I'm building a freely available data set from a 50000 student survey of attitudes towards experimental physics. Typically this kind of data is never available without working at the institution where it's gathered. We figured that's BS and really against NSF guidelines for data collection anyways.