Patterico's Pontifications

8/11/2008

Jurors, Likely Misled by L.A. Times, Acquit Man Accused of Sexually Assaulting an Elderly Woman

Filed under: Crime,Dog Trainer,General — Patterico @ 9:21 pm



Just as I feared, those misleading L.A. Times articles on DNA, cold hits, and statistics are starting to have an effect in the courtroom.

In May, the L.A. Times ran a front-page DNA article that completely botched the statistics on “cold hit” cases, falsely telling people (on Page One of the Sunday paper) that the odds of a particular convicted defendant being innocent were actually 1 in 3. The paper had absolutely no basis to make this statement the way they phrased it. I called them on it, as did Eugene Volokh. But the reporters and editors stubbornly refused to correct the error.

Then in July, the paper ran another Page One Sunday blockbuster, telling readers on the front page that a researcher had run searches in an Arizona DNA database that revealed “dozens” of matches, “each seeming to defy impossible odds.” Only on page A20 were readers told that most of these Incredibly Unexpected Matches were, in fact, “to be expected statistically” because each sample was matched to every other sample in the database — exponentially increasing the total number of comparisons, and “greatly increasing the odds” of finding a match. (There was a discrepancy between the expected number of matches and the reality, but this discrepancy was not anywhere as stunning as suggested on Page One. According to one researcher, statistics suggested there would be about 100 matches out of the 2 billion comparisons done in the Arizona database. In reality, there were actually 144 matches, instead of the expected 100.)

Even educated people got confused reading these articles. This can be explained, in part, by the fact that statistics is a very complicated subject. But part of the explanation lies in the articles’ overly dramatic presentation, coupled with their often misleading and sometimes flatly false statements concerning the relevant statistics.

I suspected that most educated people who trust the paper came away with only a vague notion of the math involved — but with a firm impression that There Is Something Wrong With DNA and that Jurors Are Being Lied To. I worried that jury verdicts in Los Angeles would be affected.

It’s starting to happen. These irresponsible articles have begun to bear fruit.

My wife tells me a very distressing story about a case recently tried in a courthouse in this County in which the defendant was charged with sexually assaulting a woman in her eighties. He was accused of scamming his way into her house and forcibly orally copulating her. He was prosecuted almost entirely on the basis of a cold hit. (There were, in fact, cold hits to the same DNA profile from samples taken from two other sexually assaulted elderly women. But the jury didn’t hear about those other sexual assaults. One of the other sexually assaulted women had died by the time of trial. The other refused to testify — and under state law could not be forced to.)

When the jury came back with a verdict, everyone expected that the defendant would be found guilty. After all, the DNA had not been degraded. After the cold hit, a sample of the defendant’s DNA was compared to the evidence sample and matched at all 13 loci. My wife doesn’t know the exact random match probability, but she believes it was 1 in several trillion (if not more). I am unaware of any proven case in the history of DNA analysis where two samples matched at 13 loci and turned out to be two different people (with the obvious exception of identical twins, of course). (The most recent L.A. Times article darkly hinted at such a possibility, but it has never been shown to happen, as far as I am aware.)

The jury came back not guilty.

The accused man would be walking the streets — except that he is being held on an immigration hold. (But that’s another story.)

And the foreman of the jury told the D.A. afterwards that he felt the D.A. had failed to prove the case with DNA, “in light of recent controversies about DNA.” (Controversies that had been specifically ruled inadmissible by the judge.) He didn’t mention the L.A. Times specifically, but he was a local law professor. It stands to reason that he likely followed the front-page articles above with some degree of interest.

I wasn’t there, and I don’t know precisely what the evidence was. But I’m hearing that the judge was shocked, and that at least one well-respected defense attorney in the courthouse feels that justice was not done.

If you want to know why I get so upset at these articles and their sloppy and misleading language, this is why.

Actual dangerous people are going to end up walking the streets because the editors and reporters at the L.A. Times feel a compulsion to overdramatize their DNA findings, and occasionally misstate the statistics in seriously meaningful ways.

It is, of course, incredibly unlikely that the reporters and editors involved in putting these articles together will ever be directly affected by verdicts like the one I just told you about. Even in the topsy-turvy statistical world of the Los Angeles Times, where impossible odds are portrayed as commonplace, it is still incredibly unlikely that the man just freed by this verdict will sexually assault, say, the elderly mother of one of the editors or reporters.

But I’d be willing to bet he’ll assault someone‘s mom.

Are you proud, L.A. Times reporters? Are you happy to see that you’re Making a Difference?

Good. Hey, guys, I might eventually have an even better story for you. If this guy sexually assaults another elderly lady, I’ll try to get you an exclusive interview with her.

Sound good? Great! Hey, I understand. It’s all about the story.

143 Responses to “Jurors, Likely Misled by L.A. Times, Acquit Man Accused of Sexually Assaulting an Elderly Woman”

  1. Gut wrenching and crucial story. The LA Times could stand for something good in a city that could use a lot more good in it. Instead, it’s part of the problem. If the paper were to disappear, the people would be better informed by the alternatives to it.

    All the the pursuit of the idea that Republicans are law and order types, and therefore law and order are bad, and DNA evidence should be shot down ridiculously.

    This monster is going to hurt someone else. This kind of thing emboldens that type. That poor victim… to see justice fail her so spectacularly. And that law professor, of all people, to make a decision based on inadmissible evidence, is a jackass. He thought he was so smart that he didn’t need to follow the rules, and since he was foreman, I assume he dominated the jury… telling them all about how DNA is a bunch of BS when it’s not. I’m sure he’s fantasizing about that 12 angry men play. But no… resistance to the law is only noble when you’re right, and this was dead wrong.

    Juan (4cdfb7)

  2. No, these reporters actually believe that DNA evidence is wrong, or bad, or evil or some such thing. An invasion of our precious bodily fluids if you will. And, like, it made OJ look guilty so it must be a racist plot.

    There’s a guy in England who keeps posting stories about cellphones killing bees, or causing three-headed chickens. Same kind of moron.

    Kevin Murphy (0b2493)

  3. I wonder if you can disbar a law professor.

    Kevin Murphy (0b2493)

  4. Patterico – What would the recourse be for a victim of a crime like this if it could be shown that the inaccurate LA Times article contributed to the decision to acquit by the jury? It’s not as though the paper didn’t receive any notice of corrections. It’s not libel, of course, but could the continued misrepresentation of facts, after the notification of inaccuracy, constitute some kind of negligence?

    Just sayin’.

    Plus, great post, as always.

    Apogee (366e8b)

  5. The rest of the jurors were probably intimidated by what they thought was the law professor’s superior understanding of the law and let him lead them into this verdict. It’s understandable but it’s still a shame.

    DRJ (a5243f)

  6. I’m censoring a lot of stuff in this post. I’m just trying to get across the minimum necessary to make the point. Over beers I’d say much more, but not on a blog.

    Patterico (cc3b34)

  7. I’d say more in person, too, but I’d order a Margarita.

    DRJ (a5243f)

  8. Isn’t this a black and white issue? You’d think some group of statisticians would go on the record here.

    Josh (f6ffef)

  9. So, the LAT publishes, with great fanfare and chest beating, half-baked statistics which have not been accepted by appropriate authority. But, they refuse to mention Edwards’ foibles without substantial corroborative evidence?

    Nice.

    Ed (841b4a)

  10. #6 & #7 – Anytime – Beers or Margaritas. Or both.

    Apogee (366e8b)

  11. It would be a pleasure, Apogee.

    DRJ (a5243f)

  12. DRJ, wouldn’t the potential for that be a good reason for one side or the other to get him tossed off the jury?

    aphrael (9e8ccd)

  13. The People thank and ask the court to excuse the law professor.

    Alta Bob (69331c)

  14. In the Dog Trainer world DNA evidence is good if it leads to acquittal, suspect if it points to conviction.

    This case is another example of the harm this journal of opinion does to our community.

    Stu707 (6e4ad5)

  15. The hilarious part, or one of them, is that the really important discrepancy in the article is the difference between the expected number of matches (100) and the actual number (144) — BUT, to assess the importance of that discrepancy, you have to know something that the article DOESN’T EVEN BOTHER TO PROVIDE. Namely, the standard error associated with that number of 100.

    For example, in 20 coin tosses, you expect 10 heads. But what if you actually get 14? Or 15? Does that result mean that you have to question your assumption that coins land heads 50% of the time? To know that, you have to know how odd it is to get that result of 14 or 15 heads.

    I agree with the thrust of the article that the FBI should open up these databases and that more knowledge is better. But I’m still furious with the sensationalistic way that the article portrayed the problem.

    Patterico (cc3b34)

  16. The People thank and ask the court to excuse the law professor.

    Alta Bob, if I can go off on a tangent here, I thought I was the only prosecutor in the world who exercised peremptories that way.

    The way I was taught was to say: “The People ask the Court to thank and excuse Juror Number x.” But, being a stubborn cuss, I decided long ago that I have the right to thank the juror myself, even if only the Court can excuse him or her. So I have always said: “The People thank and ask the Court to excuse Juror Number x.”

    Awesome to see that I’m not the only one.

    Patterico (cc3b34)

  17. aphrael,

    Jury selection is an art, not a science, and I can only give you my West Texas perspective. I don’t think there’s any reason to disqualify a law professor for cause unless he represents to the court that he can’t fairly decide the case or there is another disqualifying reason that would apply to any potential juror.

    However, the parties could excuse the law professor and any potential juror through the use of a peremptory challenge. The conventional wisdom is that highly educated jurors, especially lawyers and law professors, are dangerous because they may dominate deliberations and could have an undue impact on their fellow jurors. In addition, such jurors often are selected jury foreman, which makes them even more important than the rest of the panel. On the other hand, it’s sometimes worth the risk to retain a lawyer-juror if you think he may be an advocate for your position.

    In addition, I think lawyer-jurors tend to second-guess and share with the jurors what the parties’ lawyers do in court. I’ve sat on juries with lawyers before and I can tell you from personal experience that happens.

    I don’t know about criminal cases but in civil cases, especially complex civil cases, parties often want educated jurors that can absorb the information presented in the case, but not all professions are viewed the same. I would worry about a law professor but there are only so many challenges available in cases and perhaps this juror seemed like an acceptable pick, as indeed he might have been since we don’t know all the story.

