Comparing undergraduates' stigma toward peers with criminal records and mental health disorders.

Skip to main content An official website of the United States government Here's how you know Here's how you know Official websites use .gov A .gov website belongs to an official government organization in the United States. Secure .gov websites use HTTPS A lock ( Lock Locked padlock icon ) or https:// means you've safely connected to the .gov website. Share sensitive information only on official, secure websites. Search Log in Dashboard Publications Account settings Log out Search… Search NCBI Primary site navigation Search Logged in as: Dashboard Publications Account settings Log in Search PMC Full-Text Archive Search in PMC Journal List User Guide PERMALINK Copy As a library, NLM provides access to scientific literature. Inclusion in an NLM database does not imply endorsement of, or agreement with, the contents by NLM or the National Institutes of Health. Learn more: PMC Disclaimer | PMC Copyright Notice BMC Psychol . 2026 Mar 2;14:475. doi: 10.1186/s40359-026-04243-9 Search in PMC Search in PubMed View in NLM Catalog Add to search Comparing undergraduates’ stigma toward peers with criminal records and mental health disorders Kieran J Schendel Kieran J Schendel 1 Department of Psychology, SUNY Old Westbury, Old Westbury, NY 11568 USA Find articles by Kieran J Schendel 1 , Erik M Benau Erik M Benau 1 Department of Psychology, SUNY Old Westbury, Old Westbury, NY 11568 USA 2 Doctoral Program in Clinical Psychology, Long Island University, Post Campus, Brookville, NY 11468 USA Find articles by Erik M Benau 1, 2, ✉ Author information Article notes Copyright and License information 1 Department of Psychology, SUNY Old Westbury, Old Westbury, NY 11568 USA 2 Doctoral Program in Clinical Psychology, Long Island University, Post Campus, Brookville, NY 11468 USA ✉ Corresponding author. Received 2025 Sep 1; Accepted 2026 Feb 23; Collection date 2026. © The Author(s) 2026 Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/ . PMC Copyright notice PMCID: PMC13059198  PMID: 41772712 Abstract Background Stigma remains a significant barrier to educational attainment for individuals with criminal records or mental illness, yet little is known about how college students perceive peers with these stigmatized backgrounds. Methods Undergraduates ( N = 307) were randomly assigned to one of three conditions to assess attitudes toward peers with: (a) no disclosed background (control), (b) a mental health diagnosis, or (c) a criminal record. They completed a version of the Situational Attitudes Scale (SAS), wherein they read a series of short vignettes and rated their emotional reactions (e.g., anger, fear) to each scenario. Participants also responded to open-ended questions about societal and personal attitudes toward individuals in their assigned condition. Results SAS scores were statistically similar between the control and mental illness conditions, whereas the criminal background condition scored significantly lower than both. Sentiment analysis of open-ended responses revealed a nuanced pattern: responses to the criminal background condition were rated significantly less positive than the other two, whereas the mental illness condition was also rated lower than the control. Respondents in the Control and Mental Illness conditions rated their own treatment of their target groups more positively than society (possibly a false uniqueness effect), whereas sentiment was similar across prompts for the Criminal condition. Conclusion Stigma toward students with criminal backgrounds may differ in both form and intensity than toward other groups. These results underscore the value of combining direct and indirect measures of stigma to capture subtle but meaningful variations in how bias is expressed. Given the critical role of educational attainment as both a social determinant of health and a protective factor against recidivism, addressing campus climate is essential—not only to support students with stigmatized backgrounds, but also to advance equity and well-being within academic institutions and the communities they serve. Keywords: Stigma, Situational Attribution, Sentiment Analysis, Higher Education, Peer attitudes Educational attainment is a well-established social determinant of health [ 1 ]. In particular, higher education fosters economic mobility and privilege by providing opportunities to acquire soft skills (e.g., critical thinking), technical training, social bonds and networking, and long-term planning [ 1 , 2 ]. These benefits often surpass other social determinants—including race and baseline socioeconomic status—in predicting psychosocial and health outcomes [ 3 , 4 ]. Accordingly, higher education is one of the most reliable protective factors against criminal justice involvement [ 5 ]. Those who attain higher education after conviction show similar socioeconomic benefits and, in turn, reduce their likelihood of repeated justice system involvement and recidivism [ 5 – 8 ]. These educational gains support desistance not only through increased human capital but through strengthened social ties and emerging prosocial identities [ 6 – 8 ]. In addition to the impact on the wellbeing of the individual and the families and communities they reside [ 9 ], recidivism is a significant economic burden with billions of dollars incurred annually through policing, court proceedings, re-incarceration, and reduced productivity [ 10 , 11 ]. Therefore, access to higher education represents a major opportunity not just to address recidivism, but to promote wellbeing for the individual and the communities in which they reside. As access expands, stigma within higher education environments may inhibit these same desistance-supportive mechanisms, heightening the importance of understanding campus attitudes [ 8 , 12 ]. There are increasing opportunities for access to higher education for those with criminal histories. As of 2021, there were 396 higher education programs operating within U.S. prisons, supported by 572 academic institutions, offering 104 associate’s and 38 bachelor’s degree pathways [ 13 ]. It is estimated that, to date, tens of thousands of incarcerated individuals have completed such a program [ 14 ]. These programs are set to grow with the overturning of a ban on Pell grants—federal need-based aid for low-income undergraduates that do not require repayment—that the U.S. Congress approved in 2021 [ 15 ]. The benefits of these programs provide both intrapersonal and social empowerment [ 16 ]. Outside of prisons, however, access to higher education for those with criminal histories remains difficult to attain. There are numerous structural and personal barriers for individuals with criminal histories to complete undergraduate education once released. Structurally, as of 2019, almost 81% of private institutions and 55% of public institutions collected criminal background information, yet fewer than half implemented policies specifying how such information should be used in admissions decisions [ 17 ]. Indeed, interviews with admissions officers suggest that criminal backgrounds are handled subjectively and often arbitrarily [ 18 ]. Experimental research with samples of admissions officers suggests that a criminal record reduces likelihood of admissions, which is compounded by implications of minoritized race [ 19 ]. These decisions are made at the administrative level, yet they may reflect or be shaped by broader campus sentiment. Students generally express openness to admitting individuals with criminal histories to their institutions, though this openness decreases when a criminal background intersects with other marginalized identities, such as minoritized race [ 20 , 21 ]. Notably, there is little or no evidence that blanket exclusion of individuals with criminal records improves campus safety [ 22 ]. In addition to these structural barriers to entry, personal barriers such as stigma—a negative attitude toward a devalued group that becomes the basis for labeling, stereotyping, separation, and discrimination [ 23 ]—can negatively affect the retention of students with criminal backgrounds and, ultimately, their degree attainment [ 24 ]. Unlike many forms of bias, stigma toward the formerly incarcerated is often presumed rather than debated. Criminological frameworks highlight that criminal record stigma operates through linked processes of marking , stereotyping , and status loss (i.e., the “mark of a criminal record”), whereby individuals are framed as dangerous, morally deficient, or untrustworthy [ 25 – 29 ]. Consistent with these perceptions, students with criminal backgrounds report perceived stigma, social isolation, and discrimination within academic and professional settings, all of which hinder student retention and academic success [ 25 , 30 , 31 ]. Such experiences can arise even when a student’s record is not widely known, as anticipated judgments and stigma can generate negative social and academic experiences comparable to those produced by actual discrimination [ 30 , 31 ]. In other words, campus and social environments can either support or undermine the reintegration, retention, and flourishing of students with criminal backgrounds. Surprisingly little is known about what views undergraduates and other campus stakeholders hold about current students or classmates with criminal histories. Qualitative data suggest that students would be appreciative of the opportunity to learn from students with experiences of incarceration [ 32 ]. Quantitatively, however, attitudes toward students with such histories have been rated significantly—and substantially—lower than those without by peers [ 33 , 34 ], faculty [ 35 ], and administrators [ 27 ]. Though these studies lay important groundwork, it remains unclear whether stigma toward formerly