Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!linus!decvax!decwrl!glacier!oliveb!hplabs!hao!seismo!harvard!cmcl2!phri!roy
From: roy@phri.UUCP (Roy Smith)
Newsgroups: net.news,net.news.group
Subject: Re: net readership poll (discussion)
Message-ID: <2300@phri.UUCP>
Date: Sun, 30-Mar-86 12:15:51 EST
Article-I.D.: phri.2300
Posted: Sun Mar 30 12:15:51 1986
Date-Received: Wed, 2-Apr-86 08:20:07 EST
References: <5192@glacier.ARPA> <1994@hao.UUCP> <5249@glacier.ARPA>
Reply-To: roy@phri.UUCP (Roy Smith)
Organization: Public Health Research Inst. (NY, NY)
Lines: 30
Xref: linus net.news:3913 net.news.group:4635
Summary: Non-readable .newsrc files skew results?


	In article <89@zeus.UUCP> bobr@zeus.UUCP (Robert Reed) points out 2
ways that arbitron will miscount: non-readable .newsrc files and people
with accounts on several machines at a site, but who usually only read news
at one of them.  He claims that "in either of these cases the effect is
systemic, reducing the percentages with without skewing towards any
particular newsgroup."

	Looking around my site (about 100 users, maybe a dozen of which
read news on a regular basis) I can divide the population into two groups.
The sophisticated ones (by and large) know how to change the mode of their
files, and can deal with the idea of accounts on multiple machines.  The
non-sophisticated users (the majority) have enough trouble figuring out how
nroff works.  The first group also has the people most likely to read the
technical groups.  Thus, I claim that both of the types of errors described
by Robert discriminate against technical groups.

	Now, don't get me wrong.  This is just anecdotal evidence with
absolutely no hard data to back it up.  At any rate, I suspect that the
effect is quite small and not worth worrying about too much.

	I fully support Brian Reid's efforts to gather statistics, and have
confidence that all the minor problems that have been pointed out will end
up not making much difference (as long as the sample set is large enough).
Will 2.10.3 support some kind of "sendstats" control message when it
finally gets released?
-- 
Roy Smith, {allegra,philabs}!phri!roy
System Administrator, Public Health Research Institute
455 First Avenue, New York, NY 10016