655837

research-article2016
CJBXXX10.1177/0093854816655837CRIMINAL JUSTICE AND BEHAVIORRade et al. / PUBLIC ATTITUDES TOWARD EX-OFFENDERS

A Meta-Analysis Of Public Attitudes

Toward Ex-Offenders

CANDALYN B. RADE
SARAH L. DESMARAIS
ROGER E. MITCHELL
North Carolina State University

Ex-offenders face barriers to community reintegration including negative attitudes held by members of the public. This meta-
analysis summarizes the extant research on the correlates of public attitudes toward ex-offenders—namely, public, ex-
offender, and community characteristics—and the moderating effects of sexual offense history. A systematic search of four
databases (PsycINFO, Web of Science, National Criminal Justice Reference Service [NCJRS], and ProQuest Dissertation &
Theses) identified 19 records, consisting of 9,355 participants. Results revealed small associations between correlate vari-
ables and attitudes, suggesting that people are more similar than different in their attitudes toward ex-offenders. Indeed, only
political ideology, interpersonal contact, and sexual offense history emerged as significant correlates. Moderation analyses
revealed differences in public attitudes toward ex-offenders based upon the year a record was produced. Findings reveal the
need for additional research examining moderators of public attitudes toward ex-offenders and suggest that interventions
should explore ways to incorporate interpersonal contact and reduce stigma related to criminal histories.

Keywords: public attitudes; public opinion; stigma; ex-offender; meta-analysis

A vast number of offenders reenter the community following incarceration each year. In
2011, for example, close to two thirds of a million people were released from U.S. state
and federal prisons, and one of every 50 American adult residents was under the supervision
of probation or parole (Carson & Sabol, 2012; Maruschak & Parks, 2012). Yet, men and
women who reenter the community after incarceration face many barriers to successful
reintegration. These may be attributable, at least in part, to negative attitudes held by mem-
bers of the public regarding ex-offenders (e.g., Brooks, Visher, & Naser, 2006; Clear, Rose,
& Ryder, 2001; Wakefield & Uggen, 2010). Indeed, many studies have identified that

Authors’ Note: Funding for this study was provided, in part, by the North Carolina State University
Psychology Emeritus Fund. North Carolina State University had no further role in study design; in the collec-
tion, analysis, and interpretation of data; in the writing of the report; or in the decision to submit the paper for
publication. We thank Dr. Mark Wilson for his support and meta-analytic expertise and Jordan Gregory, Mirela
Scott, and Brittanie Moore for their research assistance. In addition, we thank Drs. Gaylene Armstrong,
Cassandra Atkin-Plunk, Sarah Brown, Keri Burchfield, Erin Comartin, Paul Hirschfield, Jill Levenson,
Christina Mancini, Devah Pager, David Perkins, Alex Piquero, Joshua Raines, and Gwenda Willis for their
generous willingness to provide additional data for inclusion in this study. Correspondence concerning this
article should be addressed to Candalyn B. Rade, Department of Psychology, North Carolina State University,
Campus Box 7650, Raleigh, NC 27695-7650; e-mail: cbrade@ncsu.edu.
CRIMINAL JUSTICE AND BEHAVIOR, 2016, Vol. 43, No. 9, September 2016, 1260­–1280.
DOI: 10.1177/0093854816655837
© 2016 International Association for Correctional and Forensic Psychology

1260
Downloaded from cjb.sagepub.com by guest on October 12, 2016
 Rade et al. / PUBLIC ATTITUDES TOWARD EX-OFFENDERS 1261

members of the public commonly hold negative attitudes toward and desire social distance
from ex-offenders (e.g., Hirschfield & Piquero, 2010; Leverentz, 2011; Manza, Brooks, &
Uggen, 2004; S. Park, 2009), often resulting in social rejection, discrimination, and loss of
social status (Phelan, Link, & Dovidio, 2008). These negative attitudes can contribute to the
development of policy restrictions and barriers for ex-offenders in domains such as educa-
tion, employment, health, housing, and voting rights (Clear et al., 2001; Pager, 2003;
Pogorzelski, Wolff, Pan, & Blitz, 2005; Schnittker & Bacak, 2013; Varghese, Hardin, Bauer,
& Morgan, 2010; Wakefield & Uggen, 2010). Furthermore, subgroups of ex-offenders may
experience variable rates of barriers due to negative public attitudes. This is evident in the
case of sex offenders, who are among one of the most discriminated against groups of ex-
offenders (Tewksbury & Lees, 2006; Viki, Fullerton, Raggett, Tait, & Wiltshire, 2012).

Prejudice, Stigma, And Social Distance

Prejudice, stigma, and social distance theories provide a foundation for understanding
negative attitudes toward ex-offenders (see Allport, 1954; Dovidio, Glick, & Rudman,
2005; Goffman, 1963). Prejudice is described as “antipathy” toward groups or individuals
because of incorrect beliefs or generalizations (Allport, 1954, p. 10). Similarly, stigma is the
process through which individuals or groups are rejected by others based on differences in
physical characteristics, personal character, identity, or flaws (Goffman, 1963). Stigma and
prejudice theories can be considered as complementary, emphasizing normal sociocultural
processes, and generally can be applied to discriminated, dehumanized, or devalued groups,
including ex-offenders (Phelan et al., 2008). As a result, ex-offenders often experience dif-
ferential treatment because of their status as an “ex-offender.” Discrimination against ex-
offenders also may occur in more subtle ways as described by social distance theories, such
that ingroup members (i.e., non-offenders) desire social distance from outgroup members
(i.e., ex-offenders; R. E. Park & Burgess, 1921). Sex offenders are particularly subject to the
effects of social distance, as ex-offenders with a history of sexual offenses are more com-
monly dehumanized and socially excluded (e.g., Viki et al., 2012).
Stigma and prejudice toward, and desired social distance from ex-offenders can be mani-
fested as barriers to accessing services in the community (e.g., housing, education), exclu-
sion from social settings and rights (e.g., voting), and inequality when seeking employment
(Pager, 2003; Pogorzelski et al., 2005; Schnittker & Bacak, 2013; Wakefield & Uggen,
2010). To illustrate, negative employer attitudes and stigma toward ex-offenders serve as a
barrier to obtaining employment (Clear et al., 2001; Pager, 2003; Varghese et al., 2010).
Employment can decrease the likelihood of recidivism and increase the likelihood of suc-
cessful community reentry (Graffam, Shinkfield, Lavelle, & McPherson, 2004; Uggen,
2000; Visher, Debus, & Yahner, 2008); as such, reducing stigma and discrimination toward
ex-offenders seeking employment may be one strategy for decreasing reentry barriers and
improving community reintegration.

Correlates Of Attitudes Toward Ex-Offenders

Identification of correlates associated with negative attitudes toward ex-offenders may
assist efforts to reduce stigma and facilitate successful reentry. Based on our systematic
review of the literature, the most frequently assessed correlates can be categorized as repre-
senting public characteristics, ex-offender characteristics, and characteristics of the local

Downloaded from cjb.sagepub.com by guest on October 12, 2016

1262 Criminal Justice and Behavior

community. We briefly summarize trends in the research on correlates within each of these
categories below.
Prior research has found four public characteristics associated with attitudes toward ex-
offenders: sex, race/ethnicity, political affiliation or ideology, and interpersonal contact.
Many studies show that women compared with men demonstrate less favorable attitudes
toward ex-offenders (Leverentz, 2011; Mancini, Shields, Mears, & Beaver, 2010; Willis,
Malinen, & Johnston, 2013). However, some research does not find differences in attitudes
between women and men (Hirschfield & Piquero, 2010; Locke, 2010; S. Park, 2009).
Similarly, White or non-minority participants typically—but not always—report less favor-
able attitudes toward ex-offenders compared with minority participants (Hirschfield &
Piquero, 2010; Leverentz, 2011; Mancini et al., 2010; but see Comartin, Kernsmith, &
Kernsmith, 2009; Dawson Edwards, 2007). Identification as politically conservative
(Hirschfield & Piquero, 2010; Locke, 2010; Mancini et al., 2010) or a Republican (S. Park,
2009) is frequently associated with less favorable attitudes toward ex-offenders compared
with identification as politically liberal or a Democrat. However, again, the research find-
ings are mixed (e.g., Dawson Edwards, 2007; Leverentz, 2011). Finally, interpersonal con-
tact theory (Allport, 1954; Pettigrew & Tropp, 2006) posits a negative relationship between
interpersonal contact and prejudice, such that increased contact with a member of an out-
group is associated with decreased prejudice, stigma, and/or desired social distance from
outgroup members. Consistent with interpersonal contact theory, members of the public
who report having contact with ex-offenders (Gibson, Roberson, & Daniel, 2009; Hirschfield
& Piquero, 2010; but see Dreiling, 2011), even sex offenders (Viki et al., 2012; Willis,
Levenson, & Ward, 2010), generally report more favorable attitudes. In contrast, age,
income, and education (e.g., Hirschfield & Piquero, 2010; Mancini et al., 2010; S. Park,
2009) typically have not been significantly associated with attitudes toward ex-offenders,
with some exceptions (e.g., Brown, 1999; Comartin et al., 2009; Willis et al., 2013).
Four characteristics of ex-offenders commonly have been found to be associated with public
attitudes in the extant research: (a) criminal history or type of offense, (b) race/ethnicity, (c)
participation in a rehabilitation program, and (d) the presence of mental illness. First, members
of the public generally hold more negative attitudes toward ex-offenders with a history of vio-
lent crimes, felonies, and sexual offenses compared with those with a history of non-violent
crimes, misdemeanors, and no sexual offenses (Hardcastle, Bartholomew, & Graffam, 2011;
Perkins, Raines, Tschopp, & Warner, 2009; Rogers, Hirst, & Davies, 2011; but see Martinez,
2011). By way of illustration, members of the public generally report significantly less support
for voting right reinstatement for violent ex-felons compared with ex-felons (Manza et al.,
2004). Second, members of the public appear to hold more negative attitudes toward ex-
offenders belonging to a minority race or ethnicity compared with ex-offenders of the majority
race or ethnicity (Pager, 2003). Third, members of the public tend to report more negative
attitudes toward ex-offenders who have not participated in a rehabilitation program compared
with ex-offenders who have participated (Hardcastle et al., 2011; Rogers et al., 2011). Fourth,
members of the public, on average, demonstrate more negative attitudes toward ex-offenders
with mental illnesses compared with ex-offenders without mental illnesses (LeBel, 2008;
Locke, 2010). Taken together, research suggests that most ex-offenders face multiple stigmas
that can increase barriers encountered during reentry (LeBel, 2012).
Finally, prior research has identified community size and crime prevalence as possible
correlates of attitudes toward ex-offenders, although community characteristics are studied