incarcerated individuals is unique to that population or reflects broader patterns of bias toward outgroup members more generally. To this end, the current study included a comparison group depicting individuals with mental illness. Despite some belief that views toward mental illness have become less negative over time, stigma remains prevalent and, in some cases, has intensified [ 36 ]. People with mental illness experience stigma in ways that often parallel those with criminal records: both groups are stereotyped as dangerous, unstable, and/or incompetent [ 37 ]. For individuals with mental illness specifically, these perceptions have been linked to increased social distance, reduced help-seeking, and higher dropout risk among students [ 38 , 39 ]. Together, these groups seem intuitive to compare in order to determine whether stigma toward those with criminal backgrounds is unique or a result of general outgroup biases. Crucially, our intent is not to establish a hierarchy of oppression, as doing so can reinforce the very systems of exclusion we seek to disrupt [ 40 ]. Instead, we believe that better understanding these differences can inform more precise and equitable strategies for inclusion. In particular, mental illness stigma has (rightly) received significant attention in higher education, with well-established intervention frameworks [ 41 ]. If stigma toward justice-involved students operates in similar ways, this comparison may help guide adaptation and tailoring of those efforts; if it operates differently, it may indicate the need for novel, targeted interventions. Further, we recognize that experiences of criminalization and mental illness often intersect, particularly in marginalized and minoritized communities [ 42 , 43 ]. 1 This overlap does not undermine the validity of the comparison; rather, it underscores its importance, as each label can contribute uniquely—or additively—to how students are perceived and stigmatized on campus. Present study The present study had three aims. First, we sought to replicate and extend prior findings on personal stigma—negative attitudes—held by students toward peers with criminal histories. The second aim was to compare these attitudes to those directed at another stigmatized group: students with mental illness. Third, we wanted to determine whether self-reported attitudes were consistent with the language used to describe attitudes toward these groups (i.e., natural language sentiment). To address these goals, we recruited a crowdsourced sample of current undergraduates who were randomly assigned to evaluate hypothetical students described as having a criminal background, mental illness, or no stated history of either. Participants completed both quantitative ratings and open-ended written responses. Sentiment analysis was applied to the writing samples to assess implicit attitudes. We hypothesized that students with criminal backgrounds would be evaluated most negatively, followed by those with mental illness, and that this pattern would be similar in both self-report and sentiment analysis. Methods Participants Participants were recruited in June 2024 from Prolific, a crowdsourcing platform that specializes in academic research. Only U.S. residents whose screening measures indicated they were current undergraduates were able to access the study. A total of 350 participants selected the study and consented to begin. Forty-three (12%) participants were omitted from analyses if they did not confirm they were undergraduate, incomplete data, did not confirm English was their primary language, or were missing more than one of four attention checks, leaving a total of 307 included participants. An a priori power analysis was conducted using G*Power 3.1 to plan the sample size for the original analytic approach, which treated the design as a one-way ANCOVA with three groups and a set of 14 covariates. Assuming a medium effect size (e.g., f = 0.25), α = 0.05, and 1– β = 0.80 for the omnibus F -test, the target sample size was approximately 90 participants per cell. Our final sample exceeded this target, yielding an achieved power of 1– β = 0.84 for detecting effects of this magnitude. Although our primary analyses now employ linear mixed-effects models without covariates, this a priori analysis provides a reasonable benchmark for the effect sizes the study was designed to detect. Participant demographics are presented in Table 1 and experiential characteristics are presented in Table 2 . With the exception of academic year, none of the individual-difference variables significantly differed across conditions. The age distribution was right-skewed, with a small number of older respondents elevating the mean ( M = 26.6, SD = 8.6); the median age of 23 years better reflects the sample. This range is broadly consistent with contemporary U.S. undergraduate populations, which increasingly include students over 22 years of age [ 44 ]. Table 1. Demographic and student status information for the sample Variable Control 88 (29%) Mental Illness 113 (37%) Criminal 106 (35%) Total 307 Demographic Information Age [Mean (SD)] 28.0 (9.8) 26.3 (8.3) 25.8 (7.6) 26.6 (8.6) Gender Man a 35 (39.8) 37 (32.7) 36 (34.0) 108 (35.2) Woman b 47 (53.4) 65 (57.5) 64 (60.4) 176 (57.3) Non-binary 3 (3.4) 9 (8.0) 5 (4.7) 17 (5.5) Other/Missing 3 (3.4) 2 (1.8) 1 (0.9) 6 (2.0) Race White 43 (48.9) 56 (49.6) 52 (49.1) 151 (49.2) Asian 7 (8.0) 19 (16.8) 19 (17.9) 45 (14.7) Black 15 (17.0) 18 (15.9) 15 (14.2) 48 (15.6) Hispanic 10 (11.4) 6 (5.3) 7 (6.6) 23 (7.5) Middle Eastern 0 (0.0) 1 (0.9) 0 (0.0) 1 (0.3) AI/AN 1 (1.1) 0 (0.0) 0 (0.0) 1 (0.3) Multi Racial 11 (12.5) 13 (11.5) 11 (10.4) 35 (11.4) Refuse/Missing 1 (1.1) 0 (0.0) 2 (1.9) 3 (1.0) Student Status and School Information Full Time 64 (72.7) 81 (71.7) 88 (83.0) 233 (75.9) Public or Private Public 66 (75.0) 86 (76.1) 77 (72.6) 229 (74.6) Private 20 (22.7) 26 (23.0) 28 (26.4) 74 (24.1) Unsure 2 (2.3) 1 (0.9) 1 (0.9) 4 (1.3) Type of Institution Two Year 20 (22.7) 21 (18.6) 21 (19.8) 62 (20.2) Small Four Year c 4 (4.5) 5 (4.4) 4 (3.8) 13 (4.2) Medium Four Year d 26 (29.5) 33 (29.2) 25 (23.6) 84 (27.4) Major University e 38 (43.2) 52 (46.0) 54 (50.9) 144 (46.9) Professional/Trade 0 (0.0) 2 (1.8) 2 (1.9) 4 (1.3) Year f First 7 (8.0) 16 (14.2) 7 (6.6) 30 (9.8) Second 16 (18.2) 30 (26.5) 11 (10.4) 57 (18.6) Third 23 (26.1) 17 (15.0) 27 (25.5) 67 (21.8) Fourth 38 (43.2) 46 (40.7) 60 (56.6) 144 (46.9) Other/Unsure 4 (4.5) 4 (3.6) 1 (0.9) 9 (3.0) Open in a new tab Values are N (%) unless noted; percentages may not sum to 100% due to rounding; AI/AN = American Indian/Alaska Native a Includes seven transgender men b Includes three transgender women c “Four-year school with no graduate programs (sometimes referred to as a liberal arts college or similar)” d “Four-year school with some graduate programs” e “Major university with many graduate and/or health science degree programs” f Significant discrepancy (p = .008), see text Table 2. Experiential and Psychosocial Descriptors Variable Range Control N = 88 (29%) Mental Illness N = 113 (37%) Criminal N = 106 (35%) Total N = 307 Experiential Variables [Mdn (IQR) or N (%)] Psych. Dx. (N [%] Yes) 34 (39.1) 36 (32.1) 43 (41.0) 113 (37.2) Been incarcerated (N [%] Yes) 4 (4.5) 10 (8.8) 2 (1.9) 16 (5.2) Known Incarcerated 1–5 a 1.0 (1.0 to 2.0) 1.0 (1.0 to 2.0) 1.0 (1.0 to 2.0) 1.0 (1.0 to 2.0) Neighborhood Crime 1–5 b 3.0 (2.0 to 4.0) 3.0 (2.0 to 3.0) 3.0 (2.0 to 3.0) 3.0 (2.0 to 3.0) Victimization 1–4 c 2.0 (0.0 to 3.0) 2.0 (0.0 to 3.0) 1.0 (0.0 to 2.0) 1.0 (0.0 to 3.0) Psychosocial Variables [M (SD)] PQ+ 3–15 11.6 (1.6) 11.4 (1.8) 11.2 (2.1) 11.4 (1.8) NQ– 3–15 6.5 (2.4) 6.9 (2.5) 6.5 (2.4) 6.7 (2.4) Light Triad 12–60 35.4 (4.4) 35.0 (4.6) 35.4 (4.3) 35.3 (4.5) SD4 28–140 70.1 (16.1) 71.9 (14.3) 67.5 (13.3) 69.9 (14.6) VSRWA 5–45 24.3 (9.4) 23.0 (9.6) 23.8 (9.3) 23.7 (9.4) SINS 1–7 5.9 (1.5) 6.1 (1.0) 5.9 (1.2) 6.0 (1.2) SITES 1–5 4.3 (0.8) 4.2 (0.7) 4.3 (0.7) 4.3 (0.7) SISES 1–7 4.0 (1.8) 3.8 (1.7) 3.8 (1.7) 3.9 (1.7) SRMH 1–6 3.5 (1.2) 3.6 (1.2) 3.4 (1.1) 3.5 (1.2) Open in a new tab Psych. Dx = Percentage reporting a psychiatric diagnosis; Been incarcerated = specified as having been in jail or prison for more than one month; Known incarcerated = How many close known others who were incarcerated; Victimization = Self or close other being victims of violent or non-violent crimes; PQ+/NQ– = Positive Quality enhancement and Negative Quality minimization scales (respectively) of the KSE-G (see text); SD4 = Short Dark Tetrad; VSRWA = Very Short Right Wing Authoritarianism; SINS = Single Item Narcissism Scale; SITES = Single Item Trait Empathy Scale; SISES = Single Item Self-Esteem Scale; SRMH = Single-item Self-Reported Mental Health; Each variable was statistically comparable across groups; percentages may not sum to 100 due to rounding; though the SD4 was developed to yield four distinct, 7-item subscales (Machiavellianism, narcissism, psychopathy, and sadism), for parsimony, we used the total score in the present analyses to reflect global “dark” personality traits; the Light Triad is comprised of three, 4-item subscales (Kantianism, humanism, and faith in humanity) as a foil to the Dark Triad but we included a sum score of the instrument as the original publication allows for either use of subscales or a single-scale measure; there were no significant differences amongst groups on any of these variables a 1 (none) to 5 (more than five) b 1 (rarely) to 5 (all the time) c Tally of “yes” to four items asking about self or close others being the victim of violent or non-violent crime Measures Stigma We included a modified version of the Situational Attitude Scale (SAS). The SAS was originally developed by Sedlacek and Brooks Jr [ 45 ]. We adapted a version that was modified by McTier and colleagues to measure stigma pertaining to students with