matching analysis. We exclude the jail matching analysis
because the matched sample was too small (N = 41) to allow for mul-
tivariate analyses
*p < .05; **p < .01; ***p < .001
b SE
Panel A: probation sentence length
Female − 0.084* 0.042
Circuit dummies (not shown) – –
Year dummies (not shown) – –
Intercept 4.444*** 0.103
Log pseudolikelihood − 8667.846
Pseudo R-squared 0.028
N = 1802
Panel B: intensive probation sentence length
Female − 0.078 0.047
Circuit dummies (not shown) – –
Year dummies (not shown) – –
Intercept 3.292*** 0.277
Log pseudolikelihood − 2047.107
Pseudo R-squared 0.022
N = 601
Panel C: prison sentence length
Female − 0.235** 0.070
Circuit dummies (not shown) – –
Year dummies (not shown) – –
Intercept 4.884*** 0.156
Log pseudolikelihood − 8156.879
Pseudo R-squared 0.011
N = 1468

Journal of Quantitative Criminology
1 3
Discussion andConclusion
The purpose of this study was to assess theoretical arguments about the role of gender in
perceptions of sex offenders and application of social control mechanisms in response to
sex offending. Three key findings emerge from this analysis. First, convicted female fel-
ony sex offenders, on average, look substantially different than convicted male felony sex
offenders. As a group, female sex offenders presented to the court as younger, more likely
to be white, and less serious offenders overall than male sex offenders. These differences
illustrate the salience of sexual offending being a gendered process (Cortoni and Gannon
2016; Williams and Bierie 2015), and underscore the need to account for a range of differ-
ences between males and females when estimating gender effects in sentencing, as well as
the potential benefits of using rigorous matching procedures like precision matching when
possible.
These findings also suggest that perceptions of female sex offending as less dangerous
may reflect, in part, documented differences in patterns of offending. These gendered per-
ceptions appear to hold influence over court decisions even when female defendants have
committed similarly serious sexual crimes. With that said, it is also important to note that
charges related to prostitution or commercialized vice were not part of this analysis. These
types of charges are more likely to be applied to female defendants, and police and court
processes centered on those cases have been criticized for being overly punitive towards
women (e.g., Pfeffer etal. 2018). Therefore, we were unable to evaluate gendered sentenc-
ing patterns for offenses that are perceived as “female-oriented.” Still, the evidence from
this study suggests that male sex offenders demonstrate to the court more traditionally rec-
ognized markers of dangers, and are punished accordingly.
Second, when perfectly matched across a host of demographic and legally relevant
variables, females convicted of sex offenses were significantly less likely than males to
go to prison. In other words, even when convicted male and female sex offenders’ crimi-
nal acts look the same, they are treated differently. This finding sheds light on the pos-
sibility that male sex offenders may represent to the court the “true” sexual predator—that
male sex offenders are more threatening to communities and are more inclined prey on
women and children or to otherwise pose substantial threat to community members if they
were released back into communities. Women, by contrast, receive supervision and com-
munity-based sanctions at substantially higher rates. This finding may indicate a greater
57.7
18.5
81.4
62.7
20.0
102.9
0
20
40
60
80
100
120
Probation Intensive ProbationPrison
Months
Females Males
Fig. 2 Predicted sentence length, by gender, across sanction groups

Journal of Quantitative Criminology
1 3
willingness of the courts to trust female sex offenders under the surveillance of the proba-
tion system. In this view, women are not viewed by the courts as sexual predators or as
significant threats if returned to their communities.
Third, the results indicated that the pattern of male sex offenders receiving harsher pun-
ishment outcomes than females holds across offense type and victim type. That is, regard-
less of the degree of seriousness or violence of the sex offense, and regardless of whether
the offense was committed against an adult or a minor, females were significantly less
likely to spend time in prison than to be assigned a community sanction. Even in cases
of rape involving minor victims—the proverbial “worst of the worst” within the offender
population (Ievins and Crewe 2015)—courts were significantly less likely to send women
to prison than men.
