Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: $Revision: 1.6.2.14 $; site uiucdcs.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!mhuxl!ihnp4!inuxc!pur-ee!uiucdcs!mcewan
From: mcewan@uiucdcs.UUCP
Newsgroups: net.women
Subject: Re: Re: rape statistics - (nf)
Message-ID: <31600087@uiucdcs.UUCP>
Date: Tue, 11-Sep-84 15:06:00 EDT
Article-I.D.: uiucdcs.31600087
Posted: Tue Sep 11 15:06:00 1984
Date-Received: Fri, 14-Sep-84 06:53:19 EDT
References: <4040@tekecs.UUCP>
Lines: 37
Nf-ID: #R:tekecs:-404000:uiucdcs:31600087:000:1608
Nf-From: uiucdcs!mcewan    Sep 11 14:06:00 1984

#R:tekecs:-404000:uiucdcs:31600087:000:1608
uiucdcs!mcewan    Sep 11 14:06:00 1984

I think everyone is missing something. If I remember correctly, the
500,000 figure was the total female population of the city. If there's
only a small range of ages in which women are raped (which I'm not
convinced is the case) then you have to reduce that 500,000 figure
to the number of women in that age range. However, if you don't make
any assumptions at all about age, use the total number of females,
and use the average life span as the window, you get the same result!
The original posting that suggested a 50 year window was too conservative.

Having said that, I now make a 180 degree turn and say that I think
that the above calculation is too simplistic. For one thing, it
assumes that each victim will only be raped once (I'm not sure if
the 700 figure was number of rapes or victims, so we may have a
source of error right there.)  If, for example, we assume that
the probability of getting raped is the same regardless of age or
previous attacks, the percentage would be

	1 - (1- ((700/500000) * fudge factor)) ^ expected life span

which is considerable different. A more accurate figure could be
obtained by getting a breakdown of ages for the rape statistics
and population, and calculating

	1 - product i=1..whatever {1 - ((# women age i reporting rapes) /
		(# women of age i)) * fudge factor(i)}

A lot would depend on the fudge factors, which would probably have a
lot of errors (there is a certain amount of guesswork involved) which
makes this calculation highly suspect.

			Scott McEwan
			pur-ee!uiucdcs!mcewan

"Just because something is obvious doesn't mean that it's true."