Via Distributed Republic, a Salary Report from PayScale.com that lists both the starting median salary and the mid-career median salary by school and field of study (though not both at the same time). A few thoughts:
As a relative matter, these numbers are about what one would expect. Ivy League, Standford, MIT, etc. dominate state schools. Engineering dominates Liberal Arts. Interestingly, "engineering schools" (a category which includes Stanford and Cornell but not Harvard or Yale, even though all of them offer engineering degrees) dominate the "starting salaries" rankings while the Ivy League dominates the "mid career salaries" rankings. This confirms my impression that engineering buys a stable middle-class life rather than a chance at great wealth.
In absolute terms, however . . . frankly, I choke at some of these numbers, and I would be curious to see a histogram of the salary data rather than just the median. I've been very skeptical of salary reports like these ever since Half Sigma (I think) showed how Law schools, by the way they skew their samples, cook the books to make their graduates appear more successful than they really are. Those book-cooking methods are presumably available to all schools across all disciplines. Take, for instance, Georgia Institute of Technology. PayScale.com lists the median starting salary of its graduates as $59k and the median mid-career salary as $105k. This is roughly in line with the data for engineering graduates: $60k and $102k respectively for EE, for instance. The problem is that Georgia Tech doesn't just offer degrees in engineering. It has colleges of architecture, management, and liberal arts, fields with starting salaries far below those of engineering. The salary data for these majors ought to weigh Georgia Tech's averages down, but somehow in this survey they don't. Further, do these numbers include the unemployed? Those working outside the field? That's why the distribution, not just the median, is useful.
Does this kind of hanky-panky bias the relative results? Or is it a tide that has lifted all boats? For instance, do state schools take more latitude in inflating their numbers than the Ivy League? Less latitude? Is there something about the way the data is collected that inflates apparent salaries in engineering more than those in social work? Less? I'm guessing not, but does anyone have any insight into this question?
Considering the extent to which the Fedrul Gummint underwrites so much of higher education, it ought to insist on uniform, meaningful, difficult-to-game metrics and reporting standards for this kind of information. I think students, parents, and taxpayers deserve to know in detail what their career profile is likely to look like, given their school, major, and academic performance therein. Of course, count on higher ed's lackeys in Congress to help colleges hide this information. How much better to shovel money at their political supporters while mindlessly parroting the propaganda about the benefits of a college education.