Re: [MLUG - DISCUSSION] web development language of choice

[Mike Miller <EMAIL:PROTECTED>]

On Thu, 3 Jan 2002, Jonathan King wrote:

> On Thu, 3 Jan 2002, Mike Miller wrote:
>
> > On Thu, 3 Jan 2002, Jonathan King wrote:
> >
> > > [web polls can be biased even if pollsters aren't (deleted by jking)]
> >
> > Brilliant analysis!  This is the big problem with this sort of poll.
>
> Yeesh, if this counts as brilliant analysis, then that would explain a
> lot about what you see going on these days...

Perhaps.  It wasn't brilliant because you pointed out a possible problem,
it was brilliant because you found a way to ascertain a cause and
direction of bias by searching newsgroups.  I haven't heard of anyone
doing that before, but it is a very good idea.  The usual method in
academic circles is to dismiss all such polls as "unscientific."  That's
always safe but your method gives us a second source of data that can help
us to understand how internet polling really works, why it is unreliable
and it gives us some information about the amount and direction of the
bias.  That's pretty good, I think.  If I were still teaching research
methods, I would quantify the number of hits for different news groups and
use this as an example in my lecture of survey methodology.  It makes a
very clear point.

Mike

--
To unsubscribe, go to http://mlug.missouri.edu/members/edit.php

Archives are available at http://mlug.missouri.edu/list-archives/