Yes, the correlation with task performance was .57 (95% CI = .43 -.68), yet the relation with OCBs was much smaller, OCBO= .30 (95% CI = .12 - .46) and OCBI .25 (95% CI = .07-..42). Although poor credit scores might be thought to predict organizational deviance, it wasn't significantly related to either production deviance or interpersonal deviance.
Yes, it was a very small convenience sample (n = 113 for the outcome correlations) and only one study, which is why I don't understand why it was published in JAP, despite the novelty of the topic. At best, it was a nice pilot study and a motivation for the authors to have conducted a better study before the results were published. The premiere journals in I/O seem not to care about sampling. That's too bad. It will be interesting to see if the .57 does hold up, but there is reason to believe it won't. Even if there is an effect in the population (which hasn't been defined), small sample studies that can muster a statistically significant effect will generally overestimate the effect size (Schmidt, 1992). And the fact that it is a sample of convenience and used only one available measure of performance further undermines the study's ability to provide a population assessment of predictive validity. Bringing up causality I guess is a moot issue because prediction is really the only important issue for selection tests. Nonetheless, are poor finances really the cause of poor job performance or is the relation spurious due to more likely common causes not assessed in this study? Perhaps a history of marital problems resulting in a recent divorce or an ill spouse or child or an ill parent or mental health problems or poor economy and job insecurity or.... Can't many of these issues culminate in a low FICO score and poor performance? Could these common causes be used in the selection process?
I also am an interested observer with a high credit score and no vested interest in selection tests. Nonetheless, I agree with efforts to ban the use of credit scores as selection tools regardless of their ultimate putative validity, at least for most jobs. Policy is often made without data or in direct contradiction to it. Often for the wrong reasons, but not always. In this particular case, I think the validity of credit scores as predictors of job performance is generally a moot issue. Perhaps they have some utility for those tending the corporate till. For example, even though off-the-job drug use and drug tests do not have any meaningful causal relation to job performance, broadly defined, they seem like a reasonable test for police officers and applicants to such jobs. If an individual is going to enforce laws, it is reasonable to hold the person to a higher standard even if recreational drug use would not impair their job performance.
The basic issue is just because personal data are available and their availability to others is beyond a person's control, this doesn't mean that organizations (public and private) have an inherent right to use this information in any situation and for any purpose. It may very well turn out that credit scores have low validities as do many selection tests, such as integrity tests and drug testing.
There was the argument that if credit scores are banned from use as selection tools with a potential validity of .57 or before their validity is even known, then perhaps selection tests with lower validities should also be banned. This is a strawman argument (see paragraph above). The issue is not to ban every selection tool with low validities. The issue might be more reasonably framed in two ways. First, the issue is avoiding decisions based on a single test has low validity. If they are all of lowish validity, then we should use more than one of them. Correct? I'm not a selection researcher, so I could be wrong. But certain tests are likely to overshadow others. No matter how well one scores on tests of ability or general intelligence, let an employer get a positive drug test result (marijuana on a Saturday night at home two weeks ago) or a dichotomous test result deeming a person potentially undesirable based on an integrity test, what will the hiring decision be? How many tests were involved in the final decision? Perhaps one test with a poor predictive validity. Second, just because some information can predict performance, does not mean others have an automatic and uncontested right to use it. A legislative decision to ban the use of credit scores for employee selection need not be based in the validity of the test. The issue is the incredibly import one of invasion of privacy, and the narrow conditions under which it might be reasonable to do so. It would be nice if organizations, and particularly the testing industry, could be trusted to police themselves, but this is naive. Though motivations underlying government policy often are affected by governmental vested interests.
Schmidt, F. L. (1992). What do data really mean? Research findings, metaanalysis, and cumulative knowledge in psychology. American Psychologist,
47, 1173-1181.
****************************************************************
Michael R. Frone, Ph.D.
Senior Research Scientist
Research Institute on Addictions
State University of New York at Buffalo
1021 Main Street
Buffalo, New York 14203
Office: 716-887-2519
Fax: 716-887-2477
E-mail: frone@ria.buffalo.edu
Internet: http://www.ria.buffalo.edu/profiles/frone.html
***************************************************************
I was just about to mention the same article. For people who cannot access it, the correlation between supervisor-rated task performance and FICO (credit) scores was .57(!). The correlation between conscientiousness and task performance in the same study was .26. FICO scores also provided (a lot of) incremental validity over personality measures. Yes it was a small sample and only one study, but that's what we have and where we start.
I think it's going to be fascinating to watch the politics of this unfold. We've already got states banning it and Steve Cohen (D-TN) has introduced federal legislation. All of this seems premature with no validity data. If the .57 holds up, should we start banning selection tools with lower validity? Of course, this would be nearly _every_ selection tool. Where does that take us?
Interested observer with good credit and no financial interests in any selection tools,
Tom
*****************************
Thomas A. Timmerman, Ph.D., SPHR
Professor of Business Management
Director, MBA Studies
Tennessee Technological University
Cookeville, TN 38506
931-372-3600
ttimmerman@tntech.edu
*****************************
From: Organizational Behavior Division Listserv [mailto:OB@AOMLISTS.PACE.EDU] On Behalf Of Babatunde Ogunfowora
Sent: Monday, October 31, 2011 10:07 AM
To: OB@AOMLISTS.PACE.EDU
Subject: Re: [OB-LIST] use of applicant credit history
Has anyone sent out this recent, online first paper published in JAP to the group?
An empirical investigation of dispositional antecedents and performance-related outcomes of credit scores.
Bernerth, Jeremy B.; Taylor, Shannon G.; Walker, H. Jack; Whitman, Daniel S.
doi: 10.1037/a0026055