DNA and Guilt
the L.A. Times has an interesting article on partial DNA searches, prompting three posts by Patterico (so far) and others by Eugene Volokh and Radley Balko. In a nutshell, John Puckett was convicted in 2004 of raping and murdering Diana Sylvester in 1972, mostly on account of a partial DNA match. Like many of the older DNA cases, it wasn’t possible to run a full DNA match, which is essentially failsafe, but just a partial one, which has roughly a 1 in 1.1 million chance of matching the wrong person.
Sounds pretty damning, doesn’t it? I mean really, if the odds of a false match are really 1.1 million to one, what are the chances they matched the wrong guy? Pretty high, actually, if you searched 1.1 million times. Buy enough lottery tickets, and you will win. In Puckett’s case, they didn’t search 1.1 million records but did search 338,000, resulting in roughly one-in-three odds that someone would get falsely matched, or about 1 in 4 that exactly one person would. The actual odds are a bit lower than that once you control for the uncertain odds that the killer was in fact in that database; presumably, if he was, he certainly would have gotten a hit, while the odds are only 1 in 3 that a second person also would have. Only one person was matched, so we can be certain that either Puckett was matched because he was the killer, or he got unlucky based on 1 in 4 odds, times whatever the odds were that the killer was not in the database. Without knowing the odds of the killer being in the database it’s tough to say how serious that error was in Puckett’s case, but easy to say exactly how serious it is in any case like Puckett’s where we don’t know for a fact that the defendant was the only match (or the only match to a person who doesn’t have a 100% airtight alibi): 1 in 3.
Some would argue that a partial DNA match that shows 1 in 1.1 million odds against a previously identified suspect, but only 1 in 3 odds against a suspect for whom the DNA semi-match was itself the basis of the suspicion, should not be admissible in court. I disagree. Anything that says you’re twice as likely to be guilty as innocent is highly probative of the charge. It is crucial, however, that such evidence be presented for what it is: enough to make you think he likely did it, but without other, unrelated corroborating evidence, not nearly enough to extinguish reasonable doubt.
I should note that the same math problem, known as the “prosecutor’s fallacy,” likely occurs every day even with full DNA matches. There, the error is equally lame in theory but harmless in practice. Without knowing exactly how long the odds have to be in order to surpass reasonable doubt, I am pretty confident that that number is somewhere north of 3 but south of 1 million. So if a prosecutor tells you that the odds of a false full match are 1 in 1 quintillion, but neglects to tell you that he found the guy by combing through a database of 1 million individuals, all that means is that the odds have fallen “all the way down” to 1 in 1 trillion. No big whoop.
May 11th, 2008 at 2:29 am
In Puckett’s case, they didn’t search 1.1 million records but did search 3.8
typo: 3.8 should be .338
May 11th, 2008 at 12:55 pm
D’oh! Fixed now. Thanks.
May 11th, 2008 at 1:37 pm
I’m not sure it’s right in your post to say: “Like many of the older DNA cases, it wasn’t possible to run a full DNA match, which is essentially failsafe, but just a partial one, which has roughly a 1 in 1.1 million chance of matching the wrong person.” The random match probability is the frequency of occurrence of the profile in a population of unrelated individuals. When you start using language like “chance of matching the wrong person” it can be interpreted by people as “we got a match and there’s only a 1 in 1.1 million chance of matching the wrong person” which would be wrong. So I would caution against expressing it that way, or as “the odds of a false match.”
May 11th, 2008 at 2:47 pm
I see your point. Mine is that there are two separate questions here: (1) how likely a random match was to occur in the first place, and (2) given that a match did occur, how likely is it that it arose as the result of a random match. This post seeks only to tackle (1). IOW, I’m not disputing that the Times got the prosecutor’s fallacy wrong. Just clarifying that when it comes to asking how likely a false match was to occur in the first place, the odds were indeed 1 and 3, as the Times reported. Are we in agreement up to that point?
As for your misgivings about my reference to “the odds of a false match,” I’m not wedded to the phrase so if you have better terminology to recommend, I’m all ears. The only ground rule is that whatever terminology we agree upon must necessarily preserve the distinction between a match to the original donor (which I’ve thus far described as a “true” match) and a match to an unrelated individual who just happens to share 5 1/2 DNA indicators with the original donor (which I’ve thus far described as a “false” one).
May 11th, 2008 at 2:58 pm
“Just clarifying that when it comes to asking how likely a false match was to occur in the first place, the odds were indeed 1 and 3, as the Times reported.”
