Assessment of Child Neglect in Community & Agency Samples
Summary. This project builds on our prior work funded by an NIH study which addressed two interrelated and crucial aspects of research that have received inadequate attention in studies of neglect: measure development and estimates of prevalence. The first major goal of the study was to develop child self-report versions of a Multidimensional Neglectful Behavior Scale (MNBS-CR). We developed an instrument which measures four primary dimensions of neglect: Physical, Emotional, Supervisory and Cognitive. A multi-media Audio Computer Assisted Self Administered Interview (ACASI) computer program was developed to test the child report version of the scale, programmed to reflect the age and gender of the child, and the gender of the primary caretaker.
Goals and Objectives
The central aims of this continuation study are to continue and expand on the original study goals on measurement development:
- Test, and revise two versions of a standardized instrument to measure neglect—a parent self-report to be administered to a community sample of families, and a child self-report measure which is being administered to a clinical sample of children, ages 6-15;
- Establish the reliability, and validity of the instruments: -- Preliminary results of work completed on the computerized child self-report measure of neglect are promising. A goal of this phase is to expand the pool of children with multiple forms of identified neglect; establish reliability and validity on a more diverse population, and test the measure on non-clinical populations to increase generalizability;
- Improve the utility of the instrument for clinical usage by programming the computer version of the child self-report instrument to provide the tester with scores at the end of each child’s assessment. Empirical data will be used to establish meaningful scores for neglect domains, total neglect scores, and child depression that can provide a basis for intervention decisions;
- Add a social desirability response bias scale to the MNBS-CR.
Data collection has been ongoing since January 2002. The clinical sample currently consists of 268 children, most from the Spurwink Child Abuse Program (CAP), a forensic child abuse program in Portland Maine, tested with the MNBS-Child Report (MNS-CR) and the Peabody Picture Vocabulary Test (PPVT III). Fifty percent of the children are in the younger age group, 52% are female, and 7% are from minority groups. 60% of the clinical sample was still living with their biological parent at the time of testing. Analysis of record data on parents of children in the clinical sample revealed the following: Developmental disabilities were present in 20% of mothers; Mental illness (mainly depressive disorders) diagnosed in 49% of parents; 43% substance abuse problems; and 72% had histories of domestic violence. Rates for identified parental problems are higher than those previously found for this study, as more children from Maine DHS, the primary site, have entered the study.
We have tested approximately 100 children from NH community areas. The current community sample now consists of 170 children from NH school programs where parents gave active consent for the children to participate in the research. For the total community sample, 71% of the children are in the younger age group, 55% are female, and 20% are from minority groups.
Recruitment for both the community and clinical samples is ongoing.
Social Desirability Scale. Preliminary results of the Social Desirability Measure show high rates of social desirability responding, particularly among young children in the clinical sample. Preliminary results also show that children with high social desirability scores report lower levels of neglect.
Reliability Analysis for Neglect Domain Scale Scores. Separate analyses were done for the two age versions of the MNBS-CR. Among older children in the clinical sample, the full version of the MNBS-CR has high internal consistency reliability (alpha of .95). For the separate domains of Neglect, the alpha coefficient ranged from .85 to .73. Overall, analyses of younger children’s reliability scores were similar but generally lower (alphas of .75 to .30).
Factor Analysis. Two factor analyses were previously performed on the MNBS-CR for the clinical sample only. Of note is the finding that for both age groups items did not tend to load on factors according to the conceptually distinct areas of neglect reflected in the four scales that were included. This might argue for the use of only the Total Neglect Scale rather than scales based on more specific manifestations of neglect. Two factors accounted for the largest proportion of the variance in older (37%) and younger children (27%). A primary factor of severe neglect included items reflecting parental substance abuse, abandonment, and severe physical neglect in older and younger children. A secondary factor included items of parental conflict and violence along with emotional and cognitive neglect for older children, and an array of severe neglect for younger children. The results are important because they support the utility of the MNBS as a screening tool for child neglect, and as a basis for intervention.
Correlations with Child Outcomes. For all children, depression remains significantly correlated with total neglect, and other domains. Neglected children in the younger group scored significantly lower on the PPVT than children without neglect concerns in the family and PPVT standard scores were related to physical neglect) among 6-9 year olds. 10-15 year olds with neglect histories scored slightly lower on the PPVT than those individuals without neglect concerns. Clinician-rated variables also correlated with children’s self-reported neglect, including parental history of alcohol and substance abuse.
Comparisons of Neglect Scale Scores between Clinical and Community Samples of Children. Children in the clinical sample tended to score higher on neglect scales than members of the community sample. Among 10-15 year olds, children in the clinical sample experienced significantly more alcohol exposure than community children; furthermore, children in the clinical sample scored higher on all scales relative to the community sample. When data were examined for the clinical subsample of children with identified neglect and compared to reports from the community sample, the results supported the discriminant validity of validity of the measure. Significant differences were identified between groups for all domains, among both younger and older children.