Maybe they don't believe the data? And it's very tedious to check. If that's the case, I suggest to check the data for a few individual universities. Cornell and Princeton are slight outliers, but otherwise individual universities don't deviate much from the average. That should give some credence to the numbers.