Reply #92: Interesting

[Nederland]

Your comments are, as always Febble, insightful.

I lost touch with the whole TIA love fest even before he got banned. I guess my frustration with him was his refusal recognize that before you start calculating a MOE you have to demonstrate some sort of confidence that your sample is truly random. It would be one thing if he and I disagreed about whether or not a sample was random, but TIA acted like it wasn't even an issue. I still remember the time he calculated the MOE for a DU poll he put up, as if a self selected sample of DU members was somehow representative of, well, anything remotely random. As you mention, in the 2004 election the easiest way to do eliminate sampling error is to drop down to the precinct level, but at the time TIA never grasped even that. I suppose I should be glad he is making some progress on the self education front.