The hiring teams have one of the toughest jobs in the organization. They have a herculean task of closing requisitions at the wild pace at which businesses open them. Added to that, they also face immense pressure to get the right people into the organization. Given the circumstances, it is only fair that to bring analytics into the picture to make the lives of talent acquisition professionals a little easier.
When we think about hiring, we usually think of it as finding the right match for the job and the organization. Analytics thinks of hiring as ‘predicting performance’. What analytics tries to do is predict the candidate’s probable job performance given the environmental conditions.
Let’s start by looking at methods that organizations use to help predict performance. Structured or unstructured interviews, job knowledge tests, personality tests, reference checks, cognitive ability tests and work samples are a few approaches to detect potential hires. It is important for the hiring teams to know which of these are most effective. A quick look at the co-relation graph of some of the most widely used methods will tell us that work samples are the best tools with structured interviews and cognitive ability tests come a close second. Reference checks are the worst predictors of performance.
A closer look at the numbers reveal a more disturbing truth. Even our best predictors of performance have a co-relation factor of only 0.54. So how do we make sure that we are screening in the right candidates?
Ryan & Tippins (2004)
Analytics helps us answer this question by going beyond tests and helping us analyze what makes people effective in a particular job. It gives us an idea of the variables that we are trying to predict through our tests. It is important to do this since what predicts performance also varies from organization to organization. Google had been famous for asking for college transcripts only to realize that once a person had been out of college for more than a couple of years, GPA was a very poor predictor of performance. On the other hand, an investment bank realized that GPA for them was a very strong predictor of performance.
A good way to begin the search for the right variables is by looking at the characteristics of the best and worst performers in your organization and testing those characteristics for statistical significance. It is of course better to compare apples with apples and thus compare characteristics within the same cohort and job. A step further and something most organizations should look to do is leverage the power of multivariate regression to try to separate influences of different characteristics. This of course isn’t 100% accurate since a part of the sample set would have left the organization. One could apply selection correction to account for those who were hired and those who left the organization. However, this step is difficult and extremely sensitive to some statistical properties and hence one could stick to multivariate regression.
But there is bad news. The best combination of various tests and methods still leave much of performance unexplained. At best, we may be close to 30-40% accurate. There is worse news and it is that these tests & methods are still better than human judgment (Ryan and Tippins,2004).
I’d like to end with a story from Daniel Kahneman’s book Thinking Fast, Thinking Slow about his first job with the Israeli defense forces. His job as a psychologist was to try to assess candidates for officer training. The idea was that psychologists would observe how candidates would perform and behave in the execution of a given task and accordingly take a call. The interesting thing is that the team knew how good their assessments were and discovered that there was absolutely no co-relation between who they though would do well based on a seemingly good set of tasks and how they actually performed. What was scary to Kahneman was that even though they knew that the tests were in no way predictive, yet for each individual, they felt very confident in making the prediction. He calls this the illusion of validity and it is something that we must all be wary of.
While we all realize the immense potential of the application of analytics try to remember that they are only as powerful as we allow it to be.
HRBP and blogger at HRpartnerstory
Ankita Poddar is an HR professional based out of India. Ankita’s experience as an HRBP gives her an opportunity to interact closely with the business leaders, innovate and execute running programs in the field of employee engagement with focus on rewards & recognition, communication, performance management, incentive schemes, ESAT surveys and more. Ankita holds a Post-Graduate Diploma in Management with a focus on Human Resources and a Bachelor of Engineering in Mechanical Engineering.