Tom Wood of the National Association of Scholars has done a masterful job analyzing the human-capital and cognitive-skills economics literature in an article on the “sheepskin effect.” His question: Does college education lead to greater job success because it teaches important cognitive skills, or because it is a “signaling” device that employers use to screen applicants, regardless of whether earning their degree has taught them much or not?
A lot of data must be massaged. To illustrate: One study discovers that college graduates are disproportionately successful compared to non-graduates who have the same years of schooling but no college degree. Wood encapsulates the finding by saying that “the earnings of individuals holding a BA credential were 33 percent higher than matched individuals who had the same years of schooling but no BA credential.”
This could mean that the degree itself  increases earnings (validating the “sheepskin effect”). But it could also mean that the students who fail to complete a degree are less able and drop out, realizing that they won’t benefit from the additional education. An alternative explanation is that those students have success-thwarting characteristics, such as lack of diligence, that are communicated to employers by failing to get a degree.  
Wood’s conclusions — that the sheepskin effect is small — seems a little too pat in the light of the complex findings that he surveys, but one can draw that conclusion. A few more questions arise, of course: Is the value of the degree changing as the pool of college graduates grows? And is the College Board estimate that the typical college graduate earns 62 percent more than a high-school graduate a valid one? Although many take it as gospel, the Spellings Commission’s Charles Miller has questioned some of its assumptions. Finally, the legal difficulties facing employers who use aptitude tests (the latest example being Ricci v. DeStefano) may have artificially increased the sheepskin effect, as Bryan O’Keefe and Richard Vedder argue here.