Closing 8,000+ stores for an afternoon feels like a sizable response, and I am going to suspend my cynicism for this and recognize that they do give a damn (even if it is influenced by PR).
If you don't spend the night in jail, then you have one data point suggesting not PR.
Have a reasonable sample of people repeat the experiment in multiple locations and multiple times and carefully record and report the results to you.
Collate the data points and report on the number of people arrested during the experiment.
Feel free to include Starbucks in a tweet about your results when you publish.
This whole thing is going to have huge variation between locations. Suburban or drive - to locations often don't ever lock the bathrooms, while ones in cities are more likely to.
The question is are they making a real sacrifice at that amount or paying a small sum for the PR?
Simply put, why close business when these trainings can be done staggered or off store hours? That's my line of thinking.