Okay, in the right order:
Testing @ 2500 rpm-no load, as compared to a dyno run, would result in a lower exhaust temperature.
Age & mileage will have a deteriorating effect on the converter.
Exhaust leaks would scew the readings ( diluted ), probably in your favour.
A large leak would be noticed by the analyzer & the operator.
In summation, a " lazy " Oxygen sensor could possibly be the main culprit of your problem.
Do you have the means to check the livelyness ( cross-counts, rich-lean swings ) of your sensor ?
If not,can you have someone do it for you ?