Comment by vacuity

You may be right. I made too strong of a statement; there is too much variability even if some good predictive features are used. I believe that it is not too difficult to identify groups where one group has significantly higher risk of being abusive than another group. In particular, people tend to be sexually abused by people close to them[0]. Especially in the context of family, there are probably some people who know and cover up the abuse. I don't believe that most people can hide their inner selves from everyone. For instance, I sometimes hear that celebrities who are rotten on the inside were actually known to be so for years by staff and some ordinary people. If we could conduct honest interviews of people, I imagine a lot of not-so-secrets would come out.

[0] https://womens-safety.com/blog/rapists-often-familiar-faces-...