New Statistics,Old Schools: An Overview of Current Introductory Undergraduate and Graduate Statistics Pedagogy Practices

Abstract

Background

A philosophical shift in statistics regarding emphasis on “New Statistics” (NS; Cumming, G. (2014). The new statistics: Why and how. Psychological Science, 25(1), 7-29.) over conventional null hypothesis significance testing (NHST) raises the question of appropriate material coverage in undergraduate statistics courses.

Objective

We examined current practices in statistics pedagogy at the graduate and undergraduate levels for both NS and NHST.

Method

Using an online survey of a nationwide sample of current graduate students (n = 452) and graduate faculty (n = 162), we examined statistics pedagogy and perceptions of best approaches for teaching undergraduate statistics.

Results

In undergraduate statistics courses, coverage of NS material involves modest instruction in effect sizes and confidence intervals, while NHST remains dominant. Graduate courses have more balanced coverage. Effect size estimation was regarded as the most important NS knowledge for success in graduate school and the topic most in need of increased undergraduate coverage.

Conclusion

Undergraduate statistics courses could increase NS coverage, particularly effect size estimation, to better align with and prepare students for graduate work.

Teaching Implications

This research summarizes graduate program expectations and graduate student experiences regarding undergraduate statistics that current instructors can use to shape the content of their classes.