The other potential major hole is the underlying assumption that there are no structural changes to adult working lives - that since the working lives of past generations were linear, it will continue to be linear for the current and future generations.
And in my opinion, the world has changed a lot even in the past 5-10 years.
I am willing to bet that at least half of you will change to a completely different career within 5-10 years after b-school, and that such a change will be discontinuous (i.e. your past experience won't help you at all, meaning you have to start over). And of that half of you that do make the switch, a good majority will be in a non-business career. Some of you will change by choice (you get sick of business, or you found another calling), and some won't be by choice (you plateau and get laid off, forcing you to change). Moreover, I am willing to bet that most of you won't stay in a business-related career on a continuous basis over you entire working life (i.e. you start off post-MBA in finance, then become a teacher, then take a few years off to focus on your kids because your spouse's career is really taking off, then maybe an elected official in your municipality, then back to business as an independent consultant, etc.).
The shelf life of a career especially in business is getting shorter and shorter. As such, you really can't use long-term compensation data if your working life is going to be a series of discontinuous changes. And that is going to apply to enough grads (dare I say the majority) that looking at long-term compensation data won't really tell you anything.