,
9 tweets,
2 min read
Read on Twitter
"Similar" is doing a tremendous amount of work. An alternate or companion hypothesis is the data is missing areas where these students my be dissimilar. There's limits to this kind of analysis, It'd be nice if those doing it admitted them, rather than making such sweeping claims.
Of course there's things institutions can do to help their students graduate, but to flat declare that disparities in graduation rate fall exclusively on the institutions is to be willfully blind to other phenomena that may be at play. It's bad data science and bad analysis.
Let's take one factoid and see if we can theorize differences other than what the institutions are doing. This is literally off the top of my head, meaning, to look closely and critically at data doesn't require a ton of time. /thread 
It's possible more dropouts is better if people are leaving college for steady, secure employment. Perhaps the economic upturn has made good paying jobs in Las Vegas easier to come by, drawing people out of school into the labor force. Maybe not. Who knows?
It's possible that students at UNLV are more mobile, more likely to transfer. To my knowledge the graduation rate data doesn't track this. The cost of living is probably much higher in Vegas meaning the burden of paying for school is greater, even if students have similar incomes
I'd want to do an analysis of the demography surrounding the institution and compare that between those that do better than expected v. worse. I'm willing to bet there's patterns there.
Some of my hypotheses may be contradictory, but this is the point. 1. We don't know in what ways these students may be dissimilar. 2. We don't know the specific material and economic conditions in which they're trying to earn a degree. These are gaping holes in the analysis.
We need to know that stuff in order to then truly understand what impact policies and programs may have inside institutions. Unfortunately, articles like this one steer us away from doing that deeper analysis, and we're stuck in an endless boom/bust cycle of policy panaceas.
This should go without saying, but a policy that works well in one institutional context may not work in another. Institutions need to be able to craft approaches that fit with their students/mission/community. Share what others are doing, but there's no one-size-fits all.
