Skip to main content
. 2020 Nov 23;31(12):1585–1601. doi: 10.1177/0956797620957625

Fig. 6.

Fig. 6.

Interaction of age group, direction of influence (prosocial vs. selfish), and influence (influenced vs. not influenced) on response times (RTs). In the top graph, age is shown as a categorical variable. Colored squares show the fixed-effects estimates of RTs from the trial-level linear mixed model. Error bars are the corresponding 95% confidence intervals. Symbols indicate significant and marginally significant differences between and within groups (†p < .10, *p < .05, ***p < .001, Bonferroni corrected). For statistics of all contrasts, see Tables S5a and S5b in the Supplemental Material available online. In the bottom graph, age is shown as a continuous variable. Circles are grand medians of RTs. Circle area is proportional to the number of participants; the key shows three examples for reference. The lines show the overall polynomial trends of age as estimated by a trial-level generalized linear mixed model on the log of RTs (then back-transformed to the response scale). The shaded area is the 95% confidence interval. Asterisks indicate significant components of the trends (dashed lines; *p < .05, ***p < .001). For statistics of all trend contrasts, see Table S6 in the Supplemental Material.