Digressions? Digressions!
90. Law: It’s About Police, Not Precision
The Illustrated Guide to Law
The comic that teaches what the law is, how it really works, and why.
Digressions? Digressions!
I care
It probably makes a bit of a difference to the defendant, too.
Poor DC. He never seems to get anywhere in suppression hearings.
Then again, we’ve seen that Pi generally doesn’t bring something to trial when it’s going to be suppressed.
More to the point, at a suppression hearing, the DC is supposed to challenge everything. He doesn’t expect all of his challenges to win, but if he throws out enough of them, maybe something will stick and knock out a crucial element of the case.
Which is why Pi generally only brings evidence that’s reasonably tight in the first place – if she brings something patently objectionable, it’s going to fall.
So if something is common knowledge to the courts, but not to juries, then expert testimony is not allowed? What purpose does that rule serve?
Seconded. It’s not the court as a whole that’s returning the guilty/not guilty verdict, it’s the jury. Unless the jury instructions specifically include “Eyewitness accounts have been found to be wrong half the time”, the part of the court can’t be relied on to make the decision correctly.
Because evidence that eyewitnesses in general are not perfectly reliable is not evidence that this particular eyewitness is unreliable.
The time to cast doubt on the eyewitness is during cross, not through expert testimony. You need to show that that particular witness may be making a mistake. If you know the ways that memory is unreliable, then you apply that knowledge to cross examination.
Actually yes, eyewitnesses being unreliable is evidence that any particular witness is unreliable. The human brain relies on heuristics and stereotypes, so in most cases what we believe the base rate of accuracy is for any given source of information is going to be assigned a greater weight than the idiosyncratic facts that we can observe in any particular case. So if juror’s think that eyewitness testimony has a higher base rate of accuracy than it actually does, then they will put more weight on it than they actually should and produce more false convictions. If that doesn’t make the general reliability of eyewitness testimony relevant to any particular case, then I don’t know what would.
So, you’re saying when I see my mother in a crowd, I’m more likely than not wrong?
Nah, we all saw that porno. Definitely her.
That is different. You have known your mother your entire life, presumably. Neither witness knew the defendant before the crime. They each only saw the criminal once before they made their identifications. So taking the rate of false positives at 50% into account, we should conclude that neither identification is to be trusted. I didn’t mean to say that the rate of false positives is the only thing that matters. I only meant to say that it matters in all cases at least a little bit. To see what I mean just take the human beings out of the equation. If I gave you a device meant to predict whether or not a tornado would hit your city within 24 hours and told you that half of all of its positives were false and it told you that a tornado was going to hit in 1 hour, would you believe the results?
I don’t get it. If it’s understood that eyewitness testimony is wrong “half the time,” meaning it’s as likely to be wrong as it is to be right, then why allow it at all? It’s like allowing someone to flip a coin, saying heads is guilty, tails is not guilty, and letting them use that result to build their case. It’s a waste of time at best if the jurors all rightfully discount it, and an outright hindrance to getting to the truth if the jurors end up giving it any weight at all against the real evidence.
Being wrong half the time is not the same as being wrong at random.
It might be, for instance, that it’s possible to evaluate an eyewitness’ credibility by examining things like what details they actually remembered and what the circumstances were (much like the things discussed in the strip so far). If you can make an evaluation, then you can filter out some or all of the wrong eyewitnesses, even if that’s half of them to start with.
Or to put it in oversimplified terms: half the M&Ms in a bag are green, orange, or yellow, and half of them aren’t. Pull one out at random, and you’re not sure what color you get; 50-50 chance that it’s a “correct” one. But if you take the time to look at the M&Ms, it’s easy sort out the half of them that are those three colors.
It is of course much harder to judge witness credibility than the color of a piece of candy. The question is whether it’s so difficult as to have no value even after you do the evaluation. The judge points out that that evaluation is a responsbility of the jury; DC would no doubt argue that juries are terrible at telling whether or not an eyewitness is likely to be correct.
Because being wrong half the time is still an improvement over being wrong 99% of the time, as just picking random people who fit the description could be, so the witness is legitimately evidence in favor of guilt despite its flaws. It just isn’t as good as juries tend to think it is.
“Evidence that eyewitnesses are wrong is not evidence that this eyewitness is wrong.”
