Thursday, September 12, 2013

Hiring interviews: You're probably doing them wrong

I'll have a research update for you as soon as I get enough content, but today I wanted to give you a preview of a talk I'll be giving in San Francisco on September 23rd titled "Hiring interviews: You're probably doing them wrong.  A motivational talk."  The topic was inspired somewhat by Google's semi-recent announcement regarding their internal research on the issue.

The sad truth is that even though interviews are one of the most popular forms of personnel assessment, they are often done wrong.  Not necessarily through any ill intent, but because of two main factors:

(a) they're harder to develop than many people think


(b) most people think they're great interviewers

On the first point, interviews are deceptively simple.  Many people assume that if they can talk to other people, they can interview someone.

Wrong and wrong.

Interviewing isn't talking to people.  Well, okay, I suppose literally it is.  But interviewing really is about MEASURING people.

If I asked you to use a set of measuring spoons to give me 1 tsp of sugar, you would have no problem, right?

But what if I asked you to use a Halloway P36 spectrometer* to measure photon radiation?  You might need some help.  It's all about what you're measuring.

On the second point, research has well established that people are generally very bad at accurately reporting their skill levels--across a wide variety of disciplines.

To make matters worse, research has also established that interviewers tend to get addicted to bad interviews: when things turn out poorly, they tend to blame outside factors rather than the interview format.

But there is good news.

Specifically, we know how to do interviews the right way.  Namely by structuring them.  What does this mean?  There are several key features.  Here are a few:

1.  Use high quality question formats.  This means behavioral, hypotheticals, knowledge-based, and background questions.  Not puzzle questions or gimme questions.

2.  Be consistent.  Each candidate should be asked the same questions in the same order (with limited variations for follow-up if needed).

3.  Use a detailed rating scale.  When there is no criteria to compare answers to, scoring tends to be inconsistent, reducing the utility of the whole process.

4.  Base the questions on an analysis of the job.  I probably should have made this #1.  Everything you do in your selection process, including interviews, should be driven by the key knowledge, skills, abilities, and other characteristics required for good job performance.

5.  Train the interviewers.  Because of the wide variety of biases that plague interviewers, it's critical that they be aware of these tendencies and guard against them.

These are just some of the ways we can avoid bad interviews.  Of course another strategy to increase your success is to look beyond interviews at things like tests that more accurately mirror the job (i.e., performance tests).  But even following these guidelines can radically improve your results.

In other words, there is hope.

* I made this up.  If this is a real instrument I'm better than I thought!

1 comment:

Gaurav Kapil said...


Aptly put.

I think it is not only about others, interviewers should also be aware of their own attribution errors. I am about to complete Social Psychology Course at Coursera by Prof. Scott Plous. It was an eye opener.