Many small companies choose their new employees based on a single interview. Unfortunately research shows that interviews have extremely poor validity when it comes to selecting the right employees. (In case you've forgotten, "validity" means an instrument measures what it's supposed to measure). Interviews are supposed to measure the likelihood that a candidate will be a good fit for our organization and our job vacancy. Casual interviews are pretty good at measuring interviewing skills, but they don't do very well at measuring how well a candidate will perform in the job we're trying to fill.
Studies show that we are subject to so many biases when it comes to a typical interview process that we are probably just as likely to pick a star employee from a stack of resumes, sight-unseen, than from bringing five people in for interviews.
So, how do we improve the validity of our process?
First, screen resumes against a predetermined set of KSAs (knowledge, skills and abilities) only. Consider having an administrative employee scrub information that might lead to biases from the resumes' themselves. For example, as a Carolina grad, I might have a negative opinion about a resume from a Duke grad. Another common bias is unintentionally rating a resume lower based on the candidate's name (assumptions rooted in gender, age and ethnic stereotypes). If you can't see the university name or the candidate's name, you are more likely to evaluate the resume on the KSAs alone.
Second, develop good interview questions! Write questions that have a likelihood of predicting the behaviors that you want. Use an interview guide and ask each candidate the same questions. Practice interviewing your current stars and see how they answer those same questions.
Third, use assessments! Utilize skills assessments - having them demonstrate skills they claim to have. This can range from taking a keyboarding test, to taking them out in the warehouse and allowing them to demonstrate their ability to operate your forklift. (There are some do's and don'ts regarding "auditions," so make sure you don't cross a line). Behavioral assessments will show if their personality style is ideal for the demands of the position. Driving Forces/Motivators assessments will show if their motivators align with the rewards and demands of the position. Competencies assessments will show what types of core competencies the candidate brings to the team.
Small employers frequently tell me, "I'm not spending $50 or $100 on a test - I know how to hire!" Considering the high cost of a poor hire and the science that shows utilizing valid assessments increases the validity of the selection process exponentially, that might be a penny-wise, dollar-foolish commitment.
Fourth, check references! LinkedIn is a great resource to see referrals and endorsements on many of your candidates. It's also a venue to conduct "unofficial" reference checks. If the candidate is connected to someone you know, you might be able to gain valuable insight. Many people dismiss the references the candidate submits - but I've gotten honest references from these in the past, so don't neglect them either. The best predictor of future success is past success!
Fifth, make sure your background screening criteria match your risk profile. Some organizations screen too tightly here and miss out on potentially good workers due to invalid criteria. If your company provides residential services and a candidate has a recent breaking-and-entering conviction, that candidate clearly doesn't fit your risk profile. But screening out a candidate who has a 12-year-old misdemeanor possession conviction might be an invalid selection criteria if you have an opening on a production line or on a commercial construction site.
Finally, if you're doing all these things but you're still losing people, it's probably not your selection process. It's likely something else, like job design, compensation, benefits and perks, or most likely, your supervision and management style. But that's a subject for another blog!
No comments:
Post a Comment