ESR Check Logo
The Background Check Authority ®
ESR Support: 888.999.4474

Criminal Records

Review of Carnegie Mellon Study on Redemption for Purposes of Employment

June 17, 2012

By Attorney Lester Rosen, founder and CEO of Employment Screening Resources (ESR)
One study cited by the U.S. Equal Employment Opportunity Commission (EEOC) in its Enforcement Guidance on the Consideration of Arrest and Conviction Records in Employment Decisions Under Title VII of the Civil Rights Act of 1964 is a 2009 study released by researchers from Carnegie Mellon University. The study by Dr. Alfred Blumstein and Kiminori Nakamura entitled "Redemption in the Presence of Widespread Criminal Background Checks" attempted to devise a model to quantify what most people assume intuitively – that the relevance of a criminal record for employment recedes over time when a person is not re-arrested. The study looks to develop a methodology to measure how much time must pass before an applicant with a criminal record is no greater risk than an applicant without a criminal record. 
The study focused on the problems faced by ex-offenders in obtaining employment due to the perception that once a person has offended in the past, they are more likely to do so in the future. The study sought to explore empirically when a past criminal act was no longer relevant so a person could be considered to be “redeemed” for purposes of seeking employment.
The study was based upon data from the state of New York for individuals first arrested within New York State at ages 16, 18, and 20 for three specific crimes: robbery, burglary, and aggravated assault. The study was seeking to establish two values.  First, the study was aimed at determining how many years must pass before the risk of re-offending was no greater than for other individuals of the same age. The second value was how many years must past before the chance of re-offending was no greater than that of the general population that never offended.
The study found that with time, a person with a criminal record was no greater threat than persons without a criminal record. Depending upon the offenses included in the study and the age at which the offenses were committed, the study suggested that after approximately 4½ to 8 years without further arrests, an offender had a minimal risk of re-offending. Of course, the more violent the offense, the longer the time would be required before a person could be considered “redeemed.” Serious crimes such as murder, rape, or child molestation were not part of this particular study, but presumably more serious crimes would have a different result.
Although there is a temptation for the press to take “sound bites” from the findings, the authors were clear that much more study was needed, and there are substantial issues still to be addressed. The authors’ characterization that the study represents a “significant step forward in area where so little is known empirically” is well-taken.
The study does not have nearly enough data to reach conclusions from which policy recommendations can be effectively made, and has a number of drawbacks that can affect its reliability. Many of these limitations were acknowledged by the authors.
  • First, the study was based only on records from New York State. Records maintained by federal law enforcement were not available, and it is clear from the study that the authors wanted to examine federal records and those of other states as well.  
  • Second, the study notes a number of issues that the researchers indicated needed further consideration. For example, the researchers were unable to determine whether offenders re-offended out of state. Without access to that data, it is certain the study was not totally accurate in terms of how many re-offended, which would have skewed the result in favor of ex-offenders.  
Another very critical issue raised by the researchers is that further study is needed to differentiate between arrests only and those cases that ended up with a conviction. This is most critical since overall it would appear cases resulting in a conviction would be more serious than an arrest only. Including individuals who were never convicted could significantly skew the time required for a criminal record to lose relevance to a number that is artificially low. 
In addition, arrests not resulting in convictions are not generally as relevant to employment decisions, since the FCRA limits the reporting of arrests to seven years and many states prohibit reporting arrests without convictions at all. Even if arrests are considered, the EEOC cautions employers to give the arrest no weight unless the employer can ascertain the underlying facts. 
Another critical area the authors indicate needs to be taken into account in further studies is the amount of time an individual spent in custody. If someone committed a serious crime five years ago and just got out of custody, it is doubtful the person could be considered “redeemed” for purposes of all types of employment.
Yet another issue is the fact more studies are needed to research a much wider range of offenses before coming to conclusions that can be used as the basis for policy recommendations. 
In reviewing the study, it appears there are two unstated assumptions that need to be folded into the analysis as well. First is the apparent assumption that criminals in 1980 are the same as criminals in 2009. The problem with this assumption is there has arguably been an explosion in streets gangs, drug sales, new potent drugs, and gun violence in the past few decades. Before utilizing a study for policy changes, this question should be explored.
The second assumption is just because a person was not re-arrested that consequently he or she has stayed clean and is closer to a point of redemption. This overlooks the fact there are large numbers of unsolved crimes, and not everyone is arrested for every crime that he or she commits. It is entirely possible these individuals committed new crimes for which they were not caught. Of course, it is also possible that people with no criminal records may have committed crimes where they were never arrested as well, but one cannot assume without additional study that both groups get away with crimes at the same rate. 
