ANOVA (analysis of variance) is a common statistical method that is often included in the undergraduate curriculum of science majors. This activity provides an interactive and an engaging way for students to apply both one-way and two-way ANOVA. The activity is laid out as an instructor-guided analysis of a problem from the scientific literature, in order to help familiarize students with these methods, before turning over the analysis to students, who work to produce solutions to the problem, and ultimately write a report on their findings.
We have included three different iterations of this activity: 1-Day JMP Activity, 2-Day JMP Activity, 1-Day R Activity. The two 1-Day versions have a partial problem-based learning (PBL) structure and only take one lab period to complete. The 2-Day version has a full PBL structure and takes two lab periods to complete and is built around JMP.
The “problem” in this activity (Figure 1) is an interaction effect that prevents the interpretation of a two-way ANOVA, bringing some of the conclusions of De Kort et al. (2020a) into question. Developing solutions to this problem requires students to apply statistical concepts, perform data manipulation on the published dataset (De Kort et al., 2020b), write code in the 1-Day R version, and conduct additional statistical procedures (one-way ANOVA, two-way ANOVA, testing assumptions, post-hoc tests).
Figure 1: Illustration of the problem in De Kort et al. 2020a.
The objectives of this activity are as follows:
Writing hypotheses for ANOVA
Assessing data for assumptions of ANOVA
Conducting one-way and two-way ANOVA with Post-Hoc tests
Using statistical methods to critique the primary literature
Developing coding skills in the R language (1-Day R Activity only)
Developing scientific writing skills that incorporate statistical methods
This is an engaging and thought-provoking assignment relative to traditional application-based activities. Students gain valuable experience collaborating with peers to develop solutions to a real problem while reinforcing statistical concepts. Due to the nature of the problem being assessed, this activity also emphasizes that statistics are an integrated part of experimental design, peer review, and the process of science, not just a troublesome step to test hypotheses after an experiment concludes. Importantly, this activity is amenable to multiple learning modalities (e.g., in-person, online).
The number of columns in the original dataset (De Kort et al., 2020b) was reduced for these activities.
The linear models created within JMP and RStudio are different from one another, yet the same conclusions should be reached about the data.
Cite this work
Researchers should cite this work as follows:
Hopkins, N., Felker, H., Dickerson, C., O'Brien, T., Whittinghill, D., Ruhl, N. (2021). Climate change and Strawberries: A real-life problem-based activity for teaching ANOVA. QUBES Educational Resources. doi:10.25334/VSMB-7667