Impact of engineering students' attitudes on achievement in statistics: A structural equation model analysis

The purpose of this study was to examine impact of engineering students' attitudes on their achievement in statistics. Two hundred and sixty-four undergraduate engineering students from a large Southwestern university were included in the study.;A structural equation model based largely on Expectancy Value theory was developed and tested. A second nested model was also tested for model fit. The following constructs were included in both models. Previous Success was defined by self-reported prior academic achievement. Cognitive Competence was defined as the individual's self-evaluation of ability in relationship to statistics courses. Value was defined as the worth the participant placed upon statistics. Difficulty was defined as the individual's evaluation of the difficulty of statistics. Affect was defined as the participant's positive and negative feelings about statistics. Cognitive Competence, Value, Difficulty and Affect were measured using the Survey of Attitudes Toward Statistics (Schau, 1995). Achievement was measured by the participant's performance quizzes, a midterm and a final in their required introductory statistics course.;The first structural model (Saturated model) utilized Previous Success as an exogenous variable, followed by Difficulty, Cognitive Competence, Affect and Value. These five constructs were allowed to impact Achievement both directly and through indirect paths. This model fit the data well. However, several of the paths were statistically insignificant. Therefore a second model (Pruned Model) with these insignificant paths eliminated also was tested. In the Pruned Model, Previous Success and Difficulty were exogenous variables. Cognitive Competence, Affect and Value were endogenous and did not all impact Achievement directly. The Pruned model exhibited an equally good fit to the data.;In the Pruned model, Previous Success had a large total effect on Achievement. Difficulty, Cognitive Competence and Affect had medium total effects on Achievement. Value had no total effect on Achievement, due to the opposite value and equal magnitudes of its direct and indirect effects. The Pruned Model predicted 76% of the variance in Achievement. Results indicate that both prior achievement and attitudes toward statistics impact achievement in statistics courses.