From a chemical standpoint - how would a developer work for two rolls and "die" for one? The molecules move around in the solution, even more during agitation. The developer only developing two out of three rolls is highly unlikely. With stand dev, variance between films might technically be possible, but even then, the timescale diffusion happens on is too short for only one film to be affected by it.
What might technically be possible, on the other hand, is that the three films develop at vastly different rates (build density in varying speeds), such that two might look okay-ish while one looks blank, in the case the developer stops working midway between development. This feels unlikely though, and Occam's razor would say the film itself is the most likely issue (would have to be determined on its own to prove whether it is or not).
Any chance one of the students might have poured used developer back into the stock Xtol container, possibly along with some kind of contamination?
Assuming the tanks are the same as last year, that seems unlikely.
On those tanks though: I was never quite sure how effective they are in combating oxidation of the developer? I know people keep xtol (and derivatives) in wine bags, but always felt like anything in the ones you had last year would slowly oxidise (slower than in ordinary bottles that are regularly opened and closed, but still oxidising). The "sudden death" of ascorbic acid developers like xtol is why always testing them beforehand is reccomended, for instance like described here:
https://www.photrio.com/forum/threads/testing-xtol.96449/
Can it be something with the thinness of the film?
That's just the thickness of the material the film is coated on - so that's unlikely. If you were to put the
same thickness of emulsion on glass, acetate, PET, or any other non-porous support that does not soak up any liquids, it
should develop to the same densities with the same exposure, with the same time (if agitation, temp etc were exactly equal). Maybe even paper would give the same result, but I'm not sure in which ways the porosity or tendency to soak up liquids would affect the development.