Fig. 3: Environmental effects on deciding to help others are consistent across studies.

Results from study 2 (n = 219) and study 3 (n = 54). For all three studies, there was a significant interaction between recipient, environment, and expected value. a, b (studies 2 and 3) At higher expected values, participants were more likely to act to help others in poor foraging environments compared to rich ones (study 2: three-way interaction OR = 2.55 [1.67, 4.01], z = 4.21, p < 0.001; study 3: three-way interaction OR = 1.92 [1.32, 2.80], z = 3.40, p < 0.001). When the decision to act benefited oneself, the environment had less of an impact. The shaded bands around the lines represent 95% confidence intervals of the mean. c, d (studies 2 and 3) On average, participants chose to act to earn rewards more for themselves than for others (study 2: self vs other: OR = 0.10 [0.07, 0.13], z = 13.52, p < 0.001; study 3: self vs other: OR = 0.20 [0.13, 0.31], z = 7.72, p < 0.001). e, f (studies 2 and 3) Participants were more likely to choose to act for a given opportunity when it appeared in a poor environment as compared to a rich one (study 2: poor vs rich: OR = 0.29 [0.22, 0.37], z = 9.59, p < 0.001; study 3: poor vs rich: OR = 0.24 [0.19, 0.30], z = 12.56, p < 0.001). Plotted here are the probabilities of accepting opportunities with the same mean expected value across both environments. Each dot represents a participant’s average for that condition. Error bars represent the standard error of the mean; ^***p < 0.001; all tests were two-sided Wald Z-tests.