The Five Canadas of Climate Change: Using audience segmentation to inform communication on climate policy

that the Canadian population can be divided into five distinct segments, offering potential tar- gets for climate change communication: the Alarmed (25%), Concerned (45%), Disengaged (5%), Doubtful (17%) and Dismissive (8%). These audiences reacted to emphasis framing in very different—and sometimes unintended—ways. In particular, “Invest in solutions” and “Relative price” messages had a negative impact on attitudes toward carbon taxes among more engaged audiences (i.e., the Alarmed) when policy stringency was low, and a positive impact on attitudes among the Concerned when policy stringency was higher. The “Relative price” was the only message positively affecting both the Concerned and the Doubtful, suggesting that communicating around the specific consequences of carbon taxes for the prices of essen- tial goods (like energy) may be a fruitful way to broaden support for carbon pricing among moderately engaged audiences. This may particularly be important as the costs of climate pol- icy rise as well as in contexts where cost of living considerations are top of mind. In extending findings from the audience segmentation literature to the Canadian case, and by exploring how each segment responds to different messages about carbon taxes, our study helps better understand heterogeneity in Canadian attitudes toward climate change as well as the potential promise and pitfalls involved when attempting to communicate about a contro- versial policy across a broad range of differently engaged audience segments. In this sense, our

study contributes to the literature on climate change policy and communication by highlight- ing the importance of audience-based data in supporting climate research, policy and commu- nication and offering a starting point for further work aimed at developing messages that speak to different climate change audiences. While exploratory in nature, our analysis is limited in a number of important respects. First, our findings are based on a relatively small sample of the five audience groups. In analyz- ing how these audiences respond to the treatments tested in our experiment, we had to rely on a small number of observations. Caution should be exercised when interpreting these results, notably for segments for which we have the smallest samples (i.e., Disengaged and Dismissive audiences). The null results reported in this analysis might reflect a lack of statistical power, particularly for these smaller audiences. In order to detect medium-sized effects with a power of 80% and an alpha level of 0.05, it would require a sample of 179 respondents in each audi- ence segment. Given that this is the first segmentation study of its kind for Canada, it was diffi- cult to anticipate the size of the audiences a priori. While we were able to meet this threshold for three segments (i.e., the Alarmed, Concerned and Doubtful), more research with larger samples would be required to determine the effectiveness of different frames across all audi- ences. For instance, future research could use data available from social media platforms as an

exciting (and less costly) research possibility to classify climate change audiences from larger samples. Second, our analysis provided a high-resolution portrait of attitudes toward climate change at one specific point in time, offering little insight into the evolution of climate opinions and audiences, as well as limited information on the durability of the observed framing effects. Many things have changed since 2017 in terms of national and international discourses on cli- mate change, likely bringing about changes in the composition of the segments. For instance, the Six America’s studies in the United States documented a clear trend toward rising Alarm in the United States [81]. Although we can only speculate as to whether Canada followed a similar trend, our study provides a baseline against which future research can be compared. Other studies could replicate this segmentation using longitudinal data to provide a more dynamic picture of Canadian attitudes towards climate change. Finally, there are other impor- tant messages, and other factors that may condition messaging effects that were not included in this analysis, such as the source of the message and the availability of competing frames or arguments, which we leave for future work. PLOS ONE The Five Canadas of Climate Change PLOS ONE | https://doi.org/10.1371/journal.pone.0273977 November 23, 2022 17 / 21