criminal backgrounds [ 35 ]. The SAS is based on attribution theory, which posits that people explain others’ behavior by inferring underlying causes—often using minimal or imperfect information [ 46 ]. This framework prompts respondents to project their attitudes into concrete, high-salience scenarios rather than making abstract judgments and, in the absence of contextual detail, these projections likely occur through emotional reasoning and other heuristics [ 45 ]. Our version maintained the structure of the SAS—nine brief vignettes describing scenarios involving a fellow student—while manipulating the background of the student across three between-subjects conditions: control (no stated background), criminal (specified felony conviction), or mental illness (specified highly stigmatized diagnosis). For the control and criminal history conditions, our modifications were based closely on prior work [ 35 ]. These vignettes featured four violent, four nonviolent, and one general description of criminal backgrounds. For the mental health conditions, we replaced criminal histories with specific diagnoses that were selected for having high ratings of stigma in previous work [ 47 ]. For example: “[Control: A classmate; Criminal: A classmate who had previously notified you about an assault conviction; Mental Illness: A classmate who had previously notified you about having antisocial personality disorder] emails you to request a meeting outside of the university.” In this example, the bracketed information preceded the situation, and in the brackets, the prime differed based on assigned condition; participants assigned to the control condition only saw the control prompts, etc. Participants rated five emotional dimensions using five-point semantic differential items adapted from Ott and McTier [ 35 ]: Anger: Very Angry (1) to Very Happy (5); Surprise: Very Surprised (1) to Very Calm (5); Safety: Very Safe (1) to Very Unsafe (5); Disappointment: Very Disappointed (1) to Very Pleased (5); Fear: Very Afraid (1) to Not at All Afraid (5). To ensure that higher scores consistently reflected more positive emotional responses, items anchored by safe were reverse scored. Each vignette text is presented in Table 3 . In all three versions of the study, participants were encouraged to look up the definition of a term if they did not know what it meant. Table 3. Each vignette in the SAS Your university announces a new program for [students; students who have been convicted of crimes; students who have a major mental illness]. The funding for this program will come from existing programs. You see a [classmate; classmate who has been convicted of drug possession; classmate who told you about having schizophrenia] driving a luxury car. A [classmate; classmate who had previously informed you about a conviction for menacing (threats) and intimidation; classmate who had previously informed you about having Delusional Disorder] earns an A in your class. A [classmate; classmate who told you about a prior conviction for forgery; classmate who told you about having bipolar disorder] was admitted with entrance test scores much lower than those of the general student population. A [classmate; classmate who had previously disclosed to you a rape conviction; who had previously discussed having dissociative identity disorder] has recently been arrested for sexual assault. A [classmate; classmate who had previously informed you of a burglary conviction; classmate who had previously informed you of having Gambling Disorder] cheating on an exam. A [classmate; classmate who had previously notified you about an assault conviction; classmate who had previously notified you about having antisocial personality disorder] emails you to request a meeting outside of the university. The academic accomplishments of a [classmate; classmate who had previously disclosed to you a conviction for murder; classmate who had previously told you about having post-traumatic stress disorder] in one of your classes are featured in the campus newspaper. A [classmate; classmate who had previously told you about a conviction for manslaughter; classmate who had previously told you about having a borderline personality disorder] and is majoring in your field has received a full scholarship to attend your university. Open in a new tab Each vignette version is shown with Control, Criminal, and Mental Illness primes but participants only saw one version of each within their own assigned condition Open ended questions To attain sentiment analysis, participants were asked three questions developed by the authors. Questions were modified in a parallel manner depending on the condition: (a) How do you view society’s treatment of [Control: college students; Criminal: college students with a criminal record; Mental illness: college students who have a mental illness]? Please elaborate on any personal experiences or observations that have influenced your perspective.; (b) What are your thoughts and feelings about [Control: your classmates; Criminal: classmates if you learned they had been incarcerated (for more than a month); Mental illness: classmates if you learned they had a major mental illness]?; (c) How do, or how would, you treat [Control: your classmates; Criminal: classmates if you learned they had been incarcerated (for more than a month); Mental illness: a classmate if you learned they have a major mental illness]? Descriptive statistics Demographic and experiential descriptive variables included: (a) age; (b) year in school; (c) full- versus part-time student status; (d) size of institution (two-year college; four-year college with no graduate programs; four-year college with some graduate programs; four-year college with extensive graduate or health science programs; professional/trade school) 2 ; (e) public versus private institution; (f) personal history of incarceration, defined as “the experience of going to jail or prison for more than a month” (yes/no); (g) number of known individuals with an incarceration history (none, one, two or three, four or five, more than five, prefer not to respond); (h) perceived neighborhood crime exposure on a scale from 1 ( never or almost never ) to 5 ( all the time ); (i) a count of all “yes” responses across four yes/no items assessing personal and known other experiences of violent and non-violent crime; and (j) self-reported presence of a serious mental illness diagnosis (ten major diagnoses were presented as a checklist; any reported diagnoses were recoded as “yes” and all omissions were coded as “no”). We also collected: (a) the six-item English-language version of the Kurzskala Soziale Erwünschtheit–Gamma scale (KSE-G) for social desirability [ 48 ], which yields a three-item Positive Quality Enhancement (PQ+; ω = 0.58) and three-item Negative Quality Minimization (NQ−; ω = 0.56) subscale; (b) the 28-item Short Dark Tetrad (SD4; ω = 0.89; [ 49 ]; (c) the 12-item Light Triad Scale ( ω = 0.77; [ 50 ]; (d) the six-item Very Short Right-Wing Authoritarianism Scale (VSRWA; ω = 0.78; [ 51 ]; (e) the Single Item Narcissism Scale [SINS; [ 52 ]]; (f) the Single Item Self-Esteem Scale [SISE; [ 53 ]]; (g) the Single Item Trait Empathy Scale [SITES; [ 54 ]]; (h) a single item measure of self-rated mental health [SRMH; [ 55 ]]; (i) political philosophy (developed for this study) on a sliding scale from 0 (very conservative) to 100 (very liberal). The above measures are reported for descriptive context only and were not included in the primary analyses discussed below. Procedure After selecting the study and completing informed consent, participants answered the experiential items described above. Consent forms indicated the possibility that they would be reading about students with criminal histories and providing their opinion of these groups. Participants then completed the self-report measures in a randomized order. Next, participants were randomly assigned to complete the Control ( n = 87), Mental Illness ( n = 113), or Criminal background ( n = 106) version of the SAS and the corresponding open-ended questions. Binomial tests confirmed comparable group sizes ( ps > 0.10). Finally, they completed the demographic questions described above and three quality control items (not presented). Upon completion, participants were debriefed, which included information about available mental health resources, and compensated in accordance with Prolific’s reimbursement policies at the time of data collection, up to $3 (USD). Reimbursement data were not released to researchers. Median completion time was 14.3 min ( M = 17.6, SD = 10.4, range: 5.6–94.0); Kruskal-Wallis tests confirmed the three groups did not significantly differ in completion time, χ 2 (2) = 5.17, p = .075, ϵ 2 = 0.02. All procedures were approved by the Institutional Review Board of the State University of New York at Old Westbury. Data analysis Situational Attitude Scale (SAS) SAS affect ratings were analyzed using linear mixed-effects models (LMMs), allowing us to retain all 45 observations per participant (five affect ratings × nine vignettes) and to model the nested item structure directly. Condition (Criminal, Mental Illness, Control), Affect Dimension, and their interaction were entered as fixed effects; random intercepts for participants and vignettes accounted for stable person-level variability and systematic differences across vignette content, with by-vignette random slopes for Condition included when models converged. Safe–Unsafe was reverse scored so that higher values corresponded to more positive affect for all measures. Open ended questions sentiment analysis Text responses to the three open-ended questions (see Measures section for full wording) were analyzed using Linguistic Inquiry and Word Count 22.0 (LIWC; Pennebaker Conglomerates, Austin, TX). LIWC is a widely validated linguistic