or apology required. On reflection you make a good point for the first statement.
However, if one accepts the fact that the poll was conducted for IW subscribers, then looks at whether IWE readers were excluded, the only possible conclusion is that they were not excluded if 60% of the responders were IWE readers. To believe that IWE readers were excluded, can only mean that the purpose of the survey is not accepted since you cannot expect to find a non IW subscriber who reads IWE within a group of IW subscribers.
You said: "Which is exactly what he said -- just because some of the tabbies (print readers) they sampled were also cats[1] (IWE readers), not all cats (IWE readers) are tabbies (print readers). " with this footnote "[1]: Yeah, so the analogy breaks down because not only _some_ but _all_ tabbies are cats, in the real world. But come on, man, be a little flexible -- or do you know a saying that fits _perfectly_? "
A better analogy, pick two intersecting sets where one set is not a sub set of the other. For instance, NT users and Infoworld Readers and say
Some NT users read Infoworld but not all Infoworld readers use NT or conversely some Infoworld readers use NT but not all NT users read Infoworld. Now would you argure that NT users were excluded? I mean, since not one non Inforworld reader who used NT was polled, you would have us believe that NT users were excluded? Isn't this what you are trying to do with IWE readers who do not subscribe to IW?
We both agree on whether the poll accurately reflects the opinion of IWE readers, it does not any more than the poll accurately reflects the opinion of NT users.
For questions or comments, email the Forums Editor