These findings support prior research and theoretical frameworks (e.g., focal concerns,
chivalry hypothesis) that argue that sentencing decisions are structured by legally irrelevant
characteristics, like gender (Steffensmeier etal. 1998). In this view, female offenders are
less likely to be viewed as threatening to court actors and, accordingly, are shown leniency
in sentencing. The findings from the current study support this contention—despite the fact
that a similar proportion of male and female sex offenders score high enough to receive a
prison sentence, females are instead awarded the benefits of community-based sanctions.
This analysis also extends these arguments to sex offender sentencing, more specifically.
Despite the fact that as a group, sexual abusers are the most frequently feared subclass of
offender (Embry and Lyons 2012), this perception also appears to be structured by gender.
Thus, prior theory, research, and the results of the present study combine to sup-
port a straightforward narrative about the role of gender in sex offender sentencing.
Females represent a small, but consistent proportion of sex offenders who come before
the courts. Evidence here indicates that this group may present a different threat image
than that displayed by men, across several theoretically important factors. As a group,
female sex offenders—white, young, and less criminally entrenched—are perceived
as less of a threat to society, less culpable for their crimes, and less of a burden to
judicial decision-making (Steffensmeier etal. 1998). In individual processing, we see
evidence here that courts consider such perceptions and that female sex offenders may
benefit, via a reduced likelihood of incarceration, from their group’s perceived lack of
threat. Even when female sex offenders are convicted of serious crimes that for male sex
offenders would result in incarceration, females are often returned to their community to
serve their sentence (Sandler and Freeman 2011) or serve lighter sentences (Embry and
Lyons 2012; Vandiver and Teske 2006). Alternative theories about the role of gender in
sentencing are not supported by this research. For example, prior sentencing research
has found that disparities by race and gender are less likely to occur in cases involv-
ing serious offenses (Spohn 2000; Unnever and Hembroff 1988). Prior scholarship has
argued and found that judges have less discretion in sentencing decision-making when
crimes are especially severe. These same mechanisms do not appear to be operating in
sex offender sentencing. Even in the most serious sex offense cases, females were less
likely to be imprisoned than males. In a similar manner, the results here do not support
the contention that judges would sentence females more severely because sex offenses
are viewed as gender atypical (Bernstein etal. 1977). We did not find support for this
argument. Going forward, studies that can compare the scale of gender gaps amongst

Journal of Quantitative Criminology
1 3
sex offenders to those amongst other offending groups can provide important insights
into the potential uniqueness of sex offenders, in turn, provide important advances to
knowledge about gender and sentencing.
Implications forResearch
These findings underscore the importance of gender in sentencing, even for particularly
serious offenses. They underscore, too, the possibility that get-tough punishment move-
ments created avenues to use prison more and for longer periods in vastly inconsistent
ways. This, in turn, provides opportunities for more dramatic disparities to exist across
groups. Here, for example, we identified wide variation in some instances, including
some in which males are going to prison when females remain in the community, all else
equal. In these instances, which punishment is the more appropriate one? This question
is, of course, rhetorical. The answer depends on what courts and society members view
as the appropriate goal of punishment. If it is retribution, any unfairness, arguably, is a
policy failure. If the goal is deterrence, then the existence of disparity is perhaps not as
inherently problematic. Judgment would have to be reserved pending evaluations of the
effects of these punishments on recidivism for different gender groups. Future theory
and research should work to answer these questions by considering more closely how
punishment variation has changed over time and the implications of that variation for
accomplishing core correctional goals.
These results raise at least two additional important questions that should be
addressed in future studies. First, what factors influence court actors’ perceptions of sex
offenders—specifically, perceptions of threat and culpability—and do such perceptions
differ for male and female offenders? What informs prosecutors’ and judges’ percep-
tions of sex offenders and how do their perceptions align with those of non-criminal
justice actors typically polled in prior public opinion studies?