    DRJ (a5243f)

  18. I wonder if that’s not out of prosecurors’ deference to Batson. The way I learned it was, “We excuse Juror #___ with our thanks”. But (don’t tell anybody) I’m pre-Batson.

    nk (e38352)

  19. I’m assisting on a DNA case right now. The Illinois Supreme Court may revisit the issue of the DNA databases. It might take three or so years before I can say more about the case, though.

    nk (e38352)

  20. Patterico, have you ever had the occasion to attempt to get a Law Professor excused for cause?

    I bet it would be damn near impossible.

    Having legal professionals sit in the jury room is short-circuiting our system. It’s not like they’re all bad eggs, but those who are are just too adept at reading the objections, knowing what probably was thrown out, etc etc, and render decisions based on things they have no right to consider (such as this story). But then again, it’s not like it’s just lawyers who can do that. And we can’t very well administer a test, and kick out everyone who passes.

    When I took evidence from Professor Wellborn (among the best UT has to offer), he showed me that people are unreliable, that confessions are unreliable, that hearsay is unreliable, and though he never denounced physical evidence to that point, I was left thinking that trials are epistemological shams, where at best jurors make a guess as to what nonsense they are going to believe. But that’s why God gave us DNA… so I wouldn’t have to worry about that (and it’s not just putting people in jail… it frees the innocent sometimes). The LA Times is party to a campaign to destroy what little reliability we have.

    Juan (4cdfb7)

  21. Was the prosecution out of peremptory strikes? How bad were the people they struck, to be worse than a friggin’ law professor?

    Beldar (7a5c38)

  22. I didn’t mention the foreman’s profession to ignite a pile-on regarding the exercise of peremptories — regardless of how I might feel about it. As DRJ says, jury selection is an art — and it’s an art I have botched myself more than once. I mentioned the guy’s profession only to prove that he was probably aware of the LAT stories.

    Patterico (cc3b34)

  23. I do a minor variation on your version of the peremptory: “The People thank and ask the Court to excuse Juror Number 6, Mrs. Jones.”

    As for the law professor remaining on the panel, never, never, never, right up there with engineers, members of Amnesty International, the ACLU, and anyone wearing a Free Mumia! t-shirt.

    I hope Beldar was right (that the prosecutor was out of peremptory challenges); it’s hard to imagine how bad the remaining prospective jurors would have had to be out in the audience to risk keeping the prof.

    Mike Lief (e6260e)

  24. With engineers I think it depends on the person and the case.

    Patterico (cc3b34)

  25. Let’s just say that when the People’s case relies on jurors making their decision based on the testimony and the evidence, if there’s any part of the prosecution that depends on science or some sort of mechanical equipment, it’s a rare engineer who doesn’t let it be known at some point that he thinks either that the science or methodology is bad or the equipment is flawed — and, oh, by the way, he’d have designed it / operated it differently.

    Mike Lief (e6260e)

  26. Yes, I agree with that.

    Patterico (cc3b34)

  27. I do civil litigation- in intellectual property. So, few actual trials. However, if I think the law and facts are on my side, I want the smartest folks on the jury. By the way, I was once an alternate, along with another lawyer, on a criminal battery case. For some reason, it was the LA City Attorney’s Office, rather than the DA’s office, prosecuting the case. The jury was full of highly intelligent people, including a teacher, an engineer, a supervisory executive and so on. The jury found the defendant not guilty. Without burdening this comment with more detail, the jury figured out, with no help from defense counsel or judge, that the evidence actually proved that the defendant was not guilty, as opposed to the prosecutor merely failing to make the case.
    My point is that, if we want justice done, and DAs represent ALL the people, including those accused, perhaps wrongly, of crimes, we should not shy away from having intelligent people on the jury. Now that we know that the “jury pool” may have been contaminated by the same paper that published the Kozinsky non-story, perhaps anyone relying on DNA evidence to prove a match should put on an expert to explain away the LA story.

    Ira (28a423)

  28. The accused man would be walking the streets — except that he is being held on an immigration hold.

    — In other words, very soon he will be walking the streets.

    What, regarding “reasonable” in the term reasonable doubt, do these CA juries not understand?

    Icy Truth (f4a81f)

  29. I agree this decision is an outrage, but to blame it on the L.A. Times articles in question smacks of post hoc ergo proper hoc. I don’t doubt the professor read the L.A. Times, but I also don’t doubt he has colleagues he trusts a hell of a lot more on the issue than any newspaper. Jurors have been ignoring even the most damning DNA evidence for a lot longer than that. Just ask O.J. Besides, it’s not as though the articles in question slam DNA evidence unfairly. The first article was wrong to say that the odds of that partial match having matched to an innocent (as opposed to the original odds that it would) was 1 in 3, but it’s not as though the real figure were any better. None of us know the odds of the criminal being in the database, so those odds are best described as “1 in hell-if-anyone-knows-so-let’s-convict-the-bastard-anyway.” The second article may not have revealed the non-surprise on the first page, but it did reveal it early enough in the article to appear on p. 1 of 6 on the online edition – certainly deep enough that one would expect a law professor following the story to have read that far. And neither article said anything about full DNA matches, which was the issue in this case.

    Sometimes I think that this case and the Puckett case, gross miscarriages of justice, both, are two sides of the same coin: people don’t understand DNA evidence, and oftentimes judges are no exception. Was the random probability match in this case really 1 in a trillion, or does that figure represent the likelihood that any particular record would return a match randomly rather than the more relevant likelihood that someone would? Was the difference explained to the jury? Since we’re talking about a full match this time around, the latter number would probably still be extremely high, of course, but that’s useless if not properly explained. It doesn’t take too many juries reading about 1 in 3 odds being portrayed as 1 in 1.1 million or the FBI refusing to allow Arizona searches before they stop believing anything anybody says about DNA.

    Xrlq (62cad4)

  30. Is there any way to address this Law Professor’s misconduct to get a retrial?

    PCD (5c49b0)

  31. On second thought, what is this professor’s name and school. We ought to publicize this travesty, lay it directly on that professor’s doorstep, and shame his school into heavily sanctioning him.

    PCD (5c49b0)

  32. I don’t want to turn this into a vendetta against a single juror, and so I have withheld any more information about his school, etc.

    Patterico (cc3b34)

  33. I do civil litigation- in intellectual property. So, few actual trials. However, if I think the law and facts are on my side, I want the smartest folks on the jury. By the way, I was once an alternate, along with another lawyer, on a criminal battery case. For some reason, it was the LA City Attorney’s Office, rather than the DA’s office, prosecuting the case. The jury was full of highly intelligent people, including a teacher, an engineer, a supervisory executive and so on. The jury found the defendant not guilty. Without burdening this comment with more detail, the jury figured out, with no help from defense counsel or judge, that the evidence actually proved that the defendant was not guilty, as opposed to the prosecutor merely failing to make the case.

    This is why I avoid lawyers on juries; no offense. Jurors love to think they’re more clever than everyone else and can figure out things that nobody else can. Lawyers fancy themselves more clever than others (so do doctors) and are even more susceptible to this pitfall.

    That’s why I often prefer a case with a small, easily solved puzzle to one where the evidence is clear. In the former, the jurors set their minds to solving the puzzle; once it’s “solved,” the defendant’s guilt is proven. In the latter case, the desire to prove themselves clever is frustrated, so they set their task as figuring out why such a clear case ever went to trial, and often start making up their own reasons (“I know! This must be a three strikes case!”).

    Patterico (cc3b34)

  34. #6, #7, one and all…

    I’m taking a short trip to Vegas this weekend (Thursday evening through sunday morning.

    Should you decide to appear, I might just buy that beer (or three).

    But (don’t tell anybody) I’m pre-Batson.

    God damn old people. 🙂

    The jury was full of highly intelligent people, including a teacher, an engineer, a supervisory executive and so on

    As the son of an engineer (my dad), a teacher (mom), and having interacted with their “fellows” and their bosses, I can tell you that the term used to describe those three groups is 80% mutually exclusive.

    For God’s sake, teachers in Illinois still think our Govenor is a great man, and seem ignorant of the fact he’s raided their state pension fund at least twice and has yet to pay any of it back.

    Scott Jacobs (fa5e57)

  35. I agree this decision is an outrage, but to blame it on the L.A. Times articles in question smacks of post hoc ergo proper hoc.

    Here’s another detail, which I left out of the post for space reasons: the defense attorney kept trying to get the issue of the Arizona searches before the jury, despite the court’s rulings to the contrary.

    I don’t doubt the professor read the L.A. Times, but I also don’t doubt he has colleagues he trusts a hell of a lot more on the issue than any newspaper.

    I don’t even know what that means.

    Jurors have been ignoring even the most damning DNA evidence for a lot longer than that. Just ask O.J.

    But there were genuine issues raised by the O.J. defense team about the collection and analysis of the DNA in that case. Not so here.

    Besides, it’s not as though the articles in question slam DNA evidence unfairly. The first article was wrong to say that the odds of that partial match having matched to an innocent (as opposed to the original odds that it would) was 1 in 3, but it’s not as though the real figure were any better. None of us know the odds of the criminal being in the database, so those odds are best described as “1 in hell-if-anyone-knows-so-let’s-convict-the-bastard-anyway.” The second article may not have revealed the non-surprise on the first page, but it did reveal it early enough in the article to appear on p. 1 of 6 on the online edition – certainly deep enough that one would expect a law professor following the story to have read that far.

    Sorry, but I think you give far too much credit to him. First, the discussions here have shown that even smart people don’t follow the math. Second, some of my smarter colleagues in the office often admit that they don’t read past the front page (often only looking at headlines) if they don’t have a particular interest in a particular subject. There’s no reason to believe this particular professor read the less sensationalistic parts of the story, and the stories were completely irresponsible in the way they presented the information.

    And neither article said anything about full DNA matches, which was the issue in this case.

    Not so. As I said in the post, the second article at the end darkly suggests that even a 13-point match may have duplicates in the system from other people. The article bases this on an Arizona-style search in Maryland that revealed three 13-point matches to a single profile. The article says:

    Experts say they most likely are duplicates or belong to identical twins or brothers. It’s also possible that one of the matches is between unrelated people — defying odds as remote as 1 in 1 quadrillion.