Downloaded from cjb.sagepub.com by guest on October 12, 2016

 Rade et al. / PUBLIC ATTITUDES TOWARD EX-OFFENDERS 1263

less frequently than public and ex-offender characteristics. A handful of studies suggest that
individuals from non-rural communities report less negative attitudes toward ex-offenders
compared with those living in rural communities (Hirschfield & Piquero, 2010; Leverentz,
2011; Mackey & Courtright, 2000). Findings of one study showed that citizens of non-rural
neighborhoods with high to moderate crime prevalence were more likely to report less puni-
tive attitudes generally—not attitudes toward ex-offenders, specifically—compared with
those living in rural neighborhoods with moderate to low crime prevalence (Leverentz,
2011). However, other studies have failed to find differences in attitudes as a function of
community size and crime salience (e.g., Hirschfield & Piquero, 2010; Locke, 2010).
Beyond the correlates reviewed above, there is evidence that sexual offense history has a
moderating effect on the relationship between public characteristics and attitudes toward
ex-offenders. Prior research suggests that members of the public hold more negative atti-
tudes toward ex-offenders who have been convicted of a sexual offense, such as sexual
assault (Hulsey, 1991), sexual assault against children (Hardcastle et al., 2011), or any sex-
ual offense (Manza et al., 2004; Willis et al., 2010) compared with ex-offenders convicted
of other offenses. Because members of the public appear to hold more negative attitudes
toward sex offenders compared with ex-offenders with no history of sexual offending, there
may be a stronger consensus of negative attitudes toward sex offenders, regardless of public
characteristics. Thus, effect sizes for the associations between public characteristics and
attitudes toward sex offenders may be smaller than those observed between public charac-
teristics and attitudes toward ex-offenders with no history of sexual offending.

The Present Study

For more than 20 years, researchers have sought to identify factors that may act as barri-
ers to successful community reintegration following release from correctional facilities,
including public, ex-offender, and community characteristics associated with negative pub-
lic attitudes toward ex-offenders. However, variability in findings across studies limits our
conclusions regarding the significance, strength, and direction of the relationships between
public characteristics, ex-offender characteristics, and characteristics of the local commu-
nity and attitudes toward ex-offenders. Inconsistent operationalization and measurement of
attitudes may have contributed to the contradictory findings, and small sample sizes may
have limited power to detect significant effects. As a result, there is a need for a systematic
evaluation of the current empirical literature.
To that end, the present study used a meta-analytic approach to synthesize extant litera-
ture on correlates of public attitudes toward ex-offenders and to examine the potential mod-
erating effect of sexual offense history for public characteristic correlates, with the goal of
resolving inconsistencies and highlighting avenues for future research. We additionally
examined the moderating effects of publication status, sampling methodology, year pro-
duced, country, and attitude operationalization to assess potential biases (see Liberati et al.,
2009; Moher, Liberati, Tetzlaff, Altman, & The PRISMA Group, 2009).

Method
To the extent possible, we followed the Preferred Reporting Items for Systematic reviews
and Meta-Analyses (PRISMA) statement for reporting the findings of meta-analyses
(Liberati et al., 2009; Moher et al., 2009). The PRISMA statement is comprised of a 27-item

Downloaded from cjb.sagepub.com by guest on October 12, 2016

1264 Criminal Justice and Behavior

checklist, created to encourage clarity, transparency, and rigor in meta-analysis reporting

(Liberati et al., 2009).

Inclusion Criteria

The current study included all published and unpublished empirical research that met the
following inclusion criteria: (a) sample was drawn from the general public (including stu-
dents) and did not include criminal justice professionals; (b) at least one outcome variable
assessing attitudes toward ex-offenders who have reentered the community, defined as mea-
sures of general attitudes, stigma, punitiveness, or desired social distance from ex-offenders;
(c) reported necessary data to calculate effect size (e.g., means, standard deviations, correla-
tion values, sample size) for at least one correlate variable, or ability to obtain necessary data
from the authors; (d) reported in peer-review journals, dissertations, theses, conference pre-
sentations, government reports, or unpublished/in press manuscripts; (e) written in English
or reliable translation available; and (f) produced between January 1, 1990, and December
31, 2013, to reflect the contemporary empirical literature and current state of the science. For
the purpose of this study and record inclusion, the term “ex-offenders” refers to individuals
with a criminal offense history who are no longer incarcerated and who have reentered the
community following incarceration.

Literature Search

Records for the meta-analysis were identified through systematic searches in PsycINFO,
Web of Science, National Criminal Justice Reference Service (NCJRS) Abstracts, and
ProQuest Dissertation & Theses electronic databases. Keyword searches were conducted
using a two-part search term system; all possible combinations were employed, consisting of
one criminal justice term (incarcerat*, offend*, inmate, felon*, misdemean*, crim*, prison*,
convict*, recidiv*) paired with one attitudinal term (attitud*, stigma*, punitiv*, “social dis-
tance,” contact, familiar*, “public opinion”). For each included record, cited references were
screened against inclusion criteria to ensure all relevant records were identified. Emails
inquiring about additional relevant records or data were sent to prominent authors (C. A.
Atkin-Plunk, personal communication, January 21, 2016; S. Brown, personal communica-
tion, September 2, 2014; K. B. Burchfield, personal communication, August 7, 2014; E. B.
Comartin, personal communication, July 8, 2014; J. S. Levenson, personal communication,
July 28, 2014; C. Mancini, personal communication, July 5, 2014; D. Pager, personal com-
munication, August 31, 2014; D. V. Perkins & J. A. Raines, personal communication, July 2,
2014; A. R. Piquero, personal communication, July 3, 2014; A. R. Piquero, personal com-
munication, October 8, 2014; G. M. Willis, personal communication, July 9, 2014).
Initial literature searches revealed 56,609 records, with 21,536 from PsycINFO, 18,146
from Web of Science, 9,028 from NCJRS, and 7,899 from ProQuest, many of which were
duplicates. An iterative process of reviewing reference lists, searching other databases, and
contacting prominent authors identified an additional 76 records. Each record title was eval-
uated against inclusion criteria (94%-96% inter-rater agreement; κ = .67-.78 on two samples
of 100 records). Duplicates and records that initially did not meet criteria were removed;
1,459 relevant records remained for further evaluation against inclusion criteria. Records
that did not meet inclusion criteria were excluded; Criterion 2 (at least one outcome variable
assessing attitudes toward ex-offenders who have reentered the community) and Criterion

Downloaded from cjb.sagepub.com by guest on October 12, 2016

 Rade et al. / PUBLIC ATTITUDES TOWARD EX-OFFENDERS 1265

Figure 1: Results of Systematic Literature Search

Note. NCJRS = National Criminal Justice Reference Service.

3 (reported necessary data to calculate effect size) were the most common reasons for record
exclusion. A final total of 19 records remained that met all inclusion criteria (Figure 1).
Literature search protocol is available from the authors upon request.

Variable Coding

General Procedures

For all records meeting inclusion criteria, relevant information regarding the record,
sample, and effect size were extracted and coded according to a comprehensive coding
manual (available upon request). When exact group sample sizes were not reported or avail-
able from record authors, equal group sample sizes were assumed. For example, if the
authors reported the total sample size, but did not report the number of male and female
participants, equal group sample sizes of men and women were assumed.

