Home > Uncategorized > Why the interactive “Chicago-style” economics workshop is the best style

Why the interactive “Chicago-style” economics workshop is the best style

February 10th, 2018 Leave a comment Go to comments


I have been reading various comment and posts of people who think economics research seminars are too interactive. Rather than having questions throughout, they would like to have questions all at the end, or perhaps only after 2/3 of the time has elapsed. They also complain that the questions are too critical, and claim that the questioners are just trying to make themselves look good.


I think those people are completely, absolutely, wrong. Interactive “Chicago-style” seminars are one of the best things about economics compared to other fields. Here are some considerations:

  1. The first priority is for the audience to understand the situation being modelled. Often the speaker does not realize this and omits key facts because he knows them so well himself (for example: how hard is for a workman to install solar panels?). Often, too, there are key facts he doesn’t realize are key and has never found out (how much of installation costs are governments fees and permits?).  It is not useful to go on until the audience understands the basics.


  1. The second priority is for the audience to understand the model. What are the important assumptions? Which ones are just simplifying? How the assumptions match the reality I discussed as item (1)?


  1. Third, it is useful to discuss the assumptions, because if they’re bad, the rest of the paper is worthless and we shouldn’t bother going on. Instead, the audience should help the speaker figure out how to redo his model so he can get meaningful results. It really is depressing to have a speaker continue a talk, with no more questions, after everybody, including himself, has realized that his paper has a fatal flaw.


  1. All this applies if the paper is empirical, too, but add to it the need to understand the data and the summary statistics. In particular, I’ve often seen seminars where the audience couldn’t even figure out the unit of observation without asking questions— the individual vs. the census district, etc.


  1. Speakers often interpret all these clarifying questions as critical questions. I remember I did as a graduate student—Eric Maskin asked questions I know now were just clarifying and I thought he was spotting flaws and that my paper was a mess. But they’re not. If someone asks why you made a particular assumption, he really means it. He isn’t just using a polite, roundabout, way of saying, “You made a bad assumption.” That’s one of the good things about the economics style— you don’t have to spend time decoding what someone says to find out their hidden agenda. Often, to be sure, the person will ask a follow-up question that is a criticism of your assumption, but first he wants to (a) understand what it is, and (b) hear why you made it.


  1. Once the basics are understood, it’s useful for the speaker to go on.


  1. To be sure, too many questions can derail a seminar in a bad way. One way that happens if when the speaker spends time discussing the previous literature. That is almost always a mistake. The audience doesn’t know what you’re going to say yet, so how can they relate it to what other people have done? Often, too, the speaker ends up spending a lot of time explaining and defending someone else’s work, not his own. I blame him, too, since he brought up the subject.


  1. Also, the audience may be tempted to play the “Guess the model game”, where the speaker has introduced the topic, and they start asking questions about how he is modelling it. There, the speaker must tell them that he’ll get to the model shortly. One way to avoid this game is to not spend much if any time telling the audience what your conclusions are going to be, since their natural instinct will be to wonder how you get to those conclusions.


  1. Something I think helps immeasurably is a one-page handout containing assumptions, notation, one or two key results, a figure perhaps, and maybe some summary statistics.  That way, even if the seminar gets off-track, the audience has your message.  Hardly any speaker except me likes this idea though.


  1. After the preliminaries, the speaker can show how to get from assumptions and data to the results. That is not the most important part of the seminar. An experienced audience will probably be able to predict that if they understand the theoretical assumptions and the data summary statistics. Nonetheless, questions along the way are important here, too, so the speaker isn’t going through the fourth step of a proof when he’s lost everyone on the first step. Lack of questions is also useful, because it tells the sophisticated speaker that not a single person in the room understands what he’s doing at this point.


  1. The speaker should not expect the audience to have read the paper. We are all too busy. In one seminar style, the pretence is made that everybody has read the paper and understood it. I am dubious. It works out badly when different people have different ideas of what the author was trying to say. It’s better to have him go through it, though speedily if it is easy to understand.