    Maryland officials never did the research to find out.

    Did Jason Felch or Maura Dolan show the slightest interest in trying to do this research themselves? Nah. It’s far more dramatic to leave open the tantalizing (tantalizing, that is, for anti-law enforcement liberals) possibility that even full matches are suspect. If they did the research and the experts were proved right, as they no doubt would be, where’s the drama in that?

    Granted, I’m pissed at authorities for not doing the research themselves and clearing up the issue.

    But Xrlq, due to the latent “anarchist wing of the Libertarian party” streak within you, you consistently minimize the errors and distortions by these LAT reporters on this issue. In particular, you seek to minimize their inexcusable overdramatization of the birthday problem by pointing out that they eventually get around to talking about the way the Arizona searches were done. Because this happens in the 18th paragraph, it barely makes it onto the first page as displayed on the Internet — a trivial fact upon which you consistently have placed high importance.

    Far more significant is the fact that, in the print edition, this was revealed on page A20, which you know full well many people didn’t turn to.

    Patterico (cc3b34)

  36. What is the best occupation to have (or at least claim to have) in order to be thanked and excused from jury duty?

    kaf (16e0b5)

  37. Here’s an analogy of the irresponsible way in which the article overdramatized the effect of the Arizona-style search.

    Pretend Joe Schmoe had just discovered the effect of the well-known “birthday problem.” Here’s how the L.A. Times might have told the story:

    Technical manual translator Joe Q. Schmoe was asking some folks at a party when they were born, when he stumbled across two people with the same birthday.

    The FBI estimated the odds of unrelated people sharing a birthday to be as remote as 1 in 365. But Mr. Schmoe discovered that the people at the party sharing a birthday were not related: One was black, the other white.

    In the years after his 2001 discovery, Schmoe found dozens of similar matches — each seeming to defy impossible odds. In some cases, people were found to be born on the same day in the same year — defying odds that were often around 1 in 20,000. In bizarre cases, people were found to be born within the same hour, in the face of odds around 1 in 500,000.

    [This is, by the way, true of myself and Matt Welch. We were born within minutes of each other. Also, I hasten to point out that I am making up these odds out of whole cloth, for entertainment value. — Patterico.]

    These findings by a little-known translator of technical manuals raised questions about the accuracy of federal birthday records, and ignited a legal fight over whether the nation’s birthday databases ought to be opened to wider scrutiny.

    [See Defying Unbelievable Odds, Page A20]

    [Sound effect: 90% of readers turn to the funnies and the horoscopes, while maybe 10% — I’m being generous here — blacken their fingers laboriously turning the pages aaaaaallll the way back to A20.]

    Federal officials say critics have exaggerated or misunderstood the implications of Schmoe’s discoveries.

    Indeed, experts generally agree that most — but not all — of the birthday matches were to be expected statistically because of the unusual way Schmoe searched for them.

    In a typical birthday scenario, a person might ask other people at a party whether they share that particular person’s birthday. In order to have a party where it is likely that someone will share that person’s birthday, it is necessary to have 253 people at the party. But Schmoe’s search looked for any matches among all the attendees at the parties that he attended, greatly increasing the odds of finding a match. In such cases, it is generally necessary to have only 23 people to have a likelihood of two people sharing a birthday.

    This is, of course, a classic manifestation of the “birthday problem.” Simply telling people that fact on the front page, of course, lessens the dramatic impact of this amazing discovery. Because it takes drama to place a story on the front page, we have decided to tell people about the amazing defiance of impossible odds on the front page, and save the pap about the results being mostly expected for Page A20.

    But hey, that’s not gonna mislead anyone. They revealed it in the 18th paragraph, on page A20, and that’s Good Enough for Libertarians.

    Patterico (cc3b34)

  38. What is the best occupation to have (or at least claim to have) in order to be thanked and excused from jury duty?

    Who knows? “Deputy D.A.” has always gotten me excused from criminal cases. But my wife actually served on a civil case.

    Patterico (cc3b34)

  39. 34, Scott, If you were in Laughlin Friday afternoon/evening (don’t know for sure when. it depends on driving conditons and traffic.), I’d make a point of catching you on our way from “Shakey Town” to Iowa.

    I don’t want to take the Northern route through “Lost Wages” to Denver because I don’t want to be driving a very loaded, and strange to us, rented minivan down the mountain from the Eisenhower Tunnel.

    PCD (5c49b0)

  40. 32, I am tired of Law Professors like Erwin Chimerinsky, Susan Estrich, and other perverters of legal minds screwing with the justice system just because they can.

    It is time that these professors and schools are reminded they have a real responsibility to the community and there are consequences for frustrating justice.

    PCD (5c49b0)

  41. What is the best occupation to have (or at least claim to have) in order to be thanked and excused from jury duty?

    Like Patterico, “Deputy D.A.” usually does the trick for me, but I did end up serving on a five-week civil trial.

    Mike Lief (e6260e)

  42. I’m flying direct. Only advantage of this hellhole town is a direct flight to Vegas…

    Scott Jacobs (fa5e57)

  43. Having lots of education doesn’t make one wise. Chief Justice Rose Bird was famous for saying that “not guilty” the only verdict that truely is just.

    Alta Bob (69331c)

  44. I don’t doubt the professor read the L.A. Times, but I also don’t doubt he has colleagues he trusts a hell of a lot more on the issue than any newspaper.

    I don’t even know what that means.

    That it’s likely the guy reads the L.A. Times, but unlikely that he gets most or even all of his inklings about DNA from it.

    Here’s another detail, which I left out of the post for space reasons: the defense attorney kept trying to get the issue of the Arizona searches before the jury, despite the court’s rulings to the contrary.

    Which, along with the FBI’s boneheaded policy of fighting to stop such searches from being conducted in the first place, is a brilliant strategy if the intent is to help the defense. Why act like you’ve got something to hide when you don’t? They should have allowed the evidence, explained the birthday paradox, and then brought in David Kaye to explain how many 13-point DNA matches could bbe expected to exist in the world. Simply comparing that number to the world’s population probably would be enough to convince most jurors that “yeah, it coulda happened, but what are the odds?” If not, the prosecutor could then explain that this DNA sample didn’t turn up as a result of a worldwide Arizona search, so the odds that someone in the world matches to someone else are irrelevant to the odds that any of them match to the particular DNA sample found at the crime scene. The jury would have concluded, in a manner reminiscent of Wayne’s World, that “maybe this guy was a random match, or maybe monkeys will fly out of my butt.”

    Instead, we had the noble defense attorney trying to get to The Truth (TM), which the prosecution and the courts conspired to withhold from the unwashed masses. As Chris Rock infamously said of O.J., I’m not saying the jury should have ruled as it did, but I understand it.

    But Xrlq, due to the latent “anarchist wing of the Libertarian party” streak within you, you consistently minimize the errors and distortions by these LAT reporters on this issue.

    How so? The 1 in 3 error made the case against Puckett look stronger than it was, not weaker. Upon reading the article, I thought Puckett was probably guilty, we just couldn’t be sure enough to sustain reasonable doubt. After doing the math, it became clear that a guy was convicted where we had (and still have) no way of knowing how likely he is to be guilty or not.

    In particular, you seek to minimize their inexcusable overdramatization of the birthday problem by pointing out that they eventually get around to talking about the way the Arizona searches were done. Because this happens in the 18th paragraph, it barely makes it onto the first page as displayed on the Internet — a trivial fact upon which you consistently have placed high importance.

    Ten years ago, when I read online newspapers myself but hardly anyone else did, I might have conceded that this was a trivial fact. Today, in an age where every Xth post of your own points out how few people read print news anymore, not so much. Does anyone have current figures on what percentage of L.A. Times readers read online vs. print?

    Xrlq (b71926)

  45. 43, I think I know of which Hellhole you speak of. Sometime we should get together for Pizza and pops. Or at least go to Hangover Fest in Cordova on Jan 1.

    PCD (5c49b0)

  46. To Patterico’s point in 38. My hair stylist (kinda hard to claim a cute twenty something is a barber) worked with another young lady when they discovered that they shared a birthday. They were both born in the same city, the one they both still worked in and in the same hospital, the only one in that city.
    When my stylist mentioned this to her mother she said, “Wait a minute, my roommate in the maternity ward had her baby the morning you were born and said she planned to name the baby (name of coworker)if she was a girl.
    Long chances come in all the time when they are not looked for.

    Have Blue (974cdf)

  47. “Instead, we had the noble defense attorney trying to get to The Truth (TM), which the prosecution and the courts conspired to withhold from the unwashed masses.”

    If you’re using irony here, I don’t understand your point. If you’re not, then you’re incredibly naive. Do you not understand that the defense attorney successfully moved (as defense attorneys in cold hit cases invariably do) to withhold from the jury the “truth” that her client’s sample was hit from a database?

    Patterico (cb905d)

  48. As a layman, I do think that DNA is probably the most accurate identifier available to us.
    As a small-l libertarian, I think that the authorities should be sanctioned when they screw-up an investigation (everyone is supposed to know the rules, after all); but, I think the exclusionary rules go too far when they allow the guilty to go free. It would be far better to make the offending authority individually liable to civil and/or criminal penalties for their mis-conduct.
    And, I would think that this “law professor” has to be on a NO-GO list in every DA’s office in the State of California.
    If not, then the “State” deserves to lose.

    Another Drew (bfaaf5)

  49. Which leads me to a fundamental question. You continually advocate that jurors be told about the database effect. How the f[vowel removed]ck are we supposed to do that when jurors aren’t told about the database? (At defense request — and don’t say “tell them anyway.” The case would be reversed.)

    Patterico (cb905d)

  50. Sometime we should get together for Pizza and pops

    The fact that you said “pop” and not “beer” tells me that you are a communist. 🙂

    Scott Jacobs (fa5e57)

  51. Can the family of the victim bring a civil case against the LA Times ?

    JD (75f5c3)

  52. Another Drew,

    As a small-l libertarian, you advocate that my office maintain a database of jurors to determine whether they are good or not?

    We deserve to lose if we don’t prove our case. If we prove it, but have failed to maintain a database of problem jurors, then we “deserve” to lose according to you. Do you then “deserve” to be victimized by the freed criminal? That makes about as much sense.

