Science of the Total Environment 478 (2014) 12–20

Contents lists available at ScienceDirect

Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv

Review

Rethinking the relationship between flood risk perception and flood management S. Birkholz, M. Muro, P. Jeffrey, H.M. Smith ⁎ Cranfield Water Science Institute, Cranfield University, Bedfordshire MK43 0AL, United Kingdom

H I G H L I G H T S • • • •

We critically review theoretical developments around flood risk perception. Contemporary flood management thinking is inadequately informed by risk perception. Past research over-emphasises the cognitive perceptions of those at risk. Flood risk perception research could benefit from a more constructivist agenda.

a r t i c l e

i n f o

Article history: Received 9 August 2013 Received in revised form 13 January 2014 Accepted 17 January 2014 Available online 11 February 2014 Keywords: Flood management Flood risk Risk perception Risk communication Vulnerability Resilience

a b s t r a c t Although flood risk perceptions and their concomitant motivations for behaviour have long been recognised as significant features of community resilience in the face of flooding events, there has, for some time now, been a poorly appreciated fissure in the accompanying literature. Specifically, rationalist and constructivist paradigms in the broader domain of risk perception provide different (though not always conflicting) contexts for interpreting evidence and developing theory. This contribution reviews the major constructs that have been applied to understanding flood risk perceptions and contextualises these within broader conceptual developments around risk perception theory and contemporary thinking around flood risk management. We argue that there is a need to re-examine and re-invigorate flood risk perception research, in a manner that is comprehensively underpinned by more constructivist thinking around flood risk management as well as by developments in broader risk perception research. We draw attention to an historical over-emphasis on the cognitive perceptions of those at risk to the detriment of a richer understanding of a wider range of flood risk perceptions such as those of policy-makers or of tax-payers who live outside flood affected areas as well as the linkages between these perspectives and protective measures such as state-supported flood insurance schemes. Conclusions challenge existing understandings of the relationship between risk perception and flood management, particularly where the latter relates to communication strategies and the extent to which those at risk from flooding feel responsible for taking protective actions. © 2014 Elsevier B.V. All rights reserved.

Contents 1. 2.

3.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . Major theoretical constructs used to describe flood risk perception . . . 2.1. Bounded rationality. . . . . . . . . . . . . . . . . . . . . 2.2. Protective measures — motivation and response . . . . . . . . 2.3. Vulnerability, adaptive capacity, and resilience . . . . . . . . Theoretical developments around risk perceptions . . . . . . . . . . 3.1. Rationalist approaches . . . . . . . . . . . . . . . . . . . 3.1.1. Revealed preferences and the psychometric paradigm . 3.1.2. Prospect Theory — heuristics and judgement . . . . . 3.2. Constructivist approaches . . . . . . . . . . . . . . . . . . 3.2.1. Political ecology and the social construction of risk . . 3.2.2. Cultural theory of risk . . . . . . . . . . . . . . .

⁎ Corresponding author. E-mail address: h.m.smith@cranfield.ac.uk (H.M. Smith). 0048-9697/$ – see front matter © 2014 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.scitotenv.2014.01.061

