This is not surprising if you pay close attention to the results of Festinger and Carlsmith, 1957. Three DVs, one significant, two not, with the discussion explaining why the significant measure was the best one all along.

Indeed. Others voiced some pretty strong criticism around that time about it. Here's a quote from Chapanis & Chapanis (1964):

Yeah but after they wrote the article they found it much more plausible