In my last post, I noted that statistical models typically out-perform experts at predictive tasks. What's true of baseball managers and talent scouts is true of many other professionals, too. Go here for a list of relevant studies (pp. 22-24), or just read Super Crunchers. So perhaps it's not surprising that professionals tend to disparage and resist efforts to introduce formal methods of prediction and assessment into their domains. That hostility is not limited to baseball, though the sport does provide some fabulous examples. What professional wants to believe that, despite years of training and experience, an actuary can out-perform them on tasks that are core to their professional identity?
So what about lawyers? Although much of what they do involves prediction - say, valuing a personal injury claim - lawyers historically have been insulated from this kind of competition. In part, that's because they have often have had a monopoly over the necessary information: past verdicts, settlements, etc. The lack of court system transparency and the private nature of most settlements help cement this monopoly. (In practice contexts like insurance defense work, where this monopoly is weak or non-existent, actuaries, not lawyers, do the valuation.) As with other professionals, however, it seems clear that lawyers will increasingly be challenged by statistical models of decision-making and that this will re-shape the profession to some (probably modest) degree. And as with other professionals, I suspect that lawyers won't welcome the change.
And what about law professors? Even within the legal academy, might some of us re-enact that same story of professional resistance and hostility, perhaps in response to the Supreme Court forecasting project, or empirical assessments of judicial performance? Surely not...
Comments