From the article: "One such limitation is that a community sample recruited through internet databases has limited psychopathy; future research could utilize an

From the article: "One such limitation is that a community sample recruited through internet databases has limited psychopathy; future research could utilize an offender sample to bridge that gap..." While I do not doubt the conclusions, particularly, I am not sure this is particularly persuasive evidence. It might make sense to have done this first, before the (much more expensive and time-consuming) later project of gathering data on actual psychopaths, and also checking if this is only something you see in the sub-population of males who do lots of Mechanical Turk work. Again, not that I particularly doubt the conclusion.