There is widespread awareness that decisions made in the recruitment process can discriminate against certain population groups, meaning that proportionately less candidates get through in terms of gender, ethnicity, age and other characteristics. Whilst many processes are ‘fair enough’ to be within the legal burden of proof, this is still quite a distance from achieving 100% fairness and inclusivity.
There are essentially two main causes of adverse impact in assessment. Firstly, systematic bias in screening tests, for example cognitive testing, and secondly, unconscious bias in interview and assessment exercise decisions.
In this article, we’ll look at how you can make sure online assessment is genuinely fair and inclusive.
Ability tests measuring verbal, numerical or abstract reasoning skills have long been a staple part of volume hiring, particularly in the graduate and emerging talent market. In theory, these tests help identify those individuals who will later make great managers, and research evidence shows they can predict elements of performance. However, this form of testing had been much criticised for being particularly prone to adverse impact in terms of gender and ethnicity. Cut-offs are often kept low on these tests to ensure the impact is within legal limits, however we still find systematic differences between groups around these abilities.
The fundamental problem though is driven by how these tests are being used as a screening tool. If we look at the competencies needed to be successful in a role, it’s not just cognitive problem-solving that matters, but may also include resilience, empathy, leadership skills, teamwork, creativity, etc. So, while cognitive ability may be predictive of performance, in terms of being inclusive, often organisations are screening on the one criteria that causes the worst adverse impact!
An increasingly popular solution is blended assessment. This focuses on the range of competencies you need in your candidates, combining situational judgement, behavioural style and cognitive elements in one assessment. Whilst a particular group may on average score a little lower on a couple of competencies, they may score higher on others so these affects average out.
A blended approach can greatly mitigate adverse impact compared to traditional tests, whilst giving sharp and focused prediction of what matters.
For example, a recent blended assessment project for a popular financial services graduate programme showed that although over twice as many males as females applied to the programme, the females’ success rate was similar to that of males – with a female pass-to-application rate of 47% and a male pass-to-application rate of 53%.
The second major issue is the effect of unconscious bias in interviews and assessment centres. As these activities are often paper-based, the results tend to receive little scrutiny, however they can be a significant source of adverse impact.
For instance, an interviewer may make assumptions of someone ‘looking right for the organisation’ or they had been at a previous employer or university which is ‘a good place to have been’. These assumptions are usually unconscious short-cuts people use to make decisions, but can drive decisions which are not inclusive.
So, what can be done about unconscious bias? The first step is to conduct recruiter and hiring manager skills training. Short workshops focused on this content can be an effective way to raise awareness and improve how judgements are made. Secondly, organisations moving from paper-based to digital assessment centres or interviews have a great opportunity to use the data they capture to identify where the inclusiveness of assessor judgements can be improved.
Achieving inclusive recruitment is sometimes seen as the holy grail, but there are very practical ways this is possible if we move forward from the traditional approach to assessment and embrace blended, digital assessment.