A randomised experiment to test for gender discrimination

Alison Booth and I have a new paper out, in which we test for gender discrimination in hiring by randomly sending fake CVs to apply for jobs in female-dominated occupations (waitstaff, data-entry, customer service, and sales). These occupations are about 70-80% female.

We find a modest bias in favour of female applicants. Resumes with a female name get a callback 32% of the time, while those with a male name get a callback 25% of the time.

The paper is forthcoming in Economics Letters, so it’s very short. Here’s the abstract. To get the full paper, just click on the title.

Do Employers Discriminate by Gender? A Field Experiment in Female-Dominated Occupations
Alison Booth & Andrew Leigh
We test for gender discrimination by sending fake CVs to apply for entry-level jobs. Female candidates are more likely to receive a callback, with the difference being largest in occupations that are more female-dominated.

(xposted @ Core Economics)

This entry was posted in Labour Economics. Bookmark the permalink.

8 Responses to A randomised experiment to test for gender discrimination

  1. conrad says:

    You know you’ll never get to the real bottom of this unless you start doing social psychology — I’m surprised that economists have started caring about these sorts of things.

  2. In response to Conrad, isn’t there a case for saying some economists were doing social psychology before the term was coined? A lot of economics seems to be about how people react in the large to particular situations: Adam Smith and self-interest as a driver, the role of incentives.

    I’m reminded by the research of a crack from “Up in the Air” (pretty good movie by the way), where the George Clooney character explains to his critical associate that he doesn’t endorse discrimination, just stereotyping because it’s “quicker”, that is stereotyping as an efficient means of sieving candidates.

  3. conrad says:

    John, I’m really just joking. It really should be a false dichotomy between social psychology and economics — it’s just one that happens to exist, no doubt for reasons like social psychologists calling themselves psychologists (and hence mainly living in psychology departments) and economists calling themselves economists (and hence mainly living in economics departments). Despite being a false dichotomy, this makes a big difference — you can just look at the male-female ratios of people looking at similar problems (at a guess, 80% of people doing social psychology PhDs these days are female, and 80% of economics ones are male — there’s no reason for this based on what they are looking at).

    This false dichotomy leads to a situation where economists are basically replicating stuff done by social psychologists for decades (try doing a Scopus search on Nick Haslam over at Melbourne who’s been done this sort of stuff most of his career if you’re interested. Zimbardo gives great talks on his life’s work also if you can get to see him). This is fine in experiments like this, which are interesting and culturally specific, and less fine when they use bad paradigms that social psychologists should have given up on years ago, presumably because they think that they are good to use because they have been used a lot, not because they actually know what the results from them actually mean (like the IAT task — no one in cognitive land would even consider using this as an implicit measure due to all the possible confounds, and it’s only recently that social psychologists have started to move away from it — this shows people in different areas even in the same departments don’t even bother talking to each other!).

    So the interesting thing about this is that after economists have (re)collected all this data, sometime or other, apart from looking at stuff, someone is going to try and come up with a predictive theory of it (why are human’s racist? What dimensions can we categorize racism on? How does that relate to other higher and lower level aspects of being human?), which is exactly what social psych people have been trying to do for donkey’s years, so it will be interesting to see if they come up with the same thing. This will be interesting, because the way economists tend to go about looking at problems is someone different to the way social psych people do, so perhaps they will come up with something somewhat different.

  4. conrad says:

    That should be “somewhat different”

  5. diana kornbrot says:

    wonder if you would get same result if you tried SUPERVISORS for same job descriptions. Male head teachers in primary schools are far more common than proportions of serving teachers in post for same number of years would predict


    ps i find the mutual sociology V economics head bashing/job protectionism uninformatibe and unedifying

  6. conrad says:

    “i find the mutual sociology V economics head bashing/job protectionism uninformatibe and unedifying”
    There are unseen advantages to it. These are (you probably know both already):
    1) Psychology is often far more “effect driven” than economics. So if you looked at the same problem, psychology people might think that the 50 effects that people have found in the domain are important, whereas the economics people are far more quantitative, and often use analyses rarely used in psychology. They also don’t seem to care about single effects nearly as much. So in some cases it is really interesting to see what conclusions both groups come to on similar problems or even what both groups look at (the education data is a really good example of this where you have a three way split between sociology, psychology, and economics).
    2) Some well developed areas get stuck in a certain orthodoxy, which makes it really hard to publish in those areas if you don’t share it. If you have essentially the same problem going in two different literatures of which most people only read one, it means that people with less orthodox but equally as legitimate views have a better chance of being published (a good within-psychology example of this is the “self” literature. I believe the Americans think we should have a two-part self, and the Europeans think we should have a three-part self. Apparently it’s almost impossible to publish stuff in the literature where your view differs).

  7. Simon Eagleton says:

    Interesting – I was wondering why this article is downloadable from The Age under a sub-headline to the missing Rockefeller story on the front page at theage.com.au…

  8. Cathal Kelly says:

    I don’t speak, um, either version of Belgian, but these two short videos came my way today.

Comments are closed.