A Meta-Analysis of Public Attitudes Toward Ex-Offenders

Abstract

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 variables 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 members 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 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 education, 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 differential 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 commonly dehumanized and socially excluded (e.g., Viki et al., 2012).

Stigma and prejudice toward, and desired social distance from ex-offenders can be manifested as barriers to accessing services in the community (e.g., housing, education), exclusion 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 successful 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 representing public characteristics, ex-offender characteristics, and characteristics of the local 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 favorable 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 findings are mixed (e.g., Dawson Edwards, 2007; Leverentz, 2011). Finally, interpersonal contact theory (Allport, 1954; Pettigrew & Tropp, 2006) posits a negative relationship between interpersonal contact and prejudice, such that increased contact with a member of an outgroup 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 violent 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 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 punitive 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 attitudes 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 sexual 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 characteristics 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 barriers to successful community reintegration following release from correctional facilities, including public, ex-offender, and community characteristics associated with negative public 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 community 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 literature on correlates of public attitudes toward ex-offenders and to examine the potential moderating 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 produced, 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 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 students) 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 measures 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, correlation 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 presentations, 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 distance,” 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 communication, 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 communication, 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 communication, 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 evaluated 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 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.

Figure 1:

Results of Systematic Literature Search

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 available 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.

Outcome Variables

Attitudes toward ex-offenders were operationalized to include measures of general attitudes (e.g., attitudes toward ex-offenders), stigma (including measures of prejudice), punitiveness (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 available; 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) definition of violent crimes to include murder, non-negligent manslaughter, forcible rape, robbery, 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 United States = 1) was assessed for potential biases, as attitudes toward ex-offenders may vary internationally. Finally, attitude operationalization (social distance = 0, general attitudes = 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 coding 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., publication status, population type). For continuous variables (e.g., sample size), inter-rater reliability ranged 82% to 100% agreement (ICC_3,2 = .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 parameters 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 relationships 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 correlate 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 without providing specific data to approximate an effect size were coded as r = .00 (k = 6).

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 ( $\bar{r}$ _w) and to calculate the sample-weighted observed variance, var(r), using the formulas below (Arthur, Bennett, & Huffcutt, 2001; Schmidt & Hunter, 2015).

{\bar{r}}_{w} = \frac{Σ (N_{i} \times r_{i})}{Σ N_{i}} var (r) = \frac{Σ (N_{i} \times {(r_{i} - \bar{r})}^{2})}{Σ N_{i}} .

Corrections were applied to individual effect sizes of three artificially dichotomized variables (public education, k = 6; public age, k = 2; public income, k = 5) to correct for associated 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 $\bar{r}$ _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 (PVA_se) 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 demonstrated the presence of a moderation effect (Arthur et al., 2001; Schmidt & Hunter, 2015). Meta-regression analyses were conducted to assess moderation effects for attitude operationalization 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 journals (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 history (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.

Figure 2:

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

Table 1:

Descriptive Characteristics of Included Records

Record	Year	N	k	Country	Status	Sample	Sampling 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 Kernsmith	2009	703	4	The United States	Published	General public	RDD	Punitiveness measure
Dawson Edwards	2007	218	5	The United States	Not published	Students	Convenience	Punitiveness measure
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, and Baker	2007	193	5	The United States	Published	General public	Community	Social distance measure
Locke	2010	318	8	The United States	Not published	General public	Community	Social distance measure
Mancini, Shields, Mears, and Beaver	2010	1,380	6	The United States	Published	General public	RDD	Punitiveness measure
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 published	Employers	Convenience	Social distance measure
Pager and Quillian	2005	199	1	The United States	Published	Employers	Random	Social distance measure
S. Park	2009	529	6	The United States	Not published	Students	Convenience	Attitudinal measure
Perkins, Raines, Tschopp, and Warner	2009	404	1	The United States	Published	General public	RDD	Social distance measure
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.

Measures of attitudes included a variety of categorical and continuous variables that assessed public attitudes toward ex-offenders. Outcome measures assessing desired social 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 definitions 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 attitudes 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 ( $\bar{r}$ _w = −.12, 95% confidence interval [CI] = [−0.18, −0.04]) and those who reported no previous contact with a current or ex-offender $\bar{r}$ _w = −.18, 95% CI = [−0.22, −0.13]) reported more negative attitudes toward ex-offenders compared 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.

