Jump to navigation
The Criminal & Pigmentation Marks
R. Alex Whitlock
A few years ago Devah Pager and Northwestern University
lead a study on the employment prospects of felons (a drug-related felony that landed them in prison for 18 months) when it comes to entry level jobs in Milwaukee, Wisconsin, advertised either in the local newspaper of the state's job placement posting. Her group employed four testers, two black males and two white males, who applied for various entry-level positions. The two black guys would apply for a job with one admitting to a felony and one not admitting to a felony and the two white guys did the same.
It's no surprise that the felons had diminished prospects when it came to applying for the jobs. Even when the crime they committed did concern misbehavior that would directly adversely affect their job functions and even though Milwaukee actually has laws against discrimination against ex-convicts. For the jobs they applied to, the results went as follows:
White without a conviction: 34% callback rate
White with a conviction: 17% callback rate
Black without a conviction: 14% callback rate
Black with a conviction: 5% callback rate
The aspect of the study that jumps out at me is not so much the relative callback rates between felons and non-felons but rather the rates between the races. If I'd been asked to guess, I would have figured that blacks would probably be called back between 65-80% or so as much as the whites. I would not have guessed less than 50% in a major metropolitan area and I definitely would not have guessed that
having black skin would be the statistical equivalent of having a criminal record.
It may not be as simple an issue to say "white racists are picking on black applicants" because we don't actually know the races of the people making the decisions and we're not sure why precisely they made the decisions that they did. For all we know all of the recruiting managers were black. The decision may have hinged more on a fear of the work ethic or moral terpitude rather than race explicitly. But I'm not sure how much that matters. The salient point to me is that young black men without a college degree face more of an uphill climb than do young white men in the same demographic and that's important to know regardless of who is doing the hiring. And even if the recruiting managers are more concerned about moral terpitude and work ethic, they seem to be making the determination of the applicant based
almost entirely on race which is a factor the applicant cannot change.
They controlled the study to make sure that the applicants were bright and effective communicators, they made sure that there wasn't a disparity in the resumes, and they applied for the same types of jobs with the same types of recruitment practices in the same areas. The only potential flaw I can see is that the black testers applied for the same jobs as one another but not the same precise jobs as the whites. That could account for some disparity, but given the similarity of profiles of the jobs it seems unlikely that this would account for the wide gulf. Every other objection I can think of was addressed in a section of the study itself.
In regards to the central focus of the study I agree with the study's authors that the seeming inability of ex-convicts to find work is a problem. It's easy to write them off and say that they shouldn't have committed the crime to get convicted in the first place. It's true enough. Nonetheless it creates a drag for society at large to have a segment of society that can't find jobs but has criminal experience. It's setting them up to fail.
I disagree, however, with some of the suggestions that the study makes. They seem to support laws that would allow a potential employee to conceal his or her convictions. I believe that an employee has a right to know and act on the criminal past of its applicants (so I disagree with the Milwaukee law prohibiting it). The right of a potential employer to know who they are hiring trumps a criminal's right to privacy in that regard.
But absent that I'm really not sure what we should do. I'm certainly not opposed to looking into sending fewer drug criminals to prison on a number of fronts, including the disruption that occurs in their employment history. I've seen in other ways how the judicial incarceration-and-release system sets up a class of people to fail. On the other hand I think it's a bit sanguine to suggest that if we just give them a chance they will take us up on it or do less damage. After all, they had a chance before their conviction.
Even though I don't have any firm conclusions, I found both the racial and judicial aspects of the study to be quite fascinating and worth the time it took to read the report. At the very least I will need to hold on to it to use as an object lesson as to the most clear and present dangers of marijuana usage: getting caught.
Observations
Having lived in that city and worked in its job market and having friends in the same market, the study has a major flaw in what it presents. In order to have a proper study, one would have to make sure that each of them matched the application resume nearly exactly in each case, or have all four apply for each job with the random variance.
There exists - not noted in the study, but probable - a difference between which high schools someone lists on a resume. There is a difference in how someone presents themselves; were all dressed in similar manner, or not? Were there noticeable accents? Were all well-spoken, or was slang more apparent in usage on the part of one group or another? All of these would make a noticeable difference for an in-person interview. Without them doing some serious cross-race comparisons (and noting WHO it was doing the interviewing), it'd be hard to say for certain.
