Any study that attempted to prove that kickstarter has had a negative effect on retail stores very well could prove that kickstarter hasn't had a positive effect on retail stores. So therefore if a study could be produced to prove one condition, then a study could easily be proven to demonstrate the other position.In this case, "proving a negative" means conclusively demonstrating that Kickstarter specifically has not had any harmful impact on retail sales outlets.
If you want to take it to this extreme level, then yes, it is impossible to prove. But if you want to take it to such an extreme level, let's take it the other way. "If I can prove that I have backed a kickstarter and that I would have otherwise bought an RPG product with that money from a retail store, I have proven that kickstarter has negatively impacted a retail store." I'm sure someone here can say they have spent money on kickstarter that they had otherwise set aside for a retail store. Therefore the proof that kickstarter has had an impact on retail stores is one that's easily proven.Given that this would include conclusively demonstrating that none of the money used for any Kickstarter pledges - regardless of what project is being funded, RPG-related or not (e.g. indirect competition) - would have gone to any purchases in a FLGS, that's pretty well impossible to prove.
This is of course taking it to an extreme level. If one were to take this to a more reasonable level, then a study should be able to demonstrate what, if any, impact kickstarter has had on retail stores, regardless of if those findings were kickstarter has harmed, helped or had no appreciable effect on retail stores.