Kathleen Moritz Rudasill and
Eric S. Buhs
Date of this Version
Frohn, S. F. (2017). An evaluation and revision of the Children's Behavior Questionnaire effortful control scales (Unpublished doctoral dissertation). University of Nebraska-Lincoln.
The Children’s Behavior Questionnaire (CBQ; Rothbart, Ahadi, Hershey, & Fisher, 2001) is a popular parent report measure of children’s temperament. Effortful control, which refers to processes involved in regulating reactivity to internal and external stimuli, is one factor of temperament measured by the CBQ using five scales tapping multiple dimensions. Numerous studies examining the psychometric properties of the CBQ have shown some problems with the scales, including inconsistent factor structures and measurement noninvariance. Furthermore, the way effortful control is typically defined in the literature, and even according to the CBQ’s authors, is inconsistent with how it is actually measured with the instrument. An additional concern which has not yet been addressed is whether the dimensions of effortful control are measured consistently across the ages for which the CBQ is purported to be useful (i.e., are measurement invariant). In Study 1, I evaluate the psychometric properties of the CBQ Effortful Control scales, with particular attention to age-group measurement invariance. Results provide indication that many CBQ items demonstrate various degrees of noninvariance, and provides strong evidence the CBQ Effortful Control scales need revision. In Study 2, with the help of temperament and effortful control content experts, I piloted a large item pool with a sample of parents using an online questionnaire. Using a subsample of pilot participants (N = 400) I refined the scales based on a series of decision rules based in Classical Test Theory (CTT) and Confirmatory Factor Analysis (CFA). The scales were further refined on theoretical grounds with the help of content experts. The psychometric properties of the revised and original scales were then compared in CTT and CFA frameworks with the remaining pilot sample (N = 272). The revised scales show considerable improvement over the original scales in content, length, and factor structure, and are comparable to the original versions in terms of CTT reliability. Implications, limitations, and future directions are discussed.