Downloaded from cjb.sagepub.com by guest on October 12, 2016

1266 Criminal Justice and Behavior

Outcome Variables

Attitudes toward ex-offenders were operationalized to include measures of general atti-

tudes (e.g., attitudes toward ex-offenders), stigma (including measures of prejudice), puni-
tiveness (e.g., support for punitive policies toward ex-offenders and ex-offender rights), and
social distance (e.g., willingness to hire or work with an ex-offender) toward ex-offenders
in the community, with higher values representing more negative attitudes (continuous
when available; positive attitudes = 0, negative attitudes = 1). Responses of “don’t know”
or “unsure” were excluded from all analyses, when reported.

Correlate Variables

Public, ex-offender, and community characteristics were coded for all records when
available. Public correlates included sex (female = 0, male = 1), race/ethnicity (majority = 0,
minority = 1), education (continuous when available; less education [<bachelor’s
degree] = 0; more education [≥bachelor’s degree] = 1), religious beliefs (Christian = 0,
other religion = 1), political affiliation or ideology (conservative/Republican = 0, other
political affiliation/ideology = 1), age (continuous when available; younger age [<35] = 0,
older age, [≥35] = 1), household income (continuous when available; <US$50,000 = 0,
≥US$50,000 = 1), and interpersonal contact (any type of contact, continuous when avail-
able; contact absent = 0, contact present = 1).
Ex-offender correlates included violent offense history (violent = 0, non-violent = 1; as
defined by the original record or consistent with the U.S. Department of Justice (2014) defi-
nition of violent crimes to include murder, non-negligent manslaughter, forcible rape, rob-
bery, aggravated assault), sexual offense history (sexual offense history = 0, no history of
sexual offense = 1; as defined by the original record), felony offense history (felony = 0,
misdemeanor = 1; as defined by the original record), race/ethnicity (majority = 0, minority
= 1), rehabilitation participation (no = 0, yes = 1), and presence of mental disorder (mental
disorder present = 0, mental disorder not present = 1; as defined by the original record).
Community correlates included community size (rural = 0, non-rural = 1; as defined by
the original record) and crime prevalence (low crime = 0, high crime = 1; as defined by the
original record).

Moderator Variables

Sexual offense history was assessed for moderation effects for public characteristics
when ample data were present and was coded for each record when explicitly stated and
available (sexual offense history = 0, no history of sexual offense = 1; as defined by the
original record). Five additional characteristics were tested for moderation effects to assess
any potential biases when sufficient data were available. Publication status (published in
peer-review outlet = 0, other outlet, including dissertations = 1) was assessed to evaluate the
potential bias of published records focusing on significant results. Sampling methodology
(random or representative sample = 0, other sample, including convenience samples = 1)
was assessed for potential biases, as more confidence may be placed in more well-constructed
methodologies. Year produced (dummy coded based on median year; articles produced
before 2009 = 0, articles produced during or after 2009 = 1) was tested for moderating
effects to determine whether the potential changes in attitudes over time had a biasing
effect. Country (conducted in the United States = 0, conducted in a country other than the
Downloaded from cjb.sagepub.com by guest on October 12, 2016
 Rade et al. / PUBLIC ATTITUDES TOWARD EX-OFFENDERS 1267

United States = 1) was assessed for potential biases, as attitudes toward ex-offenders may
vary internationally. Finally, attitude operationalization (social distance = 0, general atti-
tudes = 1, stigma = 2, punitiveness = 3) was tested for moderating effects to determine
potential differences in public attitudes based on multiple measurements of public
attitudes.

Inter-Rater Reliability

All records were coded by the first author. A random sample of 21% of records was
selected for coding by an additional rater to establish inter-rater reliability (n = 4 records,
k = 17 unique effect sizes). The additional rater participated in a comprehensive training on
all coding procedures, including two practice records. Reliability was calculated for all cod-
ing decisions (i.e., each variable coded was characterized as a coding decision) using
Cronbach’s alpha or Cohen’s kappa, and percentage agreement.
Excellent levels of inter-rater reliability were produced for effect size coding (intraclass
correlation coefficient (ICC)3,2 = 1.00, 100% agreement). Inter-rater reliability ranged from
67% to 100% agreement (κ = .50-1.00) for all dichotomously coded variables (e.g., publica-
tion status, population type). For continuous variables (e.g., sample size), inter-rater reli-
ability ranged 82% to 100% agreement (ICC3,2 = .62-1.00). All disagreements were resolved
by consensus. A final review of all effect sizes was conducted by the first author to verify
the accuracy of coding.

Data Analysis

All analyses were conducted using the Hunter and Schmidt random-effects model of
meta-analysis (Schmidt & Hunter, 2015) considered an accurate approach for estimating
random-effects and mean reliability estimates of categorical variables (Mason, Allam, &
Brannick, 2007). The random-effects model assumes variability among population param-
eters and utilizes a random variable of effect to calculate weights and estimate mean effect
sizes (Hunter & Schmidt, 2000; Schulze, 2004), resulting in more conservative estimates.
Correlation coefficient r was recorded as the effect size measure for all available rela-
tionships between public, ex-offender, and community characteristics with attitudes toward
ex-offenders. Continuous and categorical data were used to calculate effect sizes, based on
the format in which data were available.1 Most records (78.9%, n = 15) produced more than
one unique, independent effect size (e.g., effect size for income and education). When a
product-moment correlation coefficient was not reported or available from the record
authors, r was calculated according to the appropriate conversion formulas (Lipsey &
Wilson, 2001; Wilson, 2001a, 2001b). Means and standard deviations (k = 34), F ratios
(k = 3), or t values (k = 1) were used to calculate r for continuous data, and frequencies
(k = 17) or chi-squares (k = 1) were used to calculate r for categorical data. Several records
(n = 4) reported multiple effect sizes and used multiple outcome measures for a single cor-
relate within one sample. If a record reported the relationships between education and two
outcome measures (social distance and attitudes), for the same sample, the two effect sizes
were averaged together. When this occurred, effect sizes (k = 26) were averaged, producing
a mean effect size estimate for the sample (k = 13), resulting in 73 independent effect sizes
altogether. To reduce risk of bias further, records that reported a non-significant effect with-
out providing specific data to approximate an effect size were coded as r = .00 (k = 6).

Downloaded from cjb.sagepub.com by guest on October 12, 2016

1268 Criminal Justice and Behavior

To reduce sampling bias in aggregate estimates, individual effect sizes for each correlate
were aggregated and weighted by sample size to produce a mean sample-weighted effect
size ( r w) and to calculate the sample-weighted observed variance, var(r), using the formu-
las below (Arthur, Bennett, & Huffcutt, 2001; Schmidt & Hunter, 2015).

rw =
Σ ( Ni × ri )
var ( r ) =
(
Σ Ni × ( ri − r )
2
).
ΣN i ΣN i
Corrections were applied to individual effect sizes of three artificially dichotomized vari-
ables (public education, k = 6; public age, k = 2; public income, k = 5) to correct for associ-
ated biases (Schmidt & Hunter, 2015). Corrected and non-corrected meta-analysis results
are presented (see Tables 3 and 2, respectively). However, corrected results are discussed in
text. An effect size cutoff rule was implemented, such that any weighted effect sizes less
than r w = .10 was considered negligible to interpretation.
Homogeneity of variance was assessed using the Q statistic to detect the extent to which
observed variance is due to sampling error or artifact biases. When significant, the Q statistic in
combination with a percent of variance accounted for by the sampling error and artifacts
(PVAse) less than 75% suggest the presence of moderation (Arthur et al., 2001; Schmidt &
Hunter, 2015). Moderation analyses were conducted when sufficient data were available using
a subset approach (Schmidt & Hunter, 2015), by performing separate analyses within each
subset (e.g., sexual offense history and no history of sexual offense). Differences in mean effect
sizes, reduced variances within subsets, and credibility intervals that do not overlap demon-
strated the presence of a moderation effect (Arthur et al., 2001; Schmidt & Hunter, 2015).
Meta-regression analyses were conducted to assess moderation effects for attitude operational-
ization when sufficient data were available. All analyses were conducted in Microsoft Excel
2013, SPSS Version 20, and Comprehensive Meta-Analysis Version 3, as appropriate.

