Objective: The aim was to examine whether design features of Wave 1, 1
980 National Incidence Study (NIS) data resulted in sample selection b
ias when certain victims of maltreatment were excluded. Method: Logist
ic regression models for the probability of child abuse reports to the
child protective services (CPS) were estimated using maximum likeliho
od methods for Black (n = 511) and White (n = 2499) child abuse cases.
The models were estimated with and without correction for selection b
ias using a two-step procedure proposed by Heckman. Results: Substanti
al differences were found in the characteristics of Black and White vi
ctims by source of report and by type of maltreatment. Also found were
sizeable differences within each racial group between sampled agencie
s and nonsampled agencies. Sample selection bias affected the estimati
on of both White and Black child abuse reporting rates. In the Black s
ample, however, the effect of sample selection bias was to reduce the
statistical significance of the impacts of reporting agency and physic
al and sexual abuse on report rates. In the White sample, most signifi
cant factors in the basic model remained statistically significant wit
h correction for selection bias. Conclusions: Selection bias was found
to be caused by the exclusion of family, friends, and neighbors in th
e NIS sample design. Such exclusion has the effect of altering the int
erpretation of the determinants of child abuse reporting among Blacks,
but not among Whites. Thus, conclusions about racial differences in c
hild maltreatment must be reached cautiously, given the NIS study desi
gn. (C) 1998 Elsevier Science Ltd.