analysis tool with strong reliability and construct validity across emotional, cognitive, and social language domains [ 56 , 57 ]. It has been used extensively in stigma research and the analysis of brief, naturalistic text [ 58 , 59 ]. We focused on three LIWC dimensions relevant to evaluative tone: Tone , Affect , and Social . Tone reflects the relative balance of positive to negative emotion words, with higher values indicating more positive affective valence. Affect indexes emotional expressiveness via the overall frequency of emotion-related words (positive or negative). Social captures interpersonal and relational language (e.g., references to affiliation, helping, or group membership). Because the three indices were all positively and significantly correlated ( r s ≈ 0.17–0.31; p s < 0.001), and because higher scores on each dimension indicate more positive and/or affiliative language, we standardized each score and averaged the resulting z-scores to create a composite Sentiment score. This composite is intended to capture overall evaluative valence, with higher scores reflecting more positive sentiment. The composite served as the dependent variable in the LMM described below. Following the structure of the SAS model, the LMM included Condition (Criminal, Mental Illness, Control), Prompt (Society, Thoughts, Treatment; described above), and their interaction as fixed effects. Random intercepts for participants accounted for stable person-level differences, and random intercepts for prompts accounted for systematic variability across prompts; by-prompt random slopes for Condition were included when models converged. Both SAS and Sentiment models were estimated using restricted maximum likelihood (REML) with Satterthwaite degrees of freedom, and significance testing relied on Type-III Wald tests. We retained the standard α = 0.05 threshold, as the mixed-effects framework accounts for within-person dependence and helps mitigate Type I error inflation. Covariate adjustment was not implemented because (a) the three vignette conditions did not differ on participant-level descriptors and (b) the LMMs accounted for stable individual variability through random intercepts. All effects are reported in raw score units, corresponding to the original response scale (higher scores = more positive affect or sentiment). Estimated marginal means and Δ M values therefore reflect unstandardized differences between conditions. Effect sizes for omnibus main effects and interactions (η p ²) were calculated from the corresponding F statistics and degrees of freedom. Effect sizes for pairwise comparisons (Cohen’s d ) were calculated for using estimated marginal means and the model residual standard deviation. Results Sample characteristics As shown in Table 1 , the sample was mostly White, women, full time students, who were seniors and attending a major public university. Top three majors were: computer science (12%), business and management (11%), and psychology (11%). A plurality identified as Democrats (48%) or Independent (33%). The groups significantly varied by year in school, χ ²(10) = 23.67, p = .008, V = 0.20, likely due to overrepresentation of second-year students in the mental illness condition and third- and fourth-year students in the criminal background condition. The three groups were statistically comparable on all remaining demographic and self-report measures ( p s > 0.06). SAS Random-effects estimates indicated that approximately 7.9% of the variance in affective responding was attributable to stable between-person differences (ICC = 0.079), with the remaining variance reflecting within-person differences across vignettes and affective dimensions. There was a significant main effect of Condition, F (2, 304) = 73.70, p < .001, η p 2 = 0.326, with a large effect size. Pairwise comparisons showed that classmates with a criminal record were evaluated substantially more negatively than both control classmates, Δ M = 0.58, p < .001, d = 0.49, and classmates with mental illness, Δ M = 0.53, p < .001, d = 0.45. Control and mental-illness conditions did not significantly differ, Δ M = 0.05, p = .40, d = 0.04. There was also a significant main effect of Affect, F (4, 13,496) = 69.69, p < .001, η p ² = 0.020, with a small but meaningful effect size. Post-hoc comparisons indicated meaningful variability across dimensions. Relative to Angry–Happy, participants were more surprised (Surprised–Unsurprised: Δ M = 0.31, p < .001, d = 0.26), less disappointed (Disappointed–Pleased: Δ M = 0.11, p < .001, d = 0.09), less afraid (Afraid–Not Afraid: Δ M = 0.19, p < .001, d = 0.16), and showed no reliable difference on Safe–Unsafe (Δ M = 0.03, p = .29, d = 0.03). Taken together, participants tended to report feeling modest disappointment, low fear, and moderate surprise across vignettes. Finally, there was a significant Condition × Affect interaction, F (8, 13,496) = 4.02, p < .001, η p ² = 0.020, with a small effect size. This result indicates that the magnitude of condition differences varied slightly, but significantly across affective dimensions, and the direction remained consistent. Across all five dimensions, classmates with a criminal record were rated more negatively than both control and mental-illness classmates, Δ M s ≈ 0.24–0.75, all p s < 0.001, corresponding to small-to-moderate effect sizes, d ≈ 0.20–0.63. By contrast, control and mental-illness conditions showed no reliable differences on any dimension (all p s ≥ 0.06). The interaction therefore reflected degree, not direction, of evaluative separation. The largest condition differences emerged for Safe–Unsafe, where the Criminal condition was rated substantially less safe than both Control, Δ M = 0.73, d = 0.62, and Mental Illness, Δ M = 0.62; d = 0.52, all p s < 0.001. In terms of fear (Afraid–Not Afraid), participants remained relatively low in fear overall yet were still more likely to report fear in response to the Criminal condition than to Control, Δ M = 0.62, d = 0.52, or Mental Illness, Δ M = 0.69, d = 0.58; all p s < .001. On Angry–Happy and Disappointed–Pleased, differences were slightly smaller but remained robust, Δ M s ≈ 0.48–0.65, d s ≈ 0.40–0.55, p s < .001. Surprise also showed a consistent pattern: vignettes involving a criminal record elicited greater surprise than both Control and Mental Illness, though differences were moderate, Δ M s ≈ 0.35–0.42, d s ≈ 0.29–0.35 p s < .001. See Fig. 1 for a graphical depiction of these effects. Fig. 1. Open in a new tab Estimated marginal means for the Condition × Affect interaction on the SAS. Higher values indicate more positive affect ratings. Error bars represent ± 1 SEM Sentiment The random intercept for participant indicated that about 24% of variance in sentiment scores was between individuals (ICC = 0.24). There was a significant main effect of Condition, F (2, 307) = 45.28, p < .001, η p 2 = 0.228, such that sentiment was significantly more positive in the Control condition than in the Mental Illness, Δ M = 0.33, p < .001, d = 0.58, and Criminal history conditions, Δ M = 0.62, p < .001, d = 1.09. The Mental Illness and Criminal also significantly differed, Δ M = 0.29, p < .001, d = 0.52. There was also a significant main effect of Prompt, F (2, 614) = 22.95, p < .001, η p 2 = 0.070, such that Thoughts and Society did not significantly differ, Δ M = 0.01, p = 1.0, d = 0.02, but (personal) Treatment prompts elicited significantly more positive sentiment than both Society, Δ M = 0.28, p < .001, d = 0.48, and Thoughts, Δ M = 0.26, p < .001, d = 0.46. These findings were qualified by a significant Condition × Prompt interaction, F (4, 614) = 7.00, p < .001, η p ² = 0.044. Post-hoc tests indicated that, in the Control condition, responses were substantially more positive when describing how participants personally treat their classmates (Treatment) compared with how society treats them (Society), Δ M = 0.57, p < .001, d = 1.00. This pattern was also observed in the Mental Illness condition, although to a smaller degree, Δ M = 0.25, p = .036, d = 0.44. In contrast, the Criminal history condition did not show this “behavioral uplift,” Δ M < 0.01, p = 1.00, d < 0.01. A similar pattern emerged for Treatment relative to Thoughts in the Control condition, Δ M = 0.51, p < .001, d = 0.89, which was smaller but still significant in the Mental Illness condition, Δ M = 0.17, p = .036, d = 0.30 and was not significant in the Criminal history condition, Δ M = 0.11, p = 1.00, d = 0.19. Across conditions, Society and Thoughts prompts did not differ from one another, p s = 1.00, ds ≈ 0.11–0.19. Taken together, across all three conditions, the sentiment conveyed when participants described their personal thoughts about classmates resembled the sentiment expressed when describing how society treats them. However, when describing their own behavior, participants reported treating classmates with mental illness and those with no stated condition more positively than society. This “behavioral uplift” was absent for classmates with criminal histories. Figure 2 presents the results of this interaction as a bar graph. Table 4 presents sample statements that were coded high and low in sentiment across the three conditions. Fig. 2. Open in a new tab Estimated marginal means of the sentiment composite (average z -scores across the Society , personal Thoughts , and personal Treatment prompts) by vignette condition. Higher values indicate more positive sentiment. Control and Mental Illness conditions showed a behavioral uplift, such that prompts about one’s personal treatment of the target elicited more positive sentiment than prompts about societal perceptions or personal thoughts. This uplift was absent in the Criminal condition. Error bars represent ± 1 SEM Table 4. Verbatim sample statements from high and low sentiment ratings (positive vs. negative) in each