Results
Record Characteristics

The 19 records included in the meta-analysis produced 73 unique, independent effect

sizes, representing a total sample size of 9,355 participants (M = 492.37, SD = 546.82, Mdn =
318, range = 42-2,282). A majority of included records were published in peer-review jour-
nals (n = 15, 78.9%). Included records were produced between 1996 and 2013 (Mdn =
2009; see Figure 2), with a majority conducted in the United States (n = 16, 84.2%). Across
all records, a majority of the samples included members of the general public (n = 12,
63.2%), with a handful of student (n = 3, 15.8%) and employer (n = 4, 21.1%) samples.
Most records used rigorous and representative sampling procedures, such as random digit
dialing (n = 6, 31.6%), other types of random sampling (n = 3, 15.8%), door-to-door surveys
(n = 1, 5.3%), and recruitment within local community establishments (n = 2, 10.5%).
About half of records assessed attitudes toward ex-offenders based on sexual offense his-
tory (n = 9, 47.4%), one quarter assessed attitudes based on felony offense history (n = 5,
26.3%), and few assessed attitudes based on violent offense histories (n = 4, 21.1%).2 See
Table 1 for descriptive characteristics of all included records.
Measures of attitudes included a variety of categorical and continuous variables that
assessed public attitudes toward ex-offenders. Outcome measures assessing desired social

Downloaded from cjb.sagepub.com by guest on October 12, 2016

 Rade et al. / PUBLIC ATTITUDES TOWARD EX-OFFENDERS 1269

Figure 2: Number of Records Assessing Public Attitudes Toward Ex-Offenders Over Time

distance were most common (n = 11, 57.9%), followed by measures of negative attitudes
(n = 5, 26.3%) and punitiveness (n = 5, 26.3%).3 Although the specific operational defini-
tions varied, all outcome variables assessed public attitudes toward ex-offenders in the
community.

Correlates Of Public Attitudes

Public Characteristics

Small weighted mean correlations were found between public characteristics and atti-
tudes toward ex-offenders, with values ranging from .01 for age to −.18 for interpersonal
contact (Tables 2 and 3). Two correlates associated with public attitudes met the established
effect size threshold: political affiliation/ideology and interpersonal contact. People with
politically conservative ideologies ( r w = −.12, 95% confidence interval [CI] = [−0.18,
−0.04]) and those who reported no previous contact with a current or ex-offender r w =
−.18, 95% CI = [−0.22, −0.13]) reported more negative attitudes toward ex-offenders com-
pared with members of the public with non-conservative political beliefs and those with
reported contact (including personal experience with the criminal justice system). Religious
affiliation was not examined because effect size data were only available for one record that
met inclusion criteria.

Ex-Offender Characteristics

Small weighted mean correlations were found between ex-offender characteristics and
public attitudes toward ex-offenders. As hypothesized, the public reported more negative
attitudes toward ex-offenders with histories of sexual offending compared with ex-offenders
with no history of sexual offending ( r w = −.19, 95% CI = [−0.26, −0.12]). However, the
effect size for public attitudes toward ex-offenders with and without histories of violent
offenses was small and negligible to interpretation. Felony offense history, ex-offender
race/ethnicity, rehabilitation participation, and presence of a mental disorder were not
examined because effect size data were only available for one record that met inclusion
criteria within each category (see Table 2).

Downloaded from cjb.sagepub.com by guest on October 12, 2016

 1270
Table 1: Descriptive Characteristics of Included Records

Sampling
Record Year N k Country Status Sample Method Outcome

Albright and Denq 1996 83 2 The United States Published Employers Convenience Social distance measure
Atkin-Plunk and Armstrong 2013 100 3 The United States Published Employers Random Social distance measure
Brown 1999 312 4 The United Kingdom Published General public Random Social distance measure
Burchfield 2012 95 4 The United States Published General public Door-to-door Attitudinal measure, social
distance measure
Comartin, Kernsmith, and 2009 703 4 The United States Published General public RDD Punitiveness measure
Kernsmith
Dawson Edwards 2007 218 5 The United States Not Students Convenience Punitiveness measure
published
Gibson, Roberson, and Daniel 2009 461 1 The United States Published Students Convenience Social distance measure
Hirschfield and Piquero 2010 2,282 9 The United States Published General public RDD Attitudinal measure
Levenson, Brannon, Fortney, 2007 193 5 The United States Published General public Community Social distance measure
and Baker
Locke 2010 318 8 The United States Not General public Community Social distance measure
published
Mancini, Shields, Mears, and 2010 1,380 6 The United States Published General public RDD Punitiveness measure
Beaver
Manza, Brooks, and Uggen 2004 1,000 2 The United States Published General public RDD Punitiveness measure
Martinez 2011 42 1 The United States Not Employers Convenience Social distance measure
published

Downloaded from cjb.sagepub.com by guest on October 12, 2016

Pager and Quillian 2005 199 1 The United States Published Employers Random Social distance measure
S. Park 2009 529 6 The United States Not Students Convenience Attitudinal measure
published
Perkins, Raines, Tschopp, and 2009 404 1 The United States Published General public RDD Social distance measure
Warner
Piquero et al. 2011 400 4 The United States Published General public RDD Punitiveness measure
Rogers, Hirst, and Davies 2011 235 3 The United Kingdom Published General public Convenience Attitudinal measure
Willis, Malinen, and Johnston 2013 401 4 New Zealand Published General public Social media Attitudinal measure, social
distance measure

Note. Reported sample size (N) refers to total sample size for the record; actual sample size for individual effect sizes (k) may vary. RDD = random digit dialing.
 Table 2: Weighted Mean Effect Sizes of Negative Attitudes Toward Ex-Offenders and Moderation Analyses for Sexual Offense History

Correlates k n r w(σρ) PVA (%) 95% CIw 80% Cred. Q(k−1)

Public characteristic
Sex (female = 0) 13 6,949 −0.04 (0.07) 28.47 [−0.09, 0.02] [−0.13, 0.05] 45.75***
  Sex offender 7 3,109 −0.08 (0.03) 75.11 [−0.12, −0.04] [−0.11, −0.04] 9.34
  Non-sex offender 2 711 −0.13 (0.14) 20.03 [−0.23, −0.03] [−0.27, 0.00] 10.02**
Race/ethnicity (majority = 0) 11 6,252 −0.09 (0.05) 38.55 [−0.13, −0.04] [−0.15, −0.02] 28.59***
  Sex offender 5 2,437 −0.04 (0.06) 33.54 [−0.10, 0.02] [−0.12, 0.04] 14.94**
  Non-sex offender 2 686 −0.07 (0.03) 71.73 [−0.15, 0.01] [−0.12, −0.03] 2.80
Education 8 5,584 −0.04 (0.07) 25.32 [−0.09, 0.02] [−0.12, 0.05] 31.65***
  Sex offender 5 2,604 −0.08 (0.04) 52.79 [−0.13, −0.03] [−0.13, −0.03] 9.49*
  Non-sex offender 2 698 −0.10 (0.07) 35.34 [−0.19, −0.01] [−0.19, −0.01] 5.68*
Religious affiliation (Christian = 0) 1 — — — — — —
Political affiliation/ideology 6 4,792 −0.12 (0.06) 27.88 [−0.18, −0.08] [−0.20, −0.05] 21.55***
(conservative = 0)
  Sex offender 1 — — — — — —
  Non-sex offender 2 532 −0.05 (0.01) 95.40 [−0.14, 0.04] [−0.07, −0.03] 2.10
Age 9 5,099 0.01 (0.07) 26.95 [−0.05, 0.06] [−0.08, 0.09] 33.46***
  Sex offender 5 2,214 0.00 (0.08) 24.33 [−0.07, 0.07] [−0.11, 0.11] 20.60***
  Non-sex offender 1 — — — — — —
Income 7 4,690 0.03 (0.04) 54.93 [−0.00, 0.07] [−0.01, 0.08] 12.76*
  Sex offender 4 1,885 −0.01a 607.53 [−0.05, 0.02] a 0.66
  Non-sex offender 1 — — — — — —
Contact (no contact = 0) 4 3,659 −0.18 (0.05) 31.45 [−0.22, −0.13] [−0.24, −0.12] 12.73**
  Sex offender 1 — — — — — —
  Non-sex offender 0 — — — — — —
Ex-offender characteristic

Downloaded from cjb.sagepub.com by guest on October 12, 2016

Offense history (violent = 0) 4 1,596 0.03 (0.13) 12.82 [−0.07, 0.13] [−0.14, 0.20] 31.29***
Offense history (sexual = 0) 2 1,343 −0.19 (0.07) 21.87 [−0.26, −0.12] [−0.28, 0.10] 9.16**
Offense history (felony = 0) 1 — — — — — —
Race/ethnicity (majority = 0) 1 — — — — — —
Rehabilitation participant (no = 0) 1 — — — — — —
Mental disorder (yes = 0) 1 — — — — — —
Community characteristic
Community size (small/rural = 0) 3 3,120 −0.04 (0.03) 54.95 [−0.08, −0.00] [−0.08, −0.01] 5.47
Crime prevalence (low = 0) 1 — — — — — —

Note. Bold text indicates effect sizes that met the established weighted effect size cutoff ( r w ≥ .10). k = number of effect sizes; n = total number of participants;
r w = weighted mean effect size; σρ = standard deviation of sampling error; PVA = proportion of variance accounted for by sampling error and attenuating artifacts; 95%
CIw = confidence interval of weighted mean effect size; 80% Cred = credibility interval of corrected standard deviation; Q(k−1) = chi-square homogeneity test.

1271
aNegative residual variance resulted in inability to calculate standard deviation of sampling error.

