The Development and Validation of the Self-Efficacy in Statistical Practices Scale

The purpose of the current study was to develop a scale intended to measure undergraduate students’ self-efficacy in statistical practices. In order to apply statistical concepts and methods that students learn in the classroom to real world situations, it is important for college students to have not only statistical knowledge and skills, but also self-efficacy in using those concepts and methods. Even though there is growing attention on the importance of assessing students’ statistics self-efficacy, currently available measures have numerous limitations. Therefore, the Self-Efficacy in Statistical Practices Scale (SESPS) was developed, and the internal consistency reliability and convergent validity of the scale were assessed. Additional analyses comparing differences in statistics self-efficacy based on students’ gender, major, year of study, and number of previous statistics courses were also conducted.

Data on undergraduate students’ statistics self-efficacy, statistics anxiety, and various background characteristics were collected from a sample of students from two US universities. The underlying structure of the SESPS was analyzed using principal components analysis. The resulting 2-component solution for the SESPS provides a reliable and valid measure to assess undergraduate students’ statistics self-efficacy. The components “Self-Efficacy in Conducting Statistical Procedures” and “Self-Efficacy in Utilizing Computer Software for Statistical Procedures” contain 37 and 6 items respectively. Students’ self-efficacy in utilizing computer software for statistical procedures was statistically significantly lower than their self-efficacy in conducting statistical procedures. In group comparisons, statistically significant differences were found in students’ self-efficacy in utilizing computer software for statistical procedures. Male students’ self-efficacy was higher than females’, STEM majoring students’ self-efficacy was higher than non-STEM students’, seniors’ self-efficacy was higher than first years’ and juniors’, and self-efficacy in utilizing computer software for statistical procedures was significantly higher for those students with more previous statistics courses. No significant group differences were found for self-efficacy in conducting statistical procedures.

SESPS appears to be a viable construct with useful implications for college students, faculty, and statistics education researchers. Additional study is needed to confirm the factor structure and to further validate the scale.