Learners' motivational characteristics in statistics: A causal model.

Abstract

A total of 263 participants enrolled in three introductory statistics courses completed a two-part instrument measuring the variables of interest prior to their midterm exam. In order to assess the validity of the causal model, path analysis procedures outlined by Pedhazur (1982) were followed. Results of path analysis indicated the data fit the overidentified model well. A subsequent path analysis using a trimmed model also fit the data well. Results found that deep processing strategy use, self-efficacy, learning goals and prior experience have direct effects on achievement, and future career consequences, future graduate school consequences, and effort have indirect effects on achievement. Self-efficacy, by far, played the biggest role, directly and indirectly, in accounting for variance in many key variables related to achievement and achievement itself. Findings related to future consequences, a variable rarely investigated in statistics, provided support for theory and warrants further investigation of the role this variable plays in motivation. Suggestions for future research and the implications of these findings for teaching statistics are discussed.

This study explored learner characteristics related to motivation and cognition and their influences on cognitive engagement and achievement in statistics. Few previous studies have investigated the role of multiple variables, such as prior experience, self-efficacy, and goals in statistics, to examine how they influence statistics achievement in the context of one another. An examination of these variables together provides a better picture of the key influences motivational and cognitive engagement variables have on achievement in statistics. The present study examined the variables of (a) prior experience, (b) self-efficacy, (c) future consequences, (d) learning and performance goal orientations, (e) effort, and (f) deep and shallow processing strategy use in the contest of one another in the domain of statistics to test the proposed theoretical causal model for achievement in statistics.