No known scholarship to date has focused on identifying whether and how gendered
viewpoints emerge about sex offenders or the mechanisms by which female sex offend-
ers are perceived to represent substantially less of a threat to community members. Sim-
ilarly, limited theory and knowledge exists about the origins of citizens’ views about sex
offenders (for exception, see Pickett etal. 2013), or the extent to which gender struc-
tures these views. For example, how influential are news media accounts of singular,
traumatic sex offender cases? How many cases covered on television are typically of
male sex offenders versus female sex offenders, and how does this ratio influence public
views? Answering these questions can help to answer questions about the mechanisms
that lead to the gender disparities identified in this study.
Second, are gender disparities and sex offender sentencing decisions influenced
by other case factors not examined here? For example, future studies might consider
including additional relevant factors, such as the offense occurring in a co-offending
context or details about case victims, such as victim age, gender, and relationship to
the offender. These dimensions were unavailable for the analyses here but may con-
found the association between gender and sex offender sentencing outcomes, and thus
should be included as matching variables in future studies. Victim characteristics may

Journal of Quantitative Criminology
1 3
be especially salient considering a recently emerging line of research that has identified
disparities in criminal justice outcomes along victim-offender gender dynamics (e.g.,
same-sex vs. opposite-sex victim-offender dyads) (see, e.g., Chaffin etal. 2016).
Implications forPolicy
These results also raise important questions for sentencing policy decision-making,
especially pertaining to the types of punishments assigned to sex offenders and the effec-
tiveness of these sanctions for reducing future offending. Specifically, this study found
a general reliance on incarceration for sex offenders, but also that many sex offenders
(especially female sex offenders) serve their time in the community. As described above,
these findings do not support advocating for policies that “even the score” by becoming
more punitive for females, but rather call for future policy research to evaluate the over-
all appropriateness and effectiveness of current sex offender punishment approaches,
including whether there are antitherapeutic consequences (see Meloy etal. 2013) asso-
ciated with a reliance on incarceration, and whether gender conditions the effectiveness
of different sanction options. For example, empirical analyses of recidivism across dif-
ferent sanction outcomes for sex offenders may find important gender differences. Per-
haps male sex offenders are at a higher risk of recidivism, require greater incapacitation,
or in need of more intensive treatment that can only be provided at an institutional set-
ting (and, perhaps, court actors are somehow aware of such differences). Alternatively,
court actors may perceive that female sex offenders are more easily deterred than males
and so simply require less serious punishment.
Most broadly, this study provides an important empirical foundation for future stud-
ies that test the relative effectiveness of correctional sanctions for sex offenders and for
males compared to females. Ideally, future court decisions would be based on empirical
evidence regarding which sanction options achieve correctional goals more effectively
and less so on perceptions or cognitive biases about what a given offender requires. The
latter is likely to unnecessarily contribute to the types of disparities along extralegal
dimensions, like gender, identified here.
Acknowledgements We thank Sam Scaggs for helpful consultation with the paper’s analyses, William
Bales for consultation with the preparation of the paper’s dataset, and Daniel Mears for helpful comments
on earlier drafts.
Appendix
See Table6.

Journal of Quantitative Criminology
1 3
References
Ackerman AR, Harris AJ, Levenson JS, Zgoba K (2011) Who are the people in your neighborhood?
