Sandy, if you had what you thought was bad data, you should have dumped it. You should not have editorialized on it, should not have cited from it, should not have mentioned it -at all-. Instead, you have already released parts of the data, but only the parts you had an interest in releasing. Doesn't this strike you as a double-standard?
This alone could be cited as evidence of tainting the second poll, given your editorial on the 'non-' release of the results blasting OS/2 users for the failure of the polling. What kind of backlash would that do to this new poll? The only difference between the polls are that you don't know how tainted each of them are, and that you yourself did the tainting on the second.
-steve
For questions or comments, email the Forums Editor