Alison Booth, Elena Varganova and I have just released a study comprising three experiments to gauge racial and ethnic discrimination in Australia.

 

Does Racial and Ethnic Discrimination Vary Across Minority Groups? Evidence From Three Experiments
Alison Booth, Andrew Leigh & Elena Varganova
We conducted several large-scale field experiments to measure labor market discrimination across different minority groups in Australia – a country where one quarter of the population was born overseas. To denote ethnicity, we used distinctively Anglo-Saxon, Indigenous, Italian, Chinese, and Middle Eastern names, and our goal was a comparison across multiple ethnic groups rather than focusing on a single minority as in most other studies. Our main experiment, an audit discrimination study, involved sending over 4000 fictional resumes to employers in response to job advertisements. In all cases, we applied for entry-level jobs and submitted a CV showing that the candidate had attended high school in Australia. We found economically and statistically significant differences in callback rates, suggesting that ethnic minority candidates would need to apply for more jobs in order to receive the same number of interviews. These differences vary systematically across groups, with Italians (a more established migrant group) suffering less discrimination than Chinese and Middle Easterners (who have typically arrived more recently). We also conducted two additional experiments to form a more nuanced picture of prejudice. These were a ‘Return to Sender’ experiment and an Implicit Association Test. The results from both experiments reveal societal prejudice against minority groups, although the ranking sometimes differs from that in the audit discrimination study.

(see also a media release, brief summary of results)

A few words about each of the experiments.

The Jobseeker Study

There were several things that surprised me about the jobseeking experiment:

• Indigenous applicants seem to face less discrimination at the interview stage than Chinese or Middle Eastern applicants. To get the same number of interviews as an applicant with an Anglo-Saxon name, a Chinese applicant must submit 68% more applications, a Middle Eastern applicant must submit 64% more applications, an Indigenous applicant must submit 35% more applications, and an Italian applicant must submit 12% more applications.
• Discrimination against Chinese and Middle Eastern jobseekers is highest in Sydney, and lowest in Brisbane. As a Sydneysider, this rather dispelled my notion of tolerant Sydney vs redneck Brissie, but it’s entirely consistent with the literature on migration threat, which finds that an influx of migrants increases prejudice. In the long-term, Sydney may end up more tolerant, but the short-term effect of being the number one destination for immigrants is a rise in prejudice.
• Even in data-entry jobs, there is substantial discrimination against non-Anglo applicants. This indicates that it can’t just be customer-based discrimination, but must be either driven by coworker or employer biases.

For some real-world evidence to back up the results of the first experiment, here’s a quote from one jobseeker:

“After completing TAFE in 2005 I applied for many junior positions where no experience in sales was needed – even though I had worked for two years as a junior sales clerk. I didn’t receive any calls so I decided to legally change my name to Gabriella Hannah. I applied for the same jobs and got a call 30 minutes later.”
~ Gabriella Hannah, formerly Ragda Ali, Sydney

The chart below shows how Australian discrimination in 2007 compares with similar studies that have been done in other places and at other times. The way to read the vertical axis is that 1 means the minority candidate must submit just as many more applications to get the same number of interviews, 1.5 means the minority candidate must submit 50% more applications to get the same number of interviews, and 2 means the minority candidate must submit twice as many applications to get the same number of interviews.

So it’s worse to be Middle Eastern in Australia in 2007 than in Sweden in 2005, but it’s better to be Indigenous in Australia in 2007 than African-American in the US in 2001. Comparing our results with the 1986 study, it doesn’t look as though discrimination in Australia has fallen over time.

The Return to Sender Experiment

The second experiment was a return-to-sender experiment (a suggestion of my wife). By sending out mis-addressed letters to over 2000 Australian households, we effectively give them two options: to return the letter (costly, but the right thing to do), or put it in the bin (cheap, but slightly naughty). On average, 53% of letters with Anglo names were returned, but only 48-49% of letters with Chinese, Italian or Middle-Eastern names. In other words, 1/20th of Australian households would return a letter if it had an Anglo name, but not if it bore a Chinese, Italian or Middle-Eastern name.

The Implicit Association Test

I’m particularly indebted to the blogosphere for help on the third experiment, which is an Implicit Association Test (anyone who is curious can still take it at http://iat.org.au/). Among the bloggers who kindly directed readers towards the IAT were Larvartus Prodeo, Core Economics, Possum Pollytics, Andrew Norton, Ambit Gambit, Oz Politics, and Club Troppo.

Of course, those who read blogs aren’t a representative sample of Australians, so we reweighted the sample in two ways. The first is pretty standard – we just created demographic weights (making the age-sex-education-birthplace composition of the sample match the general population). But our second technique was more novel: we asked all IAT respondents some questions about their explicit prejudice, and then created prejudice weights, exploiting the fact that the same questions were asked of a nationally-representative sample of respondents in the 2007 Australian Survey of Social Attitudes. So far as we know, we’re the first to have combined implicit and explicit prejudice in this way.

And the results of the IAT? While some people are pro-minority, the average respondent showed a subconscious bias against each of the four minority groups. Here’s the distribution.

 

(xposted @ andrewleigh.com)