Rationalizing Law
posted by Daniel Solove
For quite a long time, extensive empirical work in psychology, sociology, and behavioral economics has been revealing that many of the law’s most cherished rules are faulty. They are based upon mistaken assumptions about human behavior. They are often flat out wrong. And yet they persist.
The work of Daniel Kahneman and Amos Tversky has shown that the human mind operates with all sorts of biases and heuristics that lead to systematic errors in judgment and perception. As Dan Ariely put it in a recent work, Predictably Irrational (2008): “[W]e are not only irrational, but predictably irrational . . . our irrationality happens the same way, again and again.” (p. xx). Richard Thaler and Cass Sunstein describe many of these systematic blunders in human judgment in their book, Nudge (2008).
As these studies increasingly make their way into legal scholarship, they are proving that many existing legal rules don’t work as they should. And this is more than a mere normative critique. The rules just fail because people don’t act or think the way the law thinks they ought to. In fact, what we’re learning about the way people act and think is often counterintuitive. It is hard to grasp and hard to deal with.
This research should be undermining many legal rules at their very foundations. Yet the legal rules don’t seem to be shaken despite their foundations being annihilated.
In many domains, when something is proven flat wrong, it is confronted and dealt with. If evidence shows that bleeding the patient isn’t a good cure for disease, then we move on and stop doing it. But in law, if the evidence shows that a rule doesn’t work, what’s the response? Often, it is to just accept the evidence with a grin and continue on. If science were like law, we’d be talking about how the earth is round yet continuing to behave as if it were flat.
I think that it is time to hold the law up to the light of contemporary understanding and bring it into the 21st century. Many legal rules are based on crusty assumptions made back in the 19th century, such as the M’Naghten rule for insanity(1843). Hardly any experts would contend that the M’Naghten rule (inability to distinguish right from wrong) is consistent with contemporary understandings in psychology. The law of evidence is replete with faulty psychological assumptions. So many areas of law don’t stand up to modern understandings of science, statistics, psychology, or empirical studies.
The example I want to focus on primarily is eyewitness testimony. Countless studies have shown that eyewitness testimony is very unreliable, yet the law routinely permits convictions based solely upon eyewitness testimony. I dug out my copy of Wrightsman’s Psychology and the Legal System (my edition is older than the one I’m linking to). Here are some interesting facts:
As many as 80% of the eyewitnesses in some studies choose someone, even when none of the lineup members is the true criminal. (p. 148)
Buckhout (1975) staged a simulated crime on a television news broadcast and asked members of the viewing audience to phone in their choise of suspects shown shortly after the crime. Out of 2145 viewer-witnesses who called in, only 14.7% were correct, an accuracy rate that was no different from the one achieved by simply guessing. (p. 148)
Also, victims of crime are less reliable in identifying the perpetrators than bystanders (p. 149). Compounding the problem:
Jurors put too much emphasis on witnesses’ statements about the confidence of their identifications. In fact, Cutler, Penrod, and Stuve (1988) found that, for laypersons, the witness’s level of confidence was the most important criterion for judging the witness’s truthfulness. . . . [But] an eyewitness’s confidence is not a strong indication of accuracy. (pp. 149-50)
This is a big problem, yet the law doesn’t deal with it. Instead, the law does just the opposite — it prevents defense attorneys from introducing evidence of all these studies that undercut the accuracy of eyewitness testimony, thus preventing jurors from being properly educated.
Perhaps the law could adopt one of the following rules:
1. Eyewitness testimony could be completely barred as more prejudicial than probative.
2. Eyewitness testimony could be allowed only if corroborated.
3. Eyewitness testimony could be allowed, but defendants would be permitted to school the jury about the studies showing its unreliability.
Moving down this path would be difficult. Imagine the case with the parent who witnesses the murder of her child right before her eyes. She is the lone witness and says she’s absolutely certain it was the defendant. Should her testimony be not allowed? To bar the parent from testifying about what she saw would be incredibly difficult to do — it seems almost inhumane.
But that’s what the studies are showing us. Reforming the law so that it reflects contemporary understandings of psychology and human behavior is a difficult and daunting task. But the law needs to be modernized. It should be rationalized.
April 14, 2010 at 12:47 am
Posted in: Criminal Law, Criminal Procedure, Economic Analysis of Law, Empirical Analysis of Law, Evidence Law, Jurisprudence, Law and Humanities, Law and Psychology, Legal Theory
Print This Post
Responses (10)
Orin Kerr - April 14, 2010 at 2:40 am
“Countless studies have shown that eyewitness testimony is very unreliable, yet the law routinely permits convictions based solely upon eyewitness testimony.”
But that’s not because the law doesn’t reflect contemporary understandings of psychology and human behavior . It’s because the law is extremely reluctant to overturn jury verdicts generally, whether based on eyewitness testimony or any other kind of evidence.
Susan Franck - April 14, 2010 at 7:08 am
I can’t help but be reminded of “Stumble, Predict, Nudge: How Behavioral Economics Informs Law and Policy” – http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1327077 – by Orly Lobel and On Amir that came out in the Columbia Law Review. Basically, it’s a lovely book review about both “Nudge” and “Predictably Irrational” that recommends a more nuanced account of the range of mechanisms avalable in behavioral law and econ (and presumably cognitive psychology) and greater focus on the limits, costs, and consequences of applying these lessons to law.