A descriptive analysis of individuals on public sex offender registries. Int J Law Psychiatry
34:149–159
Albonetti CA (1991) An integration of theories to explain judicial discretion. Soc Problems 38:247–266
Amirault J, Beauregard E (2014) The impact of aggravating and mitigating factors on the sentence sever-
ity of sex offenders: an exploration and comparison of differences between offending groups. Crim
Justice Policy Rev 25:78–104
Anderson I, Swainson V (2001) Perceived motivation for rape: gender differences in beliefs about female
and male rape. Curr Res Soc Psychol 6:107–122
Apel RJ, Sweeten G (2010) Propensity score matching in criminology and criminal justice. In: Piquero
AR, Weisburd D (eds) Handbook of quantitative criminology. Springer, New York, pp 543–562
Armstrong GS, Freeman BC (2011) Examining GPS monitoring alerts triggered by sex offenders: the
divergence of legislative goals and practical application in community corrections. J Crim Justice
39:175–182
Bales WD, Piquero AR (2012) Racial/ethnic differentials in sentencing to incarceration. Justice Q
29:742–773
Bernstein IN, Kick E, Leung JT, Schulz B (1977) Charge reduction: an intermediary stage in the process of
labelling criminal defendants. Soc Forces 56:362–384
Blackwell BS, Holleran D, Finn MA (2008) The impact of the Pennsylvania sentencing guidelines on sex
differences in sentencing. J Contemp Crim Justice 24:399–418
Budd K, Desmond SA (2014) Sex offenders and sex crime recidivism: investigating the role of sentence
length and time served. Int J Offender Ther Comp Criminol 58:1481–1499
Cain CM, Sample LL, Anderson AL (2017) Public opinion of the application of sex offender notification
laws to female sex offenders: why it is important to examine. Crim Justice Policy Rev 28:155–175
Center for Sex Offender Management (2007) Female sex offenders. http://www.csom.org/pubs/femal e_sex_
offen ders_brief .pdf. Accessed 29 Apr 2018
Chaffin M, Levenson J, Letourneau E, Stern P (2009) How safe are trick-or-treaters? An analysis of child
sex crime rates on Halloween. Sex Abuse 21:363–374
Chaffin M, Chenoweth S, Letourneau EJ (2016) Same-sex and race-based disparities in statutory rape
arrests. Interpers Violence 31:26–48
Christensen LS (2018) The new portrayal of female child sexual offenders in the print media: a qualitative
content analysis. Sex Cult 22:176–189
Table 6 Precision matching
results for negative binomial
regression analyses in Table5
Males Females
Probation
All cases 7341 237
Matched cases 1627 175
Percent matched 22 74
Intensive probation
All cases 3055 112
Matched cases 521 80
Percent matched 17 71
Jail
All cases 867 27
Matched cases 33 8
Percent matched 4 30
Prison
All cases 15,414 190
Matched cases 1315 153
Percent matched 9 81

Journal of Quantitative Criminology
1 3
Clements H, Dawson DL, Das Nair R (2014) Female-perpetrated sexual abuse: a review of victim and pro-
fessional perspectives. J Sex Aggress 20:197–215
Cohen M, Jeglic EL (2007) Sex offender legislation in the United States: what do we know? Int J Offender
Ther Comp Criminol 51:369–383
Cortoni F (2015) What is so special about female sexual offenders? Introduction to the special issue on
female sexual offenders. Sex Abuse 27:232–234
Cortoni F, Gannon TA (2016) The assessment of female sex offenders. In: Boer DP (ed) The Wiley hand-
book on the theories, assessment and treatment of sexual offending. Wiley-Blackwell, Chichester, pp
453–471
Cortoni F, Babchishin KM, Rat C (2017) The proportion of sexual offenders who are female is higher than
thought: a meta-analysis. Crim Justice Behav 44(2):145–162
Daly K (1987) Discrimination in the criminal courts: family, gender, and the problem of equal treatment.
Soc Forces 66:152–175
Denov MS (2001) A culture of denial: exploring professional perspectives on female sex offending. Can J.
