The Job Interview - Anecdotal Data

December 30, 2007 
Posted by Jay Livingston


Scatterplot
has had a discussion about campus visits for job candidates. Good timing – at Montclair, we’ve just completed a search. We reviewed several dozen applications and had two people come to campus for interviews. But why?

Interview
isn’t quite the right word. Neither is ordeal, though it comes closer. The person spends an entire day on campus: there’s morning coffee, then The Talk presenting her research, informal chats and lunch with faculty, an interview with the dean, maybe teaching a sample class, a campus tour, dinner with faculty. I’ve become convinced that these visits are useful only for seeing how you’ll get along socially, not for anything truly academic. It’s sort of like a first date or, in societies with arranged marriages, the pre-wedding meeting that a couple may have. And about as useful for predicting compatibility.

For the task part of the job, for gauging how the person will be as a scholar and teacher, the campus visit may be worse than no visit at all. That’s especially true for teaching. We used to ask candidates to do a sample class. This time around, we dropped that requirement, though for logistical reasons not methodological ones. Still, it was the right decision. A class session taught by a job applicant is anecdotal evidence, and it puzzles me that a group of social scientists would use it at all.

It’s not just anecdotal evidence, it’s unrepresentative anecdotal evidence. You have your candidate teach one session of another teacher’s course – students she’s never seen before and as many as half a dozen faculty members watching from the back of the room. Nevertheless, just as the dramatic story is often more convincing than a ream of statistics, seeing someone in person can outweigh more systematic data, even for sociologists, who should know better.

In discussing the candidate later, when someone cites the outstanding evaluations the person has received in several courses at her home university, someone else might say, “Well, she didn’t seem so good with our students,” as if this bit of anecdotal evidence wiped out the systematic evidence of the all those evaluations. As Stalin is supposed to have said, “The death of a million Russian soldiers, that is a statistic; the death a single Russian soldier, that is a tragedy.” And even among sociologists, statistics can be less compelling than tragedy.

6 comments:

  1. So is statistical evidence in this case ratemyprofessors.com?

    ReplyDelete
  2. This post is wonderful, in part as a reflection of how the process works. But also because I realize that people think about this... what it means to interview and how they make choices. I have a very narrow view about what sociology is and how sociologist practice their discipline. But I'm glad to see that my thoughts about sociology resonate beyond my narrow view/experience.

    ReplyDelete
  3. I would disagree a bit with this post about the value of teaching demonstrations. At our institution, we try to see whether candidates attempt to put into practice their teaching philosophy. I cannot tell you how many candidates say that they are invested in active learning who proceed to lecture the entire time! We are not looking for perfection, but evidence that a candidate "practices what they preach".

    ReplyDelete
  4. Anonymous,

    What I'm looking for is evidence of teaching ability. The teaching demo is evidence, but it's anecdotal evidence.

    I agree with you about the statements of teaching philosophy. They are not evidence at all, no more than writing about how to play golf is evidence that the writer actually plays well.

    ReplyDelete
  5. You've overlooked the value of the campus visit to the candidate. What inference would a candidate make from a department's unwillingness to let him meet his potential colleagues? A good visit gives both sides of the negotiation valuable information--almost certainly more so for the visitor than for the visited.

    ReplyDelete
  6. Constant,

    I agree. Both the candidate and the department check one another out. That's why I said it's like a blind date.

    ReplyDelete