>but why does that affect the quality of the data If I poll only insurance company executives on whether the country should do away with private insurance and

>but why does that affect the quality of the data If I poll only insurance company executives on whether the country should do away with private insurance and switch to universal healthcare, I haven't really done anything "data driven", have I? Taking a fat shit in my toilet "generates data", but I wouldn't call that visit to my bathroom "data driven", would I? Collecting sex poll responses from people who follow porn stars who specialize in rape and similar fantasy porn does not collect any data except what extreme porn consumers think about sexuality, which is of little use to making generalizations. >How can you evaluate whether someone counts as a "gooner" or not based on whether they follow a specific person on Twitter or answer a poll about sex? Following porn stars is absolutely gooner behavior, so yes, it makes a person a gooner.