Table 2:

Weighted Mean Effect Sizes of Negative Attitudes Toward Ex-Offenders and Moderation Analyses for Sexual Offense History

Correlates	k	n	$\bar{r}$ _w(σ_ρ)	PVA (%)	95% CI_w	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 (conservative = 0)	6	4,792	−0.12 (0.06)	27.88	[−0.18, −0.08]	[−0.20, −0.05]	21.55 ***
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	−0.01^a	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
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 ( $\bar{r}$ _w ≥ .10). k = number of effect sizes; n = total number of participants; $\bar{r}$ _w = weighted mean effect size; σ_ρ = standard deviation of sampling error; PVA = proportion of variance accounted for by sampling error and attenuating artifacts; 95% CI_w = confidence interval of weighted mean effect size; 80% Cred = credibility interval of corrected standard deviation; Q_(k−1) = chi-square homogeneity test.

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

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

Table 3:

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

Corrected Public Characteristic Correlates	k	n_c	$\bar{r}$ _wc (σ_ρc)	PVA_c (%)	95% CI_wc	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	−0.01^a	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; n_c = corrected total number of participants; $\bar{r}$ _wc = corrected weighted mean effect size; σ_ρc = corrected standard deviation of sampling error; PVA_c = corrected proportion of variance accounted for by sampling error and attenuating artifacts; 95% CI_wc = 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.

Negative residual variance resulted in inability to calculate standard deviation of sampling error and credibility interval.

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

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 ( $\bar{r}$ _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).

Community Characteristics

A small weighted mean correlation was found between community size and public attitudes 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 ( $\bar{r}$ _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 PVA_se was less than the 75% benchmark (Arthur et al., 2001; Schmidt & Hunter, 2015) for all public and ex-offender correlates (Tables 2 and 3), providing 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 standard error accounted for over half of the variance (PVA_se = 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 compared 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.

Potentially Biasing Records

Though publication status, sampling methodology, and country did not emerge as statistically significant moderators, effect sizes tended to be smaller for records that were published 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 attitudes. (Full results not reported but available upon request.) In contrast, analyses of potential biases associated with year produced revealed significant moderation for public age and income correlates, with reduced variance in subsets and non-overlapping credibility intervals (Table 4). Before 2009, older respondents reported more negative attitudes toward ex-offenders compared with younger respondents ( $\bar{r}$ _wc = .15, 95% CI = [0.05, 0.24]). During and after 2009, the association is negligible ( $\bar{r}$ _wc = −.01, 95% CI = [−0.06, 0.03]). In addition, before 2009, members of the public with larger household incomes reported more negative attitudes toward ex-offenders than did those with smaller household incomes ( $\bar{r}$ _wc = .16, 95% CI = [0.03, 0.29]). During and after 2009, the association is negligible ( $\bar{r}$ _wc = .03, 95% CI = [0.00, 0.06]).

Table 4:

Moderation Analyses for Potentially Biasing Records: Year Produced

Correlates	k	n	$\bar{r}$ _w (σ_ρ)	PVA (%)	95% CI_w	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	−0.01^a	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 (majority = 0)	11	6,252	−0.09 (0.05)	38.55	[−0.13, −0.04]	[−0.15, −0.02]	28.59***
<2009	3	631	−0.11^a	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***
Education^b	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 (conservative = 0)	6	4,792	−0.12 (0.06)	27.88	[−0.18, −0.08]	[−0.20, −0.05]	21.55 ***
<2009	1	—	—	—	—	—	—
≥2009	5	4,574	−0.12 (0.04)	36.34	[−0.16, −0.07]	[−0.17, −0.06]	13.77**
Age^b	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 *
Income^b	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 contact = 0)	4	3,659	−0.18 (0.05)	31.45	[−0.22, −0.13]	[−0.24, −0.12]	12.73 **
<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 (small/rural = 0)	3	3,120	−0.04 (0.03)	54.95	[−0.08, −0.00]	[−0.08, −0.01]	5.47
<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 ( $\bar{r}$ _w ≥ .10) or significant moderation effects. k = number of effect sizes; n = total number of participants; $\bar{r}$ _w = weighted mean effect size; σ_ρ = standard deviation of sampling error; PVA = proportion of variance accounted for by sampling error and attenuating artifacts; 95% CI_w = confidence interval of weighted mean effect size; 80% Cred = credibility interval of corrected standard deviation; Q_(k−1) = chi-square homogeneity test.

Negative 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.

