= 27d567bc2d48d9f40e9ccc20ea370b15)
Lumifer comments on My new rationality/futurism podcast - Less Wrong Discussion
You are viewing a comment permalink. View the original post to see all comments and the full post content.
= 783df68a0f980790206b9ea87794c5b6)
Loading…= b715d02e85f7b3b4ae65ca3d3e450868)
Subscribe to RSS Feed
Powered by Reddit
= f037147d6e6c911a85753b9abdedda8d)
You are viewing a comment permalink. View the original post to see all comments and the full post content.
Comments (129)
Once I was at a lecture about violence against women, where the lecturer told us that 20% of women are victims of domestic violence. I asked her if she knows in which country and approximately which decade was this research done; suggesting that the results for e.g. Sweden could be different than for e.g. Afghanistan, and also maybe the results now could be different from e.g. half century ago.
She said that the research was replicated many times, and that no matter which country, or which year -- or even which definition of domestic violence was used! -- the results are always 20%. Somewhat ironically, after hearing about so much successful replication, my faith in the research actually decreased.
Maybe "20%" is some psychological attractor, where all values smaller than half naturally converge.
I also read somewhere an explanation (but haven't verified it) that the number of rapes at campus was achieved by surveying students in the first grade, adding together the results for "rape" and "attempted rape", and then multiplying the result by five (for five years of study). If that's true, even ignoring the "attempted" part, I think the linear approximation is wrong.
First, it ignores the possibility that some factors could make being raped more likely in some parts of the population (such as binge drinking, or choosing violent boyfriends, or maybe just being in a really horrible campus), so being raped at grade X may overlap strongly with being raped at grade Y, and R(X ∪ Y) < R(X) + R(Y). Second, this approach multiplies the Lizardman’s Constant by five, which coincidentally already provides the result of 20%.
I don't want to depreciate a serious issue, but I really wish that people doing research would start taking methodology more seriously.
Did it occur to you that she was just lying?
It's just as likely that she heard that statistic many times and assumed that this must be because there were many studies with that result.
"She said that the research was replicated many times, and that no matter which country, or which year -- or even which definition of domestic violence was used! -- the results are always 20%" -- that's doesn't sound like a vague assumption to me. That sounds like she's being very specific.
I guess you can call it lying if you want, but that is a thing people do very frequently without thinking about or considering that they are lying, namely making very precise statements when in fact there is something vague in their minds.
People lie (defined as intent to mislead) frequently, yes :-/
I guess lying is one option, believing a liar is another option, and... well, are there any realistic options beyond that? (Maybe something in between, like suspecting an information, but deciding to suppress the feeling in the name of the greater good and being on the right side of history.)
But how do I distinguish between these two options, in real time? Ask "hey, lady, it occured to be that you are either lying or stupid -- and because I don't want to uncharitably accuse you of something that you are not, could you please help me solve this dilemma?" I don't suppose that would work.
I tried communicating with mindkilled people in the past, it didn't go well. (I get accused of something; they congratulate themselves for disarming an evil person.) Now I usually suppress the urge.
If there is a forum where people could rationally communicate this kind of concerns, I don't know about it. Christina Hoff Sommers try to address the problem of fake statistics in her book, in return she got her Wikipedia page vandalized. I don't expect to do better.
Sure. There is a very popular option of "I will look only here, I will not look there and even if I accidentally glance over there, I will quickly avert my gaze and feel guilty about my transgression".
Deliberate ignorance combined with cherry-picking evidence can get you very very far.
If only there were some way to think about the two options in something like probabilities... :-)
You mean to tried to deconvert them? No surprise it didn't go well.
Note that the aims of the conversation can be quite limited, though. It's up to you to define your goals and they don't have to be "convince that person that her belief is wrong".
I don't think she's mindkilled. Basically in this case you both haven't read the literature but she's in a position where she's not willing to admit to ot having read the research because that means she would lose status.
A good question might be: "Then how does it come that Louis Harris et al only found 2% of woman to have been raped?"
Skeptics.Stackexchange is a good forum for this purpose. I think it makes sense to open questions there whenever I can boil down the claim to a specific form.
My question Is a woman who dresses sexually suggestively more likely to get raped? for example also produced good answers.