Social norms and climate-friendly behavior of adolescents

social preferences of children and adolescents [52–54] and to learn about their development (for an overview, see [55]). Approval for the study was obtained from the school principals and the state school author- ity. During school hours, students were invited to participate in a survey study on their atti- tudes toward climate protection. Participation in the survey was of course voluntary and the participation rate was 87%. As we have no information about the socio-economic characteris- tics of the non-participants, we cannot formally test whether a self-selection bias is present in our sample. Yet, given the relatively high participation rate, we consider potential self-selection effects to be limited. Students who did not want or who were not allowed by their parents to participate were supervised by their teachers in the meantime. Students who participated received monetary compensation, which provided all students the same opportunity to donate. We obtained written informed consent for participation and anonymized data use from all participants. For students from Grades 8 and 9, we also obtained the written consent of one parent. A total of 628 adolescents participated, ranging in age from 11 to 17 (Mean: 14.5 years, SD: 1.04, Median: 15), and 55% were female. In each school, the survey was conducted on one day during consecutive school lessons. All classes of the same grade, usually five classes with between 20 and 30 students each, met in the assembly hall during the designated lesson. After the students were welcomed, they received

instructions on the process via a standardized presentation and learned that they would be compensated for their participation with 8 €. This corresponds to the average amount adoles- cents of this age group receive as weekly pocket money [56]. Students found the pen and paper surveys as prepared packages on spaced out chairs. The survey consisted of three parts, and students were asked to answer it in the pre-determined order. In the first part, students were informed that they could donate between 0 and 8 € of their compensation for participation for CO2 offsetting. The instructions described the NGO (myclimate) and what activities their donation would finance. The description also showed how students could monitor the transfer of the donation to the organization (we provided the donation receipts on a web page, listing all individual donation amounts with the correspond- ing participation code). The monitoring information was also used to implement our treat- ment, as described in the next section. Once students had made their donation decision, they were asked to pack this part in an envelope so that they could not return to it. In the second part of the survey, students learned that they could earn one additional euro if they correctly guessed the average donation in their grade and school (belief elicitation 1). In the third part, students were asked to estimate how socially appropriate their schoolmates of the same grade rated each of the possible donation amounts on a four-point scale ranging

from very socially inappropriate to very socially appropriate (belief elicitation 2). Since the con- cept of social appropriateness was for the younger age groups not trivial to grasp, we provided the following additional explanation: “By socially appropriate behavior, we mean behavior that is considered morally acceptable by the majority of people. If the majority is upset about a par- ticular behavior of a person, then that behavior is considered socially inappropriate”. We tested the understanding of this concept during our pre-test. Following the incentivized norm elicitation of Krupka and Weber [57], students could earn again one additional euro if they correctly guessed the most common answer given by their fel- low students of the same grade. At the end of the study, we determined with a public lottery which donation amount from belief elicitation 2 was used to determine the payouts. The two belief elicitations constitute our dependent variables empirical and normative expectations. At the end of the third part, students filled out a post-experimental questionnaire with questions on (i) their individual characteristics, such as age, gender, time at school, friends, neighborhood of residence, and languages spoken at home, as well as scales measuring PLOS ONE Social norms and climate-friendly behavior of adolescents PLOS ONE | https://doi.org/10.1371/journal.pone.0266847 April 27, 2022 4 / 16