Thanks to Andrea, Michael, Sarah, Art, Jill & Cynthia for giving me an opportunity to participate on the blog. I’ve enjoyed the blog as a reader for several years. One of the blog topics that has interested me recently is the subject of gender in negotiations. Andrea convinced me that I need to dedicate more time to covering gender in my class, so I’ve been keeping an eye out for good research on the subject. This week I ran across a new article by Elizabeth Loftus (always an insightful empirical scholar) and colleagues titled “Insightful or Wishful: Lawyers’ Ability to Predict Litigation Outcomes.” The article reports on an empirical study designed to assess lawyers’ accuracy in predicting the likelihood that they would meet the litigation goals they set. The authors asked 481 attorneys to set a minimum goal for what they hoped to achieve in a case set for trial, and then to give a confidence estimate for the likelihood that they would achieve that goal (a process referred to as “calibration”). They predicted (as we all would) that lawyers would be overconfident in their predictions. Not surprisingly, given the wealth of research into overconfidence in a wide range of contexts, the lawyers were overconfident—whether the outcome was trial or settlement, the lawyers held a belief in their ability to achieve their goals that proved unjustified.

That result was expected, but two of their specific findings were less mundane. First, overconfidence was not reduced by experience. Lawyers with more experience had not learned to calibrate their expectations more conservatively. (That may be the result of another finding: that lawyers’ subjective perceptions of their success exceeded their actual achievement of their objective goals. In other words, they didn’t believe they had failed to meet their goals, even when they had. Self-serving bias, anyone?) Second, by a statistically significant margin, female attorneys’ confidence judgments were more likely to be accurate than their male counterparts’ (even though the male attorneys had more experience on average). If a female attorney gave herself a 66%-75% chance of achieving her goal, she was likely to achieve it, whereas male attorneys who made similar confidence judgments were likely to fall short.

Takeaways? Well, clients should discount their male attorneys’ estimates of likelihood of success if they express confidence above 66%, and discount their female attorneys’ estimates above 75%. Attorneys, especially male attorneys, should seek third-party input to test their predictions about case outcomes. Mediators can be an important resource in that regard, and may want to adjust the extent to which they “reality-test” depending on the degree of confidence the lawyers assert. Of course, it may prove difficult to disabuse those overconfident and self-serving lawyers of their faith in their predictive abilities—which just reinforces the need to have clients participate actively in mediation.