Skip to main content
. 2026 Sep 7;14:1896650. doi: 10.3389/fpubh.2026.1896650

Figure 3.

Scatter plot with error bars showing random intercepts for 2,591 class-year units, ranked along the x-axis. Blue dots represent intervals entirely below zero, red above zero, and gray include zero. Points are grouped and color-coded by interval type as detailed in the legend.

Caterpillar plot of class-year random intercepts (Model 2, empirical-Bayes estimates with approximate 95% intervals). Intervals entirely above zero (n = 529) and entirely below zero (n = 565) are distinguished from intervals that include zero (n = 1,497). These intervals are descriptive summaries of the estimated random effects and are not multiple-comparison-corrected significance tests.