Note-taking for Principles of Empirical Research with Catherine Voulgarides
Pager, Devah. (2007). MARKED: Race, Crime, and Finding Work in an Era of Mass Incarceration. Chicago: University of Chicago Press.
Page 21 of Pager’s book includes this chart, showing annual prison admissions for drugs by race in the United States. In the 1980s, we imprisoned roughly the same numbers of black and white people for drugs. There are about six times as many white people as black people in the population generally, so unless you believe that black people do drugs at six times the rate white people do, there would appear to have been some racism at work.
Then in the late 80s, there was an incredible jump in the number of black prisoners, leading to the present situation, where there are between two and three times as many black people in prison for drugs as white people. While America has become less racist in some respects, this statistic tells us that we are not making as much progress as we like to imagine. These facts also have implications for the history of hip-hop. You can see the rise of both gangsta and overtly socially conscious rap in large part as a response to the devastating effect of this sentencing disparity.
Before the 1970s, the prison population held steady at around one per thousand Americans. As of 2004, it was almost five times that. The total number of people under criminal justice supervision, including parole, is more than twenty people per thousand. That is a hell of a lot of Americans. I was familiar with the immense costs of imprisoning all those people, both in human and financial terms, but I wasn’t aware of the drag that it creates on the economy generally. I was surprised to learn that we don’t count prisoners when we calculate our unemployment statistics. It turns out that America’s supposedly low unemployment rates compared to Europe are substantially a function of our not counting prisoners as unemployed. Oops.
Every year, 600,000 inmates get released. A depressingly large percentage of them end up back in prison, and it’s not difficult to imagine why. Even with the best will in the world, the employment prospects for ex-cons are grim. Everyone can agree that outcomes for former prisoners are better if they’re employed, but the problem is, who’s willing to employ them? Five times as many inmates get released each year as there are new fast-food positions available. Entire sectors are closed off to people with criminal records.
Why aren’t ex-cons employed? Pager lists three classes of explanation.
- Selection: the kinds of people who wind up in prison don’t want to work or don’t have the skills. Incarceration and unemployment are correlated but not causally related, because if prison didn’t exist, these people would be unemployed anyway.
- Transformation: prison experience makes you less employable. It creates gaps in your work history, inflicts psychological trauma, and disrupts personal ties. This is complicated, though, because for some people, prison can also be a stabilizing force, and can offer education opportunities. Research is lacking in this area.
- Credentialing: the stigma of incarceration is the main barrier to employment.
Ex-cons face both formal restrictions to employment and informal ones. Some of them make sense. No one thinks ex-cons should be working in schools or day care centers. But in some states, ex-cons are also banned entirely from working as septic tank cleaners, barbers, or real estate agents. Ex-cons are also barred from the entire public sector. This is especially significant for black people, for whom civil service has historically been a key vector for upward social mobility.
Pager is interested in how racism intersects with employers’ attitudes toward people with criminal records. Racism is socially unacceptable, but discrimination against ex-cons is perfectly fine, because it’s at least partially justifiable. Employers have good reason to be skeptical of ex-cons–the costs of a bad hire are way higher than the costs of missing a good candidate.
How do we study the labor market consequences of incarceration? One method is to conduct surveys of employer attitudes. This is easy to do, and you can obtain large sample sets which you can differentiate out by any number of factors. However, surveys have their limitations. You need to trust your respondents to reply accurately and honestly. People willingly express negative views of ex-cons, but don’t like to admit that they’re racist. Small changes in question wording can get produce wildly different results, which might reflect the fact that employers are ambivalent in their attitudes. In general, surveys are better for comparing attitudes among groups in a relative way, and for showing changes over time, not for showing absolute numbers.
A second method is to use interviews and ethnography. Thick description gets at the complexity of real life, but also precludes isolating specific causal variables. On the more quantitative side, you can do statistical analysis of large-scale longitudinal surveys tracking people over time. However, not much of this data exists, and even when it does, analyzing it doesn’t isolate causal pathways.
Pager’s method for her study is to conduct experiments, using two different techniques. The first is correspondence tests, where you send out pairs of fake resumes, which are identical except for criminal record or whatever other factor you want to test. This is a cost effective and easy technique, but it doesn’t let you test for the effects of race. You can give your fake job seekers stereotypically “black” names, but employers use those as a proxy for socioeconomic status, so you can’t really isolate the racial aspect. Also, the correspondence method only lets you test for jobs for which you apply by mail, but most entry-level jobs require you to apply in person.
Pager’s study also uses in-person audits, with actors posing as job applicants. This technique is not easy at all. It requires a lot of casting and training of testers. However, the results can be quite nuanced. You get wonderfully generalizable aggregate data, though you can’t establish attitudes of any one employer since they’re usually only considering one “applicant.” In-person audits are considered the gold standard for this kind of research, but the method does have its vulnerabilities. It’s hard to control for all of the variables. Experimenter effects are a risk as well; if tester expects to be treated badly by a particular employer, it can become a self-fulfilling prophecy. This risk can be addressed by training, and by comparing in-person with the results of correspondence tests to see whether they differ.
Pager’s results are show predictably grim outcomes for ex-cons. Having a criminal record reduces job callbacks by fifty percent for white testers, and sixty percent for black testers. Any kind of contact with the criminal justice system adversely affects your employment prospects. Being acquitted of a crime doesn’t harm you as badly as a conviction does, but it still makes you less likely to get a callback.
The really surprising and demoralizing result is that a white applicant with a felony drug conviction is more likely to get a callback than a black applicant with no criminal record at all. Let that sink in. Personal contact with employers makes black non-offenders way more likely to get a callback because it makes employers see a specific person, not just a stereotype. But personal contact actually makes black offenders less likely to get a callback. We have a lot of work to do, America.