I've seen a bunch of links in the last week or two to this story from a few years ago about a study on the availability of raw data from old papers. I'm not sure why it's come around again, but ...