In the scenario where you got only one match, you have to be very careful with your language so as not to imply that these odds apply to the one match in front of you. The best way to do this is to say that, in a database of people unrelated to the donor, there is still a 1 in 3 chance of a match — which, because of your initial assumption, would necessarily be a false match.
May 11th, 2008 at 3:27 pm
I agree, but now you’re into the next issue, which is the subject of your newer post: confusing the odds of a false hit going in with the odds that a false hit actually occurred, based on what we know now. Still, if all we know is that there was a 1/3 chance of generating a false hit in the first place, and we have no clue how likely a true hit was to occur, I’m not sure what else we have to go on besides the 1 in 3 figure. You can only measure the odds according to the variables you know.
Also, by far the biggest problem in this particular case (IMNSHO) is that the jury wasn’t told this was a database search at all. Given that, they couldn’t properly have considered that this was the only hit, or therefore that an extremely high probability that the killer was in the database they never heard of may have precluded the possibility of this being a false hit altogether. That could be the key on appeal, though, if the odds were long enough that the killer definitely was in the database to preclude reasonable doubt on that basis. It would be a fun decision to read, though, as the gist would be “that dumbass prosecutor pulled some wild shit out of his ass, but this was a harmless error because as luck would have it, the real stats the dumbass prosecutor should have quoted were pretty damning, too.”
May 11th, 2008 at 7:32 pm
The problem with reference to a database is simple actually. The main reason that someone’s DNA is in a database is because of either PC 296 or because they were a bad enough criminal to get typed before 296.
So…the odd’s are all skewed, because the assembled population is not random at all, it’s a population of previously taken samples because of a prior conviction for a felony. I suppose say there’s a 1 in 4 chance that any profile in the database would match, but be fair to the jury and also tell them why he’s in the database in the first place. But that’d never happen because it’s “too prejudicial.”
May 11th, 2008 at 8:05 pm
Just to clarify, however… I do think the best course of action is to tell the Jury as much as we can, and let the warts be known ahead of time. I think the database relevance isn’t that much because a priori it’s impossible to know if the included party is guilty or not. So…what to do…
I think it’s best to explain that 1 in 1.1 million means that at the date of the match approximately 6000 people would be returned as a match to that sample. It’s important to know whether those matches would be correlated with gender/race/etc. If not (that is they are statistically independent), then go ahead and use a description (if given) (for example white ~20-60 year old males) to eliminate as many as relevantly possible (it probably elimanates nearly 90% of the total population, but for simplicity let’s say 75%, and so that leaves 1500 people. And just say…1500 people in the world would share that match. Then, let the jury figure out what that means… I suppose the defense attorney could argue that based on the DNA alone that there is a 1499 in 1500 chance that their defendant is innocent.
Then, allow the prosecutor to come around and point out that almost no evidence is good enough to eliminate as many as 6.99985 billion people. But that’s what DNA in this case has done. So how do we know it’s this one…point out that the defendant has propensity evidence (admissable under EC 1108), that he can be put in the area on the same day, that the marks match his MO, and the lack of forced entry match. Then point out that the job of any jury is to determine whether or not a case has been proven beyond a reasonable doubt, not all possible doubt, for that is never possible. And my guess is…the jury is still going to convict.
May 11th, 2008 at 9:43 pm
Perhaps so, but it’s a sad irony when a rule intended to protect criminal defendants from stupid juries assuming too much turns around and screws criminal defendants by “protecting” them against intelligent juries having any clue what is going on where a proper understanding of the data would actually have helped them. This is particular so in Puckett’s case, as evidence of his priors came in anyway under a rather strained theory of showing an M.O. (he and the killer had both falsely described rapes they committed as “making love,” while we all know every other rapist in the world would have had the decency to honestly describe his rape as a “rape”). If the jury had been told it was a database search but was not told it was a database of previously convicted sex offenders, a rational juror could have thought to himself “Gee, I realize that the odds of somebody randomly matching this guy’s DNA are 1 in 3, but if we did get one of those hits, what are the odds that that one unlucky person would ‘just happen’ to be another convicted rapist?’” If the same juror had been told this was a database of sex offenders, he’d have easily answered his own question: “damned close to 1.”
May 11th, 2008 at 9:55 pm
“This is particular so in Puckett’s case, as evidence of his priors came in anyway under a rather strained theory of showing an M.O. (he and the killer had both falsely described rapes they committed as “making love,” while we all know every other rapist in the world would have had the decency to honestly describe his rape as a “rape”).”
And here I thought there was more to the M.O.
But I guess I just thought that because I, ya know, read the article.
Your representation of the facts here is Balko-like in its unfairness to the prosecution.
May 12th, 2008 at 7:05 am
This is primarily a discussion about cold searches like Puckett’s, not the more general question of whether this particular prosecutor was as full of crap on other issues as he obviously was about DNA. The only reason I cited the lame bit about “we’re making love” supposedly showing an M.O. is because Joel and I were discussing the supposedly prejudicial impact of telling the jury that Puckett was in a database of prior sex offenders. Once evidence of past offenses comes in under any theory, the “don’t tell the jury about his priors” train has left the station anyway. That was my point, but congratulations for missing it and harping on an irrelevant detail instead.
That said, having just re-read the article in its entirety (which, as I’m sure you know, highlighted the “we’re making love” theory with a neat little heading of its own), I am curious to know what this other, supposedly far more conclusive M.O. evidence is supposed to be. We know Puckett’s M.O. was to pose as a cop; we also know that apart from there being no evidence of forced entry, there is zero evidence that Sylvester’s killer did. We know that while Sylvester’s killer murdered his victim relatively quickly after telling the housekeeper to go away because they were “making love,” while Puckett never threatened to kill his known victims (as one might expect of anyone vicious enough to be capable of such a crime, let alone one who had actually done so five years earlier), typically discarded his weapon after commandering the vehicle, apologized to one and offered another a job. Are we to believe that a one-time rapist-murder reformed on his own and became a kinder, gentler rapist five years later, with no intervention from the courts or the police? Yeah, really strong M.O. evidence, that.
The rest of the article contains no additional evidence of M.O. It does, however, offer some pretty significant evidence against Puckett’s guilt, to wit:
In fact, after re-reading the article in full, I’m gravitating toward the position that this prosecutor isn’t just a bumbling, incompetent boob but a Nifongesque liar:
Pretty damned high, actually, if you know (as he himself did but the jury conveniently did not) that the police found the guy in a huge database of known sex offenders. This is the forensic equivalent of Mike Nifong’s “no wrong answers” police line-up. If this article is anything close to the truth, the bastard should be disbarred.
May 12th, 2008 at 9:58 am
A couple thoughts…
First, to me one of the more “darning” pieces of evidence was that (the way I read the article) the defendant was in the area on the day of the incident. Now it might have been a different day but the article leaves me with the impression that it was the same day. If you put Puckett in contact with her that same day and get a match…I think that is immensely probative.
Second, the “but he didn’t kill before” evidence that you suggest X, doesn’t really seem as strong for the not-guilty point, because it seems clear that rape is pretty much second only to murder (not completely mind you, but how many kidnapping for rape/ransom do we really see) in its seriousness…the jump from rapist to murderer just does not seem that great especially as she may have been more capable of contacting police given that it occurred in her residence.
Third, guilty people volunteer all sorts of stuff, including their DNA, especially if the police could obtain a Search Warrant really easy like, so often yeah they’ll consent to a DNA search (you can’t hide that evidence in the 12 hours it takes to go and come back with a warrant.)
Lastly, I myself am immensely curious if the defendant testified or not as that likely would make a huge difference. While not testifying can not be used against him, it’s important to remember that not testifying does not help the defendant. (That just meaning that if the defendant doesn’t testify he can’t explain his story and the jury is left with just the evidence before it.)
I do agree that the “What are the odds?” argument is gratitous since in fact the odds of the individual being in the database and then being a serial rapist are highly related. Which as you point out, the odds of that being the case are practically 1 to 1 if the database is …database of convicted serial rapists.
May 12th, 2008 at 10:10 am
Drat, always one more thought…Also, one other piece of evidence (and again this all assumes that the DNA and other identifiers are independant) is that the defendant and the purpetrator both had O blood. While nothing to significant…if you run though the white males 20-60 with o blood you’ve restricted even further your population of suspect down to nearly 750.
I do think it has to be said that Mr. Puckett had many opportunities to be exonerated, each and every one he failed.
May 12th, 2008 at 12:12 pm
How so? All we have is the word of a dishonest prosecutor that Puckett “happened to be in San Francisco in 1972,” not even the same neighborhood. Even if we knew for a fact that he was in San Francisco sometime in 1972 (highly probable, given that he lived in the Bay Area), so what? Literally millions visit to that city every year. In the case of partial DNA matches like this one, every millionth customer is another random match, so there almost certainly were others whose profiles would have matched the crime scene if they had been included in the search. Was Baker among them? We’ll never know.
