That’s true. But a rescheduled election and likely Strange win has the double bonus of pissing off the Trumpsters who’d see it as the establishment attacking Moore. Maybe the Dems come out ahead no matter what.
I was going to let this go now that you’ve moved on to other more creepy arguments, but I can’t stand by and watch statistics get mangled like this. You don’t get to take a study, throw in your own assumptions to make a few calculations, and pretend the result has any scientific basis.
You can’t take a study that says there’s a 50% false positive rate and say “even I don’t believe this number but the real rate must be high, so let’s just say it’s 30%.” You can’t cherry pick the top of a range, throw in a guess that 50% are inconclusive, and pretend that the study actually demonstrated 20%. That just isn’t how scientific studies and statistics work. If you want to argue the number is 20%, find a study with that actual conclusion, and then be prepared to argue why it is the only correct one and every other one is an outlier.
That wasn’t my claim, my claim is that Lisak’s study isn’t an estimate of the rate of false rape accusations, it’s an estimate of the rate of false accusations where the truth value is known. In order to convert Lisak’s “2 to 10% estimate” to an estimate of how many rape accusations are actually false you have to know what fraction is thrown out as inconclusive.
If we sort accusations into three categories (true = A, false = B, inconclusive = C) then “2 to 10%” is an estimate of A/(A+B), not A/(A + B+ C).
I don’t claim to have any scientific or statistically valid estimate of the rate of false rape accusations, because I’ve never come across a study that attempts to answer that question. But while I think the Kanin study is really good, I intuitively and unscientifically have a feeling it’s too high. My unscientific estimate, in the absence of anything better, is that the rate of false accusations is probably somewhere between 10% and 40%.
I’ll see if I can find a more recent study though.
TroutMan, let me put that a little differently: one can’t conclude that if 2% of rape allegations are found to be false, the remaining 98% are true. Some are simply inconclusive, or are dropped at an early stage in the proceedings.
This is a decent study from, I think, last year that re-analyzes some of the previous studies used by Lisak, etc. and states (where possible) how many of the allegations ended up being classified as “equivocal”. And they state that the figures used in the Lisak paper are conservative, i.e. they’re a lower bound on the actual false allegation rate. If you remove the “equivocal” cases from the analysis, their estimate of the false allegation rate goes up from 5% to 8%. Which is, of course, a lot lower than Kanin’s estimate, so that makes me a bit more skeptical of Kanin’s findings than I was before I looked at the article. Let me see if I can find some more studies where they do this kind of breakdown, and I might end up shifting my prior a little bit.
Who gives a fuck? There is no accusation of rape here and the accusation against Roy Moore is very well supported and unlikely to be false.
In other news we have a poll from Opinion Savvy conducted after WaPo story dropped.
“If the election to determine the next US Senator for Alabama were held today and your choices
were: Democratic nominee and former United States attorney Doug Jones; and Republican
nominee and former Alabama Supreme Court justice Roy Moore, for whom would you vote?”
Jones 46.0%
Moore 46.4%
Other 3.8%
Undecided 3.7%
“If the election to determine the next US Senator for Alabama were held today and your choices
were: Democratic nominee and former United States attorney Doug Jones; Republican nominee
and former Alabama Supreme Court justice Roy Moore; and current US Senator for Alabama
Luther Strange as a write-in candidate, for whom would you vote?”
Jones 43.6%
Moore 41.3%
Strange 12.3%
Other 1.4%
Undecided 1.5%
It’s just one data point, but it does look like this story has moved the needle. At the very least we’re probably looking at the closest senate election in Alabama in a long time.
Correction, I looked at the Lisak 2010 table again. The adjusted average false rate when you remove equivocal category goes from 5% up to 9.5%.
nm
As outrageous as GOP behavior has become, I’m inclined to agree with the above. Let them keep digging the hole. Alabama is a lost cause and I don’t mean just because they lost the Civil War. Doug Jones winning Alabama would be like Scott Brown winning Massachusetts. It would be a stunner but the elation would be short-lived. But if the GOP is perceived as rigging an election then that’ll just fuel the outrage among progressives leading up to the mid-terms.
Never forget about the error bars, which are about 4% in this case.
I don’t know. It might send the message that there was a level of crazy that couldn’t win even in Alabama: if it’s possible to turn off enough moderates to have an effect in Alabama, it’s something that could happen anywhere, and it’s a sign the moderates need to be considered: it’s not a good thing to have the crazy win the primary.
The first poll is out showing Jones with a lead, 46-42, which is within the 4pt MOE.
I suspect there’s a lot of “Oh, no, I’d never vote for that sick creep [quietly votes for him]” going on though.
Totally agree with this. It’s a cultural characteristic in the South, the two-facedness. The whole, “Bless your heart!” thing. Not saying it only happens in the South, but it really does seem to be part of the southern identity.
But I can’t buy the idea that in Alabama people feel like admitting to voting for Moore is more embarrassing than voting for a pro-choice democrat. I’d expect the deception to go the opposite way.
That’s sure a reasonable alternative way to look at it. And mind-bending in the worst sense if true.
But nothing would surprise me at this point. I hope you’re right!
Nothing is the only thing that would surprise me. Everything else is likely.
I largely agree with your point but I also think this issue just may depress the vote on the R side somewhat. It might be easier on the conscience of some evangelicals to just “forget” to go to the polls on election day than actively do something to put a GOP perv or a pro-choice Dem in office.
Here is an article discussing how the Democratic opponent is handling the situation. It’s an example of the point I was trying to make: that Democrats will do better in this situation not attacking Moore, but rather letting Republicans deal with the seamy side of things.
I dunno - I would expect them to find **admitting **to voting to a child molester more embarrassing or uncomfortable than admitting to voting for a Democrat, but **actually **voting for a Democrat more embarrassing or uncomfortable than actually voting for a Republican child molester. Because that way they get to support their tribe while not having to be accountable for the more unsavoury implications of that support.
This very well may be possible. I remember in my own case on a few occasions in my life I had a choice of two candidates, one of which whose policies I disagreed with utterly, the other I thought was a total sleeze. I ended up not voting.
I will never, ever forget about the errors bars. Good looking out.