>Can we rule out that they actually noticed declining sales of some titles as Let's Play videos about them popped up?
Let's run with that reasoning for a minute. That means that they did a study and found not only that the Let's Plays were costing them sales, but also that the lost revenue was small enough that it would be more profitable to keep the videos up and intercept the ad revenue, than to have them taken down entirely.
What that means is that the effect size they observed must have been tiny. As a consequence, to have observed a statistically significant correlation, they would have needed an extremely large sample size. Keep in mind that these are some very noisy variables they are supposed to have measured.
So in conclusion, yes I think we can rule that possibility out. Or at least, if they did notice declining sales, they were irrational for thinking that was sufficient evidence to take this action.
Let's run with that reasoning for a minute. That means that they did a study and found not only that the Let's Plays were costing them sales, but also that the lost revenue was small enough that it would be more profitable to keep the videos up and intercept the ad revenue, than to have them taken down entirely.
What that means is that the effect size they observed must have been tiny. As a consequence, to have observed a statistically significant correlation, they would have needed an extremely large sample size. Keep in mind that these are some very noisy variables they are supposed to have measured.
So in conclusion, yes I think we can rule that possibility out. Or at least, if they did notice declining sales, they were irrational for thinking that was sufficient evidence to take this action.