The Ecological Cumulative Risk Model
This paper provides a theoretical and empirical introduction to the Ecological Cumulative Risk Model, an alternative to traditional additive models of cumulative risk (CR). The model is based upon Bronfenbrenner‟s Ecological Systems Theory (Bronfenbrenner, 1979) which posits that development occurs across a number of settings, each with varying proximity to the child. The model is intended as a compromise between additive and multiplicative measurement models of risk. Using the NICHD Study of Early Child Care and Youth Development I categorize a number of risk factors into one of six settings (i.e. demographic, parenting, neighborhood). Factor analysis is used to validate the underlying structure of these groupings. All risk factors were determined from 3rd grade measures and prior while all outcome variables (academic skills, externalizing behaviors, internalizing behaviors, and social skills) were measured at 4th grade. The predictive power of these settings/domains was contrasted against the predictive power of a traditional cumulative risk model. Thus, a total CR score within each setting was calculated as well as an overall CR score. Results indicate that the Ecological CR Model explains approximately 1% more variance across dependent variables compared to the traditional/overall approach. An advantage of dividing risk factors into domains is the ability to model interaction effects, even when using a cumulative risk measurement model. Of the thirty interaction effects that were tested, only two were statistically significant. Finally, structural equation modeling was used to validate the Ecological Domains Model. SEM analyses confirmed that the Ecological Cumulative Risk Model fit the data better than a lump sum approach. Furthermore, evidence of mediation through risk domains is provided; parenting risk partially mediates the effects of demographic risk on all outcome variables. I conclude that the Ecological Cumulative Risk Model is valuable for examining the processes through which risk operates (i.e. proximal domains mediate the impact of more distal risk domains). On the other hand, lack of interaction effects suggests that an additive approach is more viable than a multiplicative model. More research, particularly with a higher risk sample, is needed to further understand the utility of this measurement model.
cumulative risk; ecological domains; multiple risk
Evans, Gary William
Dunifon, Rachel E.; Casasola, Marianella
Ph.D. of Developmental Psychology
Doctor of Philosophy
dissertation or thesis