Discussion

Negative public attitudes toward ex-offenders are a known barrier to reentry and successful 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 affiliation/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 correlate 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 conservative or Republican report more negative attitudes toward ex-offenders. Findings are consistent 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 dispositional 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 produced the largest effect size across records. This finding is consistent with studies of attitudes toward sex offenders held by other groups, including criminal justice professionals, who report more negative attitudes toward sex offenders compared with ex-offenders with no history of sexual offending (see Weekes, Pelletier, & Beaudette, 1995). Moderation analyses 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 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 attitudes 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 identify 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 development 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 identified 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 identified 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 attitudes, efforts to improve public attitudes toward ex-offenders should not necessarily target or tailor material to certain socio-demographic groups. Rather, findings suggest that interventions 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 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 positive attitudes toward ex-offenders. Interventions that incorporate elements of contact may be more effective; however, the causal direction is unclear. Research with other stigmatized populations, such as adults with mental illnesses, suggests that contact-based interventions 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 attitudes toward ex-offenders should consider the type (e.g., face-to-face, vicarious) and quality (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 attitudes 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 assessing 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.

Conclusions and Future Directions

This study marks the first meta-analysis of the correlates and moderators of public attitudes toward ex-offenders. Findings bring to light the limited nature of the research on correlates 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 legislation 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.

Footnotes

Acknowledgements

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.

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 collection, analysis, and interpretation of data; in the writing of the report; or in the decision to submit the paper for publication.

Notes

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, implementation, 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.

References

*Albright

Denq

(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

Bennett

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 attitudes 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

(1999). Public attitudes toward the treatment of sex offenders. Legal and Criminological Psychology, 4, 239-252. doi:10.1348/135532599167879

Brown

(2009). Attitudes towards sexual offenders and their rehabilitation: A special case? In Wood

J. L.

Gannon

T. A.

(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

10.

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

11.

Clement

van Nieuwenhuizen

Kassam

Flach

Lazarus

de Castro

. . . Thronicroft

(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

12.

Cohen

(1992). A power primer. Psychological Bulletin, 112, 155-159. doi:10.1037/0033-2909.112.1.155

13.

*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

14.

Conti

Morrison

Pantaleo

(2013). All the wiser: Dialogic space, destigmatization, and teacher-activist recruitment. The Prison Journal, 93, 163-188. doi:10.1177/0032885512472654

15.

Corrigan

P. W.

Morris

S. B.

Michaels

P. J.

Rafacz

J. D.

Rüsch

(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

16.

*Dawson Edwards

B. C.

(2007). Determinants of college students’ opinions towards felon voting rights: An exploratory study (Unpublished doctoral dissertation). Virginia Commonwealth University, Richmond.

17.

Dovidio

J. F.

Glick

Rudman

L. A.

(2005). Introduction: Reflecting on the nature of prejudice: Fifty years after Allport. In Dovidio

J. F.

Glick

Rudman

L. A.

(Eds.), On the nature of prejudice: Fifty years after Allport (pp. 1-16). Malden, MA: Blackwell.

18.

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.

19.

Federal Prisoner Reentry Initiative of 2009, 42 U.S.C. § 17541.

20.

*Gibson

Roberson

Daniel

(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

21.

Gideon

Loveland

(2011). Public attitudes toward rehabilitation and reintegration. In Gideon

Sung

H.-E.

(Eds.), Rethinking corrections: Rehabilitation, reentry, and reintegration (pp. 19-36). Thousand Oaks, CA: SAGE.

22.

Goffman

(1963). Stigma: Notes on the management of spoiled identity. Englewood Cliffs, NJ: Prentice Hall.

23.

Graffam

Shinkfield

Lavelle

McPherson

(2004). Variables affecting successful reintegration as perceived by offenders and professionals. Journal of Offender Rehabilitation, 40, 147-171. doi:10.1300/J076v40n01_08

24.

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

25.

Hardcastle

Bartholomew

Graffam

(2011). Legislative and community support for offender reintegration in Victoria. Deakin Law Review, 16, 111-132.

26.

*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

27.

Hulsey

L. F.

(1991). Attitudes of employers with respect to hiring released prisoners (Unpublished doctoral dissertation). Texas A&M University, College Station.

28.

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

29.

Jacobs

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

30.

Just Leadership USA. (2016). About us. Retrieved from https://www.justleadershipusa.org/about-us/

31.

Kleban

Jeglic

(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

32.

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

33.

LeBel

T. P.

(2008). Perceptions of and responses to stigma. Sociology Compass, 2, 409-432. doi:10.1111/j.1751-9020.2007.00081.x

34.

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

35.

Legal Services for Prisoners With Children. (2015). All of Us or None. Retrieved from http://www.prisonerswithchildren.org/our-projects/allofus-or-none/

36.

*Levenson

J. S.

Brannon

Y. N.

Fortney

Baker

(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

37.

