Dominic Lawson, writing in the Sunday Times yesterday (here) about the case of 'baby P.' drew attention to the elevated propensity of step-parents to abuse/kill their step-children.
He cited the report (Daly and Wilson) here, which identifies an elevated risk of abuse (above that perpetrated by a genetic parent) of more than one hundred times.
Two points of interest.
1. What has to be explained is why a step-parent would have any interest in caring for a step child at all (zero genetic relationship). The authors suppose that it's a necessary down-payment for the benefits of a relationship with the child's other natural parent.
2. Is the effect identical in traditional, hunter-gatherer societies? The paper present research which shows that the nature of the step-parent - step-child relationship is identical, with significant lower levels of care given to step-child by a step-parent in such societies.
I think it's the continuing pernicious influence of what Steven Pinker called The Blank Slate theory of social behaviour (the theoretical basis for much political correctness) which prevents the public policy implications of all of this being drawn out.
Logically speaking we either accept that these children are going to be abused all-too-frequently by step-parents, or we set the abuse-threshhold much lower and move the step-children out at once to live permanently with adopting parents.
It's another story as to why adopting parents might be trusted not to abuse. Something to do with the elective natue of adoption, combined with - as motivation - the 'failure of discrimination' mentioned by Daly and Wilson at the start of their paper.