I recently watched “Why hiring is broken” by Kate Glazebrook from TEDxGoodenoughCollege. Yet another person saying resumes are broken (which is true), but this time with research to back it up. They don’t really do a good job predicting performance and they allow biases to creep in.
What good is a resume really? As a recruiter, we only take 6 – 10 seconds to review the resume before deciding if we will consider them further or not. At the same time, what biases are creeping in? For example, research shows that non-white sounding names do get less call backs.
So if we are truly interested in diversity (and not discriminating) and hiring the best applicants, we need comparable information to judge each applicant equally on.
The resume is not very good for this. A mediocre or poor candidate could write a great resume. A great potential high performing employee could write a poor resume. Or it could just start out poorly and through the halo/horn effect we will unconsciously see the rest of the resume in a poorer light. Maybe they are applying to various positions and don’t customize for each position. So would they really be a poorer performer because of any of this?
Per the TEDx talk, there are a few things we could do.
We could take the time to anonymize the applications. Take out people’s names. Take out foreign universities and colleges. Take out the things (all identifiers) that would bias use against someone that does not correlate with performance.
We could chunk the information we are trying to compare, so that we are comparing apples to apples across applications. Basically, ask a question and then simply judge all the applicants based upon who had the best answers to the question.
Harness the power of multiple review, but not too many people in the crowd. Research shows the optimal size of a review team is three people.
Lastly, to get predictive information of how someone would perform, we need to give tests. Simply test what counts. A uniform test for all applicants, so that the answers can be compared. Something that is relevant to the day to day job duties. If you are a software tester, perhaps a bit of code with a several errors and basically seeing who gets the most errors found in the shortest time.
Per the TedX talk, they tested this by blind testing “Applied Sift” vs “CV Sift”. “Applied Sift” showed a strong statistical with future performance (using their platform that used the above key points) VS “CV Shift” (typical resume review) did not show much correlation to future performance. More surprising is that 60% of the applicants hired through “Applied Sift” would have been rejected using “CV Sift” and they would have ended up with a less diverse workforce.
From watching this, I can see potentially a way to process applicants without asking for a resume or cover letter at all. To apply, perhaps candidates would just answer a series of short questions. Nothing too time consuming. Then based on those answers, you could send to the top 10% or so a more in-depth questionnaire and job related testing (i.e. write some code for programmers, source some names via LinkedIn for a recruiter, etc.). Names and other identifiers could be removed and just an application number assigned. Then when you want to start in-person interviewing, then you can get the name of the top candidates and contact information.
You could directly compare applicants purely by the answers they give and judging who gave the best answers to non-vague questions. Basically, comparing apples to apples. This process could take care of much of whether someone is experienced and technical enough – the in-person interviews could be culture fit and soft skills.
I like seeing researchers like this test recruiting processes. Because much of what we do currently may not necessarily have a direct correlation with performance. This is what we should be striving for.
See the Tedx Talk below…