I think that is a very interesting observation. I do wonder why, though. Considering another commenter mentioned how he felt that he would only consider hiring 4 students out of a large research university, it sounds as if the US Computer Science higher-ed is failing in a major way, and much more so than other fields of concentration. I've been on both sides, as a hiring manager and as an instructor at a fairly competent university where implementation of the ACM curriculum's more industry-focused units was debated ad nauseam. CS101-type classes in the universities and colleges I've been around fail a good part of each class. If Jane Margolis and Lenore Blum's work on women in Computer Science is to be believed, those classes might actually be too harsh of a filter (not due to rigor but rather the imposing and overbearing influence of those of us who learned programming when we were kids). Yet here we are confronted with evidence that the filter might not be good enough.