Suppose you had a house to rent and you were deciding among a number of interested parties, including married and single people. Suppose, too, that the married and single people were the same with regard to their jobs, interests, and what they like about the house. Do you think your choice of tenants would have anything to do with their marital status?
That's what Wendy Morris and Stacey Sinclair and I investigated in a series of 4 studies (with two variations on the fourth). In addition to tapping the usual participants in laboratory research - college students - we also recruited people who actually did work for rental agencies for one of the studies (Study 3).
Participants in each study read brief sketches of each of the interested parties, specifying the applicants' employment, pastimes, how they found out about the house, and what appealed to them about it. There were several versions of any given profile, varying only in the marital status of the applicants. So people reading sketches of married applicants were looking at the exact same information as the people reading profiles of single people, except for marital status.
In the first study, the choices were a married couple, a single woman, and a single man. If there were no bias, each would have been selected 33% of the time. Here are the actual results:
70% married couple
18% single woman
12% single man
The married couple was favored overwhelmingly. (This result and all the others I will describe are statistically significant.)
Perhaps the married couple was favored because there were two renters; people making the decisions could have figured that two people would be more likely than one person to have the money to cover the rent. So in the next study, the choices were a married couple, a cohabiting couple, and a pair of opposite-sex friends. Here are the results:
80% married couple
12% cohabiting couple
8% pair of friends
Again, the married couple was favored by a huge margin.
Study 3 was the same as Study 2, except this time the people making the decisions were actual rental agents. Here are their choices:
Study 3 (participants were rental agents)
61% married couple
24% cohabiting couple
15% pair of friends
The rental agents were not as wildly biased as the college students were, but they still favored the married couple by a very substantial margin.
Well, what if they assume that a married couple is more likely to stay together (and less likely to break the lease) than the cohabiting couple? In the first version of Study 4, we made it clear that both the married couple and the cohabiting couple had been together for the same length of time - 6 years. Here's what happened:
79% married couple together for 6 years
21% cohabiting couple together for 6 years
Again, the married couple was the overwhelming choice.
We tried one more variation. This time, we said that the cohabiting couple had been together far longer than the married couple - 6 years, as compared to 6 months. Now look:
71% married couple together for 6 months
29% cohabiting couple together for 6 years
The length of the relationship did not wipe out the bias. The unmarried couple was still the target of clear discrimination.
In each of the studies, we asked the participants to rate their expectations for each of the potential tenants. For example, we asked them how likely each tenant would be to keep the house clean, damage the house, make noise, pay their rent on time, or break the lease. In all instances except two, the married applicants were rated more favorably than the singles. The exceptions were the single woman and the cohabiting couple when they had been together longer than the married couple - both were rated just as positively as the married couple.
Notice that the more negative expectations for the single tenants (e.g., that they would be less likely to take good care of the house and to pay on time) could not fully account for the results. That's because even when the singles were perceived just as positively along those dimensions as the married couple was (in the case of the single woman and one of the cohabiting couples), the married couple was still favored as a renter by a very wide margin.
We also asked participants in each study to answer in an open-ended way (no rating scales) why they chose the applicant they did. By far, the most popular answer was that the choice was based on marital status. In each study, of those who chose the married couple, between 42% and 84% said simply that they choice the married couple because they were married.
If it is not obvious what's wrong with that (and I've presented this work in the past to very smart people who did not get it), imagine that the choices were between African-American and white tenants, and the decision makers said that they chose the white applicants because they were white. Wouldn't happen. Yet, across 3 studies and 2 variations of a fourth, participants unabashedly said that they chose the married couple because they were married.
There was one more study in this journal article in which we described an example of blatant discrimination against a single person, or some other more widely-recognized target of discrimination. We wanted to know whether people would recognize and object to the discrimination against singles to the same extent that they noticed and lamented discrimination against other groups. I'll tell you about those results in my next post.