MEMORIES OF CHILDHOOD NEGLECT AND ABUSE - CORROBORATION IN A SERIES OF SISTERS

Reports from 87 community-based sister-pairs, selected for high rates of neglect or abuse in childhood, have been used to establish validity of the CECA, a retrospective interview measure of childhood experienc e. Corroboration was based on independent assessments of sisters' acco unts of what happened to each other in childhood. Corroboration of sca les assessing parental neglect, physical abuse in the household and se xual abuse (either household or nonhousehold) was satisfactory, with a mean correlation for the three experiences of .60 (weighted kappa [Kw ]). Concordance reflected the degree to which experience was shared an d was judged by comparing the sisters' accounts of their own experienc e. Among sisters with shared (concordant) experience for neglect or ab use, corroboration was high (mean of .74), but for those with nonshare d (nonconcordant) experience it was largely absent (mean of .01). The degree to which the experiences of neglect or abuse were concordant wa s related to whether the perpetrator was a member of the household. Ne glect and physical abuse were by definition from household members (ma inly parents) and involved high concordance of experience. Sexual abus e occurred from many sources and in practice was commonly from an adul t living outside the household and such experiences were less likely t o be shared by sisters. An overall index indicating the presence of at least one abusive experience before the age of 17 showed a concordanc e of .64 with corroboration between sisters as high as .70. Issues inv olving retrospective recall and measurement are discussed.