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

13 13 13 14 14 15 15 15 16 17 17 17

S. Birkholz et al. / Science of the Total Environment 478 (2014) 12–20

4. Challenges for policy and management at the contested interface between perceptions and intervention 5. A renewed agenda for flood risk perception research . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1. Introduction Flooding is a significant and worsening challenge for many human societies. Between 1900 and 2006, nearly one third of all natural disasters in the world were floods, and they accounted for nearly half of all people affected by natural disasters (Adikari and Yoshitani, 2009). Furthermore, the frequency of flood events appears to be on the rise, as is the amount of damage they cause (Schanze, 2006; Adikari and Yoshitani, 2009). Recent research has linked specific flooding events, as well as a general rise in the intensity of wet weather in the Northern Hemisphere, to the effects of rising greenhouse gas levels and global climate change (Schiermeier, 2011). Coupled with this trend is the growing recognition that absolute flood prevention or protection is unattainable, which has shifted attention towards managing flood risks from a more holistic perspective (Schanze, 2006; Scott et al., 2013). This observation holds true in much of the general literature around disaster risk and management, which is increasingly focused on understanding how and why damage from disasters occurs, who is affected, and what strategies, measures and interventions can be implemented to manage and mitigate the impacts. These investigations have been intimately connected with and informed by developments in understandings of ‘risk’ and how people live with uncertainty (Tversky and Kahneman, 1982; Douglas, 1985; Blaikie et al., 1994; Slovic, 2000; Faulkner and Ball, 2007; Slovic, 2010). Definitions of ‘risk perception’ must inevitably draw from characterisations of risk itself. Simpler portrayals of risk as ‘measures of hazards’, with hazards then defined as ‘threats to people and what they value’ (Kates and Kasperson, 1983, p. 7029) have slowly been supplanted by more sophisticated definitions which portray risk as the intersection between a hazard, the exposure of people/assets to the hazard, and the vulnerability of the people/assets that are exposed (Crichton, 1999, cited in Schneiderbauer and Ehrlich, 2004). This breakdown of the three elements of risk (hazard, exposure, vulnerability) underpins current understandings of risk in a wide range of fields, including research around natural disasters (e.g. Blaikie et al., 1994; Nott, 2006). Furthermore, it usefully illustrates the relationship between behaviour and risk (i.e. certain behaviours can potentially alter exposure and/or vulnerability to a hazard). Such behavioural responses are driven by human assessments which incorporate a weighting of the perceived risk (White, 1945; Slovic et al., 1974; Kates and Kasperson, 1983; Burton et al., 1993). These intuitive judgements, through which people assess the potential impacts and consequences of a hazard and choose appropriate behavioural responses, are commonly referred to as risk perceptions (Slovic, 1987). The important role that perception plays in how individuals and communities respond to risk has achieved widespread contemporary recognition in the general risk management literature. A recent Special Issue of the Journal Risk Analysis on the subject argues that ‘perceptions of risk and risk related behaviors may amplify the social, political, and economic impact of disasters well beyond their direct consequences’ (Burns and Slovic, 2012, p. 579). The contributions also highlight the need for better understandings of the links between emotions, risk perceptions and behaviours, as a precursor to developing more effective risk communication and disaster management policies. However, when the focus shifts from general risk to flood risk management, the role of risk perception in the literature appears somewhat under-developed. Historically, flood management approaches – i.e. those which focus on physical flood defences or on improvements in monitoring and prediction – have tended to

13

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

17 18 19 19

overlook the social dimensions of flooding, such as public understanding of the hazard (Brown and Damery, 2002). Nonetheless, flood risk perception has, for some time, been recognised as a crucial factor in developing effective flood management strategies. The manner in which people (households, businesses, governance bodies, etc.) perceive and understand flood risk shapes the judgements they make and the actions they take in preparing for and responding to flood events. Research interest in the field has been growing (e.g. Messner and Meyer, 2006; Raaijmakers et al., 2008), but still warrants further development. This paper argues that there is a need to re-examine and reinvigorate flood risk perception research, in a manner that is comprehensively underpinned by contemporary thinking around flood risk management as well as by developments in broader risk perception research. We note recent contributions from Bubeck et al. (2012) and Kellens et al. (2013) which have articulately reviewed the empirical literature on flood risk perceptions. Both found significant weaknesses in current understandings, highlighting a paucity of evidence for the relationship between individual flood risk perceptions and mitigation behaviour, and a failure to make use of theoretical frameworks that are available in social science research. Our analysis seeks to address some of these weaknesses and strengthen the theoretical underpinnings of the field by reviewing the major constructs that have been applied to understanding flood risk perceptions (Section 2), and contextualising these within broader conceptual developments around risk perception theory and contemporary thinking around flood risk management (Section 3). In doing so, we highlight two broad strands of thinking in relation to risk perceptions, which we call constructivist and rationalist approaches. We show that progress in the broader domain of risk perception, which has to some degree sought to unite the insights drawn from both approaches, has not yet been reflected in research around flood risk perceptions. This presents a significant opportunity to enliven and enrich the field, in order to better inform flood management policies and strategies (Sections 4 & 5). 2. Major theoretical constructs used to describe flood risk perception 2.1. Bounded rationality Interestingly, some of the earliest explorations of risk perceptions around natural hazards were focused on flood risk with researchers seeking insights into why people choose to live on floodplains, despite a constant threat of flooding, and the adjustments they made to enable them to cope in these environments (White, 1945; Kates, 1963, 1964; Burton et al., 1968). These studies broke new ground in conceptualising the problem as being at the interface between social and natural systems (Burton et al., 1968; Kates, 1971). The basic premise behind the research was that ‘floodplain occupancy represents an interaction between the requirements of a human system with its economic, social, and geographical relationships, and a hydrologic system marked by strong elements of uncertainty’ (White, 1945, p. 436). These enquiries into why people select to live in risk-prone areas worked under the assumption that habitation choices are based on the trade-offs that exist between the benefits of living in a particular location and the associated hazards (Kates, 1963; Burton et al., 1965; White, 1972). Whilst these contributions were strongly rooted in rationalist thinking, this generation of risk perception researchers soon realised that conventional rationalist cost–benefit assessments were insufficiently nuanced to capture

