I don't really understand how this is measured? I attempted to look up some research on it, but it seems most articles that say this are referencing one conference by Jesse Schnell who basically just correlated games with demos, sales, and expected sales. What measure is used to figure out if a demo causes someone to not buy the game? I suppose if they measured presales that were cancelled after a demo, but most anticipated games don't have demos anyway so the data is already skewed in the favor of "no demos." Does it take into account outliers like FFXVI? Highly anticipated game with a demo that sold very well...

I would venture to guess that the data is skewed because lots of AAA games don't have demos and lots of indie games that might not have been purchased anyways trying to get a little markershare, but there seems to be such little research on it.

If you have an actual study on the topic, I would be very interested in seeing their method of results.