Methods This longitudinal study uses three waves of data from the Pre-Elementary Education Longitudinal Study (PEELS) secondary data set, collected from more than 3,000 CWD nationwide by the U.S. Department of Education National Center for Special Education Research (NCSER). It surveyed the characteristics of children receiving preschool special education, services received, transitions across educational levels, peer victimization, and performance over time on academic and adaptive skills assessments. The PEELS study includes four waves of data collection—school years 2003-04, 2004-05, 2005-06, and 2008-2009. In addition, the study includes a longitudinal parent/assessment/teacher sample for waves 1-4 (N=1,126). The present study analyzed three waves of data from the PEELS Longitudinal Study data. To account for the complex sampling used in the dataset, statistical analyses were conducted using Stata 10 and included Pearson correlations, a path analysis, and multi-group analyses with AMOS 17.0.
Results The path model showed an acceptable fit to the data: GFI=.99, CFI=.97, RMSEA=.04. Two pathways explained the influence of risk and protective factors for peer victimization among young CWD. First, children's environmental factors, such as low family income and spending more time in a special-education classroom setting, were associated with children's poor social behaviors, which in turn affected peer-relation difficulties, and increased peer victimization. Second, CWD from low-income families and special-education classroom settings were more likely to have poor language development and social skills, which affected children's peer-relation difficulties and peer victimization. To test whether gender moderated the relationships in the model, I compared a multigroup model (Model 1), in which all parameters were constrained to have equal values for both boys and girls, to an unconstrained multigroup model (Model 2), in which all parameters were free to vary between groups. Model 2 (the unconstrained model) provided a better fit to the data, ÷2(60) = 163.24, CFI = .95; IFI = .95; RMSEA = .04, and was in fact a substantial improvement over Model 1 according to a chi-square difference test, ÷2(47) = 140.57, p = .000. The paths from receptive-language ability to peer victimization and from developmental delay to internalizing problem behaviors were not significant for girls.
Implications Practical implications include developing programs tailored for CWD from low-income families and special-education classroom settings, providing mental health services for pre-elementary CWD, linking parents to available school and community resources to improve children's language and social skills, and promoting polices to enhance social conditions for CWD.
