A fascinating area of (ir)rationality is recruitment. I recently rediscovered my notes from Sway: The Irresistible Pull of Irrational Behavior, written by the Brafman brothers.
Much of this has been written about extensively, but is worth a reminder. The evidence they focus on primarily comes from 20 years of research by professor Allen Huffcutt. He states that the typical job interview is unstructured, and easy to compare with a first date. However, these tend not be to very successful -there is a long history of research confirming that. Think about that – if true, as mentioned, then jumping into the marriage (in a corporate sense) would mostly be based on the experience during that first date!
Managers typically try to form a first impression based on a subjective assessment of whether the candidate share his/ her interests and whether the person being interviewed would be good to work with. Does that then mean I should hire someone who scores well for a position based on more objective measures, even if I can’t stand him? The evidence suggest “Yes”, even though it feels completely (ir)rational.
You might have encountered many of these typical questions in your organizational life – all well intended, but not necessarily helpful. For example, out of the following list, Huffcutt only gives a passing mark to one question:
a. Why should I hire you?
b. What do you consider your greatest strengths and weaknesses?
c. How would you describe yourself ?
d. How would you describe yourself ?
e. What do you want to earn in 5 years from now?
f. What do you really want in life?
g. What college subject did you like best and least?
h. Why did you decide to seek a job with our company?
i. Why did you leave your last job?
j. What do you know about our company?
Can you spot the value-add question?
Huffcutt offers the following as background on the inefficiency of these questions:
- a to c: nobody will be 100% honest and there is no way to assess the truthfulness. They are prepared in advance. For example, who will state as a weakness that
“sometimes I stay up late at night watching sport and drinking which leads to late arrival at work the next morning”. Typically a weakness would be “I take work too
- d to f: requires candidates to gaze into the future, but as the Brafman brothers point out: “unless they are applying for a job at a psychic hotline, their predictions carry little value”. Everybody will come up with a nice sounding answer.
- g to i: turns the interviewer into a historian. However, often times when people revisit the past, they reinvent it.
- j is the only question Huffcutt feels adds some value, as it shows the candidate has done some research on our company. It’s still not ideal.
Other research states that employers typically rely on the interview process to accomplish objectives unrelated to employee selection. For example, employers rely on the interview to:
• Sell applicants on the job,
• Respond to the applicant’s questions, and
• Serve as a public relations tool.
There is a lots of information available to improve the interview process. Using structured questions is one often cited, and this paper from 1997 has vast information on the selection interview structure: A review of Structure in the Selection Interview. Also worth looking at is this 2014 review paper: The structured employment interview: narrative and quantitative review of the research literature.
It is also important to document/ give feedback on the candidate directly after the interview. However, out of personal experience, one should not share anything with colleagues that might encourage framing of the candidate before they interview.
For interest, here are the NYT’s notes from an interview with authors.
Given the subjective nature of recruitment it seems ripe to predictive modeling intervention, and expect to see much more of this going forward.