Skip to main content
. 2021 May 27;11:11104. doi: 10.1038/s41598-021-90656-y

Figure 2.

Figure 2

(a) Experimental task. In both the Social influence and the Indirect reciprocity task, initially participants had to produce perceptual estimates by reproducing the length of visual stimuli. They saw two consecutive light flashes (i.e., red disks) and had to touch a point, to the right of the second disk, in order to reproduce the stimulus length (s), defined as the distance between the first and the second disk. Participants were told that the very same stimuli would be presented to an alleged partner (other agent: OA) who would choose a point to reproduce the same distance. After the participants’ estimate, the other agent’s simulated estimate was shown, and then one of the two agents had the opportunity to make a final decision by choosing any position between their previous response and the other agent’s response. We computed participants’ influence (i) as the adjustment towards the partner in the final decision (a) divided by the distance between the two agents’ responses (d). (b) Experimental design. The experimental paradigm consisted of two main tasks: Social influence and Indirect reciprocity. Both tasks were based on the perceptual inference task described above, but differed in the type of interaction with the partner(s) and the identity of the agent taking the final decision. In the Social influence task, participants interacted with an alleged partner (passive agent: PA) and had to personally make final decisions (i.e. revising their estimates in light of the passive agent’s one) in all trials. In the Indirect reciprocity task, in half of the trials (decision turns) participants performed the same task with the passive agent and always made the final decision. Decision turns were alternated with observation turns, in which participants interacted with another alleged agent (active agent: AA). In these trials, the two agents made their perceptual estimates, and then the active agent took the final decision, while participants could not revise their estimate. We manipulated the active agent’s final decisions to express two different levels of susceptibility towards participants’ responses (Susceptible and Unsusceptible conditions).