CASE-DRIVEN REFORM OF APPLIED STATISTICS TEACHING THROUGH QUANTILE REGRESSION OF FOREST TREE GROWTH

In graduate applied-statistics courses, knowing how to run a regression does not necessarily prepare students to choose a model for a forestry problem or explain what its estimates mean. We describe a case-based teaching plan built around the height–diameter relationship in a mixed larch–birch stand. The source project reports measurements for 677 trees in four plots, including diameter at breast height, total height, live-crown height, and crown dimensions. The proposed course asks students to examine data quality, fit ordinary least squares (OLS) and quantile regressions, and compare findings across species and plots with different management histories. By working through 19 conditional quantiles (0.05–0.95), students can see where an average relationship leaves important variation unexplained. The teaching sequence includes preparatory exercises, collaborative analyses, and a management-oriented briefing, supported by an assessment rubric. We distinguish results mentioned in the project materials from estimates that students would need to verify with the original observations. In particular, differences between thinned and untreated plots cannot, by themselves, establish an effect of thinning. This article sets out a teaching design rather than reporting an intervention trial; classroom learning outcomes remain to be evaluated. The case illustrates how statistical computation can be taught alongside ecological interpretation, reproducibility, and caution in drawing conclusions.