Imagine facing the choice of taking a drug to treat an illness, but that drug has major side effects. In the end you might take the pill, but do so fearful of the risks but wishing for a safer alternative.
For many years now, ability testing has been ubiquitous as part of volume hiring, particularly those for graduates. Ability tests typically measure verbal, numerical or abstract reasoning tests of one kind or another. The theory is that the tests help spot those individuals who will later make great managers, and research evidence shows they can predict performance, at least in part.
However, ability testing had been much criticised for causing adverse impact. The data consistently illustrates that candidates who are female, from a minority ethnic background or who are older are, on average, less likely to score highly (known as ‘adverse impact’). This is unfair and bad for social mobility. There can also be legal risks, and in the US for instance this has led to their reduced proliferation. This is a bitter pill to swallow if using such tests comes at the price of fairness by reducing the likelihood of women, ethnic minorities or older candidates getting a job.
The solution is not simply trying to design the perfect ability test. Of the many dimensions of performance that could be measured when screening candidates, abilities are actually one of the inputs most susceptible to adverse impact. This is largely because a person’s performance on ability tests is heavily linked to the educational opportunities and social context in which they developed. Or more bluntly, the higher your socio-economic class and the better the quality of education you received, the more likely you are able to do well on the tests.
This dominant approach of screening candidates based on reasoning ability is very narrow. For a given employee population there will be several different dimensions that are critical to success, be that competencies such as resilience, communicating with others, teamwork, creativity or solving practical problems.
The assessment industry has been hooked on selling one narrow solution at the expense of addressing this broader set of competencies which it would be helpful to assess when screening candidates. Using ability tests has become almost become an automatic response for many organisations.
The way assessments are built and priced has been a factor. They are relatively cheap compared with measures of behavioural style, such as personality questionnaires, which currently cost too much per head for high volume use and are often too long and unfocused. Situational judgment tests are another option but need to be created for the context, taking some effort to put in place. Offline options such as interviews are simply too expensive and impractical at high volumes.
However, there is an alternative. Historically, tests were designed in such a way that they had to be ‘swallowed whole’ – a fairly long assessment for a particular characteristic. It is now possible to select the questions most relevant to each aspect of the role being assessed, validate these against the behaviours that differentiate high performers in that organisation, and then deliver this online. By blending different types of questions – be that behavioural style, situational or ability-based – it is possible to create rounded assessments. These can be relatively short – 25-40 minutes – and cover the full breadth of a role at the screening stage. And it doesn’t cost the earth to do this.
So there is a viable alternative, and now no need to carry on using tests which reduce diversity and cause adverse impact. Ability questions will continue to have a place, but they do not need to be administered standalone in a narrow way which exacerbates their negative effects. A blended approach is the practical solution for volume hiring programmes to not only sift out significant number of candidates, but do so both validly and fairly.
Is it the end of ability tests? May be not yet, but at least it’s the beginning of the end.