## Archive for March 10, 2013

### Hiring, Firing, and Flipping Coins

Tomorrow, I become the Director of Mathematics at Discovery Education, and I am slated to spend five hours with the Human Resources department, filling out paperwork and such. By the end of it, I suspect I’ll feel like Arlo Guthrie in *Alice’s Restaurant*:

…the sergeant came over, had some paper in his hand, held it up and said, “Kids, this piece of paper’s got 47 words, 37 sentences, 58 words, we wanna know [blah, blah, blah, blah],” and he talked for 45 minutes and nobody understood a word that he said, but we had fun filling out the forms and playing with the pencils…

I do not mean to disparage Discovery’s HR department, but my new boss has described this orientation as “a test of attrition.”

The following two days will be less bureaucratic but similarly intense. As I am the first employee hired for “the math team,” it will be my job to hire others, so my second and third days at Discovery will consist of six interviews with potential math specialists. Six! Hopefully it won’t go like this:

Three candidates are being considered for a job. The first candidate, a pure mathematician, steps into the interview room. The interviewer asks just one question, “What is 1/3 + 2/3?”

Without hesitation, the pure mathematician answers, “1.”

An applied mathematician enters next and is asked the same question. He takes out his calculator, punches some buttons, and announces, “0.99999999.”

Finally, a statistician enters the room and is asked the same question. “What is 1/3 + 2/3?” the interviewer asks.

The statistician responds, “That depends. What would you like it to be?”

My worry with so many back-to-back interviews is falling into the Gambler’s Fallacy. Just as a gambler incorrectly assumes that a string of successive heads must eventually be broken by tails, interviewers often believe that a string of successive weak candidates must eventually be broken by a strong candidate. In a recent interview with Steve Inskeep on NPR (*Deciphering Hidden Biases During Interviews*), Shankar Vedantam said…

…interviewers add the equivalent of two years of job experience to the last candidate in a row, who is weak, in order to break the streak.

Vedantam went on to say that Uri Simonsohn from the Wharton School of Business thinks interviewers should…

…have a spreadsheet where they can see all the candidates they’ve interviewed, not just on that one day, but over several days. And when you step back and actually say, “Yeah, there are an equal number of strong and weak candidates,” even though you may have the streaks of four or five really strong or really weak candidates in a row.

And who am I to argue with advice like that?

For you, I offer the **Clusters spreadsheet**, which can be used to simulate 100 coin tosses. Press F9 to generate a new set of tosses. The spreadsheet will then calculate the number of heads, the number of tails, and the longest run of either type. It shows that unexpected strings of consecutive heads and tails occur, but over the long run, the number of heads and tails are roughly equal, just as the number of strong and weak candidates who are being interviewed.

Just between you and me, though, I’m hoping for a slight imbalance toward the strong end.