The bottom-line is these drawbacks suggest it is entirely possible that the actual point of redemption may well turn out to be substantially longer than suggested in this initial study. This underlines, once again, it is extremely premature to draw policy conclusions from this one study. 
The study did recite in general terms why employers were concerned about hiring individual with criminal records, but perhaps not to the degree needed. When bad hiring decisions are made, horrendous harm can be done. The “Parade of Horribles” resulting from hiring an unsuitable candidate can range from workplace violence that results in death, grave harm, or other serious matters, to lawsuits that could ruin a business. 
Although alluded to in the study, attention should be directed to the fact there are a substantial number of state laws which in fact protect applicants from the unfair use of past criminal records. The study reports correctly that the availability of past convictions has risen dramatically due to computerization as well as the number of background checks demanded. It is equally important to emphasize the tremendous growth in state laws that limit the use of criminal records, as well as laws that prohibit the use of criminal data not resulting in convictions or cases that were judicially set aside under state law.
Ironically, the use of a mathematical approach taking into account the type of crime and age when first committed could result in discrimination.  The EEOC clearly discourages automated decision-making regarding the use of criminal records. As stated earlier, the EEOC Guidance suggests employers should not automatically reject an applicant on the basis of a conviction, but should consider whether there is a business justification by taking into account the nature and gravity of the crime, the nature of the job, and the time that has passed.  Assigning a numerical value to a person based upon a mathematical formula and treating that person accordingly may not be an improvement over the current system and in fact may be a step backwards.
The study also notes that the time period can be affected by the risk that an employer is willing to tolerate. What that appears to be driving at is the notion that risk is related to the nature of the job, which is one of the EEOC factors. Jobs with higher risk may well require a greater period of time. For example, a petty theft committed three years ago would logically have little to do with a job on a refrigerator manufacturing line. It may however affect if a person should handle a cash register. In other words, the amount of time required for the point of redemption also relates to the risks associated with the job.   The problem with this general area of research is that there is a lack of attention paid to studying how to assign the risk associated with all of the various jobs available in the U.S. Further research on risk factors would appear to be a requirement before policy can rationally be changed.
If anything, this study demonstrates that trying to predict a future “lack of dangerousness” is just as difficult as predicting future “dangerousness.” This study is critical since this entire area of research is at an embryonic state. Much more study is needed before a broad general rule can be articulated across all situations and all jobs that after a certain time, a past criminal conviction is statistically not relevant. 
The bottom line is that is it premature to contemplate policy changes based upon one study where the meaning of the data is not fully understood and where a great deal more work needs to be done. There are some policy recommendations in the study, which are consistent with a number of recommendations commonly discussed concerning this issue. However, this study is not an appropriate vehicle at this point for policy recommendations. 
The policy recommendations are also rendered less effective due to an apparent lack of appreciation by the study as to how certain things actually work in "the real world."  For example, the study discuses that certain sealed or expunged record that occur earlier in time and are no longer relevant need to be removed.  In fact, a large number of states already prohibit the use of such records, and they are removed by background screening firms from reports.  The authors discuss issues related to problems with setting aside convictions and the problems related to concealment of records. They do not appear to understand that in reality, public records are rarely physically destroyed. The records are still generally available even if there is a pardon or some other judicial set aside.
The current protections afforded to applicants are laws that either prohibit screening firms from reporting the record or prohibit employer from using them. For the most part, it is the background screening industry that protects applicants by filtering out records that are not reportable. A number of states already have some sort of seven year rule, and there are numerous restrictions as to what employers can consider.  These issues underscore the need for an interdisciplinary approach which includes experts in background screening and public records to be included in these discussions.
It should be noted that most background screeners take great effort to educate employers that criminal records cannot be used automatically to prevent employment, and the screening industry’s trade association has in fact donated money to programs to assist ex-offenders to re-enter society. No one denies that everyone needs a job to be a taxpaying, law-abiding citizen and that as a society, it is better to build more schools and hospitals than prisons. However, not every applicant is suited to every job, and the role of the screening industry is to advise employers of the facts so they can make intelligent selection decisions. 
The study Redemption in the Presence of Widespread Criminal Background Checks is available at
© 2012 Employment Screening Resources (ESR)
Certified with the National Association of Professional Background Screeners ESR's SOC 2 Audit Report confirms it meets high standards set by the American Institute of Certified Public Accountants (AICPA) for protecting customer information PCI-DDS Compliance Privacy Shield Framework Services Perfromed in the USA Safe Hiring Manual Founding Member of the National Association of Professional Background Screeners