The impact of place and legacy framing on climate action: A lifespan approach

connected they felt to their future selves [33]. A single-item measure of internal locus of con- trol asked participants to rate how much control they had over their fate from (low internal control)1-10 (high internal control) [34]. A measure of mortality salience, the Death Thought Accessibility Word Completion Task, was used to explore whether either induction was evok- ing thoughts of death and whether this was related to observed outcomes [35]. Finally, demo- graphics questions assessing age, gender, education, political affiliation, ethnicity, and whether participants identified as environmentalists were included. Methodological details including full versions of all scales, measures, and attention checks used in this study are provided in S1 File. Procedure Participants who met sampling requirements viewed an mTurk advertisement for a 20–35 minute survey and writing task. Once participants provided informed consent and opted to participate in the study, they were randomly exposed to one of three different 5–7 minute writ- ing tasks, representing three induction conditions—Place (n = 299), Legacy (n = 295), and Control (n = 394). Writing task one, the Place-based induction (Place), asked each participant to write about a place that he or she had an emotional connection to, and to imagine how extreme weather events or climate change might impact that place and change it in the future. This induction was adapted from Smyer (2017) [27]. Writing task two, the Legacy induction (Legacy), was

adapted from Zaval et al. (2015) [26], and asked participants to write about what they wanted future generations to remember them for, and in what ways they will have a positive impact on the environment or other people after they are gone. Writing task three, the control induction, asked participants to write detailed instructions on a topic unrelated to place, climate change or legacy: they were asked to write a description of how to make scrambled eggs. Complete induction prompts are provided in S2 File. Following exposure to their randomly assigned writing task, all participants completed demographic questions and our full battery of self-report measures and scales. After complet- ing all measures and questions, participants entered a code and were digitally thanked, paid, and debriefed. Once a sufficient number of participants had completed the online experiment, data collection was closed, and collected data were de-identified and downloaded for analysis. Our analytical approach was to conduct a series of 3 (Induction) x 3 (Age) factorial ANO- VAs. The independent variable ‘Induction’ had 3 levels: Legacy, Place, and Control. The inde- pendent variable ‘Age’ also had 3 levels: young adults 18–35, middle-aged adults 36–52, and older adults 53–87. 3x3 Factorial ANOVAs were run for each dependent variable: one for mean pro-environmental intention score, and one for the amount participants volunteered to donate to planting trees. We then conducted mediation analyses using two measures we had

included as manipulation checks, legacy motivation and biophilia attitudes, to assess whether they mediated the relationships between our inductions and our dependent variables. Finally, several exploratory 3x3 Factorial ANOVAs were run using the rest of the psychographic scales included in our testing battery, to identify variable relationships underlying observed effects. Results Statistical assumptions To ensure statistical assumptions necessary for a series of 3 (Induction) x 3 (Age) factorial ANOVAs had been met, data were screened for outliers, homogeneity of variance, and univar- iate normality. Histograms were visually inspected for outliers, with none identified. In all cases, Levene’s test was satisfied or standard deviation ratios did not exceed 2:1, suggesting Impact of framing on climate action PLOS ONE | https://doi.org/10.1371/journal.pone.0228963 February 25, 2020 5 / 17