As part of the Consolidated Appropriations Act of 2021, a bipartisan coalition of members of Congress struck a deal that restores eligibility for incarcerated people to secure Pell Grants to cover tuition of college classes offered in prison (Stratford Reference Stratford2020). This is the latest evidence of bipartisan support for criminal justice reform among elected policy makers that began to emerge in the George W. Bush administration and has gained momentum since (Dagan and Teles Reference Dagan and Teles2016; Percival Reference Percival2015). This pro-reform stance is a significant departure from the “tough on crime” politics that dominated both parties’ platforms throughout the 1970s, 1980s, and 1990s (Gottschalk Reference Gottschalk2006; Murakawa Reference Murakawa2014; Schoenfeld Reference Schoenfeld2018). Enns (Reference Enns2016) argued that one political factor that paved the way for elites to support reform efforts in the twenty-first century was a reduction in punitive attitudes among the mass American public.
Educational benefits for prisoners are a prime example of the ways that politics can conflict with evidence-based policy making. Researchers find that people who participate in educational programs while incarcerated are significantly less likely to recidivate and be arrested for a new crime once released than inmates who did not participate in any educational programs (Bozick et al. Reference Bozick, Steele, Davis and Turner2018; Davis et al. Reference Davis, Bozick, Steele, Saunders and Miles2013; Kim and Clark Reference Kim and Clark2013; Oakford et al. Reference Oakford, Brumfield, Goldvale, Tatum, diZerega and Patrick2019). One team of analysts estimated that reducing the need to reincarcerate recidivist offenders would save states more than $350 million per year, meaning that expanding access to higher education in prisons will yield a net financial benefit (Oakford et al. Reference Oakford, Brumfield, Goldvale, Tatum, diZerega and Patrick2019). Despite the evidence that access to education in prisons is good for both prisoners and society, Congress stripped prisoners of Pell Grant eligibility in the 1994 Violent Crime Control and Law Enforcement Act. The politics of “least eligibility” fueled support for this restriction; legislators argued that no criminal should have access to any benefit that the poorest law-abiding citizen cannot also receive—and millions of citizens never receive a college education (Clear Reference Clear1994; Lewis Reference Lewis2018).
Educational benefits for prisoners are a prime example of the ways that politics can conflict with evidence-based policy making.
In 2016, the Obama administration launched a pilot program to extend Pell Grant eligibility to incarcerated individuals at 67 colleges and universities, and the Trump administration expanded the pilot program in 2020 (Douglas-Gabriel Reference Douglas-Gabriel2020; US Department of Education 2016). The Trump administration, the US Chamber of Commerce, and even the National District Attorneys Association (2019) all expressed support for the Restoring Education and Learning Act (REAL) Act in order to make the initiative permanent. However, efforts to repeal the Pell Grant ban made little progress during several successive sessions of Congress until it was finally added to an omnibus spending bill. New York State provides a recent cautionary tale that may explain why it took federal legislators many years to enact this particular reform. Governor Andrew Cuomo was forced to abandon a proposal to pay for college classes for prison inmates with state funds in the face of substantial backlash. This included Republican members of the state’s delegation to Congress introducing the “Kids Before Cons Act” that would have blocked the expenditure of federal funds on prison educational programs (Kaplan Reference Kaplan2014). These Republican legislators had been motivated by citizen-led petitions that expressed opposition to Governor Cuomo’s proposal. Thus, despite an overall decline in mass punitive attitudes (Enns Reference Enns2016), specific reform proposals can still generate backlash among both voters and policy makers.
A few studies suggest that financial aid for prisoners to access college-level courses may be one such trigger for least-eligibility backlash among members of the public. Participants in several focus-group studies expressed the opinion that the government should subsidize inmates’ education only up to a high school degree (Brooks, Visher, and Naser Reference Brooks, Visher and Naser2006; Heumann, Pinaire, and Clark Reference Heumann, Pinaire and Clark2005; Immerwahr and Johnson Reference Immerwahr and Johnson2002). Likewise, respondents to a survey fielded in Missouri expressed substantially less support for allowing inmates to access college classes that lead to a four-year degree than they did for those that lead to a high school degree, GED, or two-year technical degree (Garland, Wodahl, and Cota Reference Garland, Wodahl and Cota2016; Garland, Wodahl, and Schuhmann Reference Garland, Wodahl and Schuhmann2013). In contrast, only 33% of respondents to a survey of South Carolinians agreed with the statement: “People who are coming out of prison should not be allowed to receive federal grants for education.” However, this statement does not address public support for granting inmates access to Pell Grants while they are still incarcerated (Ouellette, Applegate, and Vuk Reference Ouellette, Applegate and Vuk2017). Given the divergent findings and limited generalizability of these past studies, we do not yet have sufficient empirical evidence to properly delineate the American public’s “zone of acquiescence” (Stimson Reference Stimson2018) for this particular type of criminal justice reform.
