This thread goes on for pages and pages, but one of the most ferocious arguments going on seems to center on whether the authorities can, or should, use the statistical incidence of a particular sort of crime to evaluate the credibility of a purported victim’s claims.
One group seems to be saying that each crime is individual, and the statistical likelihood or unlikelihood of a particular sort of crime occurring is irrelevant to the individual case and situation at hand, and that the statistical incidence of these crimes has no place in evaluating the merit of alleged victim’s claims. The other group seems to be saying that this position is absurd, and that claims which are unsupported by the historical incidence of the type of crime being alleged should be subject to a higher level of scrutiny.
The debate in in The Pit has gotten intensely personal to the point I really can’t clearly follow the core arguments between the bursts of invective, so I’m putting it here in hope that the main tenets of each position can be more clearly elaborated.
The core question is (I think) in investigating a specific criminal complaint, how much weight (if any) should be given to the fact that the nature an alleged crime has a very low rate of occurrence as measured by crime statistics in the larger national population?
As Richard Feynman once reportedly said: “Wow! On the way to work this morning, I saw a license plate numbered 2BBX493. What are the odds of that?”
If we’re just speaking in terms of statistics, it is extremely unlikely that Dr. Feynman actually saw that license plate number. But it does not mean that it did not happen.
I’m specifically not interested in race or sex etc. which is one of the reasons I abstracted the the question in the OP to apply to all crimes that might be statistically unlikely. The generic question is (I think) if a crime has a “man bites dog” level of statistical occurrence, is it fair or reasonable to subject the claims of the alleged victim to a higher than normal level of scrutiny by the very nature of the crime being rare or unlikely
I don’t think we need stats to determine the implausibility of a scenario. No one is going to believe me if I say I’ve been raped by an alien’s ghost, nor should they. The outrageousness rests not in the poor likelihood of an alien’s ghost being capable of rape, but in the undocumented existance of both aliens and ghosts. The outrageousness is very obvious.
If a case is plausible, it deserves to be evaluated as objectively as possible. Stats do not determine plausibility. Common sense does.
Statistically rare events do happen. That’s why they’re rare, not impossible. If they are alleged crimes they should be investigated in the same way statistically common crimes are.
The odds of being killed by a serial killer are pretty low. That doesn’t mean detectives shouldn’t look at a killing being the work of a serial killer, if that’s where the evidence points.
I tried to mount a hypothetical in the other thread, but it wasn’t a great one. I’ll try to shore it up:
Two cars side by side at a stoplight. Just before it turns green one of them peels off in a blaze of burning rubber, while the other waits for the green and starts moving at normal legal speeds. Cop at the intersection sees this and pulls over the speeding car. It’s a teenage boy, and his story is that he was in a road rage from the other car from the intersection, which was driven by a stranger to him. That car had been tailgating, then sped up and changed lanes cutting him off, and otherwise had been antagonizing him for miles. So he jumped the light to get away.
Now during the investigation, the cop checks the traffic camera and gets the license plate of the other car. He runs the plate and gets the ID of the alleged road rager.
Now, at this moment in time, the cop is getting ready to contact and question the other driver. Before he does, though, he probably has some impression of whether or not the story seems plausible. My contention is that this plausibility impression will be colored differently if the other driver turned out to be a teenaged boy than if it turned out to be a 40 year old woman.
I agree fully that a the investigation should proceed regularly, regardless of the initial plausibility. I agree that every piece of evidence uncovered will be much more valuable than the initial plausibility based on demographic.
My argument is that this initial plausibility factor is reasonable. Not that it should be maintained in the face of contradicting evidence, or that the investigation should be steered to support the initial statistical bias. Merely that being cognizant of the statistical bias and temporarily coloring otherwise uninformed opinion based on that bias is reasonable.
This is a form of sampling bias. Eliminating avenues of exploration based only on the statistics will bias the statistics themselves. Suppose it’s very unlikely that 80-year-olds are murderes. If you decide not to investigate any 80-year-olds in your murder case just because it’s unlikely, you make it seem even more unlikely in the future: you’ve eliminated a case that could cause the ratio of 80-year-old murderers to increase. Eventually the probability will drop to near zero, no one will be investigating 80-year-olds in murder cases, and those old folks will be able to get away with murder. Literally.