condition Group Rating Society Treatment Personal Thoughts Personal Treatment Control High i feel like we are mostly treated good My classmates are kind and empathic individuals. I’d treat them with care and respect. Low theyre not horrible towards them but also expect a lot they come from all walks of life and diverse. I treat them as my equals. Mental Illness High They tend to dismiss it. They also treat it like it’s a phase. I am always pleased to hear anyone has achieved success especially with a mental illness. I’d show compassion and support. I’d listen and be there for them. Treat them with kindness and respect. Encourage open conversations. Promote a stigma free environment. Low I don’t feel that society addresses the mental health of students as a priority. I would view them the same way as I viewed them before. Having a major mental illness does not change my opinion. No different than I would treat anyone else. Criminal High I think it is fair. I don’t think certain criminal records should be allowed on campus. I feel like everyone makes mistakes and hopefully a lesson was learned. I would treat them the same as any other classmate. Low I believe they are treated poorly by the general population. I really don’t care as long as I don’t feel like I am being targeted. i’d exercise caution, but be discreet about it Open in a new tab These statements are presented in full and are verbatim; each was in the five highest or lowest rated statements relative to the distribution for each condition; each statement is from a different respondent Discussion The central finding of the present study is that current undergraduates expressed markedly more negative attitudes toward classmates with criminal histories compared to those with no stated background and those with disclosed psychiatric diagnoses. We hypothesized that both self-report ratings and open-ended responses would show the most positive evaluations in the control condition (i.e., classmates with no stated mental illness or criminal background), followed by those with mental illness, and then those with criminal records. This hypothesis received mixed support. Counter to expectations, self-report (SAS) ratings indicated similarly positive regard for the Control and Mental Illness groups, though the Criminal History group was evaluated substantially more negatively. The sentiment analyses were consistent with hypotheses: participants used increasingly positive, affiliative, and emotionally warm language when describing the Control group, followed by the Mental Illness group, and then the Criminal history group. Moreover, the Control group described their behavior toward peers without stigmatized backgrounds substantially more positive than their view of society’s treatment; this effect was also present, but smaller for the Mental Illness condition, and absent for the Criminal history condition. These discrepant but complementary findings point to a more nuanced pattern of bias than has been previously presented. The present results complement and extend previous findings that faculty [ 35 ], administrators [ 27 ], and fellow undergraduates [ 21 , 33 , 34 ] hold more negative views toward students with criminal histories compared to those with no stated history. Yet this is only the second study—following Overton and colleagues [ 34 ]—to ask students about hypothetical current classmates as opposed to general thoughts or admissions policies. Overton and colleagues [ 34 ] observed negative attitudes toward formerly incarcerated students in a relatively neutral context—trustworthiness in student government—and the present study additionally identified this pattern across both affectively evocative vignettes and sentiment analysis of neutral open-ended prompts. This consistency across methods, samples, and evaluative contexts provides converging evidence that campus stakeholders hold substantially negative views toward students with criminal histories. Heightened social distance, elevated fear, and other negative SAS reactions—combined with neutral or negative open-ended responses, especially the absence of more favorable self-reported treatment relative to societal treatment—converge with qualitative accounts in which formerly incarcerated students describe being perceived as dangerous, incompetent, and/or untrustworthy [ 25 , 26 , 30 , 31 ]. The present study is also the first to demonstrate that, compared to views of other stigmatized groups, attitudes toward formerly incarcerated students are profound and pervasive. By experimentally juxtaposing reactions to felony convictions and severe mental-health conditions within the same evaluative context, this study provides preliminary comparative evidence that stigma toward formerly incarcerated students does not simply reflect generalized outgroup bias or reactions to marginalization, but instead reflects uniquely severe negative evaluations relative to other groups known to experience significant stigma in academic contexts [ 37 – 39 ]. Demonstrating this relative magnitude clarifies the distinct burden associated with a criminal record in higher-education settings and identifies a profile of students who may be at heightened risk for social exclusion, biased evaluation, and, ultimately, drop-out. Although the present study was not designed to test mechanisms underlying these processes, prior work suggests that stigmatized groups that are perceived to be threatening, unpredictable, or morally culpable tend to elicit stronger negative reactions [ 23 , 60 ]. Both criminal histories and severe mental-health conditions can invoke these stereotypes, but the moral and intentionality-related components associated with criminal offenses may contribute to the comparatively more negative evaluations of formerly incarcerated students than students with severe mental health conditions. Clarifying the mechanisms that contribute to these differences remains an important direction for future research. Directly comparing stigmatized groups risks the appearance of ranking oppression, reinforcing social hierarchies, and engaging in essentialism [ 40 ], particularly when mental illness is involved [ 61 ]. Nonetheless, cautiously examining group-specific attitudes is necessary for understanding campus climate and guiding institutional responses. The mental-health conditions included here were intentionally selected for their severity, high levels of public stigma [ 47 ], relative infrequency in undergraduate populations [ 62 ], and their “invisibility” (i.e., likely requiring disclosure); these features parallel those of the criminal backgrounds used in this study. However, this design also limits generalizability: these diagnoses do not reflect the broader landscape of campus mental-health concerns, and low familiarity may lead to weaker or less stable associations and emotional reactions. Accordingly, our findings speak most directly to stigma toward less common, highly stigmatized conditions, and inferences about “mental illness” should be interpreted with this boundary in mind. Similarly, the criminal histories depicted here represent relatively rare, severe offenses and may not generalize to misdemeanor-level backgrounds that are comparatively common in the undergraduate population [ 63 ]. Future work should therefore consider incorporating milder, more typical comparison groups (e.g., severe vs. common mental illnesses; felony vs. misdemeanor records) to clarify the observed gradients of stigma. The mixed-methods design of this study offered unique insight into stigma toward students with mental illness compared to those with criminal histories. Across both methods, students with criminal histories were evaluated markedly more negatively than both control and mental-illness targets, with effect sizes approximately twice as large in the sentiment analysis than in the SAS. On the SAS, evaluations of students with mental illness were largely indistinguishable from those of control classmates across dimensions; however, sentiment analysis of written responses toward the mental-illness vignettes were more emotionally distant, critical, or indifferent than in the control condition, though still more positive than the criminal history condition. This divergence suggests that, although structured affective ratings did not reliably differentiate evaluations of students with psychiatric diagnoses from those of controls, spontaneous language revealed subtler reservations that were not captured in the rating scale. One interpretation is that structured rating formats elicit responses congruent with social norms, whereas less-constrained, open-ended formats elicit language more directly shaped by respondents’ underlying schemas and spontaneous psychological processes [ 64 , 65 ]. Contemporary younger adults—who comprised the majority of the sample—tend to endorse openness and tolerance regarding mental health concerns, even as subtler stigma persists [ 66 ]. The SAS ratings may have reflected these normative expectations. In contrast, spontaneous written responses—less constrained by these norms—revealed muted positivity and lingering discomfort or distance. Likewise, explicit attitudes toward mental illness are often more favorable than implicit evaluations [ 67 ], suggesting that greater structure in a response format can draw respondents toward socially shaped or aspirational positions, whereas open-ended formats may elicit more schema-driven reactions. The interaction patterns in both tasks offered additional insight into how participants evaluated each group. On the SAS, the Condition × Affect interaction reflected differences in magnitude rather than differences in the structure of responses: across all dimensions, the Criminal history condition was rated more negatively than the Control and Mental Illness conditions, which were statistically