Measuring college opportunity

College opportunity can be measured in any number of ways. One common measure of opportunity is the spatial availability of postsecondary institutions. Simply stated: for a given area—let’s say a county—how many colleges are nearby? Even more simply: does this county have a college? Much academic research and policy work uses these count and indicator measures when trying to account for college opportunity.

Yet these measures are not without their flaws. Indicator metrics treat counties with one college the same as those with 10, 20, 30, or more, effectively saying that being near one college is the same as being near many. While count measures take the number of colleges into account, they assume that the student college decision is bounded by the county. These boundaries ignore migration/commuter zones and the fact that county boundaries may not be salient for most students.

Another measure, the inverse log distance to surrounding schools, represents an attempt to more accurately measure college availability in an area. Using the population-weighted centroid of each county¹, the straight-line (“as the crow flies” distance) is measured to each college in the country². Taking the natural log of these numbers and summing their inverses³, a measure is constructed for each county. Higher numbers mean a higher density of schools; lower numbers mean fewer schools.

The map above visualizes each of these measures for various samples of colleges and universities. For the indicators, a county is shaded if there are any schools within the sample in the county. Hovering over the county will give the number of schools. Sliding the toggle switch gives the distance measures. These too are computed for a variety of school samples under one of two conditions. Either all relevant schools are included in the measure or only those within the same state.

Indicators

A little over 40% of all counties have at least one postsecondary institution⁴ within their borders. Here are the percentages for each category:

• Any school: 43.9%
• Public 4-year: 14.9%
• Public 2-year: 23.1%
• Private 4-year: 11.1%
• Private 2-year: 2.8%
• Proprietary 4-year: 4.1%
• Proprietary 2-year: 7.1%

Distance

The two cuts of distance measure—all sampled schools vs. only those within the same state as the county—produce two distinct types of map. Because inverse log distance measures aren’t particularly intuitive, deciles of the z-scores are mapped. Counties with values below the 50th percentile are red; those above the 50th are blue. Hues for both colors become darker as the values move away from the middle. In the distribution of all counties, those shaded dark red have the lowest density of schools whereas those shaded dark blue are at the top.

“Move to Ohio” maps

For the measures that include out-of-state schools, a number of “move to Ohio” maps are produced. While upon first glance it may seem that the Old Northwest states have the highest density of colleges in the country, these maps really measure the nationally-weighted center. The large number of colleges in New England pull the college density center away from the lower-48 geographic center.

These maps are most interesting when visualizing how different sub-samples of schools move this center. While the sub-sample of only public 4-year institutions pulls the center toward the New England states, the sub-sample of public 2-years shifts it further south and west, reflecting state systems found in those regions.

State patchwork maps

When only considering in-state schools for each county’s measure, patchwork maps are produced. Some states, such as California, Florida, and North Carolina, remain in the upper deciles of college density no matter the sub-sample. Other states, like Massachusetts, Tennessee, and Washington, move in the distribution depending on the type of schools considered.

Because the differences in college density between states is often greater than the differences within, many states have a uniform shade in these maps. A few states, however, do show variation within their borders. In some cases, like Georgia with in-state 2-year colleges, these are around city centers. In others, like in-state public 2-year colleges in Kentucky, differences occur in counties that contain that type of institution.

Notes