The equal-variance assumption requires response values to show similar variability across every group of a categorical predictor. In one-way ANOVA, diagnostic plots help assess whether this condition is reasonable. section 14.5.2
NBA wins by scoring category
Fit an ANOVA model with `aov(z_wins ~ o_pts_cat, data = nba_stats)`, then inspect Residuals vs. Fitted. Equal variance would produce similarly dispersed residuals at each fitted-value level. Here, residuals are less dispersed at the highest fitted values, suggesting a likely—but not statistically significant—violation. section 14.5.2