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

1272 Criminal Justice and Behavior

Table 3: Corrected Weighted Mean Effect Sizes of Negative Attitudes Toward Ex-Offenders

Corrected Public
Characteristic
Correlates k nc r wc (σρc) PVAc (%) 95% CIwc 80% Cred.c Q(k−1)c

Education 8 4,534 −0.03 (0.08) 23.26 [−0.09, 0.03] [−0.13, 0.06] 34.46***
Sex offender 5 1,698 −0.09 (0.06) 48.50 [−0.15, −0.03] [−0.17, −0.02] 10.23*
Non-sex offender 2 554 −0.10 (0.10) 27.98 [−0.21, 0.00] [−0.22, 0.02] 7.17**
Age 9 4,311 0.01 (0.08) 26.98 [−0.05, 0.07] [−0.09, 0.10] 33.43***
Sex offender 5 1,426 0.00 (0.11) 24.36 [−0.09, 0.09] [−0.13, 0.13] 20.60***
Non-sex offender 1 — — — — — —
Income 7 4,086 0.04 (0.04) 55.72 [−0.00, 0.08] [−0.01, 0.08] 12.59*
Sex offender 4 1,368 −0.01a 630.73 [−0.06, 0.03] a 0.64
Non-sex offender 1 — — — — — —

Note. All values and analyses corrected for artificially dichotomized independent variable. k = number of effect
sizes; nc = corrected total number of participants; r wc = corrected weighted mean effect size; σρc = corrected
standard deviation of sampling error; PVAc = corrected proportion of variance accounted for by sampling error
and attenuating artifacts; 95% CIwc = corrected confidence interval of weighted mean effect size; 80% Cred.c =
corrected credibility interval of corrected standard deviation; Q(k−1)c = corrected chi-square homogeneity test.
aNegative residual variance resulted in inability to calculate standard deviation of sampling error and credibility

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

Community Characteristics

A small weighted mean correlation was found between community size and public atti-
tudes toward ex-offenders. Members of the public who live in smaller, rural communities
compared with those living in larger, non-rural communities reported more negative attitudes
toward ex-offenders ( r w = −.04, 95% CI = [−0.08, −0.00]), but this fell below the established
benchmark (see Table 2). Effect size data for community crime prevalence were available for
only one record and, therefore, could not be analyzed as a correlate of public attitudes.

Sexual Offense History As A Moderator

Analyses revealed a significant amount of variability in all tested public and ex-offender
correlates (all Qs significant at p < .05; see Tables 2 and 3), suggesting the presence of
moderating variables. The PVAse was less than the 75% benchmark (Arthur et al., 2001;
Schmidt & Hunter, 2015) for all public and ex-offender correlates (Tables 2 and 3), provid-
ing additional support for the presence of moderators. Among the community correlates, the
homogeneity of variance test for community size was not significant (p > .05), but the stan-
dard error accounted for over half of the variance (PVAse = 54.95%), again suggesting the
presence of moderating variables (see Table 2).
When tested as a moderator, sexual offense history reduced the amount of variability
present among some correlates (see Tables 2 and 3). Across moderation analyses, effect
sizes were larger for attitudes toward ex-offenders without sexual offense histories com-
pared with those with sexual offense histories, but the differences were not significant.
Most subset correlates continued to demonstrate significant homogeneity values and some
credibility intervals contained zero. Sufficient data were not available to assess sexual
offense history as a moderator for religious affiliation, political affiliation/ideology, age,
income, and interpersonal contact.

Downloaded from cjb.sagepub.com by guest on October 12, 2016

 Rade et al. / PUBLIC ATTITUDES TOWARD EX-OFFENDERS 1273

Potentially Biasing Records

Though publication status, sampling methodology, and country did not emerge as statis-
tically significant moderators, effect sizes tended to be smaller for records that were pub-
lished in peer-review outlets, utilized representative sampling methods, and were conducted
outside of the United States. Similarly, public attitudes toward ex-offenders did not differ
significantly between operationalizations of social distance, punitiveness, and general atti-
tudes. (Full results not reported but available upon request.) In contrast, analyses of poten-
tial biases associated with year produced revealed significant moderation for public age and
income correlates, with reduced variance in subsets and non-overlapping credibility inter-
vals (Table 4). Before 2009, older respondents reported more negative attitudes toward ex-
offenders compared with younger respondents ( r wc = .15, 95% CI = [0.05, 0.24]). During
and after 2009, the association is negligible ( r wc = −.01, 95% CI = [−0.06, 0.03]). In addi-
tion, before 2009, members of the public with larger household incomes reported more
negative attitudes toward ex-offenders than did those with smaller household incomes
( r wc = .16, 95% CI = [0.03, 0.29]). During and after 2009, the association is negligible ( r wc =
.03, 95% CI = [0.00, 0.06]).

Discussion
Negative public attitudes toward ex-offenders are a known barrier to reentry and success-
ful community reintegration (e.g., Wakefield & Uggen, 2010). Research for more than 20
years has sought to understand factors associated with public attitudes toward ex-offenders
in efforts to improve reentry. However, findings regarding associations of public, ex-offender,
and community characteristics with public attitudes have been inconsistent across studies.
This meta-analysis presents a comprehensive assessment of the correlates of public attitudes
toward ex-offenders and potential moderating effects of sexual offense history, publication
status, sampling methodology, year produced, country, and attitude operationalization.
Overall, we found negligible effect sizes across frequently studied correlates of public
attitudes toward ex-offenders, with the exception of interpersonal contact, political affilia-
tion/ideology, and sexual offense history. Members of the public who report having any
type of contact with a current offender or ex-offender also report less negative attitudes
toward ex-offenders. This finding suggests that interpersonal contact is a significant corre-
late of public attitudes, despite potential differences in type (i.e., face-to-face, vicarious) or
quality (i.e., positive, negative) of the contact. Those who self-identify as politically conser-
vative or Republican report more negative attitudes toward ex-offenders. Findings are con-
sistent with research examining attitudes toward other populations, such as death row
inmates and juvenile offenders, which show that people with more conservative political
ideologies are more likely to support punitive policies that may be attributed to disposi-
tional attributions (e.g., Grasmick & McGill, 1994; Jacobs & Carmichael, 2002).
Dispositional attributions are based on the assumption that all individuals, including ex-
offenders, are responsible for their actions and, therefore, the consequences of those actions
(Grasmick & McGill, 1994; Jacobs & Carmichael, 2002).
Among the correlates of public attitudes examined herein, sexual offense history pro-
duced the largest effect size across records. This finding is consistent with studies of atti-
tudes toward sex offenders held by other groups, including criminal justice professionals,
who report more negative attitudes toward sex offenders compared with ex-offenders with

Downloaded from cjb.sagepub.com by guest on October 12, 2016

1274 Criminal Justice and Behavior

Table 4: Moderation Analyses for Potentially Biasing Records: Year Produced

Correlates k n r w (σρ) PVA (%) 95% CIw 80% Cred. Q(k−1)

Public characteristic
Sex (female = 0) 13 6,949 −0.04 (0.07) 28.47 [−0.09, 0.02] [−0.13, 0.05] 45.75***
  <2009 2 461 −0.01a 4,094.02 [−0.01, 0.09] a 0.05
  ≥2009 13 6,488 −0.04 (0.07) 24.36 [−0.10, 0.02] [−0.13, 0.05] 45.23***
Race/ethnicity 11 6,252 −0.09 (0.05) 38.55 [−0.13, −0.04] [−0.15, −0.02] 28.59***
(majority = 0)
  <2009 3 631 −0.11a 440.48 [−0.18, −0.05] a 0.68
  ≥2009 8 5,621 −0.08 (0.06) 29.36 [−0.13, −0.03] [−0.16, −0.01] 27.28***
Educationb 8 4,534 −0.03 (0.08) 23.26 [−0.09, 0.03] [−0.13, 0.06] 34.46***
  <2009 1 — — — — — —
  ≥2009 7 4,345 −0.04 (0.08) 20.60 [−0.10, 0.03] [−0.14, 0.07] 34.04***
Political affiliation 6 4,792 −0.12 (0.06) 27.88 [−0.18, −0.08] [−0.20, −0.05] 21.55***
(conservative = 0)
  <2009 1 — — — — — —
  ≥2009 5 4,574 −0.12 (0.04) 36.34 [−0.16, −0.07] [−0.17, −0.06] 13.77**
Ageb 9 4,311 0.01 (0.08) 26.98 [−0.05, 0.07] [−0.09, 0.10] 33.43***
  <2009 3 554 0.15 (0.08) 46.48 [0.05, 0.24] [0.05, 0.24] 6.49*
  ≥2009 6 3,757 −0.01 (0.05) 39.98 [−0.06, 0.03] [−0.08, 0.05] 15.03*
Incomeb 7 4,086 0.04 (0.04) 55.72 [−0.00, 0.08] [−0.01, 0.08] 12.59*
  <2009 2 239 0.16 (0.07) 63.62 [0.03, 0.29] [0.07, 0.25] 3.17
  ≥2009 5 3,847 0.03 (0.01) 87.01 [−0.00. 0.06] [0.01, 0.05] 5.75
Contact (no 4 3,659 −0.18 (0.05) 31.45 [−0.22, −0.13] [−0.24, −0.12] 12.73**
contact = 0)
  <2009 0 — — — — — —
  ≥2009 4 3,659 −0.18 (0.05) 31.45 [−0.22, −0.13] [−0.24, −0.12] 12.73**
Ex-offender characteristic
Violent (violent = 0) 4 1,596 0.03 (0.13) 12.82 [−0.07, 0.13] [−0.14, 0.20] 31.29***
  <2009 3 1,512 0.02 (0.13) 10.78 [−0.08, 0.12] [−0.14, 0.18] 27.90***
  ≥2009 1 — — — — — —
Sexual (sexual = 0) 2 1,343 −0.19 (0.07) 21.87 [−0.26, −0.12] [−0.28, 0.10] 9.16**
  <2009 2 1,343 −0.19 (0.07) 21.87 [−0.26, −0.12] [−0.28, −0.10] 9.16
  ≥2009 0 — — — — — —
Community characteristic
Community size 3 3,120 −0.04 (0.03) 54.95 [−0.08, −0.00] [−0.08, −0.01] 5.47
(small/rural = 0)
  <2009 0 — — — — — —
  ≥2009 3 3,120 −0.04 (0.03) 54.95 [−0.08, −0.00] [−0.08, −0.01] 5.47