    Patterico (95444a)

  53. 51, I am far from a communist, just a diabetic. About 15 years ago I would have said Pizza and Schnapps.

    PCD (5c49b0)

  54. Only communists are diabetic. True fact. Google it. 🙂

    Scott Jacobs (fa5e57)

  55. “Because it takes drama to place a story on the front page,..” Comment by Patterico at #38

    That’s been the style of the LA Times since editor Shelby Coffey III wanted to attract “female readers”. Drama always leads and any facts, if they are included in the story at all, can be found on page A25+. Now, facts and a better rendition of the subject story can be found on the Internet. It’s still not the medium, as the LA Times editors insist, it’s still the message. What are the chances that the LA Times will get the story wrong and omit correction? The odds are not astronomical.

    C. Norris (a6fe6c)

  56. Pop is a noise. Soda is a drink. Communists call it pop. Just sayin’

    JD (75f5c3)

  57. LASlimes lives up to it’s name.

    College professor lives up to the reputation of being the stupidest of the stupid.

    Nothing new there.

    Scrapiron (c36902)

  58. Interesting problem. My daughter is a lawyer and FBI agent but was not challenged and ended up as a jury foreman. I was in a jury panel in Newport Beach when the case was declared a mistrial in Voir Dire. It was a civil case and the plaintiff attorney asked if anyone in the jury panel, not the jury, had ever been sued. Everybody raised their hands ! He had gotten a panel of Newport Beach contractors and developers! Plus a couple of doctors. That wasn’t the reason for the mistrial but it deteriorated from there. All in all, one of the more amusing mornings I’ve spent.

    On the OJ trial, I was recovering from surgery and watched most of it on Court TV. There were problems with the DNA samples, but I was convinced the case was lost when the jury caught Marsha lying about the time line for the limo driver. I don’t think that convinced them he was not guilty but it gave them the hook to allow them to acquit him. The only part of the transcript that the jury asked to be read back was that limo driver. She tried to shorten the time OJ had to get back and change and they caught her at it.

    Mike K (155601)

  59. I did some back-of-the-envelope calculations (which I won’t include in this post, but will provide if anyone cares) and computed that the probability that two randomly selected profiles in a CODIS-type database will have a 9-locus match is about 0.00000005.

    This implies that if you take a given profile and attempt to match it against the 360,000 profiles in the California DNA database, then the expected number of 9-locus matches will be about 0.018. In turn, this implies that the probability of seeing a 9-point match for a random profile is also about 0.018.

    My understanding is that this is about the level of certainty that you would get from advanced serology studies, prior to PCR technology. Whether this would constitute reasonable doubt would depend, I suppose, on the weight of the other evidence against the defendant. I do think it shows that a 9-locus DNA match is not dispositive.

    I will concede that the probability of a random match in this analysis is likely overestimated because it doesn’t allow for the presence of relatives and duplicate records in the DNA database. However, I don’t think such an adjustment would change the estimated probability by more than a factor of 10. This would leave the probability of a random match on the order of one in one thousand.

    W. Krebs (55a367)

  60. Pat, all I’m saying is that if you think that the DA in this case proved his case, and that this individual juror improperly swayed the jury (whether or not you can prove it), then to allow him on a jury in the future is stupid.
    Would you, in a case going to trial, allow this person to be sworn as a juror?
    And, I do believe that you keep lists – everyone keeps lists, whether formal or informal.
    When I had a business, I kept lists of people I would not deal with, vendors & customers – the heartburn wasn’t worth it.

    Another Drew (bfaaf5)

  61. Clearly prosecutors need to put more engineers and scientists on juries.

    Kevin Murphy (0b2493)

  62. #25 etc–

    Yes, engineers can be a pain, even to other engineers. But there are NO engineers that fall for the birthday problem. If DNA and statistics are at issue, engineers are not going to believe BS math. You can’t complain about idiotic and innumerate BS flowing to juries and then get rid of all the folks that can do math.

    Kevin Murphy (0b2493)

  63. It’s like the only Dilbert line…

    Boss’s secretary: “My this is shocking! It appears that 20% of sick days are taken on a Monday!”

    Boss: “What kind of idiot do they think I am?”

    Secretary: “Not an Idiot Savant. They can do math…”

    Scott Jacobs (fa5e57)

  64. So all these people that they’ve released because their DNA didn’t match the crime scene; how many of them are false negatives. One of these days the logic of this argument will catch up to them.

    narciso (c36902)

  65. Um, what is a “false negative” ?

    Kevin Murphy (0b2493)

  66. 55, ok, you asked for it. May the Blue Genie come and mess up your crap game in Lost Wages.

    PCD (5c49b0)

  67. Ha!

    I don’t play Craps!! 😛

    Scott Jacobs (fa5e57)

  68. But bonus points for the “King of Cars” refference…

    Scott Jacobs (fa5e57)

  69. Um, what is a “false negative” ?

    The same thing as a “False Positive”, only the opposite – the test yields a negative result (no match) when in reality there should have been one.

    Occationally this happens on drug tests…

    Scott Jacobs (fa5e57)

  70. Patterico says:

    “But I’d be willing to bet he’ll assault someone’s mom.”

    Be careful.
    Letting one man out on the streets by mistake, is better than dozens of people being put behind bars by mistake.

    That may be the ratio, until you can prove otherwise.

    Oiram (983921)

  71. 69, may Carroll Shelby refuse your business and not allow you to drool on the cars in his showroom.

    PCD (5c49b0)

  72. Letting one man out on the streets by mistake, is better than dozens of people being put behind bars by mistake.

    I doubt you would hold that view if it were your mother who was his next victim.

    Scott Jacobs (fa5e57)

  73.      In responding to my comments in #27, Patterico in # 33 writes,

    “This is why I avoid lawyers on juries; no offense. Jurors love to think they’re more clever than everyone else and can figure out things that nobody else can. Lawyers fancy themselves more clever than others (so do doctors) and are even more susceptible to this pitfall.”

         Perhaps I was not sufficiently clear in my post. The lawyers were alternates, and did NOT take part in the deliberations. But there were intelligent, educated and experienced (three non-identical qualities) on the jury, and because of that JUSTICE was done.
         By coincidence, the week after my jury experience (which was many years ago), I was at a party and met an assistant DA. She was bummed-a jury had just returned a verdict of not guilty on a case she helped prosecute against “a really bad man.” The jury, she said, must have been full of morons. She was eager to vent, and I listened as she described her whole case. (Yeah, she was pretty.) As best as I recall, it was some sort of financial crime. I asked her if, to prove that a crime had been committed, several elements had to be proved. She said yes. I asked if the judge had instructed the jury about all the elements, and she said yes. Having listened to her story carefully (like I said, she was pretty and I wanted to seem interested), I was able to ask her what was the evidence on two of the several elements. She got that look in her eyes that said, why did I waste my time talking to such a moron, and in her spoken words said that there was “no evidence on those points,” but the jury should have “convicted because he is a really bad man.” (She spun around and walked away. It’s OK. A few weeks later I met and dated an even prettier girl.)
         Should I form the stereotype from my jury experience, my conversation with the assistant DA at the party, and some of the comments here by assistant DAs, that prosecutors really don’t want to actually have cases decided on their merits, that they rather merely demagogue verdicts?

    Ira (28a423)

  74. Scott,

    Letting one man out on the streets by mistake, is better than dozens of people being put behind bars by mistake.

    I doubt you would hold that view if it were your mother who was his next victim.

    that strikes me as being unfair and very close to an assertion that the person you are speaking to is intellectually dishonest.

    Were it my mother in law who was his next victim (my mother’s dead, so it would be very difficult for her to be a victim of anything), I would still hold the view that it’s better for one guilty person to go free than two dozen innocents to be wrongly punished; the fact that my family drew the short straw and had to pay the price for that changes nothing; the overall price paid by the community is still lower.

    (None of this should be taken to imply that I agree that the numbers are 1:24, or that I think the rate of convicted innocents to acquitted criminals in our system is a good one; I don’t know enough to form a judgement about that. But in a hypothetical with the numbers Oiram used, I think the balance of equity would be correct, even if that meant that the acquitted criminal turned around and hurt me or mine).

    aphrael (e0cdc9)

  75. Patterico, my apologies; it was not my intent to instigate a pile-on regarding peremptory challenges.

    DRJ, thank you for the explanation. 🙂

    aphrael (e0cdc9)

  76. Were it my mother in law who was his next victim (my mother’s dead, so it would be very difficult for her to be a victim of anything), I would still hold the view that it’s better for one guilty person to go free than two dozen innocents to be wrongly punished

    And there is an example of a fundimental difference between us…

    I personally would hate every since person responsible for that guy having been aquitted – jurors and defense council alike.

    Though the benifit is that the State wouldn’t have to put him on trial again…

    Because I’d kill the SOB into teeny tiny pieces…

    Scott Jacobs (fa5e57)

  77. What is the best occupation to have (or at least claim to have) in order to be thanked and excused from jury duty?

    I have lost track of the number of times I have been called for jury duty. I have never been picked. I used to think that it was because what the other lawyers and the judge knew about me but now I suspect that it also might be what they are afraid I know about them. 😉

    nk (e38352)

  78. “Instead, we had the noble defense attorney trying to get to The Truth (TM), which the prosecution and the courts conspired to withhold from the unwashed masses.”

    If you’re using irony here, I don’t understand your point.

    I was responding to your comment about the defense attorney the defense attorney trying to get the issue of the Arizona searches before the jury, while the court ruled to the contrary. That couldn’t have looked good to a juror, the defense trying to tell them something the state doesn’t want them to know about.

    Do you not understand that the defense attorney successfully moved (as defense attorneys in cold hit cases invariably do) to withhold from the jury the “truth” that her client’s sample was hit from a database?

    Actually, no, I didn’t know that, but it’s also beside the point. Unless, of course, the jury was treated to a similar spectacle of a prosecutor trying to sneak through a different truth while the defense and the judge colluded to keep them in the dark about that. In which case, neither side would have come out looking too good.

    Which leads me to a fundamental question. You continually advocate that jurors be told about the database effect. How the f[vowel removed]ck are we supposed to do that when jurors aren’t told about the database? (At defense request — and don’t say “tell them anyway.” The case would be reversed.)

