Mother Goose and Mother Nature: Designing Stories to Communicate Information About Climate Change

Abstract

Narratives have the ability to highlight climate change information in a relatable and engaging format. The purpose of the present investigation was to test the effects of five story structure types containing climate change information on perceived message effectiveness. Furthermore, we measured whether political affiliation moderated persuasive effectiveness of climate change messages among U.S. eligible voting adults (N = 594) who identified as Republican, Democrat, and Independent. The highest rated story overall was situated in the past, was realistic, and had clear moral values, indicating that certain appealing stories can serve as effective conduits for persuasive messages across the political spectrum. There were few differences between political affiliations, namely, that Democrats and Republicans rated messages differently on effectiveness when they differed on moral themes.

Keywords:

• Narratives
• Narrative Persuasion
• Climate Change
• Message Design

Acknowledgments

The authors acknowledge the USC Stevens Center for Innovation and the Diploma in Innovation program at the University of Southern California for sponsoring the research reported in this manuscript and Christina Lee and Graeme Bloomfield for their assistance with the creation of the study materials.

Notes

1. Demographics of the full sample are available upon request.

2. There were slightly more Black/African American and Asian/Asian American participants who affiliated as Democrat. In the interest of space, the comparisons between political groups are not reported here but are available upon request.

3. The participants were randomly assigned to each of the five conditions. However, there were notably fewer participants in Story Type III condition; as a result, the panel provider then recruited 49 additional participants to evaluate the Story Type III to reach an approximately equal number of participants in each condition and to allow for the ability to test for the moderating effect of political affiliation. A series of difference tests were conducted on the 119 participants of Story Type III to explore differences between those who had been randomly assigned (n = 70) to the condition versus those who were purposefully assigned (n = 49). There were no significant differences between the randomly and purposefully assigned to Story III on the key sample characteristics of sex (t = .19, p = .84), racial identification (F = 3.97, p = .05), education level (F < .01, p = .98), religious affiliation (F = .61, p = .43), political party identification (F = 1.04, p = .31), liberal-conservative leaning (t = −.80, p = .42), and political preferences (t = −.55, p = .58). There was also no significant difference between the two groups on the dependent variable of perceived effectiveness, t = .45, p = .65. These participants were also compared to all randomly assigned participants in all conditions, with no differences on the key demographic variables as well. As a result, these participants were retained to include more equivalent numbers of raters between conditions.

4. A pilot study recruited 108 undergraduate students to rate the stories. There were no significant differences between the stories on overall engagement and enjoyment. This gave us greater confidence that any story differences in this main study would be due to differences in story structure, rather than differences in story quality. Full results of the pilot study are excluded from the present manuscript in the interest of space but are available upon request.

5. The 12-item engagement measure (adapted from Busselle & Bilandzic’s [2009] narrative engagement scale) was included in data collection to allow for comparisons to the pilot test if necessary. The results of those manipulation check analyses were approximately similar to the initial pilot. Because the chief aim in this analysis is to examine the main effects of the story archetypes on the evaluation of those archetypes rather than the nuances of the message processing, the engagement results are excluded from this manuscript and are available upon request but are noted here in the interest of transparency regarding the study procedure.

Additional information

Funding

This work was supported by the USC Stevens Center for Innovation Diploma in Innovation program at the University of Southern California.