This study approached the investigation of attitudes toward statistics from the perspective of Rosenberg and Hovland's (1960) hierarchical, multicomponent model. In this model, cognition, affect, and behavior are considered interrelated first-order-factors with attitude a single second-order factor. In the present study, affective, cognitive, and behavioral components of attitude were assessed. Two other assessment tools were also administered for comparison purposes (Wise's ATA and a semantic differential measure). Data were gathered from 2 classes (n = 47) attended by master's and doctoral students in education, social work, and speech communication. Results suggest the strongest association is between affective and cognitive components but the greatest temporal stability was found for the behavioral component. Consistency in ratings of affect and cognition was not predictive of behavior nor was locus of attitude formation.
The CAUSE Research Group is supported in part by a member initiative grant from the American Statistical Association’s Section on Statistics and Data Science Education