Note. Bold text indicates effect sizes that met the established weighted effect size cutoff ( r w ≥ .10) or significant
moderation effects. k = number of effect sizes; n = total number of participants; r w = weighted mean effect size;
σρ = standard deviation of sampling error; PVA = proportion of variance accounted for by sampling error and
attenuating artifacts; 95% CIw = confidence interval of weighted mean effect size; 80% Cred = credibility interval of
corrected standard deviation; Q(k−1) = chi-square homogeneity test.
aNegative residual variance resulted in inability to calculate standard deviation of sampling error. bValues and

analyses corrected for artificially dichotomized independent variable.

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

no history of sexual offending (see Weekes, Pelletier, & Beaudette, 1995). Moderation anal-
yses revealed that effect sizes tended to be larger for attitudes toward ex-offenders without
histories of sexual offenses, compared with those with sexual offense histories. We observed
greater effects of public characteristics on attitudes toward ex-offenders with no history of

Downloaded from cjb.sagepub.com by guest on October 12, 2016

 Rade et al. / PUBLIC ATTITUDES TOWARD EX-OFFENDERS 1275

sexual offending compared with sex offenders. Women compared with men, for instance,
reported more negative attitudes toward ex-offenders with no history of sexual offending,
whereas differences in attitudes between men and women toward ex-offenders with sexual
offense histories were negligible. Subgroup analyses were not significant, but limited by the
small number of records (n = 3, 15.8%) explicitly assessing attitudes toward ex-offenders
with no history of sexual offending. Findings suggest a stronger consensus of negative atti-
tudes toward sex offenders, while attitudes toward ex-offenders with no history of sexual
offending appear to be more variable. Additional analyses of offense history failed to iden-
tify differences between public attitudes toward ex-offenders with a history of violent
offenses compared with those without a history of violent offenses.
Our investigation of potentially biasing records revealed an additional moderator: year
produced. Before 2009, older people and those with larger household incomes reported
more negative attitudes toward ex-offenders compared with younger people and those with
smaller household incomes. During and after 2009, however, the associations of age and
income with attitudes toward ex-offenders became negligible. In other words, regardless of
age and income, members of the public began reporting more similar attitudes toward ex-
offenders—although not necessarily more positive—after 2009. These findings suggest that
public attitudes toward ex-offenders change over time.
Reasons for a shift in public attitudes may be explained, in part, by changes in local or
federal policies related to ex-offender reentry. The Second Chance Act of 2007 (42 U.S.C. §
17501) is an example of a large-scale federal policy that aimed to improve reentry, reduce
recidivism, and increase public safety. The Second Chance Act also contributed to the devel-
opment of additional policy and funding mechanisms focused on improving ex-offender
reentry. One such policy is the Federal Prisoner Reentry Initiative of 2009 (42 U.S.C. §
17541), which established an inter-agency strategy for preparing ex-offenders for successful
reentry, including an emphasis on skill development, medical care, and employment, with
priority given to high risk offenders. Policies such as the Second Chance Act and the Federal
Prisoner Reentry Initiative may influence public opinion (Gideon & Loveland, 2011) and
have the potential to improve public attitudes toward ex-offenders and, ultimately, reduce
barriers to successful community reintegration. That said, given the present analysis identi-
fied that public attitudes changed over time but were not necessarily more positive, there is
a need for continued research on the ways in which such policy changes may have a positive
influence on public opinion and reentry outcomes. These findings reveal that community
context, such as local and federal policies, may be associated with public attitudes. As identi-
fied in this study, limited research has been conducted on community-level correlates. Future
research on the effects of community context on public attitudes toward ex-offenders and
reentry is warranted.
Given the generally small associations between public characteristics and public atti-
tudes, efforts to improve public attitudes toward ex-offenders should not necessarily target
or tailor material to certain socio-demographic groups. Rather, findings suggest that inter-
ventions may be more effective if they (a) seek to reduce stigma associated with specific
criminal histories and (b) incorporate interpersonal contact. First, attitudes toward ex-
offenders vary across different criminal histories, particularly across sexual offense but not
violent offense histories, and therefore, strategies for reducing stigma may require variation
or modification. A public health approach that emphasizes prevention is an example of one
suggested intervention strategy (Brown, 2009; Laws, 2000). Under this model, prevention

Downloaded from cjb.sagepub.com by guest on October 12, 2016

1276 Criminal Justice and Behavior

steps may include media campaigns, public education, training for professionals, or
increased access to community-based resources, each of which could be tailored to address
specific offenses, such as sexual offenses or drug possession offenses. Many organizations
are using these types of multifaceted targeted interventions, including advocacy initiatives,
training, and media campaigns to facilitate successful reentry, increase public awareness of
issues surrounding reintegration, and improve public perceptions of ex-offenders (e.g., Just
Leadership USA, 2016; Legal Services for Prisoners With Children, 2015). Education-
based interventions also may be tailored to preliminarily address less stigmatized ex-
offender groups (i.e., those without a history of sexual offense) or to target stigma toward
specific offenses (e.g., Kleban & Jeglic, 2012). Materials could be developed to educate
potential employers about recidivism rates, the various pathways to crime, and the diversity
of behaviors classified within a specific crime category.
Second, having interpersonal contact with an ex-offender is associated with more posi-
tive attitudes toward ex-offenders. Interventions that incorporate elements of contact may
be more effective; however, the causal direction is unclear. Research with other stigma-
tized populations, such as adults with mental illnesses, suggests that contact-based inter-
ventions can improve attitudes, although the success of these anti-stigma interventions
may depend upon characteristics of the contact (Clement et al., 2012; Corrigan, Morris,
Michaels, Rafacz, & Rüsch, 2012). In the same way, interventions such as the Inside-Out
Prison Exchange Program for college students and prisoners (Conti, Morrison, & Pantaleo,
2013; Pompa, 2013) have been designed to facilitate interpersonal contact with offenders
while incarcerated. To our knowledge, however, such contact-based interventions have not
been applied to reduce stigma toward ex-offenders living in the community specifically.
Future research and intervention addressing the influence of interpersonal contact on atti-
tudes toward ex-offenders should consider the type (e.g., face-to-face, vicarious) and qual-
ity (e.g., positive, negative) of contact.

Limitations

Findings should be considered within the context of limitations of both the design of our
meta-analysis and the included records. Even though we employed systematic efforts to
find all available research through rigorous search procedures, some relevant records may
not have been identified. We identified a relatively small number of records overall (n = 19).
Our meta-analytic review revealed a dearth of research assessing public correlates of atti-
tudes toward ex-offenders. So few empirical studies have been conducted on ex-offender
(felony criminal history, race/ethnicity, rehabilitation participation, age, mental illness) and
community characteristics (crime prevalence) that many of these correlates could not be
assessed in our meta-analysis due to insufficient data. The small number of studies assess-
ing offense history, in particular, prevented us from assessing other offense histories (e.g.,
violent offense history, felony offense history) as moderators of public attitudes. This small
number of records did have sufficient power as post hoc power analyses revealed that all
weighted effect sizes had ample sample sizes to detect small significant effects at α = .01
(power = 0.80; Cohen, 1992). Finally, we included records regardless of methodological
rigor or quality to allow for the most comprehensive analysis; even so, many records used
random sampling procedures, and potential biases were controlled for through weighted
correlations, random-effects, and artifact correction techniques.

