March 31, 2009

Lies, damned lies, and the Pacific Research Institute

Often, it takes a bit of time to uncover statistical flummery; one needs to dig deep inside arcane methods or the details of data collection. But sometimes it takes just a few clicks. Matthew Ladner's blog entry today, $243,000 per student school districts?, quoted work by the Pacific Research Institute that claimed two small districts in California each spent more than $200,000 per student. Scandalous! Criminal!

Well, it might be, but so is the credibility of anyone who quotes those statistics without taking a few minutes to look a wee bit more closely. According to the quoted press release, California spends something over $10,000 per student ($11,600 is the figure quoted for 2006-07), but the Mattole Unified and Mineral Elementary districts supposedly spent $225,256 and $242,610 per student.

My first thought was of the tiny New Jersey districts that literally had no students for odd administrative-law reasons. Or maybe these were essentially fictive districts created by companies that incorporated towns and funneled money into specialized programs for a tiny population of executive children. But the locations (Humboldt and Tehana counties) didn't fit with either hypothesis. I was curious: I clicked. And then I saw the magic words in the PRI website pages: "Revenue Received per ADA." ADA = average daily attendance and is not always the same as enrollment.  Okay, so what was up?

Mattole apparently has 905 enrollees but the ADA listed on the PRI site is 35.2. Mineral has 123 enrollees and an ADA reported by PRI as 4.7. If you divide the revenues by the enrollment, the average revenues are a much more sensible $8,761 and $9,270 per student, respectively. That sounded very odd: WHY would ADA be so low? So I checked out the districts. In Mattole, the vast majority of enrollment is in a single charter school, Mattole Valley Charter, which had 864 enrollees in 2007-08. In the Mineral Elementary district, the eScholar Academy virtual charter school had the bulk of enrollment.

If gambling on proposition bets were legal in Florida, I would bet at least a little money on charter-school enrollment NOT counting for the district's official ADA. So in any small district where the majority of enrolled students were in charter schools (or in this case, one charter school each), the official ADA would be far lower than the actual enrollment of students in schools receiving public support. 

But instead of betting any money, I'll just bet my reputation. Does anyone want to prove me wrong? 

Update: In comments, Ladner and PRI staff member Vicki Murray acknowledge that I was correct, and the original claim was incorrect.

Listen to this article
Tags: research ethics
Posted in Education policy on March 31, 2009 8:15 PM |