CELS V: The Year of the Experiment
For the last several years, I’ve posted recaps of the Annual Empirical Studies Conference. (See me, @ Cornell, @ USC). This year, as promised, will be no different. Yale hosted CELS V, and the committee did a bang up job: the food was tasty; there were no technical snafus of note; and the panels appeared to have a high degree of internal validity & congruence. Richard Brooks, Alan Gerber, Dan Kahan, Yair Listokin, Tracey Meares, and (especially) Roberta Romano are all due a round of applause, or, better yet, supersized computer monitors so they can see their data better. In this post, I’m going to provide a running diary of the conference. It will be like you were there with me, except you don’t have to suffer through my bouts of social anxiety!
Unfortunately, I missed the hottest ticket of the conference, Bruce Ackerman’s commentary on Law/Versteeg’s The Evolution and Ideology of Global Constitutionalism. From all reports, Ackerman said something like: “wrong questions, wrong data, wrong theory,” and then imploded in frustration. Instead of watching those fireworks, I was watching Yair Listokin present The Meaning of Contractual Silence: A Field Experiment [Here’s an older version of the paper]. Listokin ran a field experiment selling ipods on ebay, some with a warranty, some as-is, and some silent on the warranty term. He found that individuals paid attention to the contract, and there was some evidence that the UCC default was about what they thought silence meant. As he admitted, there were problems with the design of the study – particularly, (1) small & skewed samples; and (2) a lack of clarity about how much buyers know about ebay’s unique and self-contained dispute resolution system. As someone remarked after the presentation, it would have been interesting had Listokin sold all the customers bad ipods (instead of good ones) and studied how the contract terms influenced behavior post-“breach”. Then again, who needs that IRB hassle?
I then saw three cool experimental papers at a Law and Psychology panel. The stone-cold coolest was by Sah/Loewenstein/Cain, and presented by SOM’s Daylian Cain. The central claim in The Burden of Disclosure is that there is a “previously unrecognized perverse effect of disclosure: Disclosure of an advisor’s conflict of interest can decrease advisees’ trust in the advice while simultaneously increasing pressure to comply with that advice. This compliance pressure comes from two mechanisms: (1) recipients fear signaling distrust of the advisor, and (2) recipients feel an increased pressure to help their advisor when the advisor’s personal interests have been disclosed.” Here’s the set-up. Apparently, the experimenters have truck that they drive around town. They lure people into the truck. Once inside, subjects are paired. One member of the pair is told they can roll one of two dice. Each dice has a lottery attached to how it lands (i.e., if it lands on one, you get a starbucks card; two a snack; etc.) One of the dice-lotteries has a higher expected payoff. Now, they have the non-rolling subject advise the choosing subject on which die to pick. (The non-picker has no special knowledge.) Sometimes, the advisor has a conflict – s/he gets the good lottery if the picker picks the bad die. Not surprisingly, that conflict leads the advisor to suggest picking the bad die, and because people are suggestible, the advice is taken more often than not. But the wild manipulation happens when the advisor discloses this conflict to the picker. Notwithstanding the obvious truth that the advisor now has told the picker than they are motivated to give bad advice, the picker actually accepts the recommendation more often! As Daylain explained, the mechanism might be a form of interpersonal conflict-avoidance: we don’t want to look someone else in the face and tell them that we think they are corrupt. (Dan Simon, commenting, suggested that the mechanism might simply be a dictator game-like fairness norm.)
What’s troubling about this experiment is that disclosure is the law’s preferred solutions to many conflict problems. This perverse effect has the potential to be at play in many real-world situations – lawyers, doctors, brokers, managers, shareholders – where we expect the disclosure to clean otherwise fouled waters. And maybe it isn’t limited to the situation where the disclosing party makes a disclosure that is obviously against their self-interest. What about when journalists disclose their political contributions as a way of legitimizing them? Though this would seem to clarify their ideology & perhaps cause us to think carefully about how they construct stories, maybe it is, instead, making us trust them more. (Evidence against Olbermann?) Similarly, political regulation seems to have shifted decisively away from substantive checks toward a sunlight/disinfectant model. This research agenda suggests a host of pretty deep problems with that approach.
Moving right along, Yale apparently spent something like $500M on the poster session, because the overall production quality has now approached that of a drug-company-funded science convention. Highlights included (1) Bermant/Barness-Blakeman’s Beyond Ownership: Lessees and Idiosyncratic Valuation, which extends the endowment effect literature to mere tenants; (2) Estreicher/Heise/Nash’s Examining the Instrumental Use of Pro Bono Projects by Law Firms: Preliminary Evidence, which suggests that pro bono hours are in part a function of firm health & that over time, firms have increasingly used pro bono for training purposes; and (3) Buccafusco/Sprigman’s The Creativity Effect, reporting on an endowment/creation experiment. Yale also spent tons of money on food, but I’d caution future organizers that the make-your-own-bruschetta trend is a bad one, especially when you are in a roomful of socially and physically awkward people.
I chaired a panel Saturday, and failed to keep Dan Simon or Dan Kahan to the time they were allotted. Simon’s presentation on Simon/Stenstrom/Read’s The Spontaneous Arousal of Hot Cognitions in the Course of Deciding Criminal Cases was still pretty fun, even as I stewed at my lack of power. Unfortunately, chairing the panel meant I had to miss Cohen’s Expeditiousnesses and Delay in State Courts: An Exploration of Case Processing Time in Civil Trials, which I heard was great, and Bilz/Gold’s An Experimental Test of Civil Recourse Theory, which I know is awesome and which sheds some pretty useful light on why recourse theory might (and might not) explain private law doctrine. (The paper doesn’t appear available online – email Bilz if you are interested in reading it.) I then saw the Cohen/Lawless paper, Less Forgiven: Race and Chapter 13, which suggests that attorneys are pushing black (but not white) clients into making suboptimal choices in bankruptcy. The best part of the paper was a survey-experiment (in the nature of a name/resume study) administered to bankruptcy attorneys nationwide. This is the kind of work that ought to make national news & promote law reform.
I then presented Boyd/Hoffman, Litigating Toward Settlement, with Ted Eisenberg commenting. As it turns out, you can get through 25 slides in less than 18 minutes. You just need to be nervous enough to talk very, very fast. I saw a few more excellent papers (e.g., Pardo/Nash’s Ideological Voting in Bankruptcy) and called it a conference.
(1) More experiments & psychologists at the conference than in past years. It seemed that John Darley was on every other panel. There were relatively fewer (I thought) case-counting projects, though this could be just the panels I went to.
(2) More co-authorship between law professors & folks from other disciplines (rather than people from other disciplines coming in to talk about law, or law professors banding together).
(3) The overall quality of the methods was quite sophisticated. To the extent that the conference once billed itself as a place for people who needed help to get better … well, that’s not exactly what’s happening today. The level of polish and gloss is high. If if I were a junior scholar with incomplete data or methods, I don’t know that I’d feel comfortable presenting at CELS. I do think that schools would do well to send empirically minded junior scholars to watch the panels – you can learn tons from watching error!
(4) I only saw one paper that manifested Leiter’s claim that ELS practictioners use datasets to chase questions. And that is exactly what the paper’s commentator said, though in much nicer terms. Self-policing is alive and well.
(5) CELS VI will be at Northwestern. Will Lee Epstein make a push against powerpoint and toward whatever cool software she uses to run presentations? Now’s her chance to make it happen!
(6) After the dinner speech by Orly Ashenfelter, I know much more about wine pricing than I used to. Good cocktail party conversation information.