Downloaded from cjb.sagepub.com by guest on October 12, 2016

 Rade et al. / PUBLIC ATTITUDES TOWARD EX-OFFENDERS 1277

Conclusions And Future Directions

This study marks the first meta-analysis of the correlates and moderators of public atti-
tudes toward ex-offenders. Findings bring to light the limited nature of the research on cor-
relates of public attitudes toward ex-offenders and, particularly, the lack of research
investigating public attitudes toward ex-offenders without a sexual offense history. More
generally, there has been a focus on public characteristics to the neglect of ex-offender and
community correlates in the current empirical literature. Future research should explore the
ways in which histories of violent and non-violent offenses, as well as felony offenses,
rehabilitation participation, and mental illnesses are associated with public attitudes.
Findings also revealed small associations between those correlates that have been studied
and public attitudes toward ex-offenders. We should begin asking what else there is to learn
about public attitudes toward ex-offenders by exploring other correlates, such as implicit
beliefs and biases, the influence of family members and peers, and changes in local legisla-
tion on public attitudes. There is a need to translate these findings into practice through
development, implementation, and evaluation of anti-stigma interventions, working toward
improvement of ex-offender community reentry.

Notes
1. For example, when data for the age correlate were reported as a continuous variable in a given record, we calculated
the effect size for that record using the continuous format; however, if data for the age correlate were reported as a categorical
variable in another record, we calculated the effect size for that record using the categorical format.
2. Percentages may exceed 100% because offense histories are not mutually exclusive.
3. Percentages may exceed 100% due to the presence of more than one outcome measure in some records.

References
References marked with an asterisk indicate studies included in the meta-analysis.
*Albright, S., & Denq, F. (1996). Employer attitudes toward hiring ex-offenders. The Prison Journal, 76, 118-137.
doi:10.1177/0032855596076002002
Allport, G. W. (1954). The nature of prejudice. Reading, MA: Addison-Wesley.
Arthur, W., Bennett, W., & Huffcutt, A. I. (2001). Conducting meta-analysis using SAS. Mahwah, NJ: Lawrence Erlbaum.
*Atkin-Plunk, C. A., & Armstrong, G. S. (2013). Does the concentration of parolees in a community impact employer atti-
tudes toward the hiring of ex-offenders? Criminal Justice Policy Review, 24, 71-93. doi:10.1177/0887403411428005
Brooks, L. E., Visher, C. A., & Naser, R. L. (2006). Community residents’ perceptions of prisoner reentry in selected
Cleveland neighborhoods. Washington, DC: Justice Policy Center, Urban Institute.
*Brown, S. (1999). Public attitudes toward the treatment of sex offenders. Legal and Criminological Psychology, 4, 239-252.
doi:10.1348/135532599167879
Brown, S. (2009). Attitudes towards sexual offenders and their rehabilitation: A special case? In J. L. Wood & T. A. Gannon
(Eds.), Public opinion, crime, and criminal justice (pp. 187-213). Portland, OR: Willan.
*Burchfield, K. B. (2012). Assessing community residents’ perceptions of local registered sex offenders: Results from a pilot
survey. Deviant Behavior, 33, 241-259. doi:10.1080/01639625.2011.573396
Carson, E. A., & Sabol, W. J. (2012). Prisoners in 2011. Washington, DC: Bureau of Justice Statistics, U.S. Department of
Justice. Retrieved from http://www.bjs.gov/content/pub/pdf/p11.pdf
Clear, T. R., Rose, D. R., & Ryder, J. A. (2001). Incarceration and the community: The problem of removing and returning
offenders. Crime & Delinquency, 47, 335-351. doi:10.1177/0011128701047003003
Clement, S., van Nieuwenhuizen, A., Kassam, A., Flach, C., Lazarus, A., de Castro, M., . . . Thronicroft, G. (2012). Filmed v.
live social contact interventions to reduce stigma: Randomized controlled trial. The British Journal of Psychiatry, 201,
57-64. doi:10.1192/bjp.bp.111.093120
Cohen, J. (1992). A power primer. Psychological Bulletin, 112, 155-159. doi:10.1037/0033-2909.112.1.155
*Comartin, E. B., Kernsmith, P. D., & Kernsmith, R. M. (2009). Sanctions for sex offenders: Fear and public policy. Journal
of Offender Rehabilitation, 48, 605-619. doi:10.1080/10509670903196066
Conti, N., Morrison, L., & Pantaleo, K. (2013). All the wiser: Dialogic space, destigmatization, and teacher-activist recruit-
ment. The Prison Journal, 93, 163-188. doi:10.1177/0032885512472654

Downloaded from cjb.sagepub.com by guest on October 12, 2016

1278 Criminal Justice and Behavior

Corrigan, P. W., Morris, S. B., Michaels, P. J., Rafacz, J. D., & Rüsch, N. (2012). Challenging the public stigma of mental
illness: A meta-analysis of outcome studies. Psychiatric Services, 63, 963-973. doi:10.1176/appi.ps.005292011
*Dawson Edwards, B. C. (2007). Determinants of college students’ opinions towards felon voting rights: An exploratory
study (Unpublished doctoral dissertation). Virginia Commonwealth University, Richmond.
Dovidio, J. F., Glick, P., & Rudman, L. A. (2005). Introduction: Reflecting on the nature of prejudice: Fifty years after
Allport. In J. F. Dovidio, P. Glick, & L. A. Rudman (Eds.), On the nature of prejudice: Fifty years after Allport
(pp. 1-16). Malden, MA: Blackwell.
Dreiling, K. L. (2011). Knowledge and opinion of sex offenders and sex offender policy: Do sources of information matter?
(Unpublished doctoral dissertation). South Dakota State University, Brookings.
Federal Prisoner Reentry Initiative of 2009, 42 U.S.C. § 17541.
*Gibson, C., Roberson, D., & Daniel, A. (2009). African American women on the possibilities of a relationship with an ex-
offender. Journal of Criminal Justice, 37, 328-332. doi:10.1016/j.jcrimjus.2009.06.009
Gideon, L., & Loveland, N. (2011). Public attitudes toward rehabilitation and reintegration. In L. Gideon & H.-E. Sung (Eds.),
Rethinking corrections: Rehabilitation, reentry, and reintegration (pp. 19-36). Thousand Oaks, CA: SAGE.
Goffman, E. (1963). Stigma: Notes on the management of spoiled identity. Englewood Cliffs, NJ: Prentice Hall.
Graffam, J., Shinkfield, A., Lavelle, B., & McPherson, W. (2004). Variables affecting successful reintegration as perceived by
offenders and professionals. Journal of Offender Rehabilitation, 40, 147-171. doi:10.1300/J076v40n01_08
Grasmick, H. G., & McGill, A. L. (1994). Religion, attribution style, and punitiveness toward juvenile offenders. Criminology,
32, 23-46. doi:10.1111/j.1745-9125.1994.tb01145.x
Hardcastle, L., Bartholomew, T., & Graffam, J. (2011). Legislative and community support for offender reintegration in
Victoria. Deakin Law Review, 16, 111-132.
*Hirschfield, P. J., & Piquero, A. R. (2010). Normalization and legitimation: Modeling stigmatizing attitudes toward ex-
offenders. Criminology, 48, 27-55. doi:10.1111/j.1745-9125.2010.00179.x
Hulsey, L. F. (1991). Attitudes of employers with respect to hiring released prisoners (Unpublished doctoral dissertation).
Texas A&M University, College Station.
Hunter, J. E., & Schmidt, F. L. (2000). Fixed effects vs. random effects meta-analysis models: Implications for cumulative
research knowledge. International Journal of Selection and Assessment, 8, 275-292. doi:10.1111/1468-2389.00156
Jacobs, D., & Carmichael, J. T. (2002). The political sociology of the death penalty: A pooled time-series analysis. American
Sociological Review, 67, 109-131. doi:10.2307/3088936
Just Leadership USA. (2016). About us. Retrieved from https://www.justleadershipusa.org/about-us/
Kleban, H., & Jeglic, E. (2012). Dispelling the myths: Can psychoeducation change public attitudes towards sex offenders?
Journal of Sexual Aggression, 18, 179-193. doi:10.1080/13552600.2011.552795
Laws, D. R. (2000). Sexual offending as a public health problem: A North American perspective. Journal of Sexual Aggression,
5, 30-44. doi:10.1080/13552600008413294
LeBel, T. P. (2008). Perceptions of and responses to stigma. Sociology Compass, 2, 409-432. doi:10.1111/j.1751-
9020.2007.00081.x
LeBel, T. P. (2012). If one doesn’t get you another one will: Formerly incarcerated persons’ perceptions of discrimination.
The Prison Journal, 92, 63-87. doi:10.1177/0032885511429243
Legal Services for Prisoners With Children. (2015). All of Us or None. Retrieved from http://www.prisonerswithchildren.org/
our-projects/allofus-or-none/
*Levenson, J. S., Brannon, Y. N., Fortney, T., & Baker, J. (2007). Public perceptions about sex offenders and community
protection policies. Analyses of Social Issues and Public Policy, 7, 137-161. doi:10.1111/j.1530-2415.2007.00119.x
Leverentz, A. (2011). Neighborhood context of attitudes toward crime and reentry. Punishment & Society, 13, 64-92.
doi:10.1177/1462474510385629
Liberati, A., Altman, D. G., Tetzlaff, J., Mulrow, C., Gøtzsche, P. C., Ioannidis, J. P. A., . . . Moher, D. (2009). The
PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate healthcare interventions:
Explanation and elaboration. The British Medical Journal, 339. doi:10.1136/bmj.b2700
Lipsey, M. W., & Wilson, D. B. (2001). Practical meta-analysis. Thousand Oaks, CA: SAGE.
*Locke, C. R. (2010). Public attitudes toward mental illness: An experimental design examining the media’s impact of crime
on stigma (Unpublished doctoral dissertation). The Ohio State University, Columbus.
Mackey, D. A., & Courtright, K. E. (2000). Assessing punitiveness among college students: A comparison of criminal justice
majors with other majors. Justice Professional, 12, 423-441. doi:10.1080/1478601X.2000.9959561
*Mancini, C., Shields, R. T., Mears, D. P., & Beaver, K. M. (2010). Sex offender residence restriction laws: Parental percep-
tions and public policy. Journal of Criminal Justice, 38, 1022-1030. doi:10.1016/j.jcrimjus.2010.07.004
*Manza, J., Brooks, C., & Uggen, C. (2004). Public attitudes toward felon disenfranchisement in the United States. Public
Opinion Quarterly, 68, 275-286. doi:10.1093/poq/nfh015
*Martinez, M. S. (2011). Employer perceptions about hiring released prisoners with a criminal history (Unpublished doctoral
dissertation). Capella University, Minneapolis, MN.

