>> I don't know who originally said this but their head is in the sand.
>> Linux is becoming huge in the business world (check out the adds in the latest
>> Linux Journal). Here is an excerpt from the latest WGS linux news letter:
>> At Comdex/96, North America's largest computer trade show, WGS offered
>> a FREE Pentium computer to people for filling out a survey. One of the
>> survey questions was "Do you use Linux Yes / No". There was absolutely
>> no reason why people would lie on it as they could win either way. Our
>> preliminary count is that over 45% of the respondents said yes! We got
>> enough surveys filled out for statistical accuracy at the 250,000 attendee
>> show.
>> -- wayne
>> Wayne O. Cochran Graduate Student - Computer Science
>> web:http://www.eecs.wsu.edu/~wcochran
>> School of EE/CS Washington State University Ecclesiastes 12:12
>If most of the respondents were just choosing answers at random,
>then you would expect about 50% to say "yes". How do you know
>this isn't what happened? The fact that 45% said "yes" means nothing
>without more data.
at random. That hypothesis is less likely than the more obvious one.
It would be very difficult to hit upon a random cross section of the
delegates who behaved like that.
What might skew the figures could be the way the sample was taken. If,
for example, the questionnaire was given to people who came to some sort
of Linux stand, then it could potentially be a most un-representative
sample. If the figure is 45% of "people who expressed a general interest
in Linux", then it is probably pretty meaningless. If OTOH it was
a genuinely random sample of those attending, and they were presented
with the survey under no specific OS heading (I.e. it wasn't portrayed
as "A survey of Linux usage", but merely as a general gathering of
information) then the 45% figure could be deeply significant.
John
(Newsgroups heavily trimmed. Follow-ups trimmed even more.)
--
John Winters. Wallingford, Oxon, England.