14

S. Birkholz et al. / Science of the Total Environment 478 (2014) 12–20

human responses to natural hazards, and they turned to the idea of bounded rationality in search of more powerful conceptual models (Kates, 1971). The theory of bounded rationality suggests, in part, that ‘cognitive limitations of decision makers force them to construct a simplified model of the world to deal with it’ (Slovic et al., 1974, pp. 5). When applied to the context of choices in the face of natural hazards, individual decision-makers are seen as having to choose from a range of alternative responses. However, their choices are limited by their perception of those alternatives (Kates, 1971; White, 1972). Influenced by characteristics of the physical and biological systems on the one hand, and by the social system on the other hand, decision-makers (which importantly includes both citizens and those charged with protecting them) within hazard prone areas are assumed to express their perceptions and rating of risks through their descriptions of the risks, their articulated appraisal of the outcomes, and their actual behaviour (White, 1972). This strand of research was strongly influenced by the cognitive psychology community and efforts were made to link the two schools of thought and develop a more comprehensive understanding of adjustment choices. The progeny of this collaboration (Slovic et al., 1974), in addition to demonstrating the value of cross-disciplinary intellectual engagement, marks one of the first efforts to integrate more general theories of risk perception with research on natural hazards. The authors describe four variables that shape bounded rationality in the choices made concerning natural hazard adjustments; existence of a limited range of alternatives, misperception of risks and denial of uncertainty, crisis orientation, and the tension between individual versus collective management (Slovic et al., 1974). 2.2. Protective measures — motivation and response Protection motivation theory views self-preservation behaviour as motivated by four factors: the perceived severity of a threat, the perceived probability of the occurrence (vulnerability), the perceived effectiveness of any recommended response, and perceived ability to implement a response. Originally developed to understand how people cope with fear and threats (Rogers, 1975) the work was later extended to provide a more general theory of persuasive communication. The experiences of two recent studies, both of which have adopted protection motivation theory as an interpretive construct for empirical work on flood risk perceptions, provide some clues as to its value. Enquiring into the factors that influence German households in adopting precautionary measures to protect themselves from floods, Grothmann and Reusswig (2006) argue that decisions to take precautionary measures are influenced by four broad perceptual processes which are aligned with the components of protection motivation theory: threat appraisal, coping appraisal, threat experience appraisal and administrative measures appraisal. Through household surveys, the authors determined that these perceptual processes can interact in a variety of ways to influence behavioural responses towards flood protection. For instance, the likelihood of taking precautionary (i.e. pre-flood) action was seen to increase when both threat appraisal and coping appraisal were high, but decrease when the administrative measures appraisal was high (i.e. confidence in public measures decreases the perceived need for private measures). The authors argue that private, precautionary actions (e.g. households installing private water barriers) were some of the most effective measures for reducing actual monetary damage in the event of a flood, and that understanding the perceptual processes that motivate such responses could therefore be useful in finding ways to encourage such actions. This study also illustrates how an individual's choices to adopt protective measures are linked not just to their perceptions of the hazard, but also to their perceptions of the role of public authorities. A similar study of Dutch households found that respondents considered the government to have primary responsibility

