Researchers Tried To Explore Why “Stereotype Threat” Harms Performance, But Found It Didn’t Harm Performance At All

Is stereotype threat too context-dependent to matter all that much?

By Jesse Singal

Stereotype threat is a very evocative, disturbing idea: Imagine if simply being reminded that you are a member of a disadvantaged group, and that stereotypes hold that members of your group are bad at certain tasks, led to a self-fulfilling prophecy in which you performed worse on such tasks than you would otherwise.

That’s been the claim of stereotype threat researchers since the concept was first introduced in the mid-1990s, and it’s spread far and wide. But as seems to be the case with so many strong psychological claims of late, in recent years the picture has gotten a bit murkier. “A recent review suggested that stereotype threat has a robust but small-to-medium sized effect on performance,” wrote Alex Fradera here at the BPS Research Digest in 2017, “but a meta-analysis suggests that publication bias may be a problem in this literature, inflating the apparent size of the effect.” Adding to the confusion are some results which seem to run exactly opposite to what the theory would suspect, like the one Fradera was reporting on: In that study, female chess players were found to have performed better, not worse, against male opponents, which isn’t what the theory would have predicted.

Now, another study is poised to complicate things yet further. In a paper to be published in the European Journal of Social Psychology, and available as a preprint, a team led by Charlotte Pennington of UWE Bristol recruited female participants to test two mechanisms (reduced effort and working memory disruption) that have been offered to explain the supposed adverse performance effects of gender-related stereotype threat. They also compared different ways of inducing stereotype threat. Interesting questions, you might think, but in all cases the researchers came up empty.

Pennington and her colleagues’ study consisted of a pair of experiments in which they asked women to complete eye-tracking (both experiments) tasks, intended to test their inhibitory control, and a mental-arithmetic task (the second experiment only) designed to test their working memory. The participants were split into groups and, prior to their completion of the task, given various stereotype reminders — some, for example, were informed that some research shows “that gender differences exist on visuospatial and mathematical tasks,” and that “females are shown to perform less accurately compared to males.” These reminders were used to provoke two different kinds of stereotype threat – one was “group-as-target”  designed to induce a fear that one’s own performance may be used justify stereotypes about women’s inferior performance; the other, a “self-as-target” threat, involved mentioning that women typically perform worse on the task than men, but with the focus of the results supposedly on their own personal ability. In the second experiment, a group was exposed to a positive stereotype claim, to see if that would improve performance, and in both experiments, a control group wasn’t exposed to any stereotype-laden messaging at all.

The results of both experiments are easy to sum up: Pennington and her colleagues detected no differences between the groups on the key measures they were testing, either in terms of reduced effort or signs of working memory interference. Simply put, neither experiment provided any evidence for a stereotype effect in this setting, let alone revealing anything about the possible mechanisms or contrasting effects of group- or self-threats. Of course, plenty of other studies have provided evidence for such effects. It’s complicated.

One paragraph in Pennington et al’s paper shows just how complicated:

Research has revealed many factors that heighten individuals’ susceptibility to stereotype threat. From a methodological viewpoint, performance decrements are more likely to occur under stereotype threat when the task is difficult … However, a recent meta-analysis casts doubt on task difficulty as a significant moderator of stereotype threat … It is also proposed that stereotype threat effects are more likely to emerge when individuals attribute worth to their social group membership (i.e., group identification…), endorse the stereotype to be accurate (i.e., stereotype endorsement…), and identify strongly with the stereotyped domain … Nguyen and Ryan (2008) report in their meta-analytic review, however, that women with moderate relative to high domain identification are more affected by gender-maths stereotypes, with research further suggesting that individuals do not need to identify with the stereotyped domain or group to experience stereotype threat .… The complexity of these observed findings has led some scholars to theorise that individuals may experience unique forms of stereotype threat, which are moderated by different factors and underpinned by diverse mechanisms … [References, removed here for ease of reading, can be found in the preprint].

If what all these messy, sometimes contradictory findings suggest is true, and stereotype threat, to the extent it has an effect, is moderated by factors like explicit endorsement (or not) of the stereotype in question, or domain identification (or not), and so on – then who is to say we’ve come close to figuring out all the possible moderators? At a certain point, the simple storyline that caused the concept to spread so virally in the first place just doesn’t quite hold anymore. It’s not “You will be affected by X,” but rather, “You might be affected by X, depending on A, B, and C.”

That’s the nature of good science, of course: We hone in on the truth by figuring out the complicated ways a given phenomenon works. But sometimes this process is at odds with science communication, and can lead to preemptive, oversimplification popularisation and application. If stereotype threat is so complicated that “individuals may experience unique forms of” it, then what are we left with, exactly?

Stereotype Threat May Not Impact Women’s Inhibitory Control or Mathematical Performance: Providing Support for the Null Hypothesis

Post written by Jesse Singal (@JesseSingal) for the BPS Research Digest. Jesse is a contributing writer at BPS Research Digest and New York Magazine. He is working on a book about why shoddy behavioral-science claims sometimes go viral for Farrar, Straus and Giroux.

Article source: