# Was Amanda Knox Involved In Any Way With Her Roommate's Murder? **URL:** \ **Category:** Great Debates\ **Created:** [September 28, 2011, 5:39pm UTC](https://boards.straightdope.com/t/was-amanda-knox-involved-in-any-way-with-her-roommates-murder/597906 "2011-09-28T17:39:50Z")\ **Posts on this page:** 20\ **Page:** 13
**Author:** ![steronz](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/steronz/32/11765_2.png) [@steronz](https://boards.straightdope.com/u/steronz)\ **Post date:** [October 7, 2011, 12:52pm UTC](https://boards.straightdope.com/t/was-amanda-knox-involved-in-any-way-with-her-roommates-murder/597906/241 "2011-10-07T12:52:07Z")
> [@brazil84](#): > > I disagree, and I can give a hypothetical to show you are wrong. **DragonAsh** , by way of **F-P** , was talking about evidence with only 50% reliability. Your lengthy example completely misses this point, and is therefore irrelevant. > [@brazil84](#): > > Suppose that a man is stabbed to death in his apartment in Brooklyn Heights. The police learn that his business associate Mac took out a million dollar life insurance policy on him a couple weeks earlier. Standing alone, is this solid evidence that Mac is the killer? Of course not; people take out life insurance policies on business associates all the time. A certain percentage of those insureds will die for whatever reason shortly after the policy is bought. Unless the insurance company is only 50% sure that the policy was taken out by Mac on the dead man, I don’t see how this fits the scenario. A better example would be if the dead man’s wife took out a million dollar policy, and Mac’s secretary has suspected Mac of having an affair with her. She has no evidence besides her hunch, and when police press her, she admits that she’s somehow only 50% hunchy about the affair. Police now have an unreliable, tenuous connection between Mac and the insurance policy. > [@brazil84](#): > > The police next learn that Mac claims to have been home reading all night at his apartment in Fort Lee, New Jersey. Standing alone, is this solid evidence that Mac is the killer? Of course not, many people stay up reading and do not have solid alibis. Or, let’s say Mac fell asleep at some point and is only 50/50 on whether he was reading or sleeping. > [@brazil84](#): > > The police then check out Mac’s EZ-pass records and see that his car was driven across the bridge to New York at 1:00 am on the night in question; and that at at no time in the past did he make a similar trip. Standing alone, is this solid evidence that Mac is the killer? Of course not, there are any number of innocent reasons why Mac may have gone into New York that night and then failed to tell the police about it. Again, only relevant if the EZ-Pass company has vague records. A better example would be if an NJDOT worker comes forward saying he has video proof of Mac’s car crossing the bridge. They hand the video evidence to a visual CSI expert who can only say that there’s a 50% chance the car is Mac’s. He’s just not very sure, because the video is blurry and Mac’s car isn’t very distinct. > [@brazil84](#): > > The police then search Mac’s house and discover a knife which is later tested and found by police forensics to match the stab wounds on his business associate. It’s an unusual type of knife but not super-rare. Probably there are 10,000 similar knives in the United States. Standing alone, is this solid evidence that Mac is the killer? Of course not, since there are thousands of other similar knives floating around out there. Better would be if the forensic scientist is only 50% certain that the knife matches the murder weapon. The wounds weren’t clean, and he thinks it _might_ be the knife, but he’s only 50% sure. There’s no blood on it, so it might just be a random knife. > [@brazil84](#): > > The police interview the victim’s neighbor who states that she saw the victim give entry to and greet another man at around 2:00 am on the night in question. Shortly thereafter she heard shouting and then she saw the visitor flee the apartment building. The visitor is described as a tall slender white male, with a beige trenchcoat and a limp. Mac is a tall slender white male; has a limp; and is known to wear a beige trenchcoat. Standing alone, is this solid evidence that Mac is the killer? Of course not, there are thousands of tall slender white males with beige trenchocats and limps throughout New York City. Two witnesses see the men arguing, one picks Mac out of a lineup, the other chooses one of the stand-ins. The police combine this eyewitness testimony and decide that they’re only 50% sure that the guy arguing with the dead man was Mac. They’re not even sure if the dead man was there, or if it was just two guys arguing. > [@brazil84](#): > > Standing alone, each piece of evidence is not solid evidence that Mac is the killer. But when you put it all together, one can be reasonably confident that Mac is the killer. This would certainly be enough evidence to convict Mac and people are convicted on this kind of evidence all the time. Standing alone, each example I gave to match F-P’s 50% hypothesis is garbage. Nothing more than curiosities. Just because the police have 5 or 6 pieces of garbage doesn’t mean they have a case. It just means they’re excellent at collecting pieces of garbage. However, **DragonAsh** ’s point was that it’s very easy to fall into the trap of thinking that all those pieces of garbage add up to something. They do not. ---
**Author:** ![Acsenray](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/acsenray/32/4519_2.png) [@Acsenray](https://boards.straightdope.com/u/Acsenray)\ **Post date:** [October 7, 2011, 1:35pm UTC](https://boards.straightdope.com/t/was-amanda-knox-involved-in-any-way-with-her-roommates-murder/597906/242 "2011-10-07T13:35:33Z")
Steronz is right. You’re confusing the issue of credibility/reliability with the distinction between direct and indirect evidence. ---
**Author:** ![CaptMurdock](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/captmurdock/32/11736_2.png) [@CaptMurdock](https://boards.straightdope.com/u/CaptMurdock)\ **Post date:** [October 7, 2011, 1:47pm UTC](https://boards.straightdope.com/t/was-amanda-knox-involved-in-any-way-with-her-roommates-murder/597906/243 "2011-10-07T13:47:51Z")
> [@brazil84](#): > > I disagree, and I can give a hypothetical to show you are wrong. > > Suppose that a man is stabbed to death in his apartment in Brooklyn Heights. The police learn that his business associate Mac took out a million dollar life insurance policy on him a couple weeks earlier. Standing alone, is this solid evidence that Mac is the killer? Of course not; people take out life insurance policies on business associates all the time. A certain percentage of those insureds will die for whatever reason shortly after the policy is bought. > > The police next learn that Mac claims to have been home reading all night at his apartment in Fort Lee, New Jersey. Standing alone, is this solid evidence that Mac is the killer? Of course not, many people stay up reading and do not have solid alibis. > > The police then check out Mac’s EZ-pass records and see that his car was driven across the bridge to New York at 1:00 am on the night in question; and that at at no time in the past did he make a similar trip. Standing alone, is this solid evidence that Mac is the killer? Of course not, there are any number of innocent reasons why Mac may have gone into New York that night and then failed to tell the police about it. > > The police then search Mac’s house and discover a knife which is later tested and found by police forensics to match the stab wounds on his business associate. It’s an unusual type of knife but not super-rare. Probably there are 10,000 similar knives in the United States. Standing alone, is this solid evidence that Mac is the killer? Of course not, since there are thousands of other similar knives floating around out there. > > The police interview the victim’s neighbor who states that she saw the victim give entry to and greet another man at around 2:00 am on the night in question. Shortly thereafter she heard shouting and then she saw the visitor flee the apartment building. The visitor is described as a tall slender white male, with a beige trenchcoat and a limp. Mac is a tall slender white male; has a limp; and is known to wear a beige trenchcoat. Standing alone, is this solid evidence that Mac is the killer? Of course not, there are thousands of tall slender white males with beige trenchocats and limps throughout New York City. > > Standing alone, each piece of evidence is not solid evidence that Mac is the killer. But when you put it all together, one can be reasonably confident that Mac is the killer. This would certainly be enough evidence to convict Mac and **people are convicted on this kind of evidence all the time.** And therein lies the problem: _they should not be._ As witness [the Laurie Bembenek case,](http://en.wikipedia.org/wiki/Laurie_Bembenek) arguably the most egregious example of circumstantial evidence used to prosecute someone in recent memory. As to your hypothetical example, the question a defense attorney might raise (and a jury might ponder): [ol] [li]Was Max’s knife the murder weapon? If the police found such a knife in his possession, they would have had it analyzed for traces of blood. The absense of any such blood would definitely blow the case wide open.[/li][li]Do we know that it was Max who drove his car through the toll bridge at 1 a.m.? If the defense produces a third party who testifies that they borrowed Max’s car that night, Good-bye Ruby Tuesday.[/li][li]Eyewitness testimony has been shown, especially in recent years, to be unreliable. This witness does not even positively identify Max at the scene. How can she be sure the trenchcoat is beige? How can she be sure he walked with a limp?[/li][/ol] Max would be home in time for _Dancing With The Stars._ ---
**Author:** ![brazil84](https://avatars.discourse-cdn.com/v4/letter/b/8491ac/32.png) [@brazil84](https://boards.straightdope.com/u/brazil84)\ **Post date:** [October 7, 2011, 1:58pm UTC](https://boards.straightdope.com/t/was-amanda-knox-involved-in-any-way-with-her-roommates-murder/597906/244 "2011-10-07T13:58:23Z")
> [@CaptMurdock](#): > > And therein lies the problem: _they should not be._ I disagree. If the jury is reasonably confident that the defendant is guilty, then he should be convicted. [quote] [li]Was Max’s knife the murder weapon? If the police found such a knife in his possession, they would have had it analyzed for traces of blood. The absense of any such blood would definitely blow the case wide open.[/li][/quote] I agree, but so what? > [@](#): > > Do we know that it was Max who drove his car through the toll bridge at 1 a.m.? If the defense produces a third party who testifies that they borrowed Max’s car that night, Good-bye Ruby Tuesday. Again I agree, but so what? > [@](#): > > Eyewitness testimony has been shown, especially in recent years, to be unreliable. This witness does not even positively identify Max at the scene. How can she be sure the trenchcoat is beige? How can she be sure he walked with a limp? The same way anyone is sure of anything else they have witnessed in their life. None of this changes the basic fact that situations can and do arise where one can be reasonably confident about a certain conclusion based on numerous pieces of evidence each of which considered in isolation is inconclusive. ETA: Would you mind pointing to a few recent, prominent murder convictions where (1) the defendant insisted on his innocence; and (2) you are still reasonably confident the defendant was guilty? ---
**Author:** ![Fotheringay-Phipps](https://avatars.discourse-cdn.com/v4/letter/f/d9b06d/32.png) [@Fotheringay-Phipps](https://boards.straightdope.com/u/Fotheringay-Phipps)\ **Post date:** [October 7, 2011, 2:07pm UTC](https://boards.straightdope.com/t/was-amanda-knox-involved-in-any-way-with-her-roommates-murder/597906/245 "2011-10-07T14:07:30Z")
> [@Acsenray](#): > > Steronz is right. You’re confusing the issue of credibility/reliability with the distinction between direct and indirect evidence. They are the same thing, for this purpose. For example, if the likelihood that the victim’s DNA would end up on the knife from contamination is 50%, and the victim’s DNA was found on five different knives in different locations (i.e. we are hypothesizing 5 different contaminations), then that’s a good indicator that they are not all there from contamination - the likelihood that all were due to contamination would be 1/32. Looking at each one and dismissing it without consideration of the others is a logical error. ---
**Author:** ![Snowboarder\_Bo](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/snowboarder_bo/32/229_2.png) [@Snowboarder\_Bo](https://boards.straightdope.com/u/Snowboarder_Bo)\ **Post date:** [October 7, 2011, 2:10pm UTC](https://boards.straightdope.com/t/was-amanda-knox-involved-in-any-way-with-her-roommates-murder/597906/246 "2011-10-07T14:10:00Z")
> [@brazil84](#): > > To echo my earlier comments, it’s based on my general knowledge, common sense, and simple observations. So I don’t have a cite. > > By the way, if your position is that it’s pretty common that someone cannot provide a coherent account for their actions and whereabouts, would you mind giving a couple examples of how this might happen? In other words, you made something up and claimed it was fact. :roll eyes: I never stated that that was my position. Nice try at putting words in my mouth tho. ---
**Author:** ![brazil84](https://avatars.discourse-cdn.com/v4/letter/b/8491ac/32.png) [@brazil84](https://boards.straightdope.com/u/brazil84)\ **Post date:** [October 7, 2011, 2:11pm UTC](https://boards.straightdope.com/t/was-amanda-knox-involved-in-any-way-with-her-roommates-murder/597906/247 "2011-10-07T14:11:08Z")
> [@steronz](#): > > **DragonAsh** , by way of **F-P** , was talking about evidence with only 50% reliability. Your lengthy example completely misses this point, and is therefore irrelevant. Can you explain exactly what you mean by “50% reliability” so I know we are both on the same page? And, are you denying that situations can and do arise where one can be reasonably confident about a certain conclusion based on numerous pieces of evidence each of which considered in isolation is inconclusive? ---
**Author:** ![brazil84](https://avatars.discourse-cdn.com/v4/letter/b/8491ac/32.png) [@brazil84](https://boards.straightdope.com/u/brazil84)\ **Post date:** [October 7, 2011, 2:14pm UTC](https://boards.straightdope.com/t/was-amanda-knox-involved-in-any-way-with-her-roommates-murder/597906/248 "2011-10-07T14:14:49Z")
> [@Snowboarder\_Bo](#): > > In other words, you made something up and claimed it was fact. Lol, no. I made a statement based on my general knowledge, observations, experience, and common sense. Just like I do not have a cite for the claim that Mongolian men are in general taller than Mongolian women but I nonetheless believe it to be so even without looking for studies on the issue. > [@](#): > > I never stated that that was my position. You denied that it’s rare for a person to be unable to give a coherent account of his actions and whereabouts. Which implies you think it’s common. So I am asking you to give me some examples. ---
**Author:** ![Snowboarder\_Bo](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/snowboarder_bo/32/229_2.png) [@Snowboarder\_Bo](https://boards.straightdope.com/u/Snowboarder_Bo)\ **Post date:** [October 7, 2011, 2:17pm UTC](https://boards.straightdope.com/t/was-amanda-knox-involved-in-any-way-with-her-roommates-murder/597906/249 "2011-10-07T14:17:01Z")
> [@brazil84](#): > > Lol, no. I made a statement based on my general knowledge, observations, experience, and common sense. Just like I do not have a cite for the claim that Mongolian men are in general taller than Mongolian women but I nonetheless believe it to be so even without looking for studies on the issue. > > You denied that it’s rare for a person to be unable to give a coherent account of his actions and whereabouts. Which implies you think it’s common. So I am asking you to give me some examples. No, I never made a statement denying that. ---
**Author:** ![brazil84](https://avatars.discourse-cdn.com/v4/letter/b/8491ac/32.png) [@brazil84](https://boards.straightdope.com/u/brazil84)\ **Post date:** [October 7, 2011, 2:33pm UTC](https://boards.straightdope.com/t/was-amanda-knox-involved-in-any-way-with-her-roommates-murder/597906/250 "2011-10-07T14:33:51Z")
> [@Snowboarder\_Bo](#): > > No, I never made a statement denying that. Here’s the exchange: Me: It’s very rare that an innocent person would not be able to give a coherent account of their whereabouts and actions. You: Cite? Surely you wouldn’t make such a bald assertion of fact without a cite, would you? Me: It’s based on my general knowledge and common sense. Do you seriously dispute my claim? You: Yes. Back it up, please, with a cite (See Posts 203, 224 and 232) ---
**Author:** ![DragonAsh](https://avatars.discourse-cdn.com/v4/letter/d/db5fbb/32.png) [@DragonAsh](https://boards.straightdope.com/u/DragonAsh)\ **Post date:** [October 7, 2011, 2:36pm UTC](https://boards.straightdope.com/t/was-amanda-knox-involved-in-any-way-with-her-roommates-murder/597906/251 "2011-10-07T14:36:13Z")
> [@Fotheringay-Phipps](#): > > They are the same thing, for this purpose. > > For example, if the likelihood that the victim’s DNA would end up on the knife from contamination is 50%, and the victim’s DNA was found on five different knives in different locations (i.e. we are hypothesizing 5 different contaminations), then that’s a good indicator that they are not all there from contamination - the likelihood that all were due to contamination would be 1/32. > > Looking at each one and dismissing it without consideration of the others is a logical error. You’re doing it again. ‘Once would be a coincidence, but not all five! This can’t be a coincidence!’ You can’t group all five together and say the whole is more credible than any individual piece - not when there’s a 50% chance that any individual piece of evidence is worthless. I hope you’re never allowed anywhere near a jury trial. ---
**Author:** ![Fotheringay-Phipps](https://avatars.discourse-cdn.com/v4/letter/f/d9b06d/32.png) [@Fotheringay-Phipps](https://boards.straightdope.com/u/Fotheringay-Phipps)\ **Post date:** [October 7, 2011, 2:50pm UTC](https://boards.straightdope.com/t/was-amanda-knox-involved-in-any-way-with-her-roommates-murder/597906/252 "2011-10-07T14:50:54Z")
> [@DragonAsh](#): > > You can’t group all five together and say the whole is more credible than any individual piece - not when there’s a 50% chance that any individual piece of evidence is worthless. You absolutely can and should. [FWIW, I believe Occam’s Razor operates on a similar if not identical principle, but I’m not going to get bogged down in arguing about OR. At any rate what I’ve written is basic logic.] ---
**Author:** ![Marley23](https://avatars.discourse-cdn.com/v4/letter/m/45deac/32.png) [@Marley23](https://boards.straightdope.com/u/Marley23)\ **Post date:** [October 7, 2011, 2:52pm UTC](https://boards.straightdope.com/t/was-amanda-knox-involved-in-any-way-with-her-roommates-murder/597906/253 "2011-10-07T14:52:30Z")
> [@Gary\_Kumquat](#): > > That speaks volumes about your humanity. I allowed a couple of comments like this earlier in the thread, but I’m going to say now that it’s off limits. Commenting on a poster’s argument and its implications is fine, but when there’s too much commentary on what the argument says about the person, it’s a short step to flaming. No more of this, please. And the same goes for this: > [@DragonAsh](#): > > I hope you’re never allowed anywhere near a jury trial. And **Snowboarder Bo** , I admit I asked the same question **Gary Kumquat** did upthread, but I’m instructing you to just answer it. You’re not even being asked to defend an argument, you’re being asked to explain the basis for your opinion and I can’t think of a good reason you are refusing to post that. ---
**Author:** ![brazil84](https://avatars.discourse-cdn.com/v4/letter/b/8491ac/32.png) [@brazil84](https://boards.straightdope.com/u/brazil84)\ **Post date:** [October 7, 2011, 3:05pm UTC](https://boards.straightdope.com/t/was-amanda-knox-involved-in-any-way-with-her-roommates-murder/597906/254 "2011-10-07T15:05:57Z")
> [@Fotheringay-Phipps](#): > > You absolutely can and should. I agree. If there is a piece of evidence which, standing alone, means there is a 50/50 chance which the Defendant is guilty. And there are 4 other (independent) pieces of such evidence, then it’s very very likely that the Defendant is guilty. Consider DNA evidence. For the sake of simplicity, think about a paternity test although the same principle applies to any DNA test. When they give you a paternity test, they match 20 or 30 of the putative father’s alleles against those of the child. The chances of 1 allele matching just by chance are actually pretty good. But the chances of 20 or 30 matching just by chance are astronomically low. So that the chances the putative father is the actual father become quite high. It seems that using the logic of **DragonAsh** , one could consider each of the allele matches in isolation and dismiss each one in turn and never reach any conclusion. It’s worth keeping in mind that almost all evidence is ambiguous. Almost all the time, one can come up with plausible ways to explain away and dismiss any single piece of evidence. Which is why I asked another poster to identify a few prominent murder convictions where he was reasonably confident of the defendant’s guilt despite the defendant’s protestations of innocence. > [@](#): > > [FWIW, I believe Occam’s Razor operates on a similar if not identical principle, I agree 100%. ---
**Author:** ![Acsenray](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/acsenray/32/4519_2.png) [@Acsenray](https://boards.straightdope.com/u/Acsenray)\ **Post date:** [October 7, 2011, 3:13pm UTC](https://boards.straightdope.com/t/was-amanda-knox-involved-in-any-way-with-her-roommates-murder/597906/255 "2011-10-07T15:13:04Z")
> [@brazil84](#): > > I agree. If there is a piece of evidence which, standing alone, means there is a 50/50 chance which the Defendant is guilty. And there are 4 other (independent) pieces of such evidence, then it’s very very likely that the Defendant is guilty. Here’s the problem. We’re not talking about evidence that standing alone shows that there’s a 50-50 chance that the defendant is guilty. We’re talking about evidence about which we’re only 50 percent sure that _that piece of evidence is true at all_. > [@](#): > > It’s worth keeping in mind that almost all evidence is ambiguous. Almost all the time, one can come up with plausible ways to explain away and dismiss any single piece of evidence. Again, we’re not talking about evidence whose \*implications \*are ambiguous. We’re talking about evidence that in and of itself is not sufficiently credible to qualify as evidence at all. On the one hand you have DNA testing that produces a result that there is a 80 percent chance that Dude A is the father. On the other hand, you show that the DNA lab workers’ handling of the evidence was so bad, that we don’t even know whether they tested the DNA found at the scene of the crime or whether they smeared DNA found somewhere else all over the laboratory. There’s a huge difference. That 80 percent result no longer counts as 80 percent. It very well might count as nothing at all. ---
**Author:** ![steronz](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/steronz/32/11765_2.png) [@steronz](https://boards.straightdope.com/u/steronz)\ **Post date:** [October 7, 2011, 3:13pm UTC](https://boards.straightdope.com/t/was-amanda-knox-involved-in-any-way-with-her-roommates-murder/597906/256 "2011-10-07T15:13:26Z")
> [@brazil84](#): > > Can you explain exactly what you mean by “50% reliability” so I know we are both on the same page? “50% reliability” wasn’t my terminology, and I think putting a number on evidence that lacks reliability or credibility is a bit contrived. I tried to come up with some examples where things were “50% reliable,” and can extend that to things such as DNA evidence that’s a 50% match (not particularly good). The important thing is to acknowledge that there’s a difference between conclusive yet indirect (to steal the word from upthread) evidence, and junk evidence. The former would be EZ-Pass records that conclusively prove the suspect’s car was at x location at y time, while not conclusively proving that the suspect was in the car. In isolation it’s not a slam dunk, but it can help build a case. The latter would be a co-worker with an ax to grind saying he heard the suspect plotting the crime in the bathroom one day. I’m sure the prosecution would call the co-worker to testify and let the jury determine the credibility of the witness, but if that type of testimony is all that the prosecution had brought to the table, I’d hope that the jury would see through it and realize there wasn’t a case. > [@brazil84](#): > > And, are you denying that situations can and do arise where one can be reasonably confident about a certain conclusion based on numerous pieces of evidence each of which considered in isolation is inconclusive? Sure, it’s called the preponderance of evidence, right? What matters is that the individual pieces of evidence are credible. If everyone in the high school says that Sally is a tramp, it must be true, right? Because everyone says it. One person says it, it could just be a grudge, but if _everyone_ says it, well, tramp she she is, right? Or maybe we should sit down with every single high-schooler and figure out if they’re basing their accusations on anything more than rumor. If not, we should throw out the whole lot, no matter how many there are. ---
**Author:** ![Fotheringay-Phipps](https://avatars.discourse-cdn.com/v4/letter/f/d9b06d/32.png) [@Fotheringay-Phipps](https://boards.straightdope.com/u/Fotheringay-Phipps)\ **Post date:** [October 7, 2011, 3:19pm UTC](https://boards.straightdope.com/t/was-amanda-knox-involved-in-any-way-with-her-roommates-murder/597906/257 "2011-10-07T15:19:58Z")
> [@Acsenray](#): > > Here’s the problem. We’re not talking about evidence that standing alone shows that there’s a 50-50 chance that the defendant is guilty. We’re talking about evidence about which we’re only 50 percent sure that _that piece of evidence is true at all_. As noted earlier, this is a distinction without a difference (for our purposes). For some reason, you’ve declined to address this and explain why you think this distinction matters. ETA: the bottom line of both situations is that the likelihood that a series of unrelated plausible events took place is less than the likelihood that any one of them took place. ---
**Author:** ![Fotheringay-Phipps](https://avatars.discourse-cdn.com/v4/letter/f/d9b06d/32.png) [@Fotheringay-Phipps](https://boards.straightdope.com/u/Fotheringay-Phipps)\ **Post date:** [October 7, 2011, 3:24pm UTC](https://boards.straightdope.com/t/was-amanda-knox-involved-in-any-way-with-her-roommates-murder/597906/258 "2011-10-07T15:24:30Z")
> [@steronz](#): > > Sure, it’s called the preponderance of evidence, right? What matters is that the individual pieces of evidence are credible. If everyone in the high school says that Sally is a tramp, it must be true, right? Because everyone says it. One person says it, it could just be a grudge, but if _everyone_ says it, well, tramp she she is, right? The error here is that the various people saying it in this case are not unrelated. A lot of these people are basing their opinions on the other people who say it. So it’s all one piece of evidence of shaky value. You can’t compare that to a series of unrelated pieces of evidence. ---
**Author:** ![DragonAsh](https://avatars.discourse-cdn.com/v4/letter/d/db5fbb/32.png) [@DragonAsh](https://boards.straightdope.com/u/DragonAsh)\ **Post date:** [October 7, 2011, 3:27pm UTC](https://boards.straightdope.com/t/was-amanda-knox-involved-in-any-way-with-her-roommates-murder/597906/259 "2011-10-07T15:27:00Z")
A test for a disease is known to be accurate 99% of the time. We pull a person at random off the street and give him the test, which comes back positive for the disease. What is the probability that the person has the disease? Hint: it’s not 99%. ---
**Author:** ![Gary\_Kumquat](https://avatars.discourse-cdn.com/v4/letter/g/4491bb/32.png) [@Gary\_Kumquat](https://boards.straightdope.com/u/Gary_Kumquat)\ **Post date:** [October 7, 2011, 3:28pm UTC](https://boards.straightdope.com/t/was-amanda-knox-involved-in-any-way-with-her-roommates-murder/597906/260 "2011-10-07T15:28:11Z")
> [@Marley23](#): > > I allowed a couple of comments like this earlier in the thread, but I’m going to say now that it’s off limits. Commenting on a poster’s argument and its implications is fine, but when there’s too much commentary on what the argument says about the person, it’s a short step to flaming. No more of this, please. . I apologise. [Previous page](https://boards.straightdope.com/t/was-amanda-knox-involved-in-any-way-with-her-roommates-murder/597906.md?page=12) [Next page](https://boards.straightdope.com/t/was-amanda-knox-involved-in-any-way-with-her-roommates-murder/597906.md?page=14)