I looked this up due to a little background check on a current murder trial here in NZ. The guy had served 12 years for killing his wife and daughter, but there were a few funny inconsistencies in the prosecution story, and it has been ordered to retrial. I have long believed that a biological parent is unlikely to kill his/her son or daughter. Kill a spouse, yes. But not a biological child.
In my checking, I found the Wiki entry under "Cinderella Effect." It turns out that I seem to be correct. The Wiki entry says that studies show that, on a person for person basis, a step parent is 100 times more likely to murder stepchild than a biological parent is to murder his/her biological child. This makes it less likely that the guy up for trial is guilty, since the girl murdered was his biological daughter.
The evolutionary reasons for the Cinderalla Effect seem clear. Do you guys think my logic in relation to the murder re-trial has merit?