Recent experimental research suggests that people’s support for criminal justice reform may be sensitive to framing. Gottlieb (Reference Gottlieb2017) found that participants exposed to a message about the poor ability of incarceration to prevent recidivism or a message about the high costs of incarceration expressed significantly more support for sentencing nonviolent drug offenders and parole violators to community corrections instead of prison. These effects did not differ across demographic subgroups. Similarly, Bandara, McGinty, and Barry (Reference Bandara, McGinty and Barry2020) found that participants exposed to a message explaining that incarcerating nonviolent drug offenders does little to enhance public safety because prison does not reduce their likelihood of recidivating expressed more support for eliminating mandatory minimum sentences. This message had no effect on respondents’ support for ending bans on felons’ access to food stamps and public housing. These two studies suggest that arguments about the benefits of criminal justice reform to society may increase public support for policy change. However, framing effects appear to be limited to attitudes about particular policies; the precise scope of these effects is unclear.
We build on this nascent literature on framing effects in public opinion about criminal justice policy by analyzing the results of an experiment embedded in the Cooperative Congressional Election Study (CCES). To assess contemporary public support for extending Pell Grant eligibility to prisoners, we manipulated whether the proposal to restore their eligibility emphasized the benefit to inmates themselves or the broader benefits to society. We also tested for subgroup differences because these frames may appeal to different voters. Conservatives who are concerned about the high, economically inefficient costs of the nation’s correctional systems today may find arguments about cost savings to taxpayers particularly appealing (Dagan and Teles Reference Dagan and Teles2016). Furthermore, focusing on the aggregate benefits to society rather than benefits to individuals may circumvent least-eligibility backlash among people who believe that criminals are undeserving of taxpayer-funded benefits—beliefs that likely are racialized and concentrated among voters who harbor racial resentment, which is a key predictor of punitive attitudes (Brown and Socia Reference Brown and Socia2017; Muhammad Reference Muhammad2011; Peffley and Hurwitz Reference Peffley and Hurwitz2010; Unnever and Cullen Reference Unnever and Cullen2010).Footnote 1 Appeals along individual grounds, by contrast, may attract progressives and racial liberals who are concerned about the impacts of mass incarceration on people of color (Chudy Reference Chudy2021). In the end, the experiment’s findings suggest that both frames increase support for the policy, especially among Democrats and individuals with low racial resentment. It appears unlikely that the policy makers who restored educational benefits to incarcerated individuals will face a least-eligibility backlash from today’s voters.
DATA AND METHOD
We embedded a novel survey experiment in the 2016 CCES, which is a national online survey conducted by YouGov/Polimetrix (Ansolabehere and Rivers Reference Ansolabehere and Rivers2013; Ansolabehere and Schaffner Reference Ansolabehere and Schaffner2017). The survey is fielded in two waves, with the same panel participating before and after the general election. In addition to the survey’s Common Content questions, the 2016 CCES included 60 team modules. Our experiment was included on the pre-election wave of one module. The items used in the subgroup analyses varied across the two waves. The survey’s standard partisanship question was included on the pre-election survey and the racial-resentment items were fielded on the post-election wave.Footnote 2
In the “Control” condition, respondents received the following prompt: “The Department of Education is piloting a program to offer financial aid to incarcerated individuals interested in attending college while incarcerated. Do you support this program?”Footnote 3 In the two treatment conditions, respondents were provided additional reasoning to support the program. Before evaluating the program, respondents in the “Individuals” condition were told: “Supporters argue that the program will benefit these individuals by providing them with new skills that will reduce reincarceration.” Finally, respondents assigned to the “Society” condition read: “Supporters argue that the program will benefit society by decreasing the costs associated with reincarceration.” These frames are similar to those used by Gottlieb (Reference Gottlieb2017) and Bandara, McGinty, and Barry (Reference Bandara, McGinty and Barry2020), albeit shorter in word count. Advocates of the Second Chance Pell program promised that it would decrease recidivism, which rewards both the individuals involved and society as whole. By emphasizing different aspects of the policy, the experiment tests whether subjects and various subgroups are more supportive when primed to consider either the particularized benefits or the broader societal merits of the policy.
