My totally uneducated guess here is that TEA's methodology involves a sample of every single turnstile click--or at least did the last time they actually did in-field counts. Since Disney isn't feeding them numbers, they have to reverse-engineer them, and that's really the only way short of surveying guests as to whether they hopped. I'm guessing Disney would pretty quickly quash a third-party survey occurring at its parks...
Currently, I wouldn't be surprised if they looked more to Disney's growth numbers and adjusted their raw numbers upwards or downwards as a result.
Meaning, if their baseline is inflated because their last count (say, from 2000, 2005, 2007, or some other random year) used a methodology that counts every single click, their current end result using the percentage gains Disney has provided (for the Disney theme parks that don't release numbers) would all also be inflated.
The only Disney parks that release numbers are at Tokyo Disney Resort, and presumably they're using the 'first-click' method of internal tracking that Disney parks are purported to use, meaning that their numbers aren't inflated and DLR's and WDW's numbers are inflated (it's not possible to hop at HKDL and ain't nobody got time for hopping from DLP to WDSP, so they both get a pass from me.
).
Again, totally uneducated guess, but it certainly makes sense
to me, and helps me reconcile TDL v. MK numbers.