100 psychology experiments repeated, less than half successful (arstechnica.com)

I wonder what a different sampling of experiments would have revealed?

Previous discussion: https://news.ycombinator.com/item?id=10131387