ho3@stat-l (Timothy Hall) (02/22/85)
>From: guest@hplabs.UUCP (HP Labs Guest/guest) >Newsgroups: net.motss >Subject: Re: Heterosexual Questionnaire > > >> I find it very amusing that ...hplabs!guest was quibbling about the >> three "statistical questions" (disproportionately more heterosexuals in >> mental homes, more heterosexual child abusers ("why", he asked, "didn't >> you say *disproportionately more?"), and disproportionately more hetero- >> sexual criminals and welfare recipients). For one thing, I'd guess that >> the main point is that meaningless statistics are often used like this >> *against* homosexuals. Another thing is that these statistics are *very* >> likely, considering just the demographics - where most homosexuals are >> and where most mental patients, criminals, and welfare recipients are. >> (Of course, I wouldn't go on this alone, but I'd be less likely to get >> all critical with a questionnaire that is supposed to be, after all, >> facetious.) >> :::::::::::::::: Jym Dyer > But the statistics aren't meaningless if they're stated >as in those the two questions that said "disproportionate". ????????????????????????????????????????????????????? >Are those data true, or simply reversals of common anti-gay >prejudices? It wasn't clear that the questionaire was facetious. >(This reinforces my contention that there is no way to clearly >express any thought other than a program over the net. Was the >questionaire merely facetious?) I thought that it had >questions that it was good for me to consider. After all, >why shouldn't I turn around the prejudices and see where I stand. > > I made my stand on meaningless statistics quite clear (to me) >in my posting. Why to you so vehemently agree with me? Was it >not clear that I saw that they might be meaningless? ------------------------------------------------------------------------- >The correct response when someone hits you with a statistic is not >to say "That's meaningless!" You should determine what the caveats, >withheld information, limitations, and inapplicabilities of the >data are; then you can zing them with the stats. ------------------------------------------------------------------------- Please leave the meaning of statistics to statistical people! Tim Hall Department of Statistics Purdue University ...inuxc!pur-ee!stat-l:ho3