Susan M. Sheridan
Date of this Version
White, A. S. (2019). Using self-regulation to predict preschoolers' symptomology of disruptive behavior disorders (Unpublished doctoral dissertation). University of Nebraska-Lincoln, Nebraska.
The use of brief rating forms completed by caregivers to identify children at-risk for developing behavioral disorders is common (Lane et al., 2009). However, identifying a behavioral measure assessing child-level variables (i.e., temperamental traits) which predict later behavioral concerns has potential to improve universal screening practices in the context of a multi-tiered systems of support (MTSS) framework. Self-regulation (Rothbart & Bates, 2006) is a trait that is related to externalizing problem behaviors (e.g., Espy et al., 2011), and may be useful as a means to predict young children at risk for developing behavioral disorders. The purpose of this study is to explore the predictive validity of an established measure of self-regulation (the Head-Toes-Knees-Shoulders Task; HTKS; McClelland & Cameron, 2012), for clinically elevated externalizing behaviors (identified using clinical rating forms of externalizing behavior). It was hypothesized that assessing a stable, individual trait such as self-regulation could allow for even earlier identification and intervention among at-risk children than may be available with present screening methods. Participants were 24 preschool students and their classroom teachers. The students were administered the HTKS in their schools and their teachers each completed a rating form assessing behavioral problems across three measures (i.e., the Social Skills Improvement System, Achenbach Caregiver-Teacher Report Form, and Conners Early Childhood Behavior Scale). Surprisingly, this study did not replicate the relationships between self-regulation and behavioral concerns. Correlations between variables suggested positive relationships between the HTKS and two of the behavioral measures (i.e., opposite of the hypothesized direction). Multiple linear regression analyses exploring the relationship between continuous criterion and predictor variables were unable to reject the null hypothesis that HTKS does not predict behavioral concerns. Further, logistic regression analyses exploring a dichotomous criterion (i.e., the presence or absence of clinically-elevated behavioral problems) also failed to reject the null hypothesis of the model discriminating behavior problem status no better than chance. Follow-up Receiver Operating Characteristics (ROC) curves and comparison of the area-under-the-ROC-curve (AUC) further suggested HTKS was not an effective tool for screening in this context. Finally, the study explores its limitations and proposes additional questions for future research.
Advisor: Susan M. Sheridan