This paper examines the effect of experiencing extreme weather on personal engagement with climate change. We address this by focusing on a specific series of flooding events that occurred in the UK in late 2013/early 2014. These events were brought about by a series of major storms hitting the UK in quick succession, leading to extensive and persistent flooding events across England and Wales. The impacts were consistent with projections for increased UK flood risk under climate change (Huntingford et al. 2014), led to a major emergency response, and prompted Prime Minister David Cameron and national media to attribute the events, at least in part, to climate change (BBC Online 2014). Subsequent climate model simulations indicate that the risks of extreme precipitation in Southern England at the time were heightened significantly by anthropogenic warming (Schaller et al. 2016).
The notion that a person’s ‘experience’ with weather-related phenomena provides a potentially important route to engagement with climate change has been suggested often (Lorenzoni and Pidgeon 2006; Weber 2010; McDonald et al 2015; Reser et al 2014). This hypothesis is rooted in literature which shows that people can use direct personal experiences, in addition to secondary sources (e.g. scientists, media), to understand otherwise abstract risks. Such personal experiences may help anchor people’s understanding of climate change by making the risk more concrete and familiar (Smith and Joffe 2013; Bickerstaff and Walker 2001; Spence et al. 2012).
Many studies have now examined connections between variations in actual and perceived temperature and perceptions of climate change (e.g. Capstick and Pidgeon 2014; Zaval et al. 2014; Howe et al. 2013; Li et al. 2011). However, while climate change is expected to lead to aggregate global temperature increases, one of the principal ways it is likely to be made concrete for ordinary people at a regional level is through extreme weather events (EWEs) (Coumou and Rahmstorf 2012). This distinction is important because experiences with EWEs might hold the potential for more profound changes in people’s perceptions of climate change more generally. EWEs may act as a strong ‘signal’ or ‘focusing event’ (Renn 2011; November et al. 2009) whereby future climatic events are made more imaginable, indicating dramatic changes to familiar and local places, in turn heightening the sense of risk posed by climate change. EWEs are also often associated with changed socio-political contexts (media coverage, institutional responses etc.) which themselves constitute important influences on people’s perceptions (cf. Pidgeon et al. 2003).
The empirical evidence on this issue is mixed. In the first study to examine this question, conducted 3 years after major UK flooding in 2000, Whitmarsh (2008) reported no systematic relationship between respondents in flooded and non-flooded areas and their climate beliefs. More recent studies do however suggest that EWE experiences can increase belief in climate change and promote support for sustainable behaviour change (Myers et al. 2013; Spence et al. 2011; Broomwell et al. 2015; Taylor et al. 2014a; Lujala et al. 2015). However, a number of conceptual and methodological aspects of these previous studies deserve closer attention.
While there has been a focus on establishing empirical associations between EWE experiences and climate change perceptions, the mechanisms by which this might occur remain relatively untested. One aspect that has received some attention in the literature is the possible role of heuristics in understanding the process by which experiences might affect risk perceptions or engagement. Indeed, there are clear parallels between assumptions made about the role of EWEs impacting climate change perceptions and psychological literature on the availability heuristic, which suggests that people’s judgements about risk are influenced by the ease with which relevant events come to mind (Weber 2010; Keller et al. 2006; Taylor et al. 2014a). As such, experiences of an EWE might make climate risk more cognitively available or salient in people’s minds, particularly where such a link has also been made within the media or the statements of prominent figures. We already know that following flooding experiences, people are more likely to perceive future flood risk and to buy insurance (in effect, an adaptation to that experience) due to the increased salience of flooding (Browne and Hoyt 2000). Equally, if a person has experienced EWEs, this could lead to heightened salience of climate change, making it easier to envisage other ways in which climate change might affect them personally.
A second heuristic, potentially acting as a parallel mechanism, or mediator, of the relationship between experience and risk perceptions, is the affect heuristic (Finucane et al. 2000). Here it is proposed that experiential learning is facilitated through an emotional response, which heightens the importance of the experience and its subsequent influence on attitudes towards risk (Weber 2010; Keller et al. 2006), whereby it is easier to remember affect-laden events. The intense emotional responses known to accompany flooding experiences (e.g. Walker et al. 2011) provide a potentially powerful mechanism linking these events with responses to climate change, and in a way that is likely absent in other types of climate ‘experiences’ such as temperature changes. We would expect those with strong direct emotional experiences of extreme weather to be more likely to have heightened perceptions of the event and its possible related risks (in this case climate change).
The increased salience of climate change deriving from direct experience might also manifest in terms of other forms of engagement. For example, many studies on climate change perceptions have noted that climate change ‘issue salience’ is low when compared with other issues (e.g. the economy, migration, health; see Pidgeon 2012; Taylor et al. 2014b; Capstick et al. 2015b). Accordingly, direct experiences might influence the relative importance of climate change in relation to other priorities in life, alongside its personal salience as a risk issue. Previous research, by contrast, has tended to focus primarily on changes to epistemic beliefs (i.e. whether people accept climate change is a physical reality; see Myers et al. 2013). It would be fruitful, therefore, to take a broader view of personal engagement with climate change, representing a shift of focus – moving away from asking whether ‘seeing is believing’ and asking instead whether experiences lead to engaging and acting.