    If it were up to me, I’d amend the Evidence Code so that you could (and in fact would be required to) tell jury about the database. Barring that, the question of whether to identify the database search and the question of whether to present the odds honestly and accurately are two separate issues. There is nothing (apart from the judge’s innumeracy) to prevent a court from refusing to disclose the circumstances under which a DNA match was found, while correctly instructing a jury as to what the odds of a random match were under these undisclosed circumstances. In the Puckett case, the jury should have been told that they had found a match to 5 1/2 loci, and under the particular circumstances, the odds of a random match occurring were 1 in 3. That’s it. They didn’t need to know why the odds of a random match were 1 in 3, but they did need to know that they were. Similarly, in this case the jury shouldn’t have been told the odds were several trillion; they should have been told that under these particular circumstances, they were 1 in a few million, or whatever other number “several trillion” turns into after the database effect has been taken into account.

    Alternatively, if you think that telling the jury about the database search and telling them the actual odds of a random match resulting from that search are inextricably linked, and that there is no politically/”constitutionally” feasible way to get them admitted as a package, then the only acceptable solution would be to treat database searches (at least those resulting in partial matches) much the way we treat polygraphs today: useful as an investigative tool, but inadmissible at trial.

    Xrlq (b71926)

  79. 79, nk, do you have pictures of the judge and other lawyers from the last local Bar assn soiree or something? I remember in Omaha the lawyers always got drunk and fought like biker gangs, but never a word in the Omaha “Weird Harold”.

    PCD (5c49b0)

  80. No, sorry, PCD. I got gloriously drunk at a recent PD/defense attorney party, about forty of us plus our spouses, but I behaved well and walked home. Even if someone took pictures, I would have only looked slack-faced and sleepy-eyed. 😉

    nk (e38352)

  81. #74 If my ratio in #72 is accurate (just a guess really). Don’t you think that 12 of your friends and family could be vulnerable to spending the rest of their lives in prison? That’s a lot more lives affected than just one person.

    We’re talking about a number of that dozen with families, jobs, children.

    In prison for life or possibly the death penalty.

    I Love my Mother very much, but I’ll take my chances on a repeat offender towards her vs. 12 people being wrongly convicted….. my mother could end up wrongly convicted as well.

    Oiram (983921)

  82. You can get an innocent person out of prison.

    You can’t unrape a victim.

    Scott Jacobs (fa5e57)

  83. Oiram, that is (to me) an interesting point: for some reason the fear of being the victim of crime is more visceral and gripping, and easier for most people to believe in, than the fear of being wrongly convicted by the machinery of the justice system.

    Perhaps there isn’t enough required reading of Kafka in high schools?

    aphrael (e0cdc9)

  84. And since, statisticly, there will eventually be an innocent person convicted, by y’alls logic we should let everyone go because there’s that SMALL chance the innocent will end up in prison…

    Enjoy that tasty, tasty anarchy.

    Scott Jacobs (fa5e57)

  85. It’s wonderful how you exaggerate in #86 Scott. “y’alls logic we should let everyone go”. That’s not what I’m saying.

    A society that wrongly convicts on a huge scale is asking for anarchy and rebellion.

    Oiram (983921)

  86. #84 I would rather try to undue rape by mental counseling than trying to undue life in prison or even the death penalty.

    Oiram (983921)

  87. Scott: like anything else, there’s a balance between competing goals. I don’t think anyone in their right mind asks that there never be an innocent person convicted; that’a an unrealistic expectation. Nor do I think anyone in their right mind asks that there never be a guilty person wrongfully acquitted; that’s also an unrealistic expectation.

    The question is this: what’s the optimum point for both? And in particular, when one can be traded off against the other, what’s the balancing point for the trade-off?

    aphrael (e0cdc9)

  88. DNA and other scientific evidence is presented to a jury through the testimony of an expert witness. Each side can raise concerns or expose weaknesses in the other side’s expert testimony in cross-examination and/or through direct testimony from their own expert witness(es).

    Jurors can believe any or none of the testimony from expert witnesses. They can bring their common sense with them into the courtroom and the jury room. But they cannot discredit expert witness testimony based on their understanding of the subject matter from sources outside the courtroom. In other words, a juror cannot discredit testimony from an expert witness based on something he learned in college biology, or something he was told by his neighbor the scientist, or an article he read in the LA Times that discussed DNA.

    Did that happen here? I don’t know, but it worries me when a juror explains the verdict was due to the DA’s failure to prove its case “in light of recent controversies about DNA.” Thus, I would want to know if the expert testimony or cross-examination in this case specifically raised “recent controversies about DNA.” If not, I would be suspicious that one or more jurors discussed information they heard outside the courtroom to discredit the expert testimony.

    DRJ (a5243f)

  89. Oiram, that is (to me) an interesting point: for some reason the fear of being the victim of crime is more visceral and gripping, and easier for most people to believe in, than the fear of being wrongly convicted by the machinery of the justice system.

    That is the litmus test of a people who have faith in their government, their laws, and their system of justice, aphrael.

    nk (e38352)

  90. Good point, nk #91.

    DRJ (a5243f)

  91. Two prisoners in a cell in Russia:

    First Prisoner: How long are you in for?
    Second Prisoner: Twenty-five years.
    First Prisoner: What did you do?
    Second Prisoner: Nothing.
    First Prisoner: You are a liar. For nothing, you get ten years.

    nk (e38352)

  92. NK, at 93: *laugh*

    NK, at 91: I have faith in my government, my laws, and my system of justice; but I also believe that everything constructed by humans is fallible and that even a system constructed with the best of intentions, and manned by people with the best of intentions, will fail from time to time.

    Because we are not perfect, the things we create will not be perfect.

    aphrael (e0cdc9)

  93. 85, aphrael, the kids aren’t even required to read Captain Kangaroo, let alone Kafka.

    PCD (5c49b0)

  94. I might add: in art, the lack of perfection is often part of the beauty; in political systems, not so much.

    aphrael (e0cdc9)

  95. 87, Oiram, I believe you are just another ACLU type that overhypes a wrongful conviction to keep the guilty free. I think there are very few innocent convicted of major crimes, and too damn few of the guilty convicted.

    PCD (5c49b0)

  96. There has been a gradual loss of focus in this thread.

    Um, what is a “false negative” ?

    Comment by Kevin Murphy

    Here is the key. The difficulty people not trained in methodology have in understanding statistics and the next step, Bayes Theorem.

    There are false positive and false negatives. A false positive is a positive result when the result should be negative. The same is true of true positive and true negatives, but we don’t hear so much about them. The terms in epidemiology is sensitivity and specificity. A very sensitive test will give you all the true positives but also some false positives. A very specific test will give you all the true negatives but also some false negatives. The diagnostic value of a test is a point on a curve in which you get the most true positives and least false negatives.

    This doesn’t just apply to diagnostic tests. The curve that is used is called ROC curve because it also applies to radio waves and transmission lines, sorting our “noise” from true signal.

    Anyway, explaining that to a jury is a real problem and anyone who is dumb enough to think Letting one man out on the streets by mistake, is better than dozens of people being put behind bars by mistake.

    That may be the ratio, until you can prove otherwise. this, is probably impervious to scientific explanations.

    Mike K (155601)

  97. There has been a gradual loss of focus in this thread.

    That seems about normal for any web-based discussion of more than a dozen comments … 🙂

    aphrael (e0cdc9)

  98. “There is nothing (apart from the judge’s innumeracy) to prevent a court from refusing to disclose the circumstances under which a DNA match was found, while correctly instructing a jury as to what the odds of a random match were under these undisclosed circumstances. In the Puckett case, the jury should have been told that they had found a match to 5 1/2 loci, and under the particular circumstances, the odds of a random match occurring were 1 in 3. That’s it.”

    Please explain the exact wording of what you would tell the jury.

    This should be entertaining.

    Patterico (c5dfe1)

  99. http://obamascrapbook.com/index.htm

    Googlebomb it.
    Make it viral.
    Post it everywhere.
    Protect obama from liars and swiftboaters. They’re coming.

    husbus (c61201)

  100. By which I mean, it will be entertaining to see how you do this without falsely suggesting that millions of people in California fit the profile.

    I think the most revealing fact of this thread is fact that you didn’t know (even though I have mentioned it *many* times in these threads) that jurors never hear that cold hits came from a database. How this fact could have escaped you is beyond me, but it’s central to the analysis of why it’s not necessarily unfair for courts to refuse to tell juries about the database adjustment.

    Internalize that fact first — JURIES DON’T HEAR THAT THE HIT CAME FROM A DATABASE — and *then* rethink the issue.

    Patterico (f37983)

  101. Protect obama from liars

    Obama needs protection from the MSM and his supporters?

    This is going to be a harsh election…

    Scott Jacobs (fa5e57)

  102. 104, Obama, like Clinton, like the Daleys, only hires liars.

    PCD (5c49b0)

  103. 103

    “Internalize that fact first — JURIES DON’T HEAR THAT THE HIT CAME FROM A DATABASE — and *then* rethink the issue.”

    It seems simple enough to me. If the defense wants to tell the jury that the hit came from a database they should be allowed to do so. Are they?

    James B. Shearer (fc887e)

  104. I have other fish to fry.

    daleyrocks (d9ec17)

  105.      DRJ in comment #90 writes,

    “Jurors can believe any or none of the testimony from expert witnesses. They can bring their common sense with them into the courtroom and the jury room. But they cannot discredit expert witness testimony based on their understanding of the subject matter from sources outside the courtroom. In other words, a juror cannot discredit testimony from an expert witness based on something he learned in college biology, or something he was told by his neighbor the scientist, or an article he read in the LA Times that discussed DNA.”

         I’m shocked, shocked! DRJ’s statement is internally inconsistent. If a juror knows an expert is wrong, or even lying, based on whatever knowledge that juror has, that juror should discredit the expert. Otherwise, we would be throwing common sense out the window! Obviously, the juror would have to explain to the other jurors how he or she knows that the expert is wrong or lying, and that is why we have 12 jurors, and not just one or two.

    Ira (28a423)

  106. By which I mean, it will be entertaining to see how you do this without falsely suggesting that millions of people in California fit the profile.