Criminol 43:303–329
Denov MS (2003) The myth of innocence: sexual scripts and the recognition of child sexual abuse by female
perpetrators. J Sex Res 40:303–314
Denov MS, Cortoni F (2006) Women who sexually abuse children. In: Hilarski C, Wodarski JS (eds) Com-
prehensive mental health practice with sex offenders and their families. Taylor and Francis, Hoboken,
pp 71–99
Doerner JK, Demuth S (2014) Gender and sentencing in the federal courts: are women treated more leni-
ently? Crim Justice Policy Rev 25:242–269
Elliott IA, Eldridge HJ, Ashfield S, Beech AR (2010) Exploring risk: potential static, dynamic, protective
and treatment factors in the clinical histories of female sex offenders. J Female Violence 25:595–602
Embry R, Lyons PM Jr (2012) Sex-based sentencing: sentencing discrepancies between male and female
sex offenders. Female Criminol 7:146–162
Farkas MA, Stichman A (2002) Sex offender laws: can treatment, punishment, incapacitation, and public
safety be reconciled? Crim Justice Rev 27:256–283
Finkelhor D, Jones LM (2006) Why have child maltreatment and child victimization declined? J Soc Issues
62:685–716
Finkelhor D, Jones LM (2012) Have sexual abuse and physical abuse declined since the 1990s? Crimes
Against Children Research Center, Durham
Finkelhor D, Ormrod R, Chaffin M (2009) Juveniles who commit sex offenses against minors. Juvenile Jus-
tice Bulletin—NCJ227763. US Government Printing Office, Washington, DC
Florida Department of Corrections (2015) Florida criminal punishment code: scoresheet preparation man-
ual. http://www.dc.state .fl.us/pub/sen_cpcm/cpc_manua l.pdf. Accessed 29 Apr 2018
Freeman NJ, Sandler JC (2008) Female and male sex offenders: a comparison of recidivism patterns and
risk factors. J Interpers Violence 23(10):1394–1413
Gakhal BK, Brown S (2011) A comparison of the general public’s, forensic professionals’ and students’
attitudes towards female sex offenders. J Sex Aggress 17:105–116
Gannon TA, Rose MR, Ward T (2008) A descriptive model of the offense process for female sexual offend-
ers. Sex Abuse 20:352–374
Gannon TA, Waugh G, Taylor K, Blanchette K, O’Connor A, Blake E, O’Ciardha C (2014) Women who
sexually offend display three main offense styles: a reexamination of the descriptive model of female
sexual offending. Sex Abuse 26:207–224
Greenfeld L (1997) Sex offenses and offenders: an analysis of data on rape and sexual assault. U.S. Depart-
ment of Justice, Washington
Griswold DB (1989) Florida’s sentencing guidelines: six years later. Fed Probation 53:46
Guo S, Fraser MW (2010) Propensity score analysis: statistical methods and applications. Sage, Thousand
Oaks
Harris AJ (2009) The civil commitment of sexual predators: a policy review. In: Wright RG (ed) Sex
offender laws: failed policies, new directions. Springer, New York, pp 219–239
Hassett-Walker C, Lateano T, Di Benedetto M (2014) Do female sex offenders receive preferential treatment
in criminal charging and sentencing? Just Sys J 35:62–86
Hayes S, Carpenter B (2013) Social moralities and discursive constructions of female sex offenders. Sexu-
alities 16:159–179
Ievins A, Crewe B (2015) ‘Nobody’s better than you, nobody’s worse than you’: moral community among
prisoners convicted of sexual offences. Punishm Soc 17:482–501
Institute of Medicine (2013) New directions in child abuse and neglect research. National Academies,
Washington

Journal of Quantitative Criminology
1 3
Jenkins P (1998) Moral panic: changing concepts of the child molester in modern America. Yale University