The murder happened in 1972, the rapes in 1977. Not the other way around. A jump from rape to rape-murder would not be all that unusual; I’m sure serial rapists turn to serial rapist-killers all the time. A “reform” in the other direction, by contrast, strikes me as more than a little odd, particularly given how long Puckett’s record remained clean after that.
It’s not just gratuitous, it’s wilfully deceitful. At the David Merin said that, he knew well what the odds were that the individual found would be some sort of sexual predator, and cynically manipulated the fact that the jury didn’t know. I consider that prosecutorial misconduct.
Well, his lawyer certainly “failed” to convince the innumerate judge to properly instruct the jury of the 1 in 3 probability of a false match rather than the grossly misleading 1 in 1.1 one, I’ll grant you that. Whether the remaining evidence would have been sufficient to get from 1 in 3 to anything beyond reasonable doubt is anybody’s guess. At least one juror, quoted in the article, says it would not have been.
Here’s hoping his lawyers don’t “fail” on appeal.
May 14th, 2008 at 12:42 am
“Puckett had kidnapped his victims by holding a knife or ice pick to their necks, leaving scratches similar to those found on Sylvester’s neck — what Merin called “his signature.”"
You sorta forgot to mention that.
I can see Radley Balko leaving out that part and the ruse/no forced entry part and saying “evidence of his priors came in anyway under a rather strained theory of showing an M.O. (he and the killer had both falsely described rapes they committed as “making love,” while we all know every other rapist in the world would have had the decency to honestly describe his rape as a “rape”).”
Because every rapist tells victims or passers by that they are “making love.” And manages to gain entry without breaking in. And leaves scratches on their victim’s neck. And deposits sperm with DNA found only once in every 1.1 million people in the population. And turns out to be the only hit in a database full of rapists and murderers and other criminals from the same state.
Yeah, it ALLLLL comes down to a rapist saying his rape was making love, while we all know every other rapist in the world would have had the decency to honestly describe his rape as a “rape.” That’s SUCH an honest and full description of his MO and of the evidence against him!!
May 14th, 2008 at 11:29 am
You’re right, I did forget to mention that John Puckett and Robert Baker (the only suspect in Sylvester’s murder at the time, and from where I sit, her most likely real killer) both used knives. Whopee. Hardly any criminals ever use those things, let alone ever let them get anywhere near their victims’ necks. Especially if theyr’e planning on stabbing their victim in the heart, as Baker did to Sylvester but Puckett never even threatened to do to his subsequent victims Baker’s. Not terribly convincing evidence to me, but I guess that according to San Francisco’s answer to Mike Nifong, the fact that both assailants used knives and caused some scratches to the neck is a “signature.” O-kay.
BTW, before you predictably connipt over the fact that I just compared David Merin to Mike Nifong, re-read my prior comment about his closing statement. If you can offer a single, plausible explanation of how Merin could possibly have delivered “what are the odds” line (regarding a match to a “serial” rapist) in good faith, let’s hear it. If I don’t hear one in the next week or so, I will seriously consider filing a formal complaint against him for deliberately misleading the jury on this point. I’m not kidding; it sounds to me like he committed serious prosecutorial misconduct irrespective of Puckett’s guilt (or lack thereof).
The two are not remotely comparable. Most rapists probably don’t tell their victims that they are “making love,” but judging by Puckett’s subsequent apology to one and job offer to another, it appeared that he did because in his sick fantasy world, that is exactly what was happening. There is no evidence that Baker (or, in your fantasy world, Puckett) said any such thing to Sylvester. The phrase only came up at all because the housekeeper had confronted him in the act and he had to say something to get her to go away. What do you think any other rapist/murderer would have told her under the circumstances? “Go away, I’m busy raping and killing someone?!”
There are only two defense to rape: “I didn’t do it” and “we made love.” When caught in the act, “I didn’t do it” is not available as a defense. Ergo, Baker blurted out the only semi-plausible defense he had, and the Nifong of Frisco calls it an M.O. Lame.
You’re right that some rapists leave evidence of forced entry while others do not, but since when is mere lack of evidence of forced entry considered significant evidence of an M.O.?! Posing as a cop three times to commandeer three vehicles before raping one’s victims? Now that’s what I call an M.O. Varying from that a bit by attacking the first victim in her apartment instead? Less likely to be an M.O., but still plausible if other evidence of the cop routine could be corroborated, e.g., a fake police badge had been found at the scene, or the housekeeper had testified that rather instead of yelling “Go away, we’re making love!” as should be expected of any rapist, Baker/Puckett had instead blurted out something more unusual like “Go away, I’m making an arrest!”
Of course, if any such evidence had existed, the case probably wouldn’t have stayed cold for three decades. Certainly not if Puckett was the perp, as his investigation for the 1977 rapes would have likely led the cops to re-open the 1972 case back then.
I realize that this statement was intended to be sarcastic, but in fact, aside from the detail that not every rapist ejaculates, it’s literally true. If any rapist deposits sperm at the scene of the crime, and the evidence is preserved in a condition to allow the 5.5 DNA indicators to be compared, that DNA will partially match to 1 in every 1.1 million people in the population, and will therefore run a 1 in 3 chance of randomly matching to someone in the database who has no connection to the crime. One more piece of prosecutorial “evidence” that is in fact nothing more than a tautology.
So what? You didn’t say when they started collecting samples for that database. DNA testing wasn’t available in 1972, the last time Baker was charged with a crime, and he died six years later (when AFAIK the database as we knew it still did not exist), so it should come as little surprise that Baker never made it into the database. Given that Baker is the only known individual with any real evidence connecting him to the crime, and the only one anyone thought of as a potential suspect back then, I’d say the odds are at least 50-50, if not higher still, that Diana Sylvester’s killer is not in the database, and that any resulting match must necessarily be false. I presume you disagree vehemently with that conclusion, but have yet to hear a credible reason why I should.
Full, no, honest, yes. It was meant to be an example of a generally flimsy theory, not a grand fisking of every single phony aspect of it, of which there were quite a few. It would be one thing if I had sifted through a mountain of credible theories and hand-picked the one solitary theory that wasn’t credible. I did not do that. Nearly all of the “evidence” you cited is every bit as flimsy in its own right, and one piece (the bit about leaving DNA matching 1 in 1.1 million figure of the population) is no evidence at all. This case stinks all around. As with Louis Farrakhan’s habit of making inflammatory statements and claiming to have been quoted “out of context,” it doesn’t look better when you put the original context back in. All it does is make for a longer, more confusing post.
May 14th, 2008 at 11:02 pm
“You’re right, I did forget to mention that John Puckett and Robert Baker (the only suspect in Sylvester’s murder at the time, and from where I sit, her most likely real killer) both used knives. Whopee. Hardly any criminals ever use those things, let alone ever let them get anywhere near their victims’ necks. Especially if theyr’e planning on stabbing their victim in the heart, as Baker did to Sylvester but Puckett never even threatened to do to his subsequent victims Baker’s.”
Your evidence that Robert Baker used a knife?
Your evidence that Baker stabbed Sylvester in the heart?
I’ll be waiting right here.
May 15th, 2008 at 7:05 am
Cops found a parking ticket with drops of blood on it in Baker’s van. [I presume there was no clear innocent explanation of it, i.e., Baker didn't show signs of a paper cut, or if he did, the blood on the ticket didn't match him. If this presumption is wrong, that omission is a much bigger problem with the article than any of the statistical stuff we've been yammering about.]
They couldn’t test for DNA back then, but they could test for blood type. Diana Sylvester had Type O blood. So did whoever’s blood was on the ticket. What are the odds that two randomly selected blood samples from unrelated individuals would have matched randomly? About 19%, assuming Wikipedia’s ABO distribution chart is right, and that the the test employed by police could not distinguish O+ from O-. Otherwise, it’s more like 14% (if they were both O+) or less than 0.5% if they were both O-. Hardly a smoking gun, but it’s longer odds than this phony DNA stuff, which stood a 1 in 3 chance.
OK, maybe that’s not quite the right way to look at it. If we treat Sylvester’s blood type as a given, which we probably should, and further assume that cops employed a crude test on the blood that could only identify the blood’s general type and not its RH factor, there is now a 44% (37.4% + 6.6%) chance that the sample found on Baker’s parking ticket would match randomly. If the test did identify RH factor, and it was the expected positive, the odds of a random match were 37.4% Conclusive proof? No. Evidence almost as strong as the hocus-pocus DNA search? Certainly. And stronger still when you consider that the DNA database is culled from across the state, and landed on someone who, unlike Baker, could not even be placed in the neighborhood and wasn’t known to have been terrorizing women in the area around that time (though he would three times later, after an unexplained five-year hiatus).
