The trouble with stats like this is that they assume that all scrum and agile teams are equal. Not to mention waterfall teams and who does waterfall anyway - life just doesn't work like that!
I've used scrum and think it is very useful in managing small teams but it is basically just a fairly formal way of trying to measure how well you are performing. Does it solve any problems? No scrum / agile methods can't solve or improve anything but they can highlight issues which good teams can solve.
Looking at stats this way assumes that it is the agile method that makes the difference. The truth is that good teams that are prepared to critically look at themselves and try to improve do better, agile is simply one way to help with that.
So ignore these stats you won't get a productivity gain by adopting one method or another but it might be a useful tool in recognising shortcomings and improving.