E ach of our three categories included several components. We determined the Community Service score by measuring each school’s performance in three different areas: the percentage of their students enrolled in the Army or Navy Reserve Officer Training Corps; the percentage of their students who are currently serving in the Peace Corps; and the percentage of their federal work-study grants devoted to community service projects. A school’s Research score is based on two measurements: the total amount of an institution’s research spending, and the number of Ph.Ds awarded by the university in the sciences and engineering. For both Community Service and Research, we weighted each component equally to determine a school’s final score in the category.
The Social Mobility score was a little more complicated. We had data that told us the percentage of a school’s students on Pell Grants, which is a good measure of a school’s commitment to educating lower-income kids. But we wanted to know how many of these students graduate, and, unfortunately, schools aren’t required to track those figures. So we devised our own method of estimating that statistic.
Because lower-income students at any school are less likely to graduate than wealthier ones, the percentage of Pell Grant students is an important indicator: If a campus has a large percentage of Pell Grant students—that is to say, if it is disproportionately poor—it will tend to diminish the school’s overall graduation rate. Using data from all of our schools, we have constructed a formula (using a technique called regressional analysis) that will predict a school’s likely graduation rate given its percentage of students on Pell. Schools that outperform their forecasted rate will score better than schools that match, or worse undershoot, the mark. For instance, 37 percent of UCLA’s students receive Pell Grants from the federal government. Using our analysis, one would expect UCLA to have a much lower graduation rate (a mere 48 percent) than it does (a rather high 87 percent) making it the top performer in the category.
When it came to our methodology, we chose to pursue two primary goals. First, we did not consider any single category more important than any other. And second, the final rankings needed to reflect excellence across the full breadth of our measures, rather than rewarding an exceptionally high focus on, say, research.
All components were weighted equally when calculating the final score. In order to ensure that each measurement contributed equally to a school’s score in any given category, we standardized the data sets so that each had a mean of 0 and a standard deviation of 1. The data were also adjusted to account for statistical outliers. In our published list of universities and liberal arts colleges, the data listed reflect a school’s actual performance within that criteria. However, for the purposes of calculating the final score, no school’s performance in any single area was allowed to exceed three standard deviations from the mean of the data set.