Press, New Haven
Katz-Schiavone S, Levenson JS, Ackerman AR (2008) Myths and facts about sexual violence: public per-
ceptions and implications for prevention. J Crim Justice Pop Cult 15:291–311
Kaufman K (2010) The prevention of sexual violence and exploitation: a sourcebook. Wood & Barnes,
Oklahoma City
King LL, Roberts JJ (2017) The complexity of public attitudes toward sex crimes. Victims Offenders
12(1):71–89
Leon CS (2011) Sex fiends, perverts, and pedophiles: understanding sex crime policy in America. NY
Press, New York
Levenson JS (2009) Sex offender residence restrictions. In: Wright RG (ed) Sex offender laws: failed
policies, new directions. Springer, New York, pp 267–290
Levenson JS, Brannon YN, Fortney T, Baker J (2007) Public perceptions about sex offenders and com-
munity protection policies. Anal Soc Issues Public Policy 7:137–161
Levenson JS, Shields RT, Singleton DA (2014) Collateral punishments and sentencing policy: per-
ceptions of residence restrictions for sex offenders and drunk drivers. Crim Justice Policy Rev
25:135–158
Malinen S, Willis GM, Johnston L (2014) Might informative media reporting of sexual offending influ-
ence community members’ attitudes towards sex offenders? Psychol Crime Law 20:535–552
Mancini C (2014) Examining factors that predict public concern about the collateral consequences of
sex crime policy. Crim Justice Policy Rev 4:450–475
Mancini C, Pickett JT (2016) The good, the bad, and the incomprehensible: typifications of victims and
offenders as antecedents of beliefs about sex crime. J Interpers Violence 31:257–281
Mancini C, Shields RT, Mears DP, Beaver KM (2010) Sex offender residence restriction laws: parental
perceptions and public policy. J Crim Justice 38:1022–1030
Mancini C, Barnes JC, Mears DP (2013) It varies from state to state: an examination of sex crime laws
nationally. Crim Justice Policy Rev 24:166–198
Matthews JK, Mathews RU, Speltz K (1991) Female sexual offenders: a typology. In: Patton MQ (ed)
Family sexual abuse: frontline research and evaluation. Sage, Newbury Park, pp 199–219
McLeod DA (2015) Female offenders in child sexual abuse cases: a national picture. J Child Sex Abuse
24:97–114
Mears DP, Mancini C, Gertz M, Bratton J (2008) Sex crimes, children, and pornography: public views
and public policy. Crime Delinq 54:532–559
Mellor D, Deering R (2010) Professional response and attitudes toward female-perpetrated child sex-
ual abuse: a study of psychologists, psychiatrists, probationary psychologists and child protection
workers. Psychol Crime Law 16:415–438
Meloy M, Curtis K, Boatwright J (2013) The sponsors of sex offender bills speak up: policy makers’
perceptions of sex offenders, sex crimes, and sex offender legislation. Crim Just Behav 40:438–452
Moore LD, Padavic I (2010) Racial and ethnic disparities in girls’ sentencing in the juvenile justice sys-
tem. Female Criminol 5:263–285
Nagin DS, Cullen FT, Jonson CL (2009) Imprisonment and reoffending. Crime Justice 38:115–200
National Center for Missing and Exploited Children (2015) Map of registered sex offenders in the United
States. http://www.missi ngkid s.com/en_US/docum ents/Sex_Offen ders_Map.pdf
Petrunik MG (2002) Managing unacceptable risk: sex offenders, community response, and social policy
in the United States and Canada. Int J Offender Ther Comp Criminol 46:483–511
Pfeffer R, Ormachea P, Eagleman D (2018) Gendered outcomes in prostitution arrests in Houston, Texas.