Downloaded from cjb.sagepub.com by guest on October 12, 2016

 Rade et al. / PUBLIC ATTITUDES TOWARD EX-OFFENDERS 1279

Maruschak, L. M., & Parks, E. (2012). Probation and parole in the United States, 2011. Washington, DC: Bureau of Justice
Statistics, U.S. Department of Justice. Retrieved from http://www.bjs.gov/content/pub/pdf/ppus11.pdf
Mason, C., Allam, R., & Brannick, M. T. (2007). How to meta-analyze coefficient-of-stability estimates: Some recom-
mendations based on Monte Carlo studies. Educational and Psychological Measurement, 67, 765-783. doi:10.1177/
0013164407301532
Moher, D., Liberati, A., Tetzlaff, J., Altman, D. G., & The PRISMA Group. (2009). Preferred reporting items for systematic
reviews and meta-analyses: The PRISMA statement. Journal of Clinical Epidemiology, 62, 1006-1012. doi:10.1016/j.
jclinepi.2009.06.005
Pager, D. (2003). The mark of a criminal record. American Journal of Sociology, 108, 937-975. doi:10.1086/374403
*Pager, D., & Quillian, L. (2005). Walking the talk? What employers say versus what they do. American Sociological Review,
70, 355-380. doi:10.1177/000312240507000301
Park, R. E., & Burgess, E. W. (1921). Introduction to the science of sociology. Chicago, IL: The University of Chicago Press.
Retrieved from http://www.gutenberg.org/files/28496/28496-
*Park, S. (2009). College students’ attitudes toward prisoners and prisoner reentry (Unpublished doctoral dissertation).
Indiana University of Pennsylvania.
*Perkins, D. V., Raines, J. A., Tschopp, M. K., & Warner, T. C. (2009). Gainful employment reduces stigma toward people
recovering from schizophrenia. Community Mental Health Journal, 45, 158-162. doi:10.1007/s10597-008-9158-3
Pettigrew, T. F., & Tropp, L. R. (2006). A meta-analytic test of intergroup contact theory. Journal of Personality and Social
Psychology, 90, 751-783. doi:10.1037/0022-3514.90.5.751
Phelan, J. C., Link, B. G., & Dovidio, J. F. (2008). Stigma and prejudice: One animal or two? Social Science & Medicine, 67,
358-367. doi:10.1016/j.socscimed.2008.03.022
*Piquero, A. R., Piquero, N. L., Gertz, M., Baker, T., Batton, J., & Barnes, J. C. (2011). Race, punishment, and the Michael
Vick experience. Social Science Quarterly, 92, 535-551. doi:10.1111/j.1540-6237.2011.00781.x
Pogorzelski, W., Wolff, N., Pan, K. Y., & Blitz, C. L. (2005). Behavioral health problems, ex-offender reentry policies, and
the Second Chance Act. American Journal of Public Health, 95, 1718-1724. doi:10.2105/AJPH.2005.065805
Pompa, L. (2013). One brick at a time: The power and possibility of dialogue across the prison wall. The Prison Journal, 93,
127-134.
*Rogers, P., Hirst, L., & Davies, M. (2011). An investigation into the effect of respondent gender, victim age, and perpetrator
treatment on public attitudes towards sex offenders, sex offender treatment, and sex offender rehabilitation. Journal of
Offender Rehabilitation, 50, 511-530. doi:10.1080/10509674.2011.602472
Schmidt, F. L., & Hunter, J. E. (2015). Methods of meta-analysis: Correcting error and bias in research findings (3rd ed.).
Thousand Oaks, CA: SAGE.
Schnittker, J., & Bacak, V. (2013). A mark of disgrace or a badge of honor?: Subjective status among former inmates. Social
Problems, 60, 234-254. doi:10.1525/sp.2013.60.2.234
Schulze, R. (2004). Meta-analysis: A comparison of approaches. Cambridge, MA: Hogrefe & Huber.
Second Chance Act. (2007). Community Safety through Recidivism Prevention, 42 U.S.C. § 17501.
Tewksbury, R., & Lees, M. (2006). Perceptions of sex offender registration: Collateral consequences and community experi-
ences. Sociological Spectrum, 26, 309-334. doi:10.1080/02732170500524246
Uggen, C. (2000). Work as a turning point in the life course of criminals: A duration model of age, employment, and recidi-
vism. American Sociological Review, 65, 529-546. doi:10.2307/2657381
U.S. Department of Justice. (2014). Violent crime (Federal Bureau of Investigation—Uniform Crime Report). Retrieved from
http://www.fbi.gov/about-us/cjis/ucr/crime-in-the-u.s/2013/crime-in-the-u.s.-2013/violent-crime/violent-crime-topic-
page/violentcrimemain_final.pdf
Varghese, F. P., Hardin, E. E., Bauer, R. L., & Morgan, R. D. (2010). Attitudes toward hiring offenders: The roles of criminal
history, job qualifications, and race. International Journal of Offender Therapy and Comparative Criminology, 54, 769-
782. doi:10.1177/0306624X09344960
Viki, G. T., Fullerton, I., Raggett, H., Tait, F., & Wiltshire, S. (2012). The role of dehumanization in attitudes toward the social
exclusion and rehabilitation of sex offenders. Journal of Applied Social Psychology, 42, 2349-2367. doi:10.1111/j.1559-
1816.2012.00944.x
Visher, C. A., Debus, S., & Yahner, J. (2008). Employment after prison: A longitudinal study of releases in three states.
Washington, DC: Justice Policy Center, Urban Institute.
Wakefield, S., & Uggen, C. (2010). Incarceration and stratification. Annual Review of Sociology, 36, 387-406. doi:10.1146/
annurev.soc.012809.102551
Weekes, J. R., Pelletier, G., & Beaudette, D. (1995). Correctional officers: How do they perceive sex offenders? International
Journal of Offender Therapy and Comparative Criminology, 39, 55-61. doi:10.1177/0306624X9503900107
Willis, G. M., Levenson, J. S., & Ward, T. (2010). Desistance and attitudes towards sex offenders: Facilitation or hindrance?
Journal of Family Violence, 25, 545-556. doi:10.1007/s10896-010-9314-8
*Willis, G. M., Malinen, S., & Johnston, L. (2013). Demographic differences in public attitudes towards sex offenders.
Psychiatry, Psychology and Law, 20, 230-247. doi:10.1080/13218719.2012.658206

Downloaded from cjb.sagepub.com by guest on October 12, 2016

1280 Criminal Justice and Behavior

Wilson, D. B. (2001a). Effect size determination program [Excel spreadsheet calculator]. Retrieved from http://mason.gmu.
edu/~dwilsonb/ma.html
Wilson, D. B. (2001b). Practical meta-analysis effect size calculator. Fairfax, VA: George Mason University. Retrieved from
http://www.campbellcollaboration.org/escalc/html/EffectSizeCalculator-Home.php

Candalyn B. Rade, MS, is a doctoral candidate in the Applied Social and Community Psychology Graduate Program at North
Carolina State University. Her research focuses on strategies to improve community reentry for justice-involved adults,
including access to community-based resources, behavioral health interventions, and attitudes toward ex-offenders.

Sarah L. Desmarais, PhD, is an associate professor of psychology and coordinator of the Applied Social and Community
Psychology Graduate Program at North Carolina State University. Her current research focuses on the development, imple-
mentation, and evaluation of assessment and intervention strategies for justice-involved adolescents and adults.

Roger E. Mitchell, PhD, is an associate professor emeritus in the Department of Psychology at North Carolina State
University. His broad research interests center on the impact of psychosocial and contextual factors on health and well-being,
as well the ecological and organizational factors influencing the capacity to launch effective health-promotion initiatives.

Downloaded from cjb.sagepub.com by guest on October 12, 2016