So then Disney has no hard facts to believe the local guest spends less than the once in a lifetime guest.
For starters, they are not (primarily) comparing "local guests" with once-in-a-lifetime guests. They are comparing passholders with guests on day tickets. They do make a few additional statements about "infrequent" guests, but those are not harder to manage.
But to your larger point: of course they have hard facts---
on average. And, the larger the group of people you are measuring, the more averages tell you. Here's how I would do this if I were them.
First, you know total attendance each day (A), and you know how many of those guests entered on an annual pass (P). The rest* entered on day tickets (D). A = P + D. Second, you know how much revenue**, in total, came in that day (R). They report a metric derived from A and R in each quarterly report: per-capita in-park spending, or R/A.
Given those numbers, you want to determine how much the
average passholder contributes to R/A (call that R_p), and how much the
average day-ticket guest contributes (call that R_d). Those are your two unkowns. On a given day, we know that:
R = R_p * P + R_d * D.
But, once you have two
different days, you can approximate R_p and R_d. Each day has a different revenue number, and a different mix of Passholders and day ticket guests.
Day 1: R = R_p * P + R_d * D
Day 2: R' = R_p * P' + R_d * D'
You have two equations, and two unknowns. As long as the ratio of P/D and P'/D' are different (i.e. you have a different attendance mix) you can solve for (approximate) values of R_p and R_d. Of course, these are approximate, but because the attendance each day is large,
and because you get a different sample of P and D each day, you eventually converge on something very close to the actual averages thanks to something called the
Law of Large Numbers.
Once you have good approximations of R_p and R_d, you can predict what happens to R as you change P and D. You can make those predictions and see if they are accurate across many future days. If they are (within some bound), you can be confident that you have a good handle on R_p and R_d. And as the quarterly earnings calls tell us (over and over and over again) Disney believes R_d > R_p. Either their accountants are incompetent at basic algebra and statistics, they are lying, or they are correct.
I know which one I'm betting on.
Above I said
They do make a few additional statements about "infrequent" guests, but those are not harder to manage. Here's what I mean by that. If you assume they know which
MDE accounts come only once in a blue moon (they do), and you assume that people mostly re-use MDE accounts (they do), then you can further refine D into, say, visitors who come no more than once every five years vs. those who come more often. Now you have three unknowns, and to solve for that you need three days' worth of revenue results, not just two. But, that's not hard when you get a new sample every single day.
Undoubtedly, what Disney is actually doing to understand this is more complicated than what I've shown here with more unknowns, etc. They probably also aren't using averages but some other descriptive statistic, because the spending is probably skewed rather than normally distributed. This is why people with graduate degrees in the social sciences often take a class called "quantitative methods." But the basic idea is the same
.
-----
*: There are also comps, main gate passes, etc. That complicates things a little, but doesn't change the basic idea.
**: Revenue is a combination of admission revenue plus in-park spending. As i understand the accounting rules, AP sales revenue is not recognized immediately, but rather as a function of AP attendance, so there is an established way to track this.