🚀 go-pugleaf

In article <1992Aug7.130818.1354@cis.ohio-state.edu>
 adkins@gecko.cis.ohio-state.edu (Brian Adkins) writes:
>Excuse me if a similar post got through, but it didn't show up after 8 
>hours or so.

The lost article was probably caused by the same thing that's kept me
from digging up Yaqub's references, namely a constipated news server.

>Oh c'mon J.  Sometimes I can't tell whether you're being facetious or
>fallacious.

I prefer "snide", thank you. :-)

>I mean you ask Adnan to provide some references, and when he does you
>ask about means or medians ?

I asked about all sorts of things; his summaries of the studies
provide no evidence of a causal relationship, and his presentation of
a table of "proof" is completely useless, because he omitted all
context.  When someone posts data that they claim supports their
views, it's quite proper to ask them what the data is supposed to
*mean* (pun intended).  The difference between mean and median is
quite important in interpreting the results that were posted,
particularly if the sample size was small.

  There are other things missing.  If, for example, these people were
told that a 60 month sentence was typical for a rapist, all of their
responses indicated an increase over the norm.  If they were not
evaluated for previous knowledge of rape sentences, and those most (or
least) experienced ended up in the same group, they would bias the
results.  Yaqub didn't even deign to tell us how many people were in
each group, which alone could be sufficient to make the results
insignificant.  Hell, he hasn't even told us what constitutes
"pornography", either in the individual studies or in his opinion.

>Maybe we should discuss the variance and approximating distribution ?

The original author certainly should have.  Another reason not to rely
on Yaqub's summaries and look for yourself.

>How do you conclude that the experiment was designed to reinforce the
>experimenter's bias?

Why length of sentences?  Juries don't sentence, they convict or
acquit.  Judges set the sentence, based on knowledge that the average
citizen is not expected to have.  Why rape of a _hitchhiker_?  Why not
a housewife, a coed, a stockbroker?  Why pick a case where the victim
was engaged in unsafe activity that placed her at the mercy of
strangers, forcing the test subjects to evaluate *her* actions as well
as the rapist's?  ("she was just asking for trouble", "she shouldn't
have been out there", "hitchhiking is illegal", "I'll bet she was
wearing tight little shorts", ...)

  I can not conclude that the experimenter was consciously biased, but
the description of the study strongly suggests that bias was present.


			"...and now a word for dog lovers."
				"Kinky."
-- 
J Greely (jgreely@cis.ohio-state.edu; osu-cis!jgreely)