Fig. 4: Frequencies of graph-reading tutorial scores among participants who answered all four graph-reading questions, stratified by SES (n = 268) and education status (n = 269).

Wilcoxon rank-sum tests for unpaired samples (solid lines) were used to assess score differences by SES and education subgroups. Wilcoxon signed-rank tests for paired samples (dashed lines) were used to assess change in scores within each subgroup between tries. Summary data are available in Table S2.