The New York Times is maintaining an
infection tracker page for U.S. Colleges and Universities. It's a great example of how such dashboards are meant to inform people but often paint
incomplete, and sometimes misleading, pictures of the pandemic.
The main feature of the NYT tracker is a list of states and colleges/universities in them, and the number of infections in each (along with some historical trends). Here is what the top part looks like for yesterday, with the Arizona item expanded:
As you can see, the primary numbers are "cases." This data makes the state schools look like disaster areas, particularly Arizona State.
But As I have noted before, using the right numbers is critical. Total cases are practically never the right numbers, especially when you are making comparisons. For example, ASU is one of the largest public universities in the country, so you would expect it to have a lot more cases than a smaller school like NAU.
In an attempt to account for this, I got enrollment numbers for each of the universities listed using web searches. This can be a problem because there doesn't seem to be any central source for getting consistent enrollment numbers for colleges and universities, so I can't ensure that all the numbers are counted the same. Do they include all campuses or only the main campus? In-person students only, or online students too?
Nonetheless, I got what is available and it serves for illustration purposes. So here are the numbers as reported by the NYT and numbers normalized to cases per 1000 students:
Here the picture looks different for the top five schools. NAU rises to the top and ASU goes down to third. Midwestern and GCU trade places.
But even this obscures potentially important differences between the schools. For example, all three of the state schools have
random testing programs. Grand Canyon, according to someone I know there, does not. Students are supposed to report cases to the university, but there is at least a chance that they do not do so. At the state schools, students not reporting would likely be detected by random sampling.
The point is that the whole design of the NYT tracker is to invite comparisons between states and schools within a state. Yet it fails to give data that make such comparisons valid. A student deciding which school to attend would be misled by this data, because it has nothing to do with their chances of getting infected at a particular school.
This is disappointing because the Grey Lady knows better. Also whereas I, a humble blogger, had to crib enrollment numbers from disparate online sources, the NYT has resources to contact registrar offices and get official enrollment numbers counted in a consistent way. Instead it looks like they're taking the lazy way out.
Image by Nikolay Georgiev from Pixabay
