Are Police Records a Representative Sample of Crime?

The data you have not actualy measurement of x. It's a measurement of some complex interaction between x and your ability to observe x, right? So in this case, this first model, we're talking about somecomplex interaction between fraudiness and our ability to observe company is right? We're not actually learning something about x. Is not actually measuring fradines itself. And when we make predictions on the basis of that model, we aren't really making predictions about x,. or making predictions about something like our ability to observing x in the future.