Separating the harmful versus beneficial effects of marital disruptions on children

Abstract

Although a marital disruption can certainly be harmful for some children, it might be beneficial to other children. Analyses on how children are affected by marital disruptions typically capture the average estimated effects (or associations) of a disruption on an outcome. Thus, the harmful effects of the disruption on some children are being averaged with the neutral and beneficial effects on other children. This could mute the estimated effect, and it could prevent the detection of significant harmful (or beneficial) effects. Using achievement test scores and an index of behavioral problems in a first-difference framework, I find evidence for the standard approach having muted estimated effects and failing to detect significant effects when the same data produce significant isolated harmful effects.