    We don’t have to say 1 in 3 Californians match. All we have to say is that the odds of an innocent match occurring in this case were 1 in 3. The best approach would be to explain to the jury that there was a database search and how it works. That being off the table, the next best option would be to instruct the jury that exigent circumstances exist that make the odds of an innocent match artificially high in this case, which is why it’s so important that they forget everything they “know” about DNA matches and rely on the court’s instructions instead. If even that comes too close to revealing the dreaded database, then the only remaining option is to tell the jury what the odds are of a false match in this case, and accept the fact that they may falsely extrapolate that it also applies to other cases not before them. If they walk away with a false belief that one in three Californians share a particular DNA profile, so what? One in three Californians are not on trial. Only this one individual is, and all that matters is getting the odds right in his case.

    I think the most revealing fact of this thread is fact that you didn’t know (even though I have mentioned it *many* times in these threads) that jurors never hear that cold hits came from a database.

    I never said I didn’t know the fact is not revealed to jurors. What I said I didn’t know – and which I’m pretty sure you hadn’t said before – is that we have defense attorneys alone (or even primarily) to thank for this.

    How this fact could have escaped you is beyond me, but it’s central to the analysis of why it’s not necessarily unfair for courts to refuse to tell juries about the database adjustment.

    I don’t see how. It’s one thing not to tell the jury about a search, and quite another to allow prosecutors to quote statistics as though it never happened. Regardless of what the jury is or isn’t told, a database search either occurred or it didn’t. The odds of a match to an innocent non-relative were either 1 in 1.1 million, or they were 1 in 3. Misrepresenting 1 in 3 odds as 1 in 1.1 million is nothing short of lying with statistics. A lie doesn’t magically turn into a non-lie just because we also withheld a key fact from the jury which, if known to the jury, would have put the lie to the lie.

    Xrlq (b71926)

  107. #99 —

    OK, you can define something called a “false negative” but it really only applies when the information being evaluated is subject to corruption. the test is incomplete, or the test is fuzzier than the demanded results.

    I guess if you search a database and demand a “perfect match” you will get false negatives if the data isn’t perfect or the database is incomplete. WHich is why one would ask for lesser matches, resulting in a number of false positives.

    As I understand it that’s what happened here BUT the match they got was indeed neigh perfect (13 loci) along with some lesser matches. Certainly there were billions of negatives from not being in the database, but were any of them false?

    Kevin Murphy (805c5b)

  108. If the defense wants to tell the jury that the hit came from a database they should be allowed to do so. Are they?

    Err, no, thank you very much, Your Honor. I do not want the jury to know that my client is in a database of sex offenders. As a matter of fact, a courtesy copy of my motion in limine to keep that out was delivered to your chambers three days ago.

    nk (e69fdd)

  109. “All we have to say is that the odds of an innocent match occurring in this case were 1 in 3.”

    The second you tell the jury that, you’re telling them (falsely) that there is a 1 in 3 chance that the defendant is innocent.

    In other words, you’re explicitly suggesting that the court instruct the jury using the very LAT fallacy that set off this whole controversy.

    There would be no other rational way to interpret the language “The odds of an innocent match occurring in this case were 1 in 3.” Remember, this is a case where they don’t know there has been a database search — thanks to the supposdly critical need to protect defendants by hiding the truth from juries — and there is one and only one defendant sitting in front of them. How in the world is a juror to take that instruction, other than a statement that there is a 1 in 3 chance that this defendant is innocent?!?!

    Patterico (e6faae)

  110. What I don’t think you understand, Xrlq, is that use of these databases is nothing more than an investigative tool that brings suspects onto the radar screen — like unusable “word on the street” rumors might cause investigators to stick a suspect’s picture in a set of photos to be identified by a witness.

    The hit is not the evidence. Once authorities have the hit, they take the suspect’s DNA and compare it to the evidence sample. That’s the evidence — and even with degraded DNA there are often more matching loci than found in the database search. So if one hit is made off of seven loci. but the sample matches the evidence sample at nine loci, what statistic do you want jurors to be told? The one based on seven loci???

    It’s not a situation where the true number is 1 in 3 instead of 1 in 1.1 million. You’ve simply been taken in by the LAT spin, apparently because you never understood how these trials actually are conducted.

    Patterico (ec55ce)

  111. If they walk away with a false belief that one in three Californians share a particular DNA profile, so what? One in three Californians are not on trial. Only this one individual is, and all that matters is getting the odds right in his case.

    You’e in Bizarro-land. You think it’s hunky dory to suggest that a profile with a frequency of occurrence of 1 in 1.1 million in the general population might actually occur in 1 of 3 Californians — misleading jurors into thinking there are over 10 million possible suspects — THAT’S not lying with statistics! No sir!

    But telling them the ACTUAL frequency of occurrence, instead of an artificially adjusted value — a value that has relevance only to the meaning of a search that isn’t in evidence, isn’t used in the trial in any way, and is in fact hidden from jurors FOR THE DEFENDANT’S PROTECTION — THAT is somehow “lying with statistics.”

    In other news, up is down and black is white. That is all.

    Patterico (cc3b34)

  112. Folks, if a smart guy like Xrlq can spout stuff like this — WHOLLY on the basis of these LAT articles, which he KNOWS are flawed — what does that tell you about what we’re facing with these jurors?

    I truly despair.

    Patterico (cc3b34)

  113. Pardon my ignorance, but, if juries can’t be told that a hit was from a “database”, why are fingerprint matches allowed? Don’t they rely on a hit on a “database”?

    Another Drew (bfaaf5)

  114. Another Drew #116,

    No. Only the match between the suspect’s fingerprints taken when he’s booked and the prints lifted at the crime scene is allowed at trial. What you are talking about is an application for a search warrant or a knock on a suspect’s door. Investigation. Nothing to do with trial on the merits.

    There is an issue about the calculation of the probabilities in DNA matches but it has to do with the initial samples and their validity. I might have the chance to litigate that. But, as someone else said, the defense ox is as much in danger of being gored as the prosecution one. Faulty DNA has freed people as well as convicted them.

    nk (e69fdd)

  115. 111

    “Err, no, thank you very much, Your Honor. I do not want the jury to know that my client is in a database of sex offenders. As a matter of fact, a courtesy copy of my motion in limine to keep that out was delivered to your chambers three days ago.”

    If the defense doesn’t want the jury to know the hit came from a database then they should have to live with the single match probability. Their choice. But no one has answered my question. Is the defense allowed to tell the jury the hit came from a database if the defense wants to do so?

    James B. Shearer (fc887e)

  116. Ira #108,

    I agree that the opinions of expert witnesses cannot supplant the province of the jury. IMO that means jurors decide the weight and credibility to be given expert witness testimony. However, although jurors can accept or reject expert testimony, they cannot supplant the expert testimony offered in court with their own evidence of what the science is or the experts should have said.

    Thus, I don’t have a problem if this jury based its decision on conflicting opinions offered by the experts, and the jury decided to accept the the opinion of one expert and reject the opinion of another. I have a big problem if one or more jurors offered their own evidence about the science involved.

    DRJ (a5243f)

  117. But no one has answered my question. Is the defense allowed to tell the jury the hit came from a database if the defense wants to do so?

    No. There is an ethical duty to represent your client competently.

    nk (e69fdd)

  118. An enforceable duty.

    nk (e69fdd)

  119. nk…
    Thanks for the clarrification. I understand the commotion now.
    I personally give a great deal of credence to DNA, on both sides of the issue. If there is DNA at the crime-scene, and it matches that of the defendent, as far as I’m concerned, he’s toast.
    If, on the other hand, that DNA does not match him, the prosecutor had better have an overwhelming mountain of circumstancial evidence to tie the defendent to the crime.
    That being said, I understand the prejudicial nature of the jury knowing about a defendent’s DNA being in a data-base, just as his priors (if any) would also be prejudicial – though on many occassions, I believe (as a layman, mind you) that this aspect of the exclusionary rules can be highly counterproductive to the effective administration of justice.

    Another Drew (bfaaf5)

  120. 120 121

    You are being ridiculous. Ordinarily the defense would not tell the jury the defendent had been in jail. But if the defendent was in jail at the time of the crime it would. Similarly the defense is better off with the jury knowing the hit came from a database with a random match probability of 1/3 than with the jury just knowing a single test random match probability of 1/1000000.

    James B. Shearer (fc887e)

  121. It’s still the case that there is a random match probability of 1 in 1.1 million for the general population.

    Patterico (cc3b34)

  122. So you’d have two situations;

    1) The jury knows that the general random match probability is 1 in 1.1 million, and has no idea how the suspect got in front of them.

    2) The jury knows that the general random match probability is 1 in 1.1 million, and learns that the subject got in front of them via a database search. Good for the defense is that it now seems more possible that his presence in court is the result of chance. Bad for the defense is that the jury now knows the hit came from a pool of sex offenders.

    So they now know, not only that there might be x number of people in California who might have committed the crime, but also that everyone else from a pool of very likely offenders (sex criminals with DNA in the database) has been eliminated.

    Patterico (cc3b34)

  123. I believe the biggest problem of the LAT article was that the FBI was stopping research of the databases based on some wishful thinking they were going to get away without honestly describing the statistics and fallacies to the juries. (my interpretation of the FBI’s motives)
    Two problems. It immediately makes the DNA results suspect. What is the FBI trying to hide? (besides their laziness at having to explain the fallacies), and secondly it blocks research into possible subtleties of the probabilities of the DNA characteristics. 144 hits where only 100 was expected might be pure chance or this might indicate some of the characteristics are not quite as random across the population as thought. Perhaps some are correlated and are not independently random.

    gp (c64d75)

  124. Still, I believe that the answer to James B. Shearer’s question should be “yes.”

    There is a non-frivolous argument to be made that the jury intuitively has a “wow” factor based on the fact that there is a man sitting in front of them who has the same DNA as was found at the crime scene.

    If you know that 338,000 records were searched to get that man there — and you also know that, even in a search of complete innocents, there would be a 1 in 3 chance of a match — the “wow” factor is lessened. As it should be.

    So it seems to me that the defense could ask for the fact of the database to come in.

    That’s a tentative conclusion, off the top of my head. But it seems to make some sense.

    Patterico (cc3b34)

  125. You would think that the defense would do all it could to raise the possiblity of reasonable doubt?

    Another Drew (bfaaf5)

  126. 125

    “… Bad for the defense is that the jury now knows the hit came from a pool of sex offenders.”

