It could also very possibly be that your distribution isn’t one population but two. Embed a small over performing population within a larger normally performin

It could also very possibly be that your distribution isn’t one population but two. Embed a small over performing population within a larger normally performing /underperforming population. You don’t j ow who’s cheating, you don’t know when they are cheating, and if they were cheating I can’t imagine they would do it in their public funds. A simplistic analysis of self-reported data is not strong evfidence to me in this case.