The most used and perhaps in many cases the only used method of the first level candidate screening is sifting a CV. And this first level sifting being done by a person who barely knows anything about the role, let alone the expectations out of the individual, only based on a JD, makes the process extremely unscientific. 

This prevalent unscientific process begs the question whether there is a correlation between candidate success rate and his CV screening success rate. 

The behavioral insights team, an organization working towards applying behavioral sciences to decision making and analytics developed an alternative to human screening of CVs. The platform aims to avoid biases and serve as a scientific prelude to an assessment center. 

The validation of the tool was done by putting 3 sets of candidates through assessment centers; first a set that qualified CV sifting; second that ‘applied’ shortlisted; and third that cleared both criteria. 

The results are not counterintuitive, yet against common practice.

1. There was a positive correlation between applied score and assessment center scores; a negative correlation between CV success rate and assessment center scores
2. The same trend as above for final interview scores – a negative correlation with CV scores
3. 60% of the hired candidates would not have been hired if the first level screening was only through CVs.

The important question here is that what matters – what defines candidate success?

The research in the field, by MacArthur “genius” grant winner Angela Duckworth puts two factors in perspective. First, how soon can you learn once on the job; second, how much effort are you willing to put in once you know what to do. Again, it’s not counterintuitive, it is supported by research; and yet against common understanding of predicting performance.

The CV sifting process by a person unrelated to the task or team accomplishes none of the above two. I think we’re looking at an impending slow death of the CV. Although, as businesses continue to trail science by years, it might take its own sweet time.

Explore the tool here
Watch Kate Glazebrook describe the research by BE team here
Watch Angela Duckworth describe her research here