My problem is that there is too much vagueness in the study, as a result of not considering any of those factors.
We may have an idea about the reaction to 12 books "with twists", which isn't a statistically relevant number, but it's hard to transfer this result to the prediction of the enjoyment of OTHER works, when established how DIFFERENT reactions can be depending on the structure and implementation-

I'm not saying it was entirely pointless.
But it would need further research, and reproductions, before you could make it into a valid general observation.