Crime Delinq 64(12):1538–1567
Pickett JT, Mancini C, Mears DP (2013) Vulnerable victims, monstrous offenders, and unmanageable
risk: explaining public opinion on the social control of sex crime. Criminology 51:729–759
Planty M, Langton L, Krebs C, Berzofsky M, Smiley-McDonald H (2013) Female victims of sexual vio-
lence, 1994–2010. Bureau of Justice Statistics, Washington
Quinn JF, Forsyth CJ, Mullen-Quinn C (2004) Societal reaction to sex offenders: a review of the origins and
results of the myths surrounding their crimes and treatment amenability. Deviant Behav 25:215–232
Rogers P, Davies M (2007) Perceptions of victims and perpetrators in a depicted child sexual abuse case:
gender and age factors. J Interpers Violence 22:566–584
Sample LL (2011) The need to debate the fate of sex offender community notification laws. Criminol
Public Policy 10:265–274
Sample LL, Bray TM (2003) Are sex offenders dangerous. Criminol Public Policy 3:59–82
Sandler J, Freeman NJ (2011) Female sex offenders and the criminal justice system: a comparison of
arrests and outcomes. J Sex Aggress 17:61–76

Journal of Quantitative Criminology
1 3
Savolainen J, Hughes LA, Hurtig TM, Ebeling H, Taanila AM (2013) Does vocational schooling facili-
tate criminal offending? A study of educational tracking in Finland. Eur J Criminol 10:606–622
Scott C, del Busto E (2009) Chemical and surgical castration. In: Wright RG (ed) Sex offender laws:
failed policies, new directions. Springer, New York, pp 291–338
Simon J (2000) Megan’s law: crime and democracy in late modern America. Law Soc Inq 25:1111–1150
Smith SG, Basile KC, Gilbert LK, Merrick MT, Patel N, Walling M, Jain A (2017) National intimate
partner and sexual violence survey (NISVS): 2010–2012 state report. National Center for Injury
Prevention and Control, Centers for Disease Control and Prevention, Atlanta
Socia KM, Harris AJ (2016) Evaluating public perceptions of the risk presented by registered sex
offenders: evidence of crime control theater? Psychol Public Policy Law 22:375–385
Spohn C (2000) Thirty years of sentencing reform: the quest for a racially neutral sentencing process.
Crim Justice 3:427–501
Spohn C (2013) The effects of the offender’s race, ethnicity, and sex on federal sentencing outcomes in the
guidelines era. Law Contemp Problems 76:75–104
Spohn C, Beichner D (2000) Is preferential treatment of female offenders a thing of the past? A multisite
study of gender, race, and imprisonment. Crim Just Policy Rev 11:149–184
Steffensmeier D, Demuth S (2006) Does gender modify the effects of race–ethnicity on criminal sanc-
tioning? Sentences for male and female white, black, and Hispanic defendants. J Quant Criminol
22:241–261
Steffensmeier D, Ulmer J, Kramer J (1998) The interaction of race, gender, and age in criminal sentencing:
the punishment cost of being young, black, and male. Criminology 36:763–798
Strickland SM (2008) Female sex offenders: exploring issues of personality, trauma, and cognitive distor-
tions. J Interpers Violence 23(4):474–489
ten Bensel T, Gibbs B, Burkey CR (2016) Female sex offenders: is there a difference between solo and co-
offenders? J Interpers Violence. https ://doi.org/10.1177/08862 60516 67420 2
Terry KJ, Ackerman AR (2009) A brief history of major sex offender laws. In: Wright RG (ed) Sex offender
laws: failed policies, new directions. Springer, New York, pp 65–98
Unnever JD, Hembroff LA (1988) The prediction of racial/ethnic sentencing disparities: an expectation
states approach. J Res Crime Delinq 25:53–82
Vandiver DM, Kercher G (2004) Offender and victim characteristics of registered female sexual offenders in
Texas: a proposed typology of female sexual offenders. Sex Abuse 16:121–137
Vandiver DM, Teske R (2006) Juvenile female and male sex offenders: a comparison of offender, victim,
and judicial processing characteristics. Int J Offender Ther Comp Criminol 50:148–165
Vandiver DM, Walker JT (2002) Female sex offenders: an overview and analysis of 40 cases. Crim Justice
Rev 27:284–300
West SG, Friedman SH, Kim KD (2011) Women accused of sex offenses: a gender-based comparison.
Behav Sci Law 29(5):728–740
Williams KS, Bierie DM (2015) An incident-based comparison of female and male sexual offenders. Sex
Abuse 27:235–257
Williams R, Gillespie SM, Elliott IA, Eldridge HJ (2019) Characteristics of female solo and female co-
offenders and male solo sexual offenders against children. Sex Abuse 31:151–172
Wijkman M, Bijleveld C, Hendriks J (2010) Women don’t do such things! Characteristics of female sex
offenders and offender types. Sex Abuse 22(2):135–156
Zevitz RG (2006) Sex offender community notification: its role in recidivism and offender reintegration.
Crim Justice Stud 19:193–208
Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and
institutional affiliations.