What is a stereotype threat and how does it affect psychometric testing?
What is a Stereotype threat and how does it affect Psychometric Testing?
Our decision making is guided by our cognitive biases, and often without us knowing it. These biases aren’t all bad; they are just mental shortcuts that have been evolutionarily useful. Without them, it’s quite possible that we would be in a constant state of overstimulation and decision paralysis as we try to process the information. Too much information – Not enough time to process it!
One of the most well-known is stereotyping. This is to draw unknowable conclusions about people’s lives or character by observing accessible qualities. For example: taking how they look, dress and talk, to decide how clever or honest they are. Stereotyping has been evolutionarily useful for survival and social advancement, learning how to distinguish a friend or enemy, or knowing who is powerful and should be sucked up to for example. Now though it’s use is often outweighed by its part in creating injustice and poor decision making..
It is well known that stereotyping affects how we view and behave to others, and of course that it is neither fair or nice and we should try and control our biases when interacting with people. What is perhaps more complicated is how stereotypes that apply to ourselves impact our behaviour. In its negative form, it’s called “stereotype threat” and has substantial implications for education, recruitment and cognitive testing.
What is a stereotype threat?
Stereotype threat is where situational factors lead people to conform to the negative stereotypes of the group that they belong to, typically in relation to poor performance in some area. For a stereotype to impact your behaviour you do not necessarily have to believe it, just be aware of it.
Nalini Ambady illustrated this on 5-7-year-old Asian American girls with their math performance.
Before doing a group of maths questions the girls were asked to colour 1 out of 3 pictures.
- The First, a picture of a girl with a doll.
- The Second, Asian children eating from a bowl with chopsticks.
- The Third, a landscape (the control group).
The experiment showed that their performance varied to a statistically significant degree depending on the identity that the children were primed with; either their Asian identity, Asians being stereotyped as good at maths, or female, females being stereotyped as comparatively less strong at maths.
This is not isolated; the effects of stereotype threat have been demonstrated in over 300 published studies. This isn’t limited to academic areas and includes white males sporting performance when put into competition with black men, and women not doing as well in negotiation scenarios or driving tests.
However the most important area for us are how groups that are broadly stigmatised underachieve in classroom exams, even on tasks that have been suggested to be “culture free” and contain “pure” measures of cognitive ability, like standardised tests.
The relevance to commercial recruitment.
There are clearly multiple areas of recruitment where sterotype threat has implications. In our case when it comes to numerical reasoning tests (the stalwart of banks and consulting firms for entry) we know that men pass at a significantly higher rate than women, depending on the pass mark sometimes twice as many. However, we also know thanks to technology like Mapped that it is possible to mitigate these differences dramatically without making the tests any easier, without treating anyone differently and without affecting the performance of other groups.
How we tackle these gaps more broadly requires an understanding of why stereotype threat affects people and why it impacts performance; this is not a settled question but there a couple of plausible ideas.
One theory is that it is related to anxiety. It is an emotional reaction to feeling out of place, lacking confidence in your abilities or expecting to do badly that interferes with performance. Interestingly, these anxieties and worries have been linked to a diminished working memory capacity, which is essentially a short-term way that your brain holds information relevant to current task and is vital for problem-solving. There are concrete ways in which you can mimise anxiety, and make everyone feel at home, even in a digital environment.
Another theory, slightly different to anxiety is something called prevention focus – a state of vigilance whereby the individual is worried about doing something wrong to the extent that speed, flexibility and risk appetite is reduced, (which also impacts creativity). This is vital for most high powered jobs, and the speed aspect is highly relevant to ability testing.
Broadly speaking, not having your identity emphasised, and having a bit of reassurance goes a surprisingly long way (if it’s subtle) and not every individual is susceptible to stereotype threat. It’s something that only becomes really obvious when looking at macro trends (like running a banking graduate scheme with thousands of applicants).
One notable example is being outnumbered. Women showed performance reduction in maths tests when they completed it in the presence of a male majority and, in turn, their performance decreased in proportion to the number of fellow male test takers. This could be a very easy fix for businesses and schools to take in this case.
Overall, whilst we know that these stereotyping issues exist, the key thing to remember is that the disparities that occur in current analytical applicant testing are absolutely not reflective of ability and can be solved. For one, Mapped is an example of an aptitude test focused on diversity which takes measures to ensure that a mix of candidates reaches the final stage of the recruitment process. Less threat, more action.
References to specific articles upon request.
 Stereotype Susceptibility in Children: Effects of Identity Activation on Quantitative Performance Nalini Ambady, Margaret Shih, Amy Kim, and Todd L. Pittinsky Psychological Science Vol 12, Issue 5, pp. 385 – 390
 Stone, Lynch, Sjomerling, & Darley, 1999
 Kray, Galinsky, & Thompson, 2002
 Yeung & von Hippel, 2008
 (Steele & Aronson, 1995; Cole, Matheson, & Anisman, 2007; Good, Aronson, & Harder, 2008; Keller, 2007; Neuville & Croizet, 2007; Good, Aronson, & Inzlicht, 2003; Brown & Day, 2006; Klein, Pohl, & Ndagijimana, 2007).
 (Marx & Roman, 2002, Marx, Stapel, & Muller, 2005)
 Croizet, Després, Gauzins, Huguet, Leyens, and Méot’s (2004)
 Inzlicht & Ben-Zeev, 2000.