  1. In law, notably, but also in other fields, a common style is to have questions just at the end, with a “queue”. A moderator watches people who put up their hands, and writes them down and calls on them in order. The result is dismal. It is too late to ask any clarifying questions. It seems silly to make small but useful comments, like “You skipped a step in Lemma 3” or “Here’s why you should have included lung cancer  in your dataset.”  So what we have left are audience comments on the paper in general.  Since there’s a queue, several people will all have the same comment, which they’ll feel they have to ask even though it’s repetitious since they’ve been waiting in line to ask it.  Position in the queue is arbitrary, so question 2 will be about a completely different part of the paper than question 1, but question 3 will go back to question 1 again. Nobody is able to build off a previous question without waiting for 5 other questions in between.  The best strategy is just to raise your hand, in case you think of a question, just you can reserve a space in line. All audience members are treated equally, even though it is always the case that some of them have many more and better things to say than others do. And since it is so pre-planned, instead of real questions, we get little speeches on the general subject of the talk, and there is much more showing off than when people are expected to ask short, specific, relevant questions. Finally, often in this style we see a lot of what Kierkegaard calls “preliminary expectoration”: time-wasting formalities like “I liked this paper and think it provides a useful addition to the literature, making a number of points I hadn’t heard before and that really show an imaginative mind. But I did wonder about one thing….”, after which the questioner proceeds to show his utter  detestation for what the speaker said.


  1. In general, the speaker should think of himself as discussion leader, not lecturer. Lectures have their place— for plenary sessions at conferences, for example, where the grand old man lays out his wisdom on a subject. At seminar, though, the speaker is trying to improve his paper and the audience is trying to learn something. For neither of those goals is getting through the entire paper important.  He must not take the attitude that this is his moment of glory, and the audience is there in homage to his brilliance. Rather, whether he is brilliant or not, people are there to find out what he can teach them, and he is there to find out how they can help him make his paper better. The attitude on both sides should be, “Here is some work in progress.  What can we learn from it? What are its weak points? How can the weak points be patched up?


13. The Chicago-style workshop is one of the great strengths of economics compared to other fields. Think about teaching. Is the best teaching method a straight lecture, reading from the book? No. It is much better if students ask questions along the way, admitting it when they don’t understand something and probing the boundaries of what the teacher is claiming is true.  The students can read the book on their own, and the teacher can force them to read by assigning problem sets or quizzes. What class time is best for is explanation.

Categories: Uncategorized Tags:
  1. February 13th, 2018 at 23:23 | #1

    I’d grant the speaker five minutes, but not ten, though I think in my department that’s usually about the time they get before the first question appears.
    I really don’t remember having seen people asking questions just to look smart. Doing so would be self-defeating, after all. If you are obviously trying to look smart, you’ll make people think worse of you, not better. The same is true of just trying to be unpleasant. I can imagine some people would enjoy that, but it comes at considerable personal cost in reputation, just like passing gas.

  2. Melissa Jones
    February 12th, 2018 at 13:34 | #2

    In the spirit of your post, your model of a Chicago-style workshop assumes that the only reason seminar attendees have for participating is to discuss the paper at hand and add to the group’s collective understanding of economics. It ignores other potential incentives, e.g. signaling one’s intelligence to colleagues, establishing/signaling rank superiority over the speaker, or personal enjoyment at pointing out flaws in other people’s work. Many of the people discussing seminar “decorum” would agree that ignoring these incentives is a bad model assumption because people do in fact respond to these incentives in ways that detract from the overall goal of the seminar — to further knowledge and understanding. The problem is not that questions are too critical; it’s that they are often delivered unnecessarily rudely and sometimes with the sole goal of appearing smart, in which case, they often don’t add to the conversation. The reason to suggest not asking questions in the first five minutes is because while possible, it’s rare that in five minutes, you’ve happened upon a question that the speaker will not eventually address. And if it does turn out to be the case, you can ask it after 5 minutes. Most of us have a story of “that time when they didn’t even let the speaker get off his/her title slide!” That’s a bad reflection of our field.

  1. No trackbacks yet.