A final consideration in the current study is to examine not only personal salience and mitigation intentions, but also attitudes towards wider climate policies as well as adaptation intentions to ostensibly unrelated climate events (in particular heatwaves). It is conceivable that if an EWE experience heightens climate change issue salience among those affected, this might in turn translate into responses to other types of climate risks. Such a link between EWE experience and wider climate adaptation has received very limited attention to date (Reser et al. 2014; Blennow et al. 2012) but has potentially important implications for future resilience planning and wider public engagement (Moser 2014).
The discussion above leads to the following four hypotheses. People with direct personal experience of the 2013/14 winter flooding, when compared to a sample without direct experience, will have: (Hypothesis 1 – Issue Salience) increased issue salience of climate change in general, and increased salience relative to other issues; (Hypothesis 2 – Risk Perceptions) higher perceptions of general climate change risks; (Hypothesis 3 – Emotional Engagement and Climate Change Mitigation) higher levels of emotional engagement with the flooding events, which in turn will be positively associated with (as a mediating variable) higher climate change risk perceptions, mitigation intentions and policy support; and (Hypothesis 4 – Non-Flood Adaptation Intentions) higher intentions to engage in non-flood related climate adaptation measures (in this case regarding heatwaves).
In order to outline our approach to testing these hypotheses, it is important to draw attention to several shortcomings in the methodology used within the research literature to date. First, an important limitation has been the wide-ranging way in which experience has been conceptualised, measured and interpreted. Many studies have used ambiguous or loosely-defined constructs or left participants to judge for themselves what might constitute ‘experience’ of extreme weather (e.g. Dai et al. 2015; Rudman et al. 2013; Spence et al. 2011; Taylor et al. 2014a; van der Linden 2014; Lujala et al. 2015; Whitmarsh 2008). Other studies have asked participants to state whether or not they have personally experienced global warming (e.g. Akerlof et al. 2013; Blennow et al. 2012; Broomwell et al. 2015; Myers et al. 2013; Reser et al. 2014), which in effect confounds both experience and a global warming belief in a single item – this is particularly problematic where an attempt is made to infer a causal link between these two constructs. Likewise, items that directly include an explicit ‘climate change attribution’ element (i.e. whether an event is attributed to climate change) can shed light on the extent to which people believe climate change has manifested in their own lives (or the extent to which events are consciously attributed to climate change), but subsequent linkages with perceptions or actions on climate change are more problematic to interpret. This is because, in part, such items are particularly susceptible to response biases: those already concerned about or accepting of climate change are more likely to consider particular EWEs to represent manifestations of it (Corner et al. 2012; Capstick and Pidgeon 2014). In other words, it is entirely possible that self-report of ‘experience’ simply acts as a substitute for people’s climate change beliefs more generally. It is not surprising under such circumstances to find that this type of ‘experience’ measure is sometimes the most significant predictor of other climate change belief measures.
More generally, we also know that information relevant to climate change tends to be interpreted according to pre-existing social, cultural and political beliefs (Corner et al. 2012; McCright and Dunlap 2011; Kahan 2014). Even if measures of experience do not explicitly refer to climate change, they are still susceptible to substantial biased reporting if they are left open to interpretation. This includes measures that do not specify the type of experience (e.g. property damage, travel disruption, emotional reaction), the type of EWE (e.g. flooding, drought, storms), or that include vague or long timescales. If undefined then the participant must themselves decide what constitutes a relevant EWE experience. In the first study to show an association between self-reported flood experience and climate change attitudes, Spence et al. (2011) concede that their interpretation suffers from precisely this causality issue. People already concerned about climate change may have been more likely to report that they had been impacted by ‘flooding in their area’; a suggestion which subsequent studies have corroborated (Blennow et al. 2012).
Taking account of these methodological concerns, we argue that clear and concrete measures of experience are necessary if research is to draw valid conclusions about how those experiences might affect beliefs (rather than vice versa). Although all self-report measures have limitations, it is likely that precisely-worded measures of direct physical experience and material impacts of an event, well-defined in terms of concrete personal effects and damage, are less susceptible to biased reporting, providing the best proxy available for direct measures of ‘objective’ experienceFootnote 1.
Additionally, a much-overlooked methodological consideration is questionnaire structure, particularly the potential influence that the order of question presentation might have on responses. If one is interested in the extent to which a particular experience affects attitudes towards climate change, then measures gauging this experience should be placed subsequent to these more subjective measures. This is because the placing of a measure early on in a questionnaire has the potential to confound findings through inadvertently leading respondents to consider this experience when reporting on their climate beliefs. As a result of this consideration, we placed key measures of climate beliefs as the very first items on the survey, so that these could be elicited independently from any mention of floodingFootnote 2.
The present study seeks to examine the four hypotheses above, using a survey-based approach, while also addressing the methodological issues raised in this section. We apply a quasi-experimental design comparing a nationally representative British sample to individuals who had directly experienced the 2013/2014 UK flooding, taking care to: (a) measure key climate change perceptions using both prompted and unprompted questions before making any reference to extreme weather or flooding; and (b) employ a highly specific measure of direct flooding experience which is unlikely to be influenced by a person’s beliefs or preceding responses about climate change.