In a post today, George Mason University economist Robin Hanson reports on an interesting interview with Bill Raduchel. The whole interview is interesting, by the way. Robin highlights Raduchel’s conclusions about, when he was “assistant dean of admissions for Harvard and Radcliffe colleges,” finding good students. Raduchel found that a better way to judge potential students than to ask them questions and pay attention to their answers was to pay attention to the questions the students asked.
Robin uses this finding to support his own view–that he has emphasized many times–that the best research methodology is to ask questions. I don’t disagree. Robin didn’t say this in the post, even though when I first read his post, I thought he did, but he often talks about how the best way to come up with an interesting question is to try to solve a puzzle. If he still has this view, that is where I part company.
Solving a puzzle is one interesting, and often fruitful, way to go. But answering a question is another one. And the question need not be a puzzle.
Let me give an example of each. If you ask, “Why do movie houses charge the same price, no matter what the quality of the movie is?,” you’re asking about a puzzle. And various people have tried to answer that.
But what if you notice that manufacturing employment is falling in the United States and you wonder if it’s falling in other economies with large manufacturing sectors also? You’re not trying to solve a puzzle. You’re just wondering about a question because you’re curious. And it’s fairly straightforward to answer the question. The answer, by the way, is that in most manufacturing economies, even the ones with expanding manufacturing output, manufacturing employment is falling.
So my bottom line is that what matters is curiosity, whether it leads one to solving puzzles or leads one to asking questions.
On the curiosity front, by the way, I give Raduchel’s interviewer, Thomas Heath, a low grade. Check out this statement by Raduchel:
The questions give a much better clue to how a person thinks. Answers can be learned, can be rote. But it’s the questions. Like the questions Sean Parker asked me that day at AOL.
Sean Parker, in case you don’t know, co-founded Napster.
So what would an even modestly curious interviewer want to know? What were some of the questions Sean Parker asked?
No such luck.
READER COMMENTS
Robin Hanson
Jul 21 2013 at 7:38pm
I was also frustrated that they didn’t say more about the Parker questions. I accept that many non-puzzle questions are worth asking, and that the key thing is to ask interesting questions. Puzzles are especially useful for uncovering chinks in our theoretical armor, places where something should to be adjusted to account for puzzling data. But I accept that is far from the only kind of interesting question.
David R. Henderson
Jul 21 2013 at 11:42pm
Thanks, Robin.
Jess Riedel
Jul 22 2013 at 9:51am
So a puzzle is a particular kind of question where our theories give the empirically incorrect answer, or where two theories contradict each other. Is that what we’re talking about?
MG
Jul 22 2013 at 10:38am
I think there is a bias for asking “puzzle” questions instead of “knowledge broadening or deepening” questions. The two situations Jess comments on suggest why they are sexier: these kind of questions promise the opportunity of falsifying theories.
Yet, if we get to carried away with searching for puzzles, we will see them everywhere — underestimating the scope of “randomness” as an explanatory variable and oversestimating our ability to “control for everything else”. Hopefully, our questioning will allow us to accept these as legitimate, if humble, answers.
Foobarista
Jul 22 2013 at 5:07pm
For my own part, I hated “puzzle questions” in job interviews, and thankfully the “MSFT gotcha questions” trend is dying in Silicon Valley. Google found that there was little or no correlation between success in puzzle questions and whether someone was an effective engineer, so they’ve gone away from them.
Comments are closed.