Leverentz

(2011). Neighborhood context of attitudes toward crime and reentry. Punishment & Society, 13, 64-92. doi:10.1177/1462474510385629

38.

Liberati

Altman

D. G.

Tetzlaff

Mulrow

Gøtzsche

P. C.

Ioannidis

J. P. A.

. . . Moher

(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

39.

Lipsey

M. W.

Wilson

D. B.

(2001). Practical meta-analysis. Thousand Oaks, CA: SAGE.

40.

*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.

41.

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

42.

*Mancini

Shields

R. T.

Mears

D. P.

Beaver

K. M.

(2010). Sex offender residence restriction laws: Parental perceptions and public policy. Journal of Criminal Justice, 38, 1022-1030. doi:10.1016/j.jcrimjus.2010.07.004

43.

*Manza

Brooks

Uggen

(2004). Public attitudes toward felon disenfranchisement in the United States. Public Opinion Quarterly, 68, 275-286. doi:10.1093/poq/nfh015

44.

*Martinez

M. S.

(2011). Employer perceptions about hiring released prisoners with a criminal history (Unpublished doctoral dissertation). Capella University, Minneapolis, MN.

45.

Maruschak

L. M.

Parks

(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

46.

Mason

Allam

Brannick

M. T.

(2007). How to meta-analyze coefficient-of-stability estimates: Some recommendations based on Monte Carlo studies. Educational and Psychological Measurement, 67, 765-783. doi:10.1177/0013164407301532

47.

Moher

Liberati

Tetzlaff

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

48.

Pager

(2003). The mark of a criminal record. American Journal of Sociology, 108, 937-975. doi:10.1086/374403

49.

*Pager

Quillian

(2005). Walking the talk? What employers say versus what they do. American Sociological Review, 70, 355-380. doi:10.1177/000312240507000301

50.

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-

51.

*Park

(2009). College students’ attitudes toward prisoners and prisoner reentry (Unpublished doctoral dissertation). Indiana University of Pennsylvania.

52.

*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

53.

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

54.

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

55.

*Piquero

A. R.

Piquero

N. L.

Gertz

Baker

Batton

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

56.

Pogorzelski

Wolff

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

57.

Pompa

(2013). One brick at a time: The power and possibility of dialogue across the prison wall. The Prison Journal, 93, 127-134.

58.

*Rogers

Hirst

Davies

(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

59.

Schmidt

F. L.

Hunter

J. E.

(2015). Methods of meta-analysis: Correcting error and bias in research findings (3rd ed.). Thousand Oaks, CA: SAGE.

60.

Schnittker

Bacak

(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

61.

Schulze

(2004). Meta-analysis: A comparison of approaches. Cambridge, MA: Hogrefe & Huber.

62.

Second Chance Act. (2007). Community Safety through Recidivism Prevention, 42 U.S.C. § 17501.

63.

Tewksbury

Lees

(2006). Perceptions of sex offender registration: Collateral consequences and community experiences. Sociological Spectrum, 26, 309-334. doi:10.1080/02732170500524246

64.

Uggen

(2000). Work as a turning point in the life course of criminals: A duration model of age, employment, and recidivism. American Sociological Review, 65, 529-546. doi:10.2307/2657381

65.

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

66.

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

67.

Viki

G. T.

Fullerton

Raggett

Tait

Wiltshire

(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

68.

Visher

C. A.

Debus

Yahner

(2008). Employment after prison: A longitudinal study of releases in three states. Washington, DC: Justice Policy Center, Urban Institute.

69.

Wakefield

Uggen

(2010). Incarceration and stratification. Annual Review of Sociology, 36, 387-406. doi:10.1146/annurev.soc.012809.102551

70.

Weekes

J. R.

Pelletier

Beaudette

(1995). Correctional officers: How do they perceive sex offenders? International Journal of Offender Therapy and Comparative Criminology, 39, 55-61. doi:10.1177/0306624X9503900107

71.

Willis

G. M.

Levenson

J. S.

Ward

(2010). Desistance and attitudes towards sex offenders: Facilitation or hindrance? Journal of Family Violence, 25, 545-556. doi:10.1007/s10896-010-9314-8

72.

*Willis

G. M.

Malinen

Johnston

(2013). Demographic differences in public attitudes towards sex offenders. Psychiatry, Psychology and Law, 20, 230-247. doi:10.1080/13218719.2012.658206

73.

Wilson

D. B.

(2001a). Effect size determination program [Excel spreadsheet calculator]. Retrieved from http://mason.gmu.edu/~dwilsonb/ma.html

74.

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