The original debate had nothing to do with what action should be taken. It was about whether we, the uninvolved puclic at large, should be allowed to use the stats in our preliminary opinions regarding plausibility.
having participated in that thread, I disagree w/your contention that it had ‘nothing to do about what action should be taken’, since one of the major proponents of that thought kept insisting on suggeting it as an investigatory tool (‘in the absence of other data’).
I also disagree w/you on the usage of stats (as played out) on message board debates.
If you’re talking ‘general plausibility’, there’s more reliable factors than stats to use - for example, physics, logic. It’s implausible that a pink unicorn hit my car. that’s got nothing to do w/statistics. However, if I claim that a deer hit my car, you should look at the damage to my vehicle to best ascertain if the story is plausible or not, not statistics.
I think the use of stats is really going to depend on its purpose and relevance to the crime.
If you already sold the damaged car and told me a few years ago you hit a deer with your old Impala in Manhattan, I might whip out some stats on deer collisions in NYC and call bullshit. The onus would then be on you to provide physical evidence, accident reports, insurance claims, etc. Ultimately, your claim would be decided on the evidence if you’ve got it, but what is my most reasonable position if you have none? I might previously have known nothing about deer herds in Manhattan, but if it’s demonstrably wildly unlikely you hit a deer on 42nd Street, given the frequency of prior reported incidents, I might be justified in dismissing your claim as too implausible to take seriously.
The onus to prove something doesn’t change based on whether the claim is unusual or usual. You may be more likely to believe someone if they say they hit a deer in Alabama, but in court, the burden of proof should be the same.
The reported incidence of deer hits shouldn’t factor into your assessment of likelihood. What if no one has collected that type of data? Would you still find the claim to be unusual? Of course you would. Most people are familiar enough with the geographic distribution of deer populations to know they don’t favor urban environments, thus making them rarity on the roads. But this is common sense at work, not statistical analysis.
the instance in the linked thread is suggesting that utilizing demographic data wrt frequency of certain aspects of generalized crime data has some useful significance wrt plausability of a current event. Oddly enough, if you read the thread, the hook is being hung on data wrt “white male raping black female” = very rare and therefore the instant offense is less plausible.
I say oddly 'cause the proponents own data had yet other demographic data saying “gang rapes more frequently perpretrated by white males vs. black males” which would then make the instant offense more plausible.
however, the proponent didn’t seem to be as interested in that.
another point that we’ve attempted to share is that, as demonstrated above, there’s significant numbers of data points invovled in the instant offense, so that it would then become an issue of which stats one should look at to determine plausability (the answer ‘none’ didn’t seem to be acceptable).
“Common sense” is pretty much bollocks, a crock I’m as guilty of citing as anyone. If I’ve not got hard numbers, I’ve got nothing at all. If you’ve got no evidence, we’re totally stuck.
So what should you do if you have no stats on the incidence of deer hits in Manhatten and the person swears that they hit one? Do you believe them just the same as if the person told you they hit a deer driving in the woods? If not, why not> In the absence of numbers, what tells you something sounds unusual if not, to some extent, this “common sense” that you call bollocks.
What’s the real problem, if the guy from Manhattan has no evidence, with comparing statistics about statewide deer populations and police reports from Mahattan and Saranac Lake and being dismissive of the Manhattan claim? The whole point is to not rely on common sense, which I’d never trust in court.
But common sense is lacks reliability any circumstance. If one is to lay as a ground rule that only evidence is permissable, that’s fine, but if a victim appears to lack compelling physical evidence, yet insists there is a claim, I might try to evaluate the likelyhood of that claim based on something quantitative before dismissing it entirely and investigating no further. I’d hate to think common-sense arguments would carry equal weight in such circumstances.
Well, if you were take him to court, regardless of if he hit a deer in Manhattan, NY or Tuskegee, AL, he’d need evidence to prove his claim. So that point is moot.
Prior to taking him to court, though, wouldn’t you need something other than stats to tell you whether or not his claim sounded reasonable? Unless you carry around a book of random, trivial facts where ever you go, I’m thinking you rely a lot on your own past experiences, observations, and knowledge of the way the things are in the world. This is what I call “common sense”. You may have another name for it.