I don't need to research anything. Here's the problem: You guys act that nearly EVERYONE will change the station the moment a "bad" song comes on. It does not work that way. This isn't robotics were EVERYONE thinks exactly the same, all 450,000 of them. This isn't a flock of pigeons all flying away at once when a loud firecracker goes off.
Actually, they pretty much do. We see it in the MScore data. Bad song = audience departure.
Seriously, is this what you guys assume? Does your coveted research really tell you this? If so, it's flawed and not relevant to human reaction and thinking. Sorry.
The research done via tracking of PPM meters shows that when bad songs or burnt songs get played, a significant number of listeners leave instantly. And we know that this is an indication that, while the listeners near their radio or device tuned out quickly, the rest of the audience likely did not like the experience either and that by playing that song we damage our brand.