Nate Silver's data analyses have become popular for a reason, and here's a new one I like on the "decline" of liberal arts.

In my structure of higher education course, we discuss the putative decline of liberal arts college, liberal arts majors, and the liberal arts in general - all three separate issues that tend to be conflated, even by the scholars studying them. Silver is focused on the second of those, the number of students majoring in the liberal arts, and he succinctly points out that about the same number of young people are going to college and getting English degrees as 40 years ago. It's just that more people are going to college, and we have pretty good reason to believe they are a different population of students. He doesn't unpack that notion very much, but expanding access and intensifying credentialing induces students who might not otherwise go to college to go, and they make different sets of choices. Students aren't converting away from English; we're gaining new vocational students.

I'm going to add this piece to reading list for structure for the fall semester.

Presenting network analysis

The Chronicle of Higher Education; ran an article on how colleges select peers, both actual and aspirational, and with it they ran an interactive graphic showing what colleges select as their actual peers. (May be behind a paywall.) The data comes from always-popular IPEDS;, but the Chronicle put a fair amount of work into creating a more digestable presentation using Gephi, an open-source tool I have not yet tried.

I use the results in my Structure of Higher Education class. Students explored the data on their own, and we looked at it in class. In general, students found the graphic intuitive to use, although there were some technical issues: Some students reported that the data didn't load, or loaded very slowly. I found that it wouldn't load at all on my work PC, running college-mandated IE8 - but it worked, much to my surprise, on my iPhone. (At least I received the courtesy of an error telling me what browsers would work.)

The resulting graphic shows some things that tables couldn't easily. For example, some institutions cluster tightly, choosing each other as peers, whereas institutions do not. That's immediately visible by clicking on any institution, represented by a dot, and seeing whether the other dots that light up are near or far from it.

One minor distraction was that the entire graphic was not visible in one window. I didn't expect this on my iPhone, naturally, but even on a traditional monitor, only part of the data was visible at once, even when zoomed out. Ideally, the zoom level would allow it to all be seen at once. One potential fix would be to move institutional data from the right to the top or bottom, given the landscape orientation of the graphic. There is actually relatively little of what Tufte calls "chartjunk" taking up valuable screen space - the browser itself uses some space, but this is out of the Chronicle's control.

Overall, it's a relatively sophisticated presentation of data that really adds value above what a strictly tabular format would.