When a result deviates strongly from a statistical prediction of that result, one ought to step back and consider at least three possibilities to explain the apparently anomalous result: 1) sampling bias, 2) evaluation bias, and 3) incorrect prediction. I have a practical application of that suggestion described below.
Before proceeding, however, squeamish Readers should be aware that I use the word “asshole” (or some variant thereof) with astonishing frequency in this post: a total of thirty-three times, including in this sentence and in the footnote.
Now, the issue of statistically anomalous results weighs on my mind today because I served as a pro tem judge in small claims court this morning. I fear that now that I’ve done this a few times, I’m sliding down the path I’ve seen so many other pro tem judges slide down — deciding cases based not on the merits but rather on who I think is the bigger asshole. As I walked out of court this afternoon, I ran through all of the cases in my mind and realized that in every contested matter, I had ruled against the bigger asshole — sometimes after having to make a difficult judgment call as to which of the two parties was worse.
I’ve charted the result as best I can from my recollection of this morning’s events and graphed them to the right. Note that just as there is a spectrum of being a total bastard to being a nice guy, there is also a spectrum of winning, meaning the closeness of the result the party sought to the judgment I actually rendered.
Generally, you can see that assholes lose and nice people win — in one case, the nice person admitted liability but disputed the amount owed, and prevailed as to that argument. Because there was a net judgment againt the nice person I rate that as a loss, but very close to the center axis. (The other party in that case was not a huge asshole, by the way, just a moderate one.) But the trend is pretty clear.
Now, in a couple of cases, I also worry if I get the law wrong; it’s frighteningly easy to forget the rules about bad check penalties, late feed, contractors working without written contracts, and fiduciary duties while up on the bench and after three hours of dealing with whatever drifted into court that morning. With each of those cases, I realize later that I did reach the right result, but maybe I didn’t get there in the best possible way and only figured out the right way to get there after the fact. Point is, it’s harder than it looks up there on the bench.
Certainly, it’s easier to say, “Who’s the bigger asshole here?” and go on that basis, even if you don’t know the law on that particular subject. But that’s not justice; if the asshole is right on the merits, the asshole should win, despite his monumental display of assholery. And the point of this exercise is to resolve my self-doubt in my own abilities. Because I have taken it as a given that you might be an asshole, but if you’re right, you’ll win anyway.
If being an asshole has nothing to do with being right, then statistically, the asshole is going to be right about half the time. But today, assholes had a zero percent success record in my court. Non-assholes and/or assholes who weren’t quite as obnoxious as the other parties, did extremely well, or at least got a comfortable slice of baby. I heard about ten cases this morning and that seems to me to be getting into enough statistical significance that I’ve got to wonder about one of the three main dangers of comparing anticipated results with actual statistics.
I know I ought to be getting the law right; sometimes, though, the cases call on me to pull something out of my memory banks that hasn’t been touched since law school. I ought to be strictly applying burdens of proof and finding some objective basis for evaluating the credibility of parties who simply contradict one another. Credibility is not the same thing as a good attitude and just because someone has strong emotions does not mean that they are telling me the truth. At best, it means they believe they are telling me their truths. At worst, it means they are putting on a convincing show — which does happen sometimes. I am haunted by self-doubt, though, that something else, something very subtle, is working just below the surface of my consciousness when I sit in judgment of other peoples’ disputes. Because it seems a little too neat and clean that the biggest asshole loses every time.
Or maybe I’m just reacting to all the strong emotions that get vented. One thing I can be confident about — I tell them all to try and settle their cases because they have a 50% or better chance of walking out of my courtroom unhappy. If they don’t follow that advice, that’s not my fault.
So this leaves me with a few possibilities:
- Sample Bias: A disproportionate number of small claims cases as opposed to regular disputes have at least one asshole as a party (either plaintiff or defendant).* Therefore, I’m going to see more assholes in court than I would in the general population. “You have a dispute” + “You are an asshole” = “Your case goes to court instead of settling.” This charmingly suggests that being an asshole has nothing to do with being right or wrong; alas, it does nothing to explain why assholes lost nearly every time. But there does exist the possibility that this is simply an unrepresentative sample of data.
- Evaluation Bias: I might be applying the law correctly, and judging that people who do not know the law or otherwise fail to appreciate the incorrectness of their position are doing so not because they are ignorant but rather because they are assholes. Thus, my evaluation of the parties as assholes is colored because of my evaluation that their cases lack merit. In other words, “Getting the case wrong and therefore losing” leads necessarily to “The judge thinks you’re an asshole.”
- Incorrect Prediction: Perhaps my hypothesis, that assholery has nothing to do with the merits of one’s case, is incorrect. There could be a causal relationship, meaning that assholes are more likely to get the law and the facts wrong, and therefore present bad cases in court which wind up losing. This is the opposite of theory #2; it would be diagrammed “You’re actually an asshole” –> “You make a mistake in evaluating your position that causes you to present a loser of a case.”
- Evaluator Bias: Maybe I’m the asshole here. If so, then I’m awarding favorable results to parties whom I like and unfavorable results to parties whom I dislike. We’d diagram that as “The judge thinks you’re an asshole” –> “You lose.”
I must ponder this issue further before I take the bench again.
* A corollary to this theory is that the presence of an asshole in a dispute lowers the chances that it will resolve before trial. This corollary hypothesis strongly resonates with empirical data collected from both my law practice and my experience as a pro tem. This theory also suggests a converse — when both parties are nice people, they find a way to work things out, and therefore a lower number of cases that I hear will involve nice people on both sides. I know of no way to test this converse theory, since it does not appear to be either falsifiable or measurable — experience indicates that there is at least one asshole involved in pretty much every case I see either as a pro tem judge or as an attorney. Opinions as to the identity of that person on a case-to-case basis are obviously quite likely to differ.