    I don’t believe the database (in the Puckett case)consisted solely of sex offenders.

    James B. Shearer (fc887e)

  127. I think you’re right, James. I think, instead, it was felons. I stand corrected.

    Patterico (cc3b34)

  128. DRJ #119:

    “I have a big problem if one or more jurors offered their own evidence about the science involved.”

         But that is why we have juries. Jurors, obviously in varying degrees, bring in with them at the start of any case, intelligence, knowledge, experience and wisdom. If some expert says “up is down and black is white” (see, Patterico #114), I would “truly despair” (see Patterico #115) if jurors with reason to know better didn’t speak up.
         DRJ states, “I agree that the opinions of expert witnesses cannot supplant the province of the jury. IMO that means jurors decide the weight and credibility to be given expert witness testimony. However, although jurors can accept or reject expert testimony . . . .” Is DRJ saying that jurors should just rely on the good looks any honey voices of the experts?  

    Ira (28a423)

  129. I think, instead, it was felons.

    Not all that much better, frankly…

    Scott Jacobs (d3a6ec)

  130. yes, Ira, it is a violation of their oath to consider evidence that was not presented to them in court.

    SPQR (26be8b)

  131. Ira #131,

    I want jurors to use their common sense and judgment, too, but the issue is whether jurors can consider evidence not presented at trial. If the expert testimony in this case raised questions about DNA methodology and testing, the jury could properly consider that. If the expert testimony did not raise those questions, I don’t think the jury can consider that during deliberations — especially if it’s solely based on reading a newspaper article about problems with DNA. Jurors aren’t there to be the experts or to “solve” the case. They are there to decide the case based on the evidence.

    Here’s an example in another context: Assume there is a lawsuit involving a head-on collision that kills both drivers. There are no passengers or witnesses, and it’s unclear which driver was at fault. Both sides hire experts to recreate the accident and each presents evidence showing that the other driver crossed the center line and caused the wreck.

    The jurors are entitled to decide which expert provides the best explanation of what happened based on the evidence in the case. Each juror can believe or reject some or all of the expert testimony, but a juror cannot:

    (1). Visit the location of the wreck during the trial and tell the other jurors the next day what he saw, including that the location “looks different than the witnesses said it did.”

    (2). Read newspaper articles about the wreck that includes information not in evidence in the trial and speculate how the information would change the experts’ opinions.

    (3). Discuss a different accident reconstruction technique he read about in a magazine that was not used by the experts in this case, and use that to decide who was at fault.

    Jurors sometimes do these things because they want to decide the case the best they can and they aren’t sure who to believe or who is right. But jurors can’t and shouldn’t substitute their methodology for the experts any more than they can or should gather their own evidence.

    DRJ (a5243f)

  132.      DRJ #134: Fair enough. I agree that jurors should not do any of (1) to (3) listed in your comment.
         On the other hand, suppose that (i) the evidence clearly and convincingly showed that car no. 1 left a 70 foot skid mark on the LA residential city street leading up to the four-way-stop-sigh intersection where the collision occurred, (ii) expert no. 1, the expert for the beneficiaries of the driver of car no. 1 testifies, without any reference at all to the skid mark, that in his expert opinion car no. 1 had not at any relevant time leading up to the accident been traveling at an excessive speed and that that fact was an important factor in his accident reconstruction, and (iii) both the expert for the beneficiaries of the driver of car no. 2 and their attorney just plain fail to link the skid mark to the speed of car no. 1. (Hey, some attorneys and experts will “lie” by omission or by purposely testifying about an incomplete context, and some attorneys and experts are careless.) Should the jury not find that expert no. 1’s reconstruction is meritless in view of their common sense conclusion that car no. 1 had to have been speeding?

    Ira (28a423)

  133. Technically, either side could ask to admit any relevant scientific evidence if it met the Daubert test (Frye test in Illinois). So yes, it is a question of strategy and weighing the potential prejudice against the probative value in regard to your case. If your client is charged with burglary and has a prior conviction for burglary do you put him on the stand knowing the prosecution will read his prior conviction into the record at the conclusion of his testimony? How about a prior conviction for battery?

    nk (e69fdd)

  134. “All we have to say is that the odds of an innocent match occurring in this case were 1 in 3.”

    The second you tell the jury that, you’re telling them (falsely) that there is a 1 in 3 chance that the defendant is innocent.

    Which was true before the search was conducted, and is a hell of a lot closer to the truth after the fact than it would be to (even more falsely) tell them there is a 1 in 1.1 million chance that the defendant is innocent. Both figures are tainted by the prosecutor’s/selection fallacy, of course, but at least 1 in 3 is an accurate representation of the original odds.

    In other words, you’re explicitly suggesting that the court instruct the jury using the very LAT fallacy that set off this whole controversy.

    No, I’m not. There is nothing fallacious about saying that if you fish in a pond of 338,000 and stand a 1 in 3 chance of catching a fish, then that is the relevant figure to consider – not the nearly impossible odds that you will catch any particular fish. Unless, of course, there is something special or otherwise noteworthy about one particular fish in the pond, e.g., a radio station has offered a special prize for whoever catches Diamond Jim. Puckett is not Diamond Jim. He wasn’t an identified suspect before the trawl. He wasn’t anything; just another nameless, faceless fish in a pond of 338,000. The only reason we’re talking about him now is because he is the one that happened to turn up a match – a fact which, as you know, stood a 1 in 3 chance of occurring whether the killer was in the database or not.

    There would be no other rational way to interpret the language “The odds of an innocent match occurring in this case were 1 in 3.” Remember, this is a case where they don’t know there has been a database search — thanks to the supposdly critical need to protect defendants by hiding the truth from juries

    I’m not arguing that the truth should be hidden from juries. But it’s one thing to make a key fact inadmissible, and quite another to effectively declare that fact a non-fact and allow parties to tell lies inconsistent with it.

    — and there is one and only one defendant sitting in front of them.

    At this stage, sure. Just like Mike Nifong had “only” three defendants, not the entire Lacrosse team – after he had conducted his infamous “no wrong answers” police line-up. We started with 338,000 potential suspects, knowing that there was a 1 in 3 chance of getting the wrong guy, and a near certainty that whoever that person might be would end up on the wrong side of a criminal trial. That that person happened to be Puckett rather than someone else is an irrelvant detail (except, of course, to Puckett himself). This is the Monty Hall problem all over again, only this time innocents are going to prison over it.

    If you disagree, I have the perfect solution to solve North Carolina’s problem with cold cases. We’ve already done all we can with our database, so why don’t we borrow yours, Hawaii’s, and as many others as we need to pull from states nowhere near Carolina to ensure that every trawl comparing 5 1/2 loci will yield at least one match (if we get two or more matches, we’ll find some way to exclude all but one, whether by drawing lots, flipping coins, or whatever). Of course we can’t use the trawl itself as evidence, so we’ll conduct a second DNA test, checking again to see if the same 5 1/2 loci really do match. If they do, we’ll throw the book at the guy, telling the jury that the odds that he was wrongly selected are only 1 in 1.1 million.

    How in the world is a juror to take that instruction, other than a statement that there is a 1 in 3 chance that this defendant is innocent?!?!

    They should take it as a statement that there was originally a 1 in 3 (np) chance they would land on the wrong guy, vs. a 1 in God-only-knows-what chance that they landed on the right guy (this goes back to the elusive odds that the killer was in the database to begin with). Since we have no idea what the latter figure is, the DNA match should not have been admitted into evidence at all.

    What I don’t think you understand, Xrlq, is that use of these databases is nothing more than an investigative tool that brings suspects onto the radar screen — like unusable “word on the street” rumors might cause investigators to stick a suspect’s picture in a set of photos to be identified by a witness.

    Not true. Puckett wasn’t just identified by the trawl, and then convicted based on evidence unrelated to the DNA match it turned up. He was convicted primarily on the basis of match itself, based in large part on the jury’s false belief that the odds were only 1 in 1.1 million that the cops would/had get/got the wrong guy, when in fact they were 1 in 3 that they’d get the wrong guy, and the wrong guy just turned out to be him. So what if they conducted a second DNA test on the Same. Goddamned. Loci. That doesn’t make the second test any less tainted by the database effect that applied to the first. Your hero Kaye addresses this theory, which he rightly derides as “a foolish argument,” on pp. 4-5 of his paper.

    The hit is not the evidence. Once authorities have the hit, they take the suspect’s DNA and compare it to the evidence sample.

    Oh, please. That’s like saying “word on the street was X” wasn’t the evidence, it was just the lead. Then cops followed up on the lead to determine whether word on the street really was X, and lo and beyhod, it was, so that became the evidence.

    That’s the evidence — and even with degraded DNA there are often more matching loci than found in the database search.

    Often, sure. In Puckett’s case, no. So why are you bringing it up in this context?

    So if one hit is made off of seven loci. but the sample matches the evidence sample at nine loci, what statistic do you want jurors to be told? The one based on seven loci???

    No, of course not. Statistically, this scenario is equivalent to testing the entire database for a match to 9 loci; you just accomplished it in a cheaper manner by eliminating most of the potential matches with a (presumably cheaper) 7 loci test. Obviously, any record that doesn’t match at 7 loci isn’t going to match at all 9, so there’s no point in conducting the more expensive tests on the low-hanging fruit. But in effect, you’ve gone fishing in a database of n records to find one that matched at 9 loci, so the appropriate jury instruction would be the random match probability factor for 9 loci, multiplied by the size of the database.

    In other words, if Puckett’s record had first been pulled from a 5 1/2 loci trawl, and then found to match at 9 loci, then the 1 in 3 figure would indeed be wrong, and by pure coincidence, the 1 in 1.1 million figure would probably be about right. But that didn’t happen in his case, did it?

    It’s not a situation where the true number is 1 in 3 instead of 1 in 1.1 million. You’ve simply been taken in by the LAT spin, apparently because you never understood how these trials actually are conducted.

    Then kindly enlighten me as to how Puckett’s trial was actually conducted. We know from “the LAT spin” that the original trawl was based on 5 1/2 loci, thereby running a 1 in 3 risk of matching to an innocent non-relative. Was his DNA later matched at the same 5 1/2 loci, as “the LAT spin” suggests, or was it matched at more loci than that? Or, as I suspect, did they just run a second test that did nothing but confirm the findings of the first? In other words, are you serious advancing the very same argument your hero rightly described as “foolish?”

