Looking at the methodology it looks like a reasonably conducted survey, but it has a basic methodological flaw: it cannot distinguish between any particular mod

Looking at the methodology it looks like a reasonably conducted survey, but it has a basic methodological flaw: it cannot distinguish between any particular models, such as "a privileged group oppressing contrary beliefs" versus "the peace treaty of tolerance is functioning and bigots must suppress their beliefs to keep its protections." Further, the causes of self censorship need not be identical. Both of the two models I just posited can be in play at the same time in various subpopulations and around various issues. And there are lots of group/individual relations you can posit for this. Social science is hard. So this data set might be an interesting preliminary trial, but releasing it as a finished study is either sloppy or malicious on the part of the Cato Institute. Edit: Reading the other comments on here, I'm inclined towards malicious as it seems spun to create just the kind of rhetoric that we're seeing here.