David Kristofferson kristoff at
Tue May 17 17:57:13 EST 1994

>Judging by your post in bionet.announce, in order to calculate this
>figure you assumed that Usenet readers responded to your survey at
>the same rate as e-mail readers.  I am *sure* that Usenet readers do
>*not* respond to survey requests at the same rate as e-mail readers,
>and I would guess that your survey underestimates Usenet readership
>by a very large ammount.  Of course, I don't have any better data.

I would also not be surprised if the response rate to surveys is
different between e-mail users and news readers.  However, since the
only way that we have to gauge response rate is by looking at the
e-mail lists and seeing what fraction of those readers replied, there
does not appear to be any other route open to determine this value.
Obviously there is no comparable means of judging how many people
actually look at bionet in their newsreaders, so I actually consider
it fortunate that there is at least an approximate means to
guesstimate this figure even though it may be flawed.  As you indicate
above, the method used to determine this value was mentioned in the

It really doesn't make much difference if the fraction of e-mail users
is 30% or 50% instead of 40%.  Any of these figures still represents a
sizable fraction of the readership.  It would only matter if the
estimate was so horrendously bad that the actual percentage using
e-mail was, say 5% or less.  I doubt that this is the case.  If almost
all readers were using news software, then the overload issue would be
much less of a problem.


				Dave Kristofferson
				BIOSCI/bionet Manager

				biosci-help at

More information about the Methods mailing list