indistinguishable. The size of these differences varied by affective domain. Threat-related reactions—Safe–Unsafe and Afraid–Not Afraid—showed the largest disparities ( ds ≈ 0.52–0.62), whereas Angry–Happy and Disappointed–Pleased were comparatively smaller ( ds ≈ 0.40–0.55). Taken together, these findings indicate that, although negative emotional responses to classmates with criminal histories were broad, differences were especially pronounced for judgments tied to perceived danger. Therefore, across affective dimensions, these results provide clear numerical evidence that formerly incarcerated students are perceived chiefly as dangerous, with secondary impressions of untrustworthiness or incompetence [ 26 , 27 , 35 ]. The second interaction pattern was observed in the sentiment analysis of open-ended questions. Namely, participants described their own treatment of a classmate more positively than that of society in the Control and Mental Illness conditions, though this difference was substantially larger for the former than the latter. This “behavioral uplift” may be explained by the False Uniqueness Effect , a tendency to view oneself as more unique than one really is, often in a socially desirable way [ 68 ]. This effect has been found to be especially pronounced in the treatment of stigmatized groups [ 69 ]; however, in the present study, this effect was notably muted for the Mental Illness condition and entirely absent for the Criminal condition. Even in the relatively unconstrained format of open-ended responses—which, as discussed above, typically elicit schema-consistent sentiment and allow participants to present themselves more authentically [ 64 , 65 ]—participants did not describe treating a formerly incarcerated classmate any better or more positively than society at large. Therefore, reactions toward criminal history were not only relatively negative but also resistant to the self-enhancing tendencies that ordinarily soften reports of one’s own behavior. Taken together, these findings suggest a consistent gradient in which responses to mental illness occupied an intermediate position between Control and Criminal history across sentiment analyses, whereas explicit ratings indicated attitudes toward mental illness that were comparable to the Control condition. Implications for intervention The present findings have important implications for how colleges design interventions to improve campus climate. Efforts to support students with mental illness remain vital [ 38 ]. Moreover, many students with criminal histories also report mental health challenges [ 70 , 71 ]: addressing one without the other risks offering incomplete or ineffective support. That said, there are many well-established and effective interventions for reducing stigma around mental health for undergraduates [ 41 ]. Much less discussed are interventions and programs to reduce stigma for formerly incarcerated students. Addressing both structural and peer-level stigma are necessary for successful retention and degree completion for students with criminal records. These interventions must address practical and institutional barriers in a holistic way. Formerly incarcerated students benefit from wraparound programs that include mentorship, housing support, legal aid, and academic counseling, as seen in initiatives like Project Rebound and Berkeley Underground Scholars [ 31 ]. Reducing stigma requires not just services but also a shift in campus culture—through faculty education, admissions policy reform, and peer support [ 18 , 25 , 30 ]. Mentorship and academic support to help navigate the unspoken norms and expectations of college life (i.e., the “invisible curriculum”) may be particularly important for this population [ 72 ]. Visible administrative support and opportunities to share their stories in structured educational settings (e.g., classroom panels, campus forums, student organizations) can humanize experiences, challenge deficit narratives, and reduce feelings of dehumanization and exclusion [ 73 – 75 ]. Peer (public) stigma often unfolds through processes such as social distancing, perceived threat, and the burden of concealment, which can limit access to peer networks, impede belongingness, and diminish academic engagement [ 23 , 60 ], particularly for students with criminal records [ 30 , 31 ]. Contact-based and narrative approaches may be especially important targets for reducing peer and public stigma on campus, in turn mitigating the effects of the negative student sentiment and fear that was observed in the present study [ 75 ]. Together, greater, visible support and opportunities to address and correct negative stigma can enhance the student experience and expand access to student networks [ 30 , 76 , 77 ]. These improvements can strengthen belonging, thereby increasing the likelihood of degree attainment and the occupational, economic, and psychosocial benefits associated with it [ 6 , 8 , 31 ]. Limitations and future directions This study provides some important insights into stigma toward formerly incarcerated individuals and those with mental health diagnoses, but it is not without limitations. The sample was somewhat overrepresented by older students—both in age and program year—which may limit generalizability. However, U.S. undergraduates are increasingly mirroring the demographic profile of the broader, aging global population: students aged 25 and older now represent the fastest-growing segment of higher education [ 78 ], and national trends suggest that campuses will continue to diversify along dimensions that have historically been underrepresented in academic samples, including age, race, and justice involvement [ 44 , 79 ]. Accordingly, the age distribution in this sample may be more reflective of the emerging undergraduate population than traditional assumptions would suggest. Future work should examine whether age and program year meaningfully shape stigma, as life experience and time in higher education may influence these evaluative processes. That said, we did not include direct measures of stigma toward formerly incarcerated individuals or those with mental illness, nor did we assess participants’ familiarity with the crimes or mental health conditions presented in the SAS. Although participants were instructed to look up unfamiliar terms, this behavior was not verified. Accordingly, variability in familiarity across conditions (e.g., dissociative identity disorder) may have reduced the salience of associated stigma to certain items and attenuated condition differences. Additionally, sentiment analysis tools can struggle with sarcasm, ambiguity, brevity, and other contextual nuance; interpretation of these scores requires caution as a result. It was beyond the scope of this paper to conduct deeper thematic analyses of the open-ended responses, which may have helped identify common ideas or stereotypes underlying the sentiment scores. The LMM analyses treated individual respondent as a random effect, thereby accounting for stable person-level variability (e.g., dispositional attitudes, ideology, personality) that could otherwise confound condition effects. Because this approach partitions variance attributable to individuals directly into the model, covariate adjustment was not strictly necessary. Future work with larger samples or samples that could further elucidate these findings (e.g., older vs. younger students; those with or without mental illness and/or criminal backgrounds). It was beyond the scope of the present manuscript to evaluate key intersectional demographic (e.g., race, socioeconomic status) and psychosocial variables (e.g., psychiatric diagnosis) of the respondent or target. Given that intersectionality has been previously shown to result in markedly different stigma assessments [ 19 – 21 ], this level of analysis is a critical next step to avoid characterizing the formerly incarcerated population as a homogenous group. Importantly, direct comparisons across stigmatized groups must be approached carefully: juxtaposing marginalized groups can inadvertently create a hierarchy of oppression and risk reinforcing essentialist distinctions that run counter to the goals of stigma-reduction research [ 61 ]. Researchers should aim to dismantle stigma rather than further entrench it. Conclusion Higher education is a major social determinant of health with broad psychosocial benefits. As increasing numbers of justice-involved individuals pursue postsecondary education, addressing the attitudinal and structural barriers that shape their academic trajectories is essential for reducing rearrest and other downstream risks. This study, alongside a growing body of work, suggests that undergraduate attitudes toward stigmatized populations remain a key obstacle. Reducing stigma on campus is not only a moral imperative, but also integral to broader social change. Beyond interrupting cycles of marginalization and repeated incarceration, fostering a more supportive campus climate may also strengthen educational engagement, improve individual outcomes, and benefit all students and the wider communities that institutions of higher education serve. Acknowledgements The authors would like to thank Dr. Svetlana Jović for assistance and input into the development of this project, and to Colette Vaughan for editorial assistance. We also express gratitude to Drs. Molly Ott and Terrence McTier, Jr. for providing the Situational Attitude Scale that was included in the present study, as well as input into our design. Authors’ contributions KS and EB conceptualized and designed the study. KS conducted preliminary analyses and drafted the original versions of this manuscript. EB conducted the primary analyses and led manuscript preparation. Both authors contributed to the interpretation of findings and approved the final manuscript. Funding This study was supported by an internal grant awarded to the first author from the Panther Center for Undergraduate Research Experiences (CURE) at SUNY Old Westbury. Data availability De-identified data are available at https://osf.io/wfcrb/overview?view_only=eba6927361f54a398671ef94b3ef3ade . Declarations Ethics approval and consent to participate This study was approved as Exempt by the Institutional Review Board at SUNY Old Westbury (IRB# 2023-017) and was conducted in accordance with the ethical standards of the institutional and/or national research committee, and with the 1964 Declaration of Helsinki and its later amendments or comparable ethical standards. All participants provided informed consent prior to participation. