Cognitive ability tests can harm diversity. Our data show that differences in scores across sociodemographic groups its not inherent to ability, but instead phenomena like responses to competitive environments, risk appetite, and stereotype threat.
This can distort people's performance and changes the structure of your whole company.
That's why for our ready-to-use tests, like Mapped and the Numerical Skills Test, we use behavioural expertise to mitigate the potential for bias against traditionally under-represented groups.
Here are some of the ways we embed this behavioural expertise into the test-taking experience:
Behavioural design principle |
Why? |
The principle in action |
---|---|---|
A testing environment that encourages practice before taking the test
|
When people can practice for a test, they can learn what they're good at, as well as their weaknesses. Also, for people with fear or anxiety towards tests, practice allows them to feel more comfortable and set expectations around different components of the test (e.g. nature of the questions, how the test works with their computer and internet quality). All these are key to avoiding surprises when taking a test. There is also evidence that topic feedback correlates with improved learning. |
All candidates are encouraged to take a practice test that works like this:
|
Ease-up instructions that mitigate for stereotype threat and competitive settings
|
Evidence suggests that certain under-represented groups perform less well in threatening testing settings due to ‘stereotype threat’. Numeracy tests, for example, have been associated with gender biases due in part to social expectations that females are less competent in maths. Research has shown that such stereotype threat can be countered with the most subtle cues. For example, ‘values affirmations’ which seek to focus on positive, rather than stereotypically negative aspects of people's identity help to mitigate for stereotype threat. Numerous studies also show that, relative to men, women are more likely to opt out of competitive settings, and tend to perform less well when in competitively-framed environments. This is particularly the case where performance is assessed individually rather than in teams, and in settings where women have traditionally been a small minority. Test settings that emphasise competition by referring to needing to perform in the top X% of the cohort, can be expected to exhibit a higher opt out rate by women and reduced performance. |
All tests start with ease-up instructions that contain the following:
|
Levelling the playing field between risk takers and risk avoiders |
In test taking scenarios, female test-takes are more likely than men to skip questions than to offer a guess. This leads to a gender gap in test performance, as it's statistically shown that is better to guess an answer than to skip a question. This becomes more pervasive when wrong scores scores are penalised as it has negative consequences on risk averse candidates. Women have been found to be in general more risk averse than men. |
The design levels the playing field between risk takers and risk avoiders by doing the following:
|
Balancing the trade-offs of timed tests |
People value the use of timed tests because they resemble the reality of deadlines in life and careers. But there is also research that shows that time pressure creates anxiety and may lead to long term fear for a subject. For example, there's research that shows that women underperform men in competitive settings whenever two types of pressure are present: task stereotype (math task) and time constraints. |
The testing environment is designed in the following way:
|
Comments
0 comments
Article is closed for comments.