I'm curious how well the study results reflect actual numbers of offers, salaries, and benefits. I would expect that since it's a voluntary survey, there would be some amount of the 'high school reunion effect' going on here and that the survey results would skew higher than the actual. In the same way that those who look good and have done well are more likely to attend a high school reunion, I would assume that those students who have had relatively successful job searches would be more likely to respond to the survey.