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes 1 We use criminalization to refer to the structural, legal, and social processes through which certain behaviors and groups are disproportionately surveilled and punished, and minoritization to refer to the systemic marginalization of groups through inequities and socially imposed labels rather than inherent traits. These dynamics operate at the societal—not individual—level. 2 These categories reflect qualitative differences in institutional size and program scope rather than a hierarchy of educational status or prestige. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References 1. Suiter SV, Meadows ML. Educational Attainment and Educational Contexts as Social Determinants of Health. Prim Care: Clin Office Pract. 2023;50:579–89. 10.1016/j.pop.2023.04.007. [ DOI ] [ PubMed ] [ Google Scholar ] 2. Braveman P, Gottlieb L. The social determinants of health: It’s time to consider the causes of the causes. Public Health Rep. 2014;129:19–31. 10.1177/00333549141291s206. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 3. Atherwood S, Sánchez-Soto G. Does Social Class Matter Equally for the Timely Transition Into and Out of College? Evidence from the NLSY97. Res High Educt. 2023;64:95–128. 10.1007/s11162-022-09692-w. [ Google Scholar ] 4. Case A, Deaton A. Life expectancy in adulthood is falling for those without a BA degree, but as educational gaps have widened, racial gaps have narrowed. Proceedings of the National Academy of Sciences. 2021;118. 10.1073/pnas.2024777118. [ DOI ] [ PMC free article ] [ PubMed ] 5. Stickle B, Schuster SS. Are Schools in Prison Worth It? The Effects and Economic Returns of Prison Education. Am J Criminal Justice. 2023;48:1263–94. 10.1007/s12103-023-09747-3. [ Google Scholar ] 6. Pelletier E, Evans D. Beyond Recidivism. J Correctional Educ. 2019;70:49–68. [ Google Scholar ] 7. Esperian JH. The Effect of Prison Education Programs on Recidivism. J Correctional Educ. 2010;61:316–34. [ Google Scholar ] 8. Conway PF. Beyond Recidivism: Exploring Formerly Incarcerated Student Perspectives on the Value of Higher Education in Prison. Rev High Educ. 2023;46:453–83. 10.1353/rhe.2023.a900569. [ Google Scholar ] 9. Harper A, Ginapp C, Bardelli T, Grimshaw A, Justen M, Mohamedali A, et al. Debt, Incarceration, and Re-entry: a Scoping Review. Am J Criminal Justice. 2021;46:250–78. 10.1007/s12103-020-09559-9. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 10. Eren O, Owens E. Economic Booms and Recidivism. J Quant Criminol. 2024;40:343–72. 10.1007/s10940-023-09571-2. [ Google Scholar ] 11. Zane SN, Pupo JA, Welsh BC. Correctional treatment as an economically sound approach to reducing the high costs of recidivism: A review of the research. Criminal Behav Mental Health. 2023;33:125–38. 10.1002/cbm.2284. [ DOI ] [ PubMed ] [ Google Scholar ] 12. Colbert CJ. Beyond barriers: Supporting formerly incarcerated students in American higher education. J Prison Educ Res. 2025;9:1–18. 10.25771/DZ0E-0454. [ Google Scholar ] 13. Gaskill S, Gould MR, Price V, Castro EL, Lerman AE. The landscape of higher education in prisons, Academic Years: 2018–2021 [Internet]. Alliance for Higher Education in Prison; 2023. http://higheredinprison.org . 14. Dholakia N. How College in Prison is Changing Lives [Internet]. How college in prison is changing lives. 2023 [cited 2025 July 26]. https://www.vera.org/news/how-college-in-prison-changes-lives . Accessed 26 July 2025. 15. Finkel E. Once the norm, collegiate prison programs grow again - Community College Daily [Internet]. 2024 [cited 2025 July 26]. https://www.ccdaily.com/2024/03/once-the-norm-collegiate-prison-programs-grow-again/ . Accessed 26 July 2025. 16. McCorkel J, DeFina R. Identifying the Liberatory Possibilities of Prison Higher Education. Crit Educ. 2019;10:1–19. 10.14288/ce.v10i7.186316. [ Google Scholar ] 17. Bellinger P, Melvin L, Woods M, Quinet B, Gualtieri K, Hunt L, et al. Criminal and disciplinary history in college admissions. Washington, DC: American Association of Collegiate Registrars and Admissions Officers; 2019. [ Google Scholar ] 18. Mayer J, Smith J. College Admissions and Criminal Records: The Social Construction of Risk in the University Setting. J Criminal Justice Educ Routledge. 2024;1–19. 10.1080/10511253.2024.2326113. 19. Stewart R, Uggen C. Criminal records and college admissions: A modified experimental audit. Criminology. 2020;58:156–88. 10.1111/1745-9125.12229. [ Google Scholar ] 20. Anderson K, Varghese FP, Trower E, Sandlin L, Norwood N. Perceptions of African American College Applicants: The Roles of Race, Criminal History, and Qualifications. Race Social Probl. 2013;5:157–72. 10.1007/s12552-013-9091-0. [ Google Scholar ] 21. Burton AL, Puddy HN, Hong S, Burton VS Jr, Miller WT. College Students’ Attitudes Toward Denying Admission to Prospective Students with Criminal Records. J School Violence. 2024;23:71–81. 10.1080/15388220.2023.2269525. [ Google Scholar ] 22. Custer BD. College Admission Policies for Ex-Offender Students. J Correctional Educ. 2016;67:35–43. [ Google Scholar ] 23. Pescosolido BA, Martin JK. The Stigma Complex. Annu Rev Sociol. 2015;41:87–116. 10.1146/annurev-soc-071312-145702. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 24. Bowles TV, Brindle KA. Identifying facilitating factors and barriers to improving student retention rates in tertiary teaching courses: a systematic review. High Educ Res Dev. 2017;36:903–19. 10.1080/07294360.2016.1264927. [ Google Scholar ] 25. Ahmed R, Johnson M, Caudill C, Diedrich N, Mains D, Key A. Cons and pros: prison education through the eyes of the prison educated. Rev Communication. 2019;19:69–76. 10.1080/15358593.2018.1555645. [ Google Scholar ] 26. LeBel TP. Invisible Stripes? Formerly Incarcerated Persons’ Perceptions of Stigma. Deviant Behav. 2012;33:89–107. 10.1080/01639625.2010.538365. [ Google Scholar ] 27. McTier TS Jr., Briscoe KL, Davis TJ. College Administrators’ Beliefs and Perceptions of College Students with Criminal Records. J Student Affairs Res Pract. 2020;57:296–308. 10.1080/19496591.2019.1648273. [ Google Scholar ] 28. Moore KE, Phillips S, Kromash R, Siebert S, Roberts W, Peltier M, et al. The causes and consequences of stigma among individuals involved in the criminal legal system: A systematic review. Stigma Health. 2024;9:224–35. 10.1037/sah0000483. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 29. Denver M, Pickett JT, Bushway SD. The language of stigmatization and the mark of violence: Experimental evidence on the social construction and use of criminal record stigma. Criminology. 2017;55:664–90. 10.1111/1745-9125.12145. [ Google Scholar ] 30. Copenhaver A, Edwards-Willey TL, Byers BD. Journeys in Social Stigma: The Lives of Formerly Incarcerated Felons in Higher Education. J Correctional Educ. 2007;58:268–83. [ Google Scholar ] 31. Quach K, Cerda-Jara M, Deverux R, Smith J. Prison, College, and the Labor Market: A Critical Analysis by Formerly Incarcerated and Justice-Impacted Students. ANNALS Am Acad Political Social Sci. 2022;701:78–97. 10.1177/00027162221112772. [ Google Scholar ] 32. Halkovic A, Greene AC. Bearing Stigma, Carrying Gifts: What Colleges Can Learn from Students with Incarceration Experience. Urban Rev. 2015;47:759–82. 10.1007/s11256-015-0333-x. [ Google Scholar ] 33. Binnall JM, Scott-Hayward CS, Petersen N, Gonzalez RM. Taking roll: College students’ views of their formerly incarcerated classmates. J Criminal Justice Educ. 2022;33:347–67. 10.1080/10511253.2021.1962932. [ Google Scholar ] 34. Overton J, Fretwell MD, Dum CP. Who do you trust? College students’ attribution of stigma to peers with incarceration histories. J Experimental Criminol. 2022;18:847–70. 10.1007/s11292-021-09463-0. [ Google Scholar ] 35. Ott M, McTier TS Jr. Faculty attitudes toward college students with criminal records. J Divers High Educ. 2020;13:297–308. 10.1037/dhe0000138. [ Google Scholar ] 36. Schomerus G, Schindler S, Sander C, Baumann E, Angermeyer MC. Changes in mental illness stigma over 30 years – Improvement, persistence, or deterioration? Eur Psychiatry. 2022;65:e78. 10.1192/j.eurpsy.2022.2337. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 37. Corrigan PW, Kleinlein P. The Impact of Mental Illness Stigma. In: Corrigan PW, editor. On the stigma of mental illness: Practical strategies for research and social change. American Psychological Association; 2005;11–44. 10.1037/10887-001. 38. Eisenberg D, Downs MF, Golberstein E, Zivin K. Stigma and Help Seeking for Mental Health Among College Students. Med Care Res Rev. 2009;66:522–41. 10.1177/1077558709335173. [ DOI ] [ PubMed ] [ Google Scholar ] 39. Guarneri JA, Oberleitner DE, Connolly S. Perceived Stigma and Self-Stigma in College Students: A Literature Review and Implications for Practice and Research. Basic Appl Soc Psychol. 2019;41:48–62. 10.1080/01973533.2018.1550723. [ Google Scholar ] 40. Hatzenbuehler ML, Phelan JC, Link BG. Stigma as a Fundamental Cause of Population Health Inequalities. Am J Public Health. 2013;103:813–21. 10.2105/AJPH.2012.301069. