You are simply repeating the same assertion with trying to understand the flaw.
Let's take an example. Say we observe traffic along a long stretch of highway from a satellite in orbit. We can see that on average the vehicles move along at 60 mph.
Later we make ground observations and we find there are regulatory mechanisms for the speed of traffic. There are stop lights, speed limits, cops enforcing the rules.
None of that matters for our original observation. That is what it is. It was never dependent on the *reasons* traffic flowed the way it did. Only that it did.
The aggregate rate of change (the one you are talking about) can be easily miscomputed if we failed to account for all the factors involved in creating it, particularly those factors that have a frequency of occurrence longer than the period of our measurement.