RESULTS
Figure 1 presents results of the framing experiment.Footnote 4 First examining the full sample, we see that the average response to the proposal to offer financial aid to incarcerated people fell at the midpoint of our scale, which indicates, at worst, public indifference or ambivalence rather than opposition.Footnote 5 Both argument frames significantly increased public support relative to the control group, pushing the average response into the supportive range of the response scale. Examining the differences by partisan affiliation, we see that Democrats were significantly and substantially more supportive of extending financial aid to prisoners than both Independents and Republicans. At the baseline (i.e., the control group), Democrats expressed mild support for the proposal, on average, whereas Independents expressed ambivalence and Republicans expressed moderate opposition. Exposure to both frames significantly increased support among Democrats, pushing them to express strong support, on average. Exposure to both frames also increased support among Independents and Republicans, but the magnitude of this increase was statistically significant only for the individuals frame among Independents and only marginally significant for the society frame among Republicans. Frame exposure pushed Independents from the midpoint to weak support and Republicans from moderate opposition to mild opposition or indifference, on average.
First examining the full sample, we see that the average response to the proposal to offer financial aid to incarcerated people fell at the midpoint of our scale, which indicates, at worst, public indifference or ambivalence rather than opposition.
Figure 1 Effect of Message Frames on Pell Grant Access Support among Full Sample and Subgroups
Consistent with the racialized nature of crime and criminal justice in the United States (Brown and Socia Reference Brown and Socia2017; Muhammad Reference Muhammad2011; Peffley and Hurwitz Reference Peffley and Hurwitz2010; Unnever and Cullen Reference Unnever and Cullen2010), we found that support for extending financial aid to incarcerated people decreased as a linear function of stronger racial resentment. Exposure to both frames significantly increased support for the proposal among respondents who scored at low and moderate levels on the racial-resentment scale, but frame exposure had no statistically significant effect among respondents with high racial resentment.
DISCUSSION
The results of this research present both good news and not so good news for the legislators and advocates who supported the restoration of Pell Grant eligibility to prisoners in order to reduce recidivism and improve post-incarceration life outcomes. The average response among our subjects who were not exposed to a pro-reform frame fell around the midpoint of our response scale, suggesting ambivalence toward the proposal among the mass public. Disaggregated by subgroups, only Democrats and individuals who are low in racial resentment expressed baseline support for this reform proposal, whereas Republicans and individuals high in racial resentment expressed baseline opposition. This finding suggests that only those legislators who represent relatively liberal districts (both politically and racially) have a strong electoral incentive to advocate for Pell Grant eligibility reform. Legislators who represent moderate or conservative districts likely feel more inclined to support reforms that generate a more enthusiastic response among their constituents.
Our more heartening finding for supporters of this long-delayed reform is that all respondent groups responded favorably to pro-reform arguments that emphasized the program’s benefits to the prisoners and society more broadly. Both frames increased support for the proposal among all subgroups, although not all differences were statistically significant. Moreover, we found no evidence of backlash to either frame. This result suggests that any sitting members of Congress who are asked to justify their votes can effectively explain their support for this reform to their constituents, even with a short and simple justification. Most Americans seem to be initially indifferent to the idea of extending higher-education financial aid to inmates, but it does not take much advocacy to increase their estimation of the proposal merits. Our findings also add further evidence to the small body of literature that finds that arguments about the social benefits of criminal justice reform cause many Americans to react favorably to policy-reform proposals (Bandara, McGinty, and Barry Reference Bandara, McGinty and Barry2020; Gottlieb Reference Gottlieb2017; Pickett, Ivanov, and Wozniak Reference Pickett, Ivanov and Wozniak2020). All told, our findings suggest that proposals to restore inmates’ eligibility for financial aid fall within Americans’ “zone of acquiescence” for criminal justice reform (Wozniak Reference Wozniak2020). More important, voters exhibit malleability on the issue. The days of least-eligibility backlash against extending rehabilitative benefits to prisoners may have passed.
Most Americans seem to be initially indifferent to the idea of extending higher-education financial aid to inmates, but it does not take much advocacy to increase their estimation of the proposal merits.
DATA AVAILABILITY STATEMENT
Replication materials are available on Harvard Dataverse at https://doi.org/10.7910/DVN/OTN7GQ.
SUPPLEMENTARY MATERIALS
To view supplementary material for this article, please visit http://dx.doi.org/10.1017/S1049096521000019.
