Because it can be so difficult for a court to determine the reality behind the words it is hearing, it is tempting to turn to the assistance of an "expert" who will provide "reliable" insight into the situation confronting the bench about family dynamics. In fact, such resources carry their own risk of prejudice. An evaluation is only as good as the evaluator, and some view cases through a particular lens that may not serve the families well. Despite the severely negative impact of domestic violence on children, many psychologists, guardians ad litem and other custody evaluators ignore evidence of domestic violence and intimate partner sexual abuse or insist it does no harm to youngsters.

In a major study of custody awards to known batterers conducted in Massachusetts, researchers asked judges why they had awarded custody to abusers in particular cases. Many judges answered that domestic violence was not mentioned in the reports of the guardians ad litem. This was repeatedly confirmed by mothers who told researchers that despite providing the guardians ad litem with full information about the domestic violence and asking them to call the district attorney, police, and others who could confirm their reports, the guardians ad litem refused to do so.

One mother, disabled as a result of the violence, reported that the guardian ad litem told her "no one cares about that abuse crap." Another guardian wrote that although the mother's claims of abuse—including rape by her husband and repeated ransacking of her home—appeared valid, "there is no credible evidence that the children have been victimized by, or witnessed incidents of violence between, their father and mother" (Wellesley Centers for Women, Battered Mothers Speak Out: A Human Rights Report on Child Custody and Domestic Violence in the Massachusetts Family Courts (PDF 1.97MB), 2002).

Those who believe that children are not affected when a parent is traumatized by sexual assault—even if the children do not directly witness the incident—should not be conducting custody evaluations (Schafran, Evaluating the Evaluators: Problems with 'Outside Neutrals' (PDF 1.33MB), 2003 at 13).