I just took the "root of all sins" test and I tried to distinguish the answers of the Christians and non-Christians entirely based on shibboleths. Disordered love? Christ is a blinding searing light? Humans are finite beings who naturally desire the infinite? Maybe. But the decision was not "would a Christian have those ideas" but "would a Christian phrase the ideas that way".

Of course I can't just go count the shibboleths; it's possible that non-Christians might overcompensate and actual Christians don't talk about Jesus' blinding light much at all, at least not actual Christians of the type who answer such surveys.

But either way, I didn't feel that the most likely way to figure out which answer came from Christians was to look at the content of the answer. So I think that the test has already failed.

On top of this is the question of what type of Christian the non-Christians are trying to imitate. Are they trying to imitate average Christians, average survey-answering Christians, average blogging Christians, average Christians who are knowledgeable about Christianity? Trying to imitate the wrong kind of Christian can mean that knowing too much about Christianity can make your imitation fail.