Association for Public Policy Analysis & Management

Over 70% of American adults who applied for a job within the past five years report being asked about a criminal record on their most recent job application (Denver, Pickett and Bushway, 2016).  With approximately 16% of American adults reporting that they have a conviction record by age 23 (Bushway et al., 2014) and empirical evidence documenting employer discrimination against those with conviction records (e.g., Pager, 2007), criminal records present a significant policy issue in the American labor market.  Furthermore, the stigma and discrimination stemming from a criminal record may lead to heightened recidivism.

To counteract these potential negative effects, one line of policy reform has focused on managing the construction of criminal stigma. For example, the U.S. Department of Justice (DOJ) announced a policy change in April 2016 regarding language commonly used to describe individuals with criminal records.  Instead of terms like “(ex-)offender,” “criminal,” or “felon,” DOJ is promoting the use of ostensibly less stigmatizing descriptive terms, such as “person with a felony conviction.” The central, but as yet untested, theoretical assumption underlying the DOJ policy is that the specific sematic techniques used by social audiences in communications about deviance have important implications for the process of constructing and reproducing criminal stigma. Another line of policy reform has concentrated on restructuring the use (to prevent the misuse) of criminal stigma.  For example, the Equal Employment Opportunities Commission released federal guidelines in 2012 that direct employers to make individualized assessments, which includes balancing each job applicant’s criminal record with evidence of good conduct and rehabilitation. However, it is currently unclear whether the average American or employer is aware of the empirical factors driving recidivism probabilities and able to balance multiple, complex factors. 

The current study involves two large surveys of adults (18 and older) in the United States to conduct two experimental manipulations on the social construction of stigma.  First, we take a linguistic approach, theorizing that an important component of stigmatization is the actual attachment of stigmatizing labels—words or phrases—to persons who have committed deviant acts.  Drawing on data from a large nationally representative sample of American adults, we provide a direct test of whether descriptive language influences the general public’s perceptions of people with criminal records.  Second, we use a sample randomly selected from an internet panel (SurveyMonkey’s “Audience” panel) to examine the contextualized use of criminal stigma in decision-making. We employ an experimental manipulation that randomly assigns the job type (highway maintenance worker, home health care aide, or cashier position), conviction crime type (violent, non-violent drug, or non-violent property), and the amount of time that has passed since the conviction (1 year or 10 years) to compare how contextual features influence support for denying a hypothetical applicant a job.  Our results indicate that American adults overestimate recidivism rates among individuals with violent convictions in particular; using “offender” language exacerbates perceptions, but only for violent convictions; and perceived recidivism risk is associated with support for denying someone a job.