    Finally, in response to James’s inquiry, you write:

    So it seems to me that the defense could ask for the fact of the database to come in.

    That’s a tentative conclusion, off the top of my head. But it seems to make some sense.

    If you’re right, then I think Puckett has a separate basis for appeal: ineffective assistance of counsel. Puckett’s priors were discussed at trial anyway (they formed the basis of the flimsy “M.O.” theory, remember?) so he stood absolutely nothing to gain by keeping the database out of evidence. But he stood everything to lose by allowing a disingenuous prosecutor to argue that the odds were 1 in 1.1 million that he had been wrongly convicted, rather than 1 in 3 that someone would be, and that someone just randomly happened to be him.

    Xrlq (b71926)

  135. No offense to the great State of California, but the extent to which it allows prior bad acts in as M.O. is ridiculous.

    nk (e69fdd)

  136.      Without getting into detail about the Monty Hall problem mentioned in Xrlq’s comment #137, I have read that mathematicians debated the answer, and were only convinced AFTER extensive experimentation proved which answer is correct. I am an engineer as well as an attorney, and I fall into the group that was surprised by the outcome of the experiments.
         In any event, it appears to me that there were two statistics in the Puckett case (aka the Sylvester murder case): (i) There is a 1 in 1.1 million chance that anyone would have the 5.5 loci found in the unidentified suspect’s DNA; and (ii) there is a 1 in 3 chance that the database searched by the police would have at least a single match to the 5.5 loci.
    I am totally at a loss as to how anyone can conclude that the statistics mean that there is a 1 in 3 chance that the database search in the Puckett case would hit upon an innocent person. Like I said, I was among those surprised by the answer to the Monty Hall problem.
         However, I’m not sure that the Monty Hall analysis even applies here. In the Monty Hall problem, there is a 100% chance that the prize is in the database (the areas behind the doors). In the Puckett case, the prize (i.e., the guilty party) may not even be in the database.
         In my opinion, the second statistic is irrelevant to the ultimate finding of guilt or not. If the database had records of more people in it, the likelihood that at least a single match would be found in it would go up, and if the database had fewer records in it the likelihood would go down.
         The DNA and the database can logically be used for only two purposes, to potentially exclude someone as a suspect or identify someone as a potential suspect. As the deputy DA in the Puckett case said, there were about 8 California white men walking around with the DNA of the perpetrator. (The eyewitness had given the police a statement that the perpetrator was a white male.)
         However, he was not taking into account that non-Californians who match the DNA description may have been in San Francisco at the time of the crime. What was made known to the jurors is that several men, let’s say 8, “fit the description” provided by the DNA in the Puckett case, that Puckett is one of them, that Puckett was in San Francisco at about the time of the crime, that Puckett had committed at least later than the subject crime at least three other rapes or sexual assaults, that the MO of the three known crimes committed by Puckett may or may not have been the same as that used by the perpetrator in the Sylvester murder case, that the identities, whereabouts and criminal records of the other 7 men fitting the DNA description are unknown, that, other than Puckett’s presence in San Francisco at the time of the crime and the DNA description, no other physical evidence links Puckett to the crime, and that no effort was made to have the only living eyewitness to the crime identify Puckett’s photograph, though that eyewitness, who died 9 months after Puckett was identified by the police as a likely suspect, may not have had a sufficiently good look at the perpetrator to make any useful identification even if the actual perpetrator were put right in front of her.
         In summary, the people’s evidence in the Puckett case is that out of all the white men in California in 1972, only 8 could have committed the crime, and one of them was sitting in court. Since we know nothing about the other 7, I guess we can say that there is a 1 in 8 chance that Puckett is the person who committed the crime.

    Ira (28a423)

  137. Ira #135,

    Absolutely, but the jury’s decision to reject Expert #1’s testimony isn’t based on common sense — it’s based on the evidence they heard in the courtroom. The expert’s theory of the wreck doesn’t match the evidence produced at trial.

    The scenario I posit is the possibility that a juror discredited Expert #1’s testimony because of something negative he read in a newspaper article about DNA methodology. I think that’s improper if the juror’s problem with the methodology wasn’t raised at trial.

    DRJ (a5243f)

  138. DRJ # 140: I think you are mixing our hypotheticals. In ##134 and 135 we are talking about automobile accident reconstruction, not DNA methodology.
         I think you are trying to avoid calling a spade a spade. In our hypothetical, the jurors used their knowledge, experience and common sense to apply the evidence in a way suggested by no one in the case. The jury realized, through no help from the judge or any of the parties’ counsel or eperts that Expert No. 1’s methodology was undermined by the evidence. That should have happened in the hypothetical, and knowledgeable, intelligent jurors would be needed for that to occur.
         In connection with the case which inspired this whole thread (let’s call it the 13 Loci Case), I agree that no juror should have based his or her decision on purported “recent controversies about DNA.” That it was a law professor who said that was the basis of his decision is distressing. (Hey, Pat, tell us what law school is he from so I can steer my kid away from it in case my kid decides to go to law school.) We do NOT know whether or not the other jurors agreed with his position. We also do not know whether or not there were other jurors with sufficiently impressive educational backgrounds to stand up to this professor. This, of course, is an argument for having MORE jurors with education and experience on each panel. We need, lawyers, truck drivers, teachers, nurses, waiters, engineers, life guards and the like as jurors.
         If I were a juror on the 13 Loci Case, and if the only evidence were that the perpetrator left his semen behind, that the DNA of the semen matched the DNA of the defendant and that the likelihood of anyone else on earth having the same match was less than one in a trillion, I would vote guilty.
         By way of contrast, if I were a juror in the Puckett case discussed in this thread, I would probably vote not guilty. (Of course, that just means that I don’t think that the people proved its case.)

    Ira (28a423)

  139. Ira,

    I did mix hypotheticals. The first paragraph of my last comment was a response to your comment 135, but my second paragraph related that to the example in the post. I was trying to draw a parallel.

    DRJ (8b9d41)

  140. Ira:

    In any event, it appears to me that there were two statistics in the Puckett case (aka the Sylvester murder case): (i) There is a 1 in 1.1 million chance that anyone would have the 5.5 loci found in the unidentified suspect’s DNA;

    Wrong. There there is a 1 in 1.1 million chance that any randomly chosen individual, other than the perp himself or a blood relative, will nevertheless randomly match to the perp. That is the random match probability factor (RMP). If, rather than searching the database in its entirety, the lab tech had looked at one record only, the odds that that record would have matched would have been 1 in 1.1 million (or slightly higher than that, once you take into account the remote possibility that that one randomly selected record could have been the perp or a relative).

    In other words, if Puckett didn’t do it, and they had only searched HIS record rather than everyone else’s, the odds of a bad match in his case would be in 1 1.1 million.

    and (ii) there is a 1 in 3 chance that the database searched by the police would have at least a single match to the 5.5 loci.

    Wrong again. Based on the RMP, multiplied by the size of the database, the odds are 1 in 3 that the database search would yield at least one match to an innocent non-relative. To the extent that the perp himself and his family members may also have been in the database, the odds of the search turning up any match at all are necessarily higher than that.

    I am totally at a loss as to how anyone can conclude that the statistics mean that there is a 1 in 3 chance that the database search in the Puckett case would hit upon an innocent person.

    Then you don’t understand the RMP. At the macro level, a search of one innocent (non-realtive) record carries a 1 in 1.1 million chance of matching to that innocent. Play the same game 338,000 times, give or take a few thousand, and the odds rise to 1 in 3.

    Like I said, I was among those surprised by the answer to the Monty Hall problem.
    However, I’m not sure that the Monty Hall analysis even applies here. In the Monty Hall problem, there is a 100% chance that the prize is in the database (the areas behind the doors). In the Puckett case, the prize (i.e., the guilty party) may not even be in the database.

    That’s neither correct nor relevant. It’s incorrect because in both examples the relevant “database” is the subset of the data that is subject to the non-random process, i.e., the database search in the case of DNA, or Monty’s selection between doors in the game. When the contestant chose Door #1, he took that door out of the running. Monty can’t open Door #1 no matter what is or isn’t behind that door or any other. So in effect, when you chose Door #1, you pulled it out of the database, which now consists of Doors #2 and #3. There’s only a 2 in 3 chance that the prize is in the relevant database. One certainly exists, sure, but it could be behind Door #1, which is not being examined by Monty. Similarly, for any given perp matched at 5 1/2 loci, there certainly is a “prize” out there somewhere – 6,000 of them is more like it – but the odds are about 1 in 3 that there is one in the database.

    It’s not relevant, though, because nothing about the Monty Hall problem depends on the certainty that a prize exists anywhere. It will work for any kind of odds, any number of doors, and any kind of prize (with the caveat that if the prize is a random DNA match to an innocent non-relative of the perp, it may be a “prize” we’d rather no win). What matters is that you’ve got a random match probability factor (1 in 1.1 million for a DNA record, or 1 in 3 for a contestant), and a contestant is tricked into applying that random match probability factor to data which was not selected at random. Doors 2 and 3 had a 2/3 chance of having the prize behind them, but Monty’s selection between Doors 2 and 3 wasn’t random; it was specifically aimed at preserving the door that had the prize, if either of them did. Thus, so the odds that used to apply to Doors 2 and 3 together now apply to Door 3 alone. In the DNA case, same thing. The 338,000 records collectively (Doors 2 through 338,001, if you will) had a 1 in 3 chance of containing a “prize.” Cops didn’t choose their defendant randomly among the 338,000 records; they deliberately opened every door that clearly did not have a prize (didn’t match DNA at all, and therefore could not have matched randomly), leaving that one potential candidate standing (the one that certainly contained a DNA match, which may or may not have been a “prize”). Thus, Puckett alone now carries the odds originally associated with the entire database – just like Door #3 does in Monty’s game, and for the same reason.

    There are a gazillion variations you can play of Monty’s game. All that’s necessary for the paradox to hold are that (1) the original rules governing the distribution and/or placement of the prize are random, and (2) Monty’s intervening act looks random, but isn’t.

    Xrlq (62cad4)


Powered by WordPress.

Page loaded in: 0.1674 secs.