But I would encourage us to think beyond the domestic norms that Dan S. explores in this post – given that normative behavior expands beyond domestic borders. Indeed, it is precisely this sort of information about the behavior of people (particularly if explored in a cross-cultural manner) that can provide us with insights about the appropriate formation of international law and policy. But perhaps that is a debate for another day…
Brett Bellmore - April 14, 2010 at 7:40 am
Sounds perfectly consistent with the way the law approaches other forms of error, such as old decisions everybody understands were wrong, or even new cases like Heller, where the majority bases their ruling in part on objectively false statements about the case. (Like Scalia asserting that Miller was convicted.)
The response is always the same: Stare decisis and legal realism. Why would it be any different here? The legal system values consistency over accuracy, and virtually never admits mistakes.
dave hoffman - April 14, 2010 at 9:40 am
“Instead, the law does just the opposite — it prevents defense attorneys from introducing evidence of all these studies that undercut the accuracy of eyewitness testimony, thus preventing jurors from being properly educated.”
Is there any evidence that “introducing evidence of all these studies” would help? I’m not sure — and wonder if it simply would create a larger danger of Type II error.
Generally, this is a provocative post! I think you might have inspired an end to my blogging block.
Matt Bodie - April 14, 2010 at 11:39 am
Dan — I’m sympathetic to the general notion that we should use scientific research to improve the law. But you speak with far too much certainty here. Does the research really support the notion that eyewitness testimony should be “completely barred as more prejudicial than probative” or “allowed only if corroborated”? I doubt the authors of those studies would agree that your reforms are as scientifically obvious as the roundness of the earth. You are trying to use “science” to advocate a policy change when the “science” isn’t nearly as definitive as you suggest. After all, you’re assuming something about the costs of false positive identifications as compared with the costs of preventing prosecutions of the guilty. The research doesn’t say anything about that.
Daniel Solove - April 14, 2010 at 12:47 pm
Matt — I didn’t want to imply that the issue of what to do with eyewitness testimony was easy or clear. What’s clear, though, is that there’s overwhelming evidence there’s a problem, and the law doesn’t adequately address this problem. Indeed, the law is generally quite slow to respond to current scientific and empirical research. The purpose of my post is to urge ways to make law better confront science and deal with its findings. It is far too easy for law to remain insular and backward-looking. Imagine if the law were approached with more of a scientific mindset. What would the law be like? What rules should be changed?
In essence, the broader ambition behind my post is to figure out a process to better rationalize the law. We have the benefit today of all this research, and it is showing us that cherished assumptions are wrong, and it should be like a tsunami on the law, but it isn’t.
Matt Bodie - April 14, 2010 at 2:05 pm
I guess I’m just suspicious when someone claims that the science is “clear” when it comes to a specific social policy. A lot of corporate law scholarship in the last two decades jumped from quantitative research to specific policies, and these policies were touted as the only rational answer allowable based on the scientific evidence. Those who disagreed were labeled “quacks.” As it turns out, neither the evidence nor the policies were as well-grounded as they appeared to be at the time. Keep in mind — this “nudge” stuff is a reaction to old “scientific” dogma!
Tim - April 14, 2010 at 4:32 pm
As to #3, I worry about a People v. Collins-type problem if defense counsel is allowed to argue the issue of eyewitness reliability generally (as opposed to the reliability of a particular witness). That is, I worry that a general argument, cloaked in technical language, could cause a juror to question her (perhaps sound) intuition about a particular witness’s credibility. But I wholeheartedly agree that the legal system should take account of eyewitness reliability generally, and I too would like to see some kind of solution.
But there are good reasons not to hold our breath. The conservative temperament and preference for gradualism that characterizes the common law probably explains why this issue has yet to find its way into the judicially created law of criminal procedure by way of, say, a due process claim. Moreover, the Originalism that has a toehold on the current Supreme Court would probably generate considerable resistance. As to #1, the argument would be that it cannot be within the “original public meaning” of the Due Process clause to bar a type of evidence that was traditionally admissible at the time of the framing. As to #2, the same general reasoning might be augmented with an expressio unis-type nod to the Treason clause.
On the other hand, it takes no great understanding of the problems of public choice to appreciate why no legislative solution is likely any time soon. Probably no politician has ever gone wrong running against crime, and a measure that will be perceived as making defendants more difficult to convict could be viewed as “soft on crime.”
In short, I am pessimistic about the prospect of near-term systemic reform.
A.J. Sutter - April 14, 2010 at 11:11 pm
I’m not a criminal lawyer, but back in the 1970s I took a course called “Probabilistic inference in legal systems”. As I recall, back then arguments about probability were allowed to rebut eyewitness testimony (e.g. accused is a black man with a blue car, and witness saw such a person in such a car speeding away from the scene). In particular, one of the profs (a certain Dershowitz) had us read cases and talked about having used such arguments himself. Yet as I understand this post, you seem to be saying that arguments on general unreliability of eyewitness testimony (as distinct from arguments based on issues specific to the witness) are across-the-board inadmissible. Did I misunderstand? If not, what’s the basis for your assertion?
Charles Lowery - August 29, 2012 at 1:00 pm
I was a trial lawyer for 22 years. I did not find any of the clients, lawyers or judges irrational. I found them consistently driven by short-term self-interest almost always unattended by a concern for honesty. This behavior applies across the board: credit card abuse, obesity, smoking, cheating on a spouse and even voting. Politicians cannot fix problems for the long term because voters want goodies in the short term.
Leave a Reply