I'd think that there is no way to control for variables. You'd have to control for age of students, socioeconomic status, length of visit, availability of books, familiarity of the author before and after the visit ( a bestselling author vs. someone new) and more. For many visits, the local libraries and bookstores help promote. So how would you know what effect was from the promotion and what effect was from the visit? I can't see how a good scientific study could be done.