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 41. Waqas A, Malik S, Fida A, Abbas N, Mian N, Miryala S, et al. Interventions to reduce stigma related to mental illnesses in educational institutes: A systematic review. Psychiatr Q. 2020;91:887–903. 10.1007/s11126-020-09751-4. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 42. Baloch NA, Jennings WG. A Preliminary Investigation of the Intersection of Race and Disabilities among Inmates in the U.S. State Prison System. Int J Offender Ther Comp Criminol. 2019;63:597–609. 10.1177/0306624X18805599. [ DOI ] [ PubMed ] [ Google Scholar ] 43. Prins SJ. Prevalence of Mental Illnesses in U.S. State Prisons: A Systematic Review. Psychiatric Serv. 2014;65:862–72. 10.1176/appi.ps.201300166. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 44. DeBaun B. Freshman Enrollment Grows 5.5% in Fall 2024 in Encouraging, Surprising Change [Internet]. Freshman Enrollment Grows 5.5% in Fall 2024 in Encouraging, Surprising Change. [cited 2025 Nov 25]. https://www.ncan.org/news/692153/Freshman-Enrollment-Grows-5.5-in-Fall-2024-in-Encouraging-Surprising-Change-.htm . Accessed 25 Nov 2025. 45. Sedlacek WE, BrooksJr. GC. Measuring Racial Attitudes in a Situational Context. Psychol Rep. 1970;27:971–80. 10.2466/pr0.1970.27.3.971. [ DOI ] [ PubMed ] [ Google Scholar ] 46. Kelley HH, Michela JL. Attribution Theory and Research. Ann Rev Psychol. 1980;31:457–501. 10.1146/annurev.ps.31.020180.002325. [ DOI ] [ PubMed ] [ Google Scholar ] 47. Hazell CM, Berry C, Bogen-Johnston L, Banerjee M. Creating a hierarchy of mental health stigma: testing the effect of psychiatric diagnosis on stigma. BJPsych Open. 2022;8:e174. 10.1192/bjo.2022.578. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 48. Nießen D, Partsch MV, Kemper CJ, Rammstedt B. An English-Language Adaptation of the Social Desirability–Gamma Short Scale (KSE-G). Meas Instruments Social Sci. 2019;1:2. 10.1186/s42409-018-0005-1. [ Google Scholar ] 49. Paulhus DL, Buckels EE, Trapnell PD, Jones DN. Screening for Dark Personalities: The Short Dark Tetrad (SD4). Eur J Psychol Assess. 2021;37:208–22. 10.1027/1015-5759/a000602. [ Google Scholar ] 50. Kaufman SB, Yaden DB, Hyde E, Tsukayama E. The Light vs. Dark Triad of Personality: Contrasting Two Very Different Profiles of Human Nature. Front Psychol. 2019;10:467. 10.3389/fpsyg.2019.00467. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 51. Bizumic B, Duckitt J. Investigating right wing authoritarianism with a very short authoritarianism scale. J Social Political Psychol. 2018;6:129–50. 10.5964/jspp.v6i1.835. [ Google Scholar ] 52. Konrath S, Meier BP, Bushman BJ. Development and Validation of the Single Item Narcissism Scale (SINS). Wicherts JM, editor. PLoS ONE. 2014;9:e103469. 10.1371/journal.pone.0103469. [ DOI ] [ PMC free article ] [ PubMed ] 53. Robins RW, Hendin HM, Trzesniewski KH. Measuring Global Self-Esteem: Construct Validation of a Single-Item Measure and the Rosenberg Self-Esteem Scale. Pers Soc Psychol Bull. 2001;27:151–61. 10.1177/0146167201272002. [ Google Scholar ] 54. Konrath S, Meier BP, Bushman BJ. Development and validation of the single item trait empathy scale (SITES). J Res Pers. 2018;73:111–22. 10.1016/j.jrp.2017.11.009. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 55. Ahmad F, Jhajj AK, Stewart DE, Burghardt M, Bierman AS. Single item measures of self-rated mental health: a scoping review. BMC Health Serv Res. 2014;14:398. 10.1186/1472-6963-14-398. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 56. Boyd RL, Ashokkuma A, Seraj S, Pennebaker JW. The Development and Psychometric Properties of LIWC-22 [Internet]. University of Texas at Austin; 2022. www.liwc.app. 57. Koutsoumpis A, Oostrom JK, Holtrop D, Van Breda W, Ghassemi S, De Vries RE. The kernel of truth in text-based personality assessment: A meta-analysis of the relations between the Big Five and the Linguistic Inquiry and Word Count (LIWC). Psychol Bull. 2022;148:843–68. 10.1037/bul0000381. [ Google Scholar ] 58. Tausczik YR, Pennebaker JW. The Psychological Meaning of Words: LIWC and Computerized Text Analysis Methods. J Lang Social Psychol. 2010;29:24–54. 10.1177/0261927X09351676. [ Google Scholar ] 59. Foster AM, Talley AE. Interpersonal Goals and Language Use During the Disclosure of a Concealable Stigmatized Identity: A LIWC Analysis. J Lang Social Psychol. 2025;44:107–31. 10.1177/0261927X241294079. [ Google Scholar ] 60. Link BG, Yang LH, Phelan JC, Collins PY. Measuring Mental Illness Stigma. Schizophr Bull. 2004;30:511–41. 10.1093/oxfordjournals.schbul.a007098. [ DOI ] [ PubMed ] [ Google Scholar ] 61. Mula M. Is there a hierarchy in mental health stigma? BJPsych Open. 2023;9:e4. 10.1192/bjo.2022.618. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 62. Oswalt SB, Lederer AM, Chestnut-Steich K, Day C, Halbritter A, Ortiz D. Trends in college students’ mental health diagnoses and utilization of services, 2009–2015. J Am Coll Health. 2020;68:41–51. 10.1080/07448481.2018.1515748. [ DOI ] [ PubMed ] [ Google Scholar ] 63. Friedman MJ, Facts. As Many Americans Have Criminal Records as College Diplomas | Brennan Center for Justice [Internet]. Brenna Center for Justice. 2015 [cited 2025 Dec 6]. https://www.brennancenter.org/our-work/analysis-opinion/just-facts-many-americans-have-criminal-records-college-diplomas . Accessed 6 Dec 2025. 64. Hansen K, Świderska A. Integrating open- and closed-ended questions on attitudes towards outgroups with different methods of text analysis. Behav Res Methods. 2023;56:4802–22. 10.3758/s13428-023-02218-x. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 65. Pennebaker JW, Mehl MR, Niederhoffer KG. Psychological Aspects of Natural Language Use: Our Words, Our Selves. Ann Rev Psychol. 2003;54:547–77. 10.1146/annurev.psych.54.101601.145041. [ DOI ] [ PubMed ] [ Google Scholar ] 66. Botha F, Morris RW, Butterworth P, Glozier N. Generational differences in mental health trends in the twenty-first century. Proc Natl Acad Sci. 2023;120:e2303781120. 10.1073/pnas.2303781120. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 67. Robb J, Stone J. Implicit Bias toward People with Mental Illness: A Systematic Literature Review. J Rehabilitation. 2016;82:3–13. [ Google Scholar ] 68. Chambers JR. Explaining False Uniqueness: Why We are Both Better and Worse Than Others. Soc Pers Psychol Compass. 2008;2:878–94. 10.1111/j.1751-9004.2008.00076.x. [ Google Scholar ] 69. Sigelman CK. Social Distance from Stigmatized Groups: False Consensus and False Uniqueness Effects on Responding. Rehabil Psychol. 1991;36:139–51. [ Google Scholar ] 70. Nishar S, Brumfield E, Mandal S, Vanjani R, Soske J. It’s a revolving door: understanding the social determinants of mental health as experienced by formerly incarcerated people. Health Justice. 2023;11:26. 10.1186/s40352-023-00227-8. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 71. Schnittker J, Massoglia M, Uggen C. Out and Down: Incarceration and Psychiatric Disorders. J Health Soc Behav. 2012;53:448–64. 10.1177/0022146512453928. [ DOI ] [ PubMed ] [ Google Scholar ] 72. Tietjen G, Burnett J, Jessie BO. Onward and Upward: the Significance of Mentorship for Formerly Incarcerated Students and Academics. Crit Criminol. 2021;29:633–47. 10.1007/s10612-020-09507-3. [ Google Scholar ] 73. Sinko R, DeAngelis T, Alpajora B, Beker J, Kramer I. Experience of Stigma Post Incarceration: A Qualitative Study. The Open Journal of Occupational Therapy. WMU Department Occup Therapy. 2020;8:1–16. 10.15453/2168-6408.1610. [ Google Scholar ] 74. Frankel A, DePoy E. Thoughts Beyond Stigma-Implications for Change Reflected in the Voices of Previously Incarcerated Citizens. J Qualitative Criminal Justice Criminol. 2024;13:1–29. 10.21428/88de04a1.497f21ab. [ Google Scholar ] 75. Smith JM. The formerly incarcerated, advocacy, activism, and community reintegration. Contemp Justice Rev. 2021;24:43–63. 10.1080/10282580.2020.1755846. [ Google Scholar ] 76. Freire DS, Hurd NM. Discrimination and Mental Health Outcomes Among Underrepresented College Students: The Role of Sense of Belonging at Predominantly White Institutions. Emerg Adulthood. 2023;11:654–68. 10.1177/21676968231166967. [ Google Scholar ] 77. Shuraydi AM. Returning to Class: Suggestions for Practicing Inclusivity with Formerly Incarcerated Students. J Criminal Justice Educ. 2025;36:86–102. 10.1080/10511253.2024.2384571. [ Google Scholar ] 78. Bárány ZL, Corblet P, Buchinsky M. Older-adult college enrollment trends and benefits [Internet]. Brookings Institute; 2024. https://www.brookings.edu/articles/older-adult-college-enrollment-trends-and-benefits/ . 79. Hanson M. College Enrollment & Student Demographic Statistics [Internet]. Education Data Initiative. 2025 [cited 2026 Feb 19]. https://educationdata.org/college-enrollment-statistics . Accessed 19 Feb 2026. Associated Data This section collects any data citations, data availability statements, or supplementary materials included in this article. Data Availability Statement De-identified data are available at https://osf.io/wfcrb/overview?view_only=eba6927361f54a398671ef94b3ef3ade . Articles from BMC Psychology are provided here courtesy of BMC ACTIONS View on publisher site PDF (1.7 MB) Cite Collections Permalink PERMALINK Copy RESOURCES Similar articles Cited by other articles Links to NCBI Databases Cite Copy Download .nbib .nbib Format: AMA APA MLA NLM Add to Collections Create a new collection Add to an existing collection Name your collection * Choose a collection Unable to load your collection due to an error Please try again Add Cancel Follow NCBI NCBI on X (formerly known as Twitter) NCBI on Facebook NCBI on LinkedIn NCBI on GitHub NCBI RSS feed Connect with NLM NLM on X (formerly known as Twitter) NLM on Facebook NLM on YouTube National Library of Medicine 8600 Rockville Pike Bethesda, MD 20894 Web Policies FOIA HHS Vulnerability Disclosure Help Accessibility Careers NLM NIH HHS USA.gov Back to Top