I’d say that there’s a decent argument to be made about prior probabilities. Suppose that instead of an eyewitness, we were dealing with a psychic. If 50% of all psychics are fake, and there’s no way to tell whether or not this psychic is fake, wouldn’t that be reason to doubt his testimony?
Exactly. It’s like if DNA evidence actually had a 50% false positive rate, but typical jury members thought it was very reliable. If you tried to argue that a 50% false positive rate was not evidence that *this* test gave a false positive, it would be ridiculous. Likewise with testimony: the base rate of error is relevant to the confidence you should place in any juror. Thus, it unquestionably is evidence of reliability. The judge is saying “X isn’t evidence” when all she really has is “X doesn’t prove it’s wrong with certainty.” To take it to an extreme, imagine it was proven that only 1 in a billion people would get an eyewitness ID right. Would we really say that this fact would be irrelevant because maybe this witness is one of the 7 people on earth who will make an accurate ID?
Regarding the judge’s remaining points, if the judge really doesn’t think expert testimony is needed, and really thinks that it is scientifically established that witnesses are only 50% reliable, then the judge should be happy to add a jury instruction that tells them about how unreliable memory is and how untrustworthy eyewitness testimony is in general, so the jury can start with the correct prior probabilities. Of course, the judge isn’t going to give an instruction like that, and the prosecution would never acquiesce to it, either.
Thus, we have the judge basically saying that there is a fact that is clearly scientifically established, relevant to the jury’s opinion, and likely to be misunderstood by the average jury. Nonetheless, the judge will not allow expert testimony to correct the misunderstanding, nor issue instructions to the jury to clear up their misconceptions. And of course the defense can’t just argue that it is known that eyewitnesses are only 50% reliable, since that fact isn’t in evidence (because the judge won’t let it in).
Can the defense get away with something like taking a picture of himself, changing his tie during recess, and asking what color tie he was wearing during cross examination?
Would it even be wise to try?
How about bringing in an associate for the first part, then sending him away over recess, and asking the witness and victim to pick his associate out of a loaded photo array? (One with a different person in it, wearing the same outfit, while the actual person was either not present, or wearing something different.)
That seems extremely dangerous to try. What if the witness correctly recalls the tie, or the associate? Goes back to the old adage popularized in the film A Civil Action: Never ask a witness a question if you don’t already know what the answer will be.
This sort of thing has been done before, leading predictably to a not guilty verdict. This article discusses examples: https://www.nytimes.com/1994/07/29/us/at-the-bar-defense-lawyer-turns-the-tables-on-the-prosecutors-but-pays-a-price.html
The other result is a holding of contempt–not surprising, since it reveals some embarrassing inaccuracies in court proceedings. The judge seemed to think that you should inform the judge and prosecutor beforehand. Which, of course, ruins the experiment, since the prosecutor can easily indicate through his manner of questioning that the person in the defendant’s chair isn’t the defendant (or “prep” the witness regarding this fact). Basically, it is widely known that eyewitness ID is terribly flawed relative to its perceived accuracy to the layman, but the courts actively hinder any attempt to suggest or demonstrate this fact.
First off, don’t forget that the prosecution gets to use experts as well. If the defense is allowed to bring someone in to explain that eyewitness testimony isn’t always accurate, the prosecution also gets to bring someone in to explain why it’s accurate enough. It’s a waste of time and money, for no real benefit. And you might confuse the jury.
And when you think about it, it’s a largely irrelevant statistic to a case. It doesn’t matter how often eyewitness identifications are accurate. What matters is, “Is this one accurate?” And that’s for the jury to decide. The defense’s job is to make the witness seems like they made a mistake, and that’s not too hard to do on an eyewitness that truly isn’t sure of what they saw…check out My Cousin Vinny for several examples on how unreliable witness testimony can be easily impeached. As long as the jury is aware that eyewitness testimony isn’t foolproof (and I don’t think ANYONE believes that it is), there’s no reason to tell them that it isn’t.
Imagine a defense lawyer making the argument, “You know, 99.8% of the people have never stabbed the victim. Therefore, it’s extremely unlikely my client did. There’s only a .2% chance!” It’s pretty much the same thing – you’re asking the jury to play the odds in your favor, instead of having them focus on the truth of the matter in front of you. No judge would, nor should, allow that.
Wait, I thought the local fashion was hoodies and long-inseam sweats.