I think that the numbers on the question of felony - given the random variance applied - bear out that there is a problem of discrimination against convicted felons, but I don't know that the study's methodology is enough to prove the racial aspects as claimed.
The applicants were controlled -- They specifically found bright, articulate and charismatic people to be the applicants. These people were specifically chosen because they would make a good impression. Unless one thinks that it is impossible to find young black men that can make a good impression or that there will always be young white men that can make a significantly better impression, it's difficult to toss the study on these grounds.
The resumes and applications were controlled -- They specifically calibrated the resumes and even tried to equalize the experience between those that were convicted and those that weren't. For this to be a significant factor it would require that both black students went to an equivalently bad school and most white students went to an equivalently good school (since the team-members performed roughly equally with their partners). So it's possible that's true and it's possible the school that the black kids went to was so bad as to hinder their efforts... but as much as a prison record? I'm willing to believe that the results may be a little exaggerated due to some external factors, but this disparity is *huge*.
The interviewer is irrelevent -- Even if they were all black that doesn't help a young black man looking for a job. And in many cases there weren't even interviews. They just turned in the resumes and only underwent an interview if they were interviewed on the spot. For the most part this study just investigated "first exposure".
The results are apparently pretty similar to studies done elsewhere -- Yes, all of these studies could be frauds. Every last study could have a nugget of a question big enough to push through enough doubt to muster an objection if you are bound and determined that race is not the issue. That's a much further swim than the alternative.
The only major variance is that they were applying for different jobs. But given the sheer number of jobs (200 for the black applicants and 150 for the white applicants) and the following controls I find it extremely unlikely that this huge gulf was simply by chance:
The types of jobs were controlled -- They were all entry-level requiring nothing above a high school diploma.
The locations of the jobs were somewhat controlled -- The urban/suburban split was equal.
The jobs were picked randomly by the administrators from the same source -- It wasn't a matter of each group picking different types of jobs.
I'm not saying this study is absolutely bulletproof, but it's convincing enough that I'm going to need a lot more than "Might not be!" to share your skepticism of it. If you have a study with these kinds of controls wherein little or no (or reverse) racial discrimination is found, please put it forth. I would welcome the good news!
[this comment was slightly modified by its author]
I'm not sure you'll ever see one that reaches that conclusion - there's no grant funding to be had for it, and no researchers will start on a project like this without expecting, and planning for, the "found it" result.
If you're going to do a study like this, you need to be VERY careful on what you do. Little differences - not skin color, but other differences - can mean a big deal to an employer and be completely missed by those doing the study.
I'd have a lot more trust if they'd had all four applying for each job, or if they had had a real random-matching analysis where the pairs were changed around. I'm also not sure about whether their resumes were really matched between the pairs, or who (if not the blacks they sent) were given call-backs; did another black ultimately get the job? A black with college experience, or who was in college, or with other work experience? Part of the odd irony of the city of Milwaukee is that it is a majority-minority area (the suburbs being the reverse).
Now, UW-Milwaukee talks about "stealth depression" in inner-city Milwaukee, where <a href="
http://www.uwm.edu/Dept/CED...">unemployment numbers by race</a> do indeed present a problem. That's why I'd see it interesting to see how many blacks (as opposed to whites) were applying for a given job, and see a better study done on this, because it's my hunch that their black pair were not being passed over for more whites, but quite probably for latinos and other blacks.
/If you're going to do a study like this, you need to be VERY careful on what you do./
Yes, because there are people out there with a vested emotional interest in dismissing the results of any study that comes to conclusions that make them uncomfortable.
Add an Observation
Comment spam is an ongoing problems that we're trying to address. Previously we required people to create accounts and log in. I am thankful to say that is no longer the case. We're giving Captcha another try and are playing around with a text-based Q&A variant of Captcha. So bear with us as we try to figure out how to best get a handle ont he problem. Please note that any comment on a post more than 30 days old will go into the moderation queue, where I will get to it when I can which could be once a week.
