Redeemed Compared to Whom?

• Author: Shawn D. Bushway,Megan C. Kurlychek,Garima Siwach,Samuel E. DeWitt
• Published date: 01 August 2017
• Date: 01 August 2017
• DOI: http://doi.org/10.1111/1745-9133.12309

RESEARCH ARTICLE

REDEEMED COMPARED TO WHOM?

Redeemed Compared to Whom?

Comparing the Distributional Properties of Arrest Risk

Across Populations of Provisional Employees With and

Without a Criminal Record

Samuel E. DeWitt

University of North Carolina at Charlotte

Shawn D. Bushway

Garima Siwach

Megan C. Kurlychek

University at Albany (SUNY)

Research Summary

By using data on provisional employees with and without criminal records, we find

that existing standards of a “reasonable amount of [arrest] risk” (derived from “time

to redemption” research) for an employer to incur in hiring individuals with criminal

Data Disclaimer: These data are provided by the New York State Division of Criminal Justice Services (DCJS)

and the Department of Health (DOH). The opinions, findings, and conclusions expressed in this publication

are those of the authors and not those of DCJS or DOH. Neither New York State nor DCJS or DOH assumes

liability for its contents or use thereof.

Funding Disclaimer: This research was supported by Award 2012-MU-MU-0048 from the National Institute

of Justice, Office of Justice Programs, U.S. Department of Justice. The opinions, findings, and conclusions

expressed in this publication are those of the authors and do not necessarily reflect the views of the

Department of Justice.

The authors are extremely grateful to our New York State partners who made this project possible. We would

like to thank Terry Salo, Leslie Kellam, and their team at the Division of Criminal Justice Services; Daryl Barra

and his team at the Department of Health; and the research staff at the Department of Labor for their

encouragement, access to data, and complete support. We extend a special thank-you to Megan Denver,

who, as a member of our research team, provided helpful feedback on this article and support with the data.

All errors remain our own. Direct correspondence to Samuel E. DeWitt, Department of Criminal Justice and

Criminology, University of North Carolina at Charlotte, 5084 Colvard Hall, 9201 University City Blvd., Charlotte,

NC 28223 (e-mail: sdewitt2@uncc.edu).

DOI:10.1111/1745-9133.12309 C2017 American Society of Criminology 963

Criminology & Public Policy rVolume 16 rIssue 3

Research Article Redeemed Compared to Whom?

histories prove too onerous for many employees without records to meet, let alone those

with records. Wethen propose an alternative method of assessing arrest risk across these

populations—benchmarking—and provide several alternative standards, illustrating

that they (a) clear a demonstrable majority of employees without records and a sizable

minority of those with a criminal history and (b) do not increase the risk incurred

by employers over and above the level they already accept among employees without

records.

Policy Implications

Our findings suggest that the almost singular importance placed on “time since last”

policies when conducting criminal background checks is ill-placed as the risk of arrest

across populations of employees with and without criminal histories overlaps more

than the results of extant research would imply. Although future research needs to be

conducted to ascertain whether our findings generalize to other employment contexts,

current background check practices would be better served if they were to adopt an

approach that chooses a specific threshold for comparison, which should align with an

easily communicated and face-valid cost function. Furthermore, the selected standard

should (a) hold all individuals to the same standard and (b) be one that a demonstrable

majority of individuals without criminal records can meet.

Research suggests that criminal background checks are used during the hiring process

by nearly 50% of all firms in the United States (Holzer, Raphael, and Stoll, 2004),

with that number at almost 80% for the nation’s largest companies (Burke,2004).

This now common use of background checks has created a public concern that a substantial

proportion of the American workforce is being excluded from important employment

opportunities (Appelbaum, 2015; Harris and Keller, 2005; Rodriguezand Emsellem, 2011;

Swanson, Langfitt-Reese, and Bond, 2012). In response, policy makers have attempted to

create a balance between the needs of employers and individuals with criminal records. For

example, the Equal Employment Opportunity Commission (EEOC) recently strengthened

its guidance to employers on the proper use of criminal records for employment screening

purposes (EEOC, 2012). Under EEOC standards, employers cannot simply avoid hiring

individuals with records but must use available information to distinguish those of acceptable

risk from those who pose an unacceptable level of risk. This approach raises an important

policy question: What is a reasonable amount of risk for an employer to incur (Blumstein

and Nakamura, 2009)?

Initially,researchers answered this question by comparing the risk of arrest for someone

with a criminal record to the risk of arrest in a comparison group that typically comprises

either (a) a group of same-aged individuals without a record (Bushway, Nieuwbeerta, and

Blokland, 2011; Kurlychek, Brame, and Bushway, 2006, 2007) or (b) a simulated control

964 Criminology & Public Policy

DeWitt et al.

group of the same age based on aggregate data (Blumstein and Nakamura, 2009). This

approach creates an array of cutoffs that are age-specific—individuals are only “low risk”

when they have the same level of risk as someone of the same age who does not have a record.

As an illustrative example of this approach, consider two hypothetical job applicants,

each with a singular conviction: Person A was convicted at 20 years of age, and Person B

was convicted at 30 years of age. Both have now had 10 arrest-free years and are now 30 (A)

and 40 (B), respectively. The existing models (Blumstein and Nakamura, 2009; Bushway

et al., 2011; Kurlychek et al., 2006, 2007) posit that the proper comparison for Person A

is the average arrest risk for a sample of 30-year-olds without a criminal record and that

the proper comparison for Person B is the average arrest risk for a sample of 40-year-olds

without a record.1

This approach has two substantial flaws. First, all individuals with records are not being

held to the same standard. In this example, we know, based on the results of long-standing

research on the age–crime curve (M. Gottfredson and Hirschi, 1990; Liu, 2014; Sweeten,

Piquero, and Steinberg, 2013), that the standard for Person A is almost certainly easier to

meet (i.e., a higher risk) than is the standard for Person B (i.e., a lower risk). It is not clear

that there are any legitimate reasons why employers should want to hold individuals who

were older at the time of their last conviction to a higher standard (e.g., a lower level of risk)

than individuals who were younger at the time of their last conviction.2

Second, and more importantly, in this approach, a situation is created in which em-

ployers might routinely hire individuals without records who have higher levels of arrest

risk than individuals with records who are denied employment. For example, suppose the

firm in the previous example that hires individuals with criminal records that are 10 years

old or older also hires employees without records who are 25 years old. Those 25-year-olds

with no records have a demonstrably higher risk than does a sample of 30-year-olds with

no records (Bushway et al., 2011; Bushway and Piehl, 2007; M. Gottfredson and Hirschi,

1990). Using a 10-year threshold for someone who had their last conviction at 20 years

of age, therefore, means that individuals with records are being forced to demonstrate a

level of risk that is lower than the risk that the firm encounters when hiring 25-year-old

individuals without a criminal record. In this article, we demonstrate the importance of

this latter problem by using data from 139,697 entry-level job recipients in the long-term

health-care field who were required to submit to a criminal background check before their

employment could be considered permanent.

1. Because keeping track of these different cutoffs would be cumbersome, the approach is focused

instead on identifying the number of years without additional criminal justice involvement that it takes

to reach this level of risk. We colloquially refer to these “time since last conviction” policies as the

N

-year

rule, where 10 is the most common example of

N

.

2. In fact, research (Bushway et al., 2011) suggests that the older an individual is when they are first

convicted, the lower their recidivism risk is expected to be, in general, and the quicker they approximate

the arrest risk of same-aged individuals without a conviction history.

Volume 16 rIssue 3 965