GVDammerung
First Post
Dr. Awkward said:I'm seeing a lot of completely useless results from self-selecting groups comprising terrible sample sizes and answering questions that may or may not bias the responses. I vote in polls here, but all they tell us is how the few people who vote in ENWorld polls on subjects they are personally interested in feel about the subject. The error term on any conclusion you draw would be big enough to have its own gravity well, with unwary statisticians trapped in orbit around it.
All data is useful. Some is simply better. But that doesn't make less than ideal samplings worthless. ENWorld can be said to represent _a_ hardcore group of gamers. A such, even if the sampling is far from perfect, it still has relevance. It should just should not be read as more than what it is. At the same time, it is not worthless when so read.