Hate to break it to ya'll, but online polls are not only voluntary sampling, but they are indeed completely volunteer. They are considered "convenience" sampling, and no one takes, or should take, them at all seriously.
We use random digit dial surveys as they are the "Best" but obviously not accurate either. This is due to a principle error in sampling as the available results are already stratified by those willing to answer. The ol' lady works at Gallup on the "Health and Well Being Survey" she has been contemplating quitting for a long time, as she does have other options. She called a few minutes ago to let me know that her survey has 6 new questions, 4 of which are related to gun control, and she is contemplating quitting (not sure how serious she was)
Results of those will be out in a few days, once the poll hits it's target numbers, the data is compiled and dissected, and I hate to say it, but the answers will probably not be good.
"Emotion trumps reason. Every time."
Edit: She wanted me to note that changes to a survey on a weekend are something she has never seen in 2 years working there.
