First of all, college graduates on average are smarter and have better work habits than high school graduates. Those who graduated from college were better students in high school, for example. Thus, at least a portion of the earnings premium associated with college has nothing to do with college per se, but rather with other traits.

Second, a goodly proportion (more than 40 percent) of those attending four-year colleges full-time fail to graduate, even within six years. At some colleges, the dropout rate is strikingly higher. While college students sometimes still gain marketable skills from partial attendance, others end up taking jobs that are often given to high school graduates, making little more money but having college debts and some lost earnings accrued while unsuccessfully pursing a degree.

Third, not everyone is average. A non-swimmer trying to cross a stream that on average is three feet deep might drown because part of the stream is seven feet in depth. The same kind of thing sometimes happens to college graduates too entranced by statistics on averages. Earnings vary considerably between the graduates of different schools, and within schools, earnings differ a great deal between majors. Accounting, computer science, and engineering majors, for example, almost always make more than those majoring in education, social work, or ethnic studies.

Fourth, the number of new college graduates far exceeds the growth in the number of technical, managerial, and professional jobs where graduates traditionally have gravitated. As a consequence, we have a new phenomenon: underemployed college graduates doing jobs historically performed by those with much less education. We have, for example, more than 100, 000 janitors with college degrees, and 16, 000 degree-holding parking lot attendants.

Does this mean no one should go to college? Of course not. First of all, college is more than training for a career, and many might benefit from the social and non-purely academic aspects of advanced schooling, even if the rate of return on college as a financial investment is low. Second, high school students with certain attributes are far less likely to drop out of school, and are likely to equal or excel the average statistics.

Students who do well in high school and on college entrance exams are much more likely to graduate. Those going to private schools may pay more in tuition, but they also have lower dropout rates. Those majoring in some subjects, such as education or one of the humanities, can sometimes improve their job situation by double majoring or earning a minor in, say, economics.

As a general rule, I would say graduates in the top quarter of their class at a high-quality high school should go on to a four-year degree program, while those in the bottom quarter of their class at a high school with a mediocre educational reputation should not (opting instead for alternative methods of credentialing and training).

Those in between should consider perhaps doing a two-year program and then transferring to a four-year school. There are, of course, exceptions to this rule, but it is important for us to keep in mind that college is not for everyone.