Another distortion is introduced by the idea of averages. In the discussion on the worth of college degrees, which is intimately bound up with the nature of the graduate and her personal abilities, averages are meaningless because the statistical idea of an average makes logical sense only when applied to homogeneous units, like widgets. Heterogeneity in a population creates problems for the concept of average. When talking about human individuals, particularly in debates that intimately involve personal attributes such as knowledge and talent, averages are almost meaningless. So when the authors write something like the following,
Over a lifetime, the average college graduate earns roughly $570,000 more than the average person with a high school diploma only—a tremendous return to the average upfront investment of $102,000.3 An associate’s degree is worth approximately $170,000 more than a high school diploma.
Then the authors present a claim that is highly misleading, and not just in some hypothetical sense as far as personal qualities go, it is also misleading mathematically. Suppose, for example, that in a population of college graduates, 30 percent of those graduating from top schools and top programs make extraordinary returns; but the rest, the majority, make much less. Specifically, suppose that of 6 graduates, 2 make 250,000 a year while the rest make 11,000 a year. Your average salary for a college graduate will be 88,000 dollars a year, a profoundly misleading result because the majority of the population never sees those kinds of average returns. This bogus result is the effect of outliers: statistically irrelevant data points which distort the results. The earnings overall in our example are abnormally high because they have been distorted by the elite cohort and their outlier performance. To obtain a more realistic average, the exceptional earners would have to be removed from the sample. This is why you have be very careful when someone claims that on average college students earn more or have better jobs, most may not, except for a few lucky ones.
If we looked at median earnings in the above example, we’d find $11,000. In fact, the median earnings of college graduates are higher than earnings of those who did not graduate from college. Before you rush in your college application, however, consider that, in fact, these median earnings figures are again profoundly misleading.