I suspect mostly the point is to come up with an idea that can
get funded so some team will stay employed a while longer, and
they probably have some ideas of what responses they expect to
see, and you're right that the responses will be skewed because
they'll mostly come from people who are inclined to answer such
surveys.  But it's not as though they can force a representational
cross-section of MMORPGers to answer their survey.

And of course, if they're trying to compare MMORPGers to players
of smaller-scale games like PS3, they can hope many of the same
self-selection features will still apply, and they can still hope
to see some useful correlation.

But often these surveys feel like they have a result in mind and
are skewed so as to ensure the desired conclusion.

In any case, I figured I'd bollux their data nicely by responding,
since frankly I'm happily married, enjoy my job and RL gaming as
well as online games, which I figure (a) isn't the stereotype
they're looking for, and (b) normally wouldn't take the time to
respond to such a survey.  So fine, have my data! :-)

    -- Don.

---------------------------------------------------------------------
-- See the a.g.e/EQ1 FAQ at http://www.iCynic.com/~don/EQ/age.faq.htm
--
-- Sukrasisx, Monk 77 on E. Marr       Note: If you reply by mail,
-- Terrwini, Druid 60 on E. Marr       I'll get to it sooner if you
-- Teviron, Knight 59 on E. Marr       remove the "hyphen n s"
-- Wizbeau, Wizard 36 on E. Marr

Phishing

That's not fully correct.  A better statement would be that self
selected samples are not valid substitutes for random samples.  Whether
that is an issue or not depends on whether or not you are doing
something that needs a random sample.

Research trying to establish the existence of something, for example,
would probably not need a random sample, whereas researching trying to
establish how common something is would probably need a random sample.