The GAO doth count too much, methinks

I still haven't found a real use for it yet, but it's cool that the GAO published some of the data used to create their report. The information is in rather tedious formats (sorry, no spreadsheets or delimited text files), and you won't find the survey results at all unless you read the "Instructions for Viewing the E-Supplement" section. (I felt like I was in the DaVinci Code, unveiling hidden information.)

As I glance through it, I'm getting the same feeling that I've gotten from the report so far: the numbers don't seem right to me. I was looking at the entities participating in the E-Rate, and it seems high. They have 2,117 entities participating in NJ, but if I look at every entity that has ever applied in NJ, I come up with only 1,790. One possibility is double-counting because some entities apply as a school one year, then as a district the next. But even double-counting those entities, I still only come up with only 1,978 applicants.

What's not so easy for me to count is when a district uses an individual school as the Billed Entity on a 471. My 1,790 number counts such cases as two separate entities, but I would think they should only be counted once. My quick count found about 200 cases like that, so I'm down to around 1,600 participating entities.

Also, I think they are counting any entity that ever applied as a participating entity, instead of just looking at a single funding year. Of the 1,790 entities I counted earlier, only 136 (less than 8%) applied in all 12 years of the program. Meanwhile, 382 (21%) of the 1,790 applicants applied in only one year of the program. There is no funding year where the total number of participating entities in NJ is over 1,000.

Oh, and about 30 consortia in NJ have participated, and their members often also apply separately, leading to another double-count, but that's small potatoes.

So GAO says the participation rate for NJ is 82%. My look at the data says that in any given year, the participation rate is less than 40%.

Maybe I'm just missing something, but I think the GAO is seriously overcounting participation.