May 15th, 2008 at 9:54 am
In other words, there was reason to suspect Baker, so you thought you’d just speak as though the evidence was conclusive that he’d done it: “Especially if theyr’e planning on stabbing their victim in the heart, as Baker did to Sylvester but Puckett never even threatened to do to his subsequent victims Baker’s.”
So according to you, we know Puckett didn’t do kill Sylvester, because if we assume Baker did it instead, then he’s the only one who ever actually stabbed anyone in the heart. So we resolve the question of Puckett’s guilt by assuming Baker’s guilt, and then pointing out that Baker certainly seems more guilty if we assume him to be the guilty party.
Mmmm, that’s good logic!
You are not covering yourself in glory here, Xrlq.
May 15th, 2008 at 12:00 pm
I never said we knew for a fact that Baker did it. Quite the contrary, I described him above as “the only suspect in Sylvester’s murder at the time, and from where I sit, her most likely real killer.” If I thought his guilt was a certainty, I likely would have used stronger language than “most likely” to describe that eventuality, dontcha think?
That said, I do think it’s a damned near certainty that if either Baker or Puckett killed Sylvester, the other did not. Helen Nigodoff didn’t encounter two bearded men and then hear one (or both) of them yell “Go away! We’re having a menage à trois!” Therefore, unless you or anyone else conclusively show that Baker did not kill Sylvester, I don’t how on earth anyuone is supposed to conclusively show that Puckett did.
Further, even if one were to rule out the possibility that Baker is the killer (and I stress this is purely hypothetical, as you have yet to provide a lick of evidence that anyone should), that still doesn’t mean we can be all that confident that the real killer was in the database. That Baker himself wasn’t, despite having been convicted of a brutal forcible rape within blocks of Sylvester’s apartment and within weeks of her murder, begs the question of how many other sex predators were lurking about in 1972, who also didn’t make it into the database in question. Suppose that the real killer was neither Puckett nor Baker, but someone else who was convicted of raping and murdering another woman in 1973, sentenced to life without parole, and has been rotting in prison ever since. Would he have made it into the database? I suspect the answer is “maybe,” but well shy of an unequivocal “yes.”
May 15th, 2008 at 8:23 pm
“I never said we knew for a fact that Baker did it. . . . I likely would have used stronger language than “most likely” to describe that eventuality, dontcha think?”
Yeah . . . like saying positively that Baker stabbed Sylvester in the heart.
Ya mighta missed it because I only bolded it and didn’t also blockquote it, italicize it, and put it in all caps, but this, my friend, is your own words, which I am going to keep repeating back to you because you’re pretending like you didn’t say it:
“Especially if theyr’e planning on stabbing their victim in the heart, as Baker did to Sylvester . . .”
Seems like stronger language than “most likely” . . . dontcha think?
May 16th, 2008 at 7:28 am
It would, if those highlighted sentences had been written in a vacuum, rather than in the context of me having stated that Baker was the most likely real killer, and then having discussed that possibility in more detail.
Seeing as your last two comments consist of nothing more than cheap shots at me, I presume you’re now conceding my substantive points, particularly the one about the odds of the real killer ending up in the database being no better than 50-50? Cuz if you had any real counter-arguments to that, you’d likely have offered them up, dontcha think?
May 16th, 2008 at 9:08 am
They’re not cheap shots. They’re pointing out that your logic is based on a) assuming Baker did the murder and b) using that assumption to show Baker did the murder.
I don’t see how anyone could put odds on whether the true killer is in the database.
May 16th, 2008 at 9:53 am
In that case, no reasonable inferences at all can be drawn from the DNA match concerning the likelihood of Puckett’s innocence or guilt. All that evidence could possibly do is exactly what it did: confuse the hell out of the jury and make it look as though the match indicated overwhelming evidence of guilt, when in fact, it didn’t tell us a f’ing thing about the likelihood that Puckett was or was not the killer.
IOW, both your theory and mine invariably lead to the conclusion that Puckett did not receive a fair trial. Right?
April 18th, 2012 at 11:39 pm
[...] Probabilities Posted on May 10, 2008 by Brent Logan Interesting post on understanding DNA probabilities. No math degree required. This entry was posted in Asides and tagged dna by Brent Logan. [...]