CHILDRENS CLOSENESS TO FATHER AS REPORTED BY MOTHERS, SONS AND DAUGHTERS - EVALUATING SUBJECTIVE ASSESSMENTS WITH THE RASCH MODEL

Mothers' and children's reports of closeness to father are examined us ing an approach for reconciling discrepant survey responses. This appr oach is based on a model for the measurement of latent traits. When bo th mothers and children are asked to report on the closeness of the ch ild to his or her father, tabulations of survey responses show substan tial disagreement in the reported evaluation. We adopt a model that ex plains this disagreement in terms of(a) constant differences between m others and children, and (b) variability within the sampling populatio n in the actual closeness of father to child. Closeness appears to hav e an entirely different meaning when fathers are resident than when th ey are nonresident. But the meaning does not seem to vary (even if fee lings do) when the resident father is a stepfather OF whether the foca l child is a son or a daughter.