Multilevel predictors of climate change beliefs in Africa

adaptation and mitigation strategies [4, 9], especially in Africa—given current projections as the continent that will be most affected by climate change [1]. Despite of this, little research has focused on what predicts climate change beliefs in Africa. Recent metanalyses suggest that fewer than 5% of published articles included African countries in their sample [4, 8]. As CCBs and its predictors vary widely across regions [10, 11], the applicability of non-African research is, at least, questionable. This study aims to narrow this gap. We make three major contribu- tions to the existing literature. First, we offer a specific representation of what predicts African citizens’ climate change beliefs by using disaggregated data across second-order political boundaries within 33 countries. Second, we employ a novel machine learning methodology within the social sciences (random forest) to precisely classify the predictive power of a diverse number of variables over climate change beliefs. Finally, we take a holistic approach that incor- porates most of the previous predictors identified largely in Western countries to examine whether they hold true in Africa. The remainder of this article is structured as follows. First, we briefly review the current lit- erature on the factors that shape climate change beliefs. We then explain our data collection and operationalization process, as well as the specification of our random forest models. Finally, we present our results and discuss the main findings to offer clear patterns of what pre- dicts climate change beliefs in Africa.

What shapes climate change beliefs There is widespread recognition that climate change is real and that its observable conse- quences can already be appreciated—such as more frequent floods and droughts in Eastern Africa, extreme precipitation trends in Central Europe and dryer conditions that increase wildfires in North America [12, 13]. Nonetheless, anthropologic climate science denial persists among the public because individuals lack sufficient information, have a poor understanding of the matter, or because they associate climate science with conspiracy theories, and the like. Recent survey data from Africa suggests that only 56% of the continent’s population have heard about climate change and about 20% believe that ordinary citizens can do nothing to stop climate change [5]. What shapes these climate change beliefs? A growing number of inter- disciplinary studies suggest four possible answers. First, despite the growing availability of cli- mate change information, much of it remains inaccessible for many and those who do have access to it are unable to understand the complex science behind such phenomenon. Second, despite having access to climate change information, people’s religious beliefs or political ide- ology can lead to a clash of ideas that can often result in a biased interpretation of facts. Third, scientific information about climate change is too abstract, leaving people to interpret climate change trough their own personal experience with local climate conditions. Finally, people

have a “finite pool of worry”—more urgent concerns than climate change—pushing climate change concerns to the backburner. Barriers to climate change information. A first argument suggests that individuals do not form correct beliefs about climate change because they face barriers to access climate infor- mation. For instance, not having a cellphone or access to the internet are barriers that limit knowledge transference about climate change, reducing the amount of information that can improve CCBs [10, 11]. A similar argument suggests that such technological barriers limit access to social media platforms, which can also provide knowledge about climate change- related debates among the public [14]. Nonetheless, these platforms can also serve as echo chambers where previous beliefs are not contrasted, but continuously reinforced [15]. A sec- ond argument shifts to educational barriers. Lack of access to education leads to less knowl- edge, including a detailed understanding of the climate cycle and how humans negatively PLOS ONE Multilevel predictors of climate change beliefs in Africa PLOS ONE | https://doi.org/10.1371/journal.pone.0266387 April 5, 2022 2 / 14