The Empirical or Technical Law School Job Talk

Josh Wright —  28 October 2007

In the process of reading the number of informative (and at least one funny one) posts around the blogosphere on the AALS job market and job talks, I ran into this piece of advice from Dan Solove at Coop:

Choose a topic that many people on the faculty can talk about. The most important part of the job talk is the Q&A, where the faculty gets to see how well you think on your feet and respond to difficult questioning. Unfortunately, some job talks are doomed because they are on esoteric topics that hardly any of the faculty can discuss. If the faculty can’t debate you or engage with your topic, it is hard to have the kind of vigorous and interesting Q&A that is needed to show off your response skills. You need to leave the faculty with something it can discuss with you.

I generally agree with Dan on this point and the other advice he offers in his post. But what if you are somebody whose skill set includes quantitative analysis or formal modeling? It may prove quite difficult to present your empirical or theoretical model in a manner which is accessible to the entire faculty while remaining faithful to the content of your analysis. While this problem does not affect everybody, the number of these folks are increasing. For some casual empiricism on this point, one need only check out the number of economics or finance Ph.D.’s on law faculties around the country. Besides, there is no good reason why economists and empiricists should be left out of the job talk advice circuit.

As I see it, techies are left with the following choice: (1) choose a topic that does not reflect their skills but will appeal to a broader audience, or (2) choose a topic that does reflect those skills while doing your best to simultaneously keep the interest of those without the quantitative skills necessary to understand the nuances of the paper while demonstrating your expertise to those on the faculty that do. (1) seems useless and counterproductive to me. You are what you are at this point. Might as well do what you do best. If you’re an econometrician, statistician, or economist, there’s no point in hiding it. After all, you presumably got the job talk because the law school is interested in your skill set. So, below the fold I offer some advice on strategies for (2) (though some of the advice holds equally for non-technical talks).

1. Take 3-5 Minutes to Motivate Your Contribution — And Get to the Punchline ASAP! The constitutional law professors in the room may not care much about your regression tables (though they might!), but everybody will want to know what the main point of your talk is, what you’ve done, and why it is important. See Dan Solove’s point #1:

Oral presentations are a remarkably inefficient way of communicating information. So you don’t have the luxury of making several different points, exploring all the manifold nuances of a topic, and so on. You must isolate the key point, and provide the basic skeletal structure of your argument. Basically, state your thesis, explain it, and justify it. Don’t argue all the finer points. You don’t have time.

The most common pitfall I’ve seen for technical job talk papers at law schools is that the authors have not really thought through some pretty obvious questions: (1) what does this have to do with the law? (2) What are the most plausible alternative explanations of your empirical results? (3) What are the testable implications of your theoretical model? Think about these. Know what you’ve done. Know the literature. You don’t have to get the answers to all of these questions out in the first 5 minutes. In fact, it is best not to do so. But you should be able to isolate your contribution and communicate it to the audience clearly and in a manner that a non-technical audience will understand. The first few slides are the ones that should take you the longest to craft and the first few minutes of your talk the most time to prepare as you carefully select how to makeyour empirical approach or theoretical model accessible to the audience.

2. Get to the Results or Model Quickly. If the central contribution of your paper is a model or regression results, you should be walking through the model or sharing your results no later than 5 minutes into the talk. You’ve only got 20 minutes. If this is your central contribution — you’ve got to get there quick. I’ve seen a number of job talks fail because the presenter spent too much time discussing their research question and convincing the audience of its importance and too little time actually explaining what it is that they had done to answer it. The audience is left without an understanding of what the presenter had actually contributed unless they read the paper carefully. If you are a technical person presenting a technical paper, it is not safe to assume that the entire audience has read or understood the paper.

3. Present Your Results Clearly. Think long and hard about how you want to present your results. I generally have a very different set of slides when I present to a group of economists than I do at a predominantly legal conference or at a law school (though I always have all the slides with me at the end of the presentation just in case). What is the best way to present your data? For those looking for advice on how to be effective in presenting empirical papers to legal audiences, see anything by Bill Henderson and/ or read Andrew Gelman’s blog which frequently discusses presentatation with useful thoughts on how to make complex data speak to the audience. Send the slides to some non-techie friends. It doesn’t matter if they are lawyers or not. If they don’t understand your tables, charts, figures or graphs, work on them more. This applies to empirical results and presentations of your data. Formal theoretical models just aren’t that easy to present in an accessible manner and present a tougher call for the presenter as to whether to leave them in the slides or present an intuitive version of the model in the slides and leave the math to the paper. I have less experience with this, but have a general sense that the latter is more successful with a general law school audience.

4. Try to Channel Hyper-Technical Discussions to Q&A. Leaving out the technical minutae from your 20 minute talk can accomplish two things at once. First, you wont bore the non-techies in the room and cause them to completely disengage by talking about nuanced issues in your choice of standard errors. Second, you can exercise some degree of control over what questions you can expect in the Q&A by highlighting some issues for the audience. Don’t get me wrong. If you are in a workshop setting (like GMU) where questions can and will come during your 20 minutes (usually, after about 120 seconds), its better to nip them in the bud right then. But at many schools you will have the opportunity to get in 20 minutes uninterrupted. Highlight any technical issues you’d like to chat about in Q&A and let the audience decide if they are interested enough to bring them up. If they do, you’ve got a chance to show off your knowledge. If they don’t, it would have been a waste of time to discuss them during the talk anyway.

5. Acknowledge Strengths and Weaknesses. This is similar to Solove’s #4:

What is more important to me than whether a candidate has all the answers is whether the candidate can recognize the weak spots in an argument and has thought about the different ways to address it. So if you’re caught in a weak spot, demonstrate that you recognize it, that you’ve thought about it, that it hasn’t come as a surprise, that you have some preliminary inclinations about how to address it, but that you will have to give it more thought to fully resolve it. I can accept that. What I can’t accept is your being unaware that these weak spots exist or nonchalant as to their import and consequences.

This certainly has application to technical/ empirical job talks. Often times, Q&A will involve questions that try to push you to make generalizations from your data that are not justified. For example, a common line of questioning asks the presenter whether their results have implications for issues X, Y or Z which were not the focus of the paper. It is tempting for candidates to succumb to the temptation to allow for the possibility that their data has greater relevance than it really does. If you’ve not thought about the answer, say that you don’t know without further thought. If there are limitations to your identification strategy, highlight them for the audience in the presentation and be prepared to discuss them in Q&A. As Dan’s advice suggests, appreciation for an understanding and willingness to highlight the weaknesses of your analysis for the audience earn credibility. This is perhaps even more important to the sector of the audience that is less capable of evaluating the details of your contribution on the merits. They may defer to the economist or empiricist on the faculty on the merits, but will at least be able to appreciate that you are a serious and credible scholar.