for protection against flood damage, which helped to explain their low motivation towards adopting individual protective measures (Terpstra and Gutteling, 2008). More recently, Koerth et al. (2013) have asked what motivates coastal dwellers in Germany and Denmark to proactively respond to rising sea-levels and associated flood events. Using close proxy independent variables for two of the protection motivation components this group found that risk perception, and intended household level adaptation measures were reliable predictors of measures that require small investments in terms of efforts and costs, but not of measures requiring high levels of investment. 2.3. Vulnerability, adaptive capacity, and resilience As noted above, thinking around flood risk management has matured over recent decades with a diminished focus on engineeringbased defences and a shift towards more holistic perspectives on the drivers and impacts of flood risk and its appreciation by individuals and communities. Such thinking has benefited from stimulus and direction provided by contemporary developments in the fields of vulnerability, adaptive capacity and resilience. These are highly productive and dynamic fields of theory development and empirical study which we do not have opportunity to comprehensively review here. The following paragraphs are therefore necessarily selective in their presentation of those contributions which have a direct bearing on flood risk perception theory. The concept of vulnerability has become particularly important within natural hazard and disaster research (Cutter, 1996; Pelling, 1997; Cutter et al., 2000; King and MacGregor, 2000; Cutter et al., 2002), used as an analytical lens that encapsulates how social context shapes risk and, by extension, natural disasters such as flood events. For instance, vulnerability has been defined as ‘the characteristics of a person or a group in terms of their capacity to anticipate, cope with, resist, and recover from the impact of a natural or man-made disaster — noting that vulnerability is made up of many political–institutional, economic and socio-cultural factors’ (Schneiderbauer and Ehrlich, 2004, pp. 12). In a similar vein, the pressure-and-release model of disasters (Blaikie et al., 1994) describes vulnerability as being socially produced. This model describes the ‘progression of vulnerability’ as underlying root causes and dynamic pressures (e.g. high inequality plus rapid population growth) which ultimately foster localised unsafe conditions. When a natural hazard occurs within those conditions, a disaster can result (Blaikie et al., 1994). Vulnerability is often described as having three components: exposure to a hazard, susceptibility to harm, and adaptive capacity (Few, 2003; Messner and Meyer, 2006; Adger, 2006; Birkmann, 2006; Smit and Wandel, 2006; de Bruijn et al., 2007; López-Marrero, 2010). Risk perceptions are often implied as a key component of vulnerability assessments. For instance, in the hazards-of-place model (Cutter, 1996; Cutter et al., 2003) perceptions of risk are shown as one component of the ‘social fabric’, which interacts with the geographic context to create ‘place vulnerability’. Similarly, with specific reference to flood vulnerability, Few (2003) sees ‘crisis perceptions’ as part of the cultural arrangements of a society and important components of its ability to cope with flood events. However, while many conceptual frameworks acknowledge the importance of risk perceptions as determinants of vulnerability (e.g. Messner and Meyer, 2006; Kuhlicke et al., 2011), they rarely elaborate on flood risk perceptions in a substantive manner. Consequently an important, indeed one might argue crucial, locus of appreciation is overlooked. It is precisely at the interface between perceptions and management/ intervention that vulnerability to flooding is both experienced and contested. In concert with vulnerability, the concept of ‘adaptive capacity’ is frequently used to refer generically to an individual's or group's ability to recover from a hazardous event such as a flood. However, more

S. Birkholz et al. / Science of the Total Environment 478 (2014) 12–20

recently it has been defined as the ability to make adjustments so as to become more effective at dealing with hazards (Smit and Wandel, 2006; López-Marrero, 2010). There is an interesting gradation here in the relative meaning of recovery and adaptation with the former implying a return to the status quo, whereas the latter suggests moving beyond the status quo and becoming more equipped for dealing with a particular environment (e.g. proactively adopting protective measures). Attempts to incorporate flood risk perceptions into adaptive capacity have suggested that some cognitive factors – such as the perception that the frequency of precipitation (and therefore flood events) was decreasing, or the perception that large-scale structural protective measures were sufficient – curtail a willingness to adopt household-level protective actions. Evidence for this relationship has been found in both Latin America (López-Marrero, 2010) and Europe (Grothmann and Reusswig, 2006; Terpstra and Gutteling, 2008) with the former study also revealing that perception-related strategies such as improving residents' knowledge of the causes and likelihood of flooding, enhancing social memory of past flood occurrences, and reducing reliance on public structural measures (i.e. construction of dikes) could improve communities' adaptive capacity. These findings have also prompted some fruitful enrichment of theory, particularly with respect to ideas around ‘social capacity’ (Kuhlicke et al., 2011; Kuhlicke and Steinführer, 2013) although we would note that many embellishments retain the concept of risk perception and its link to motivation for preparedness at their heart. Community resilience has recently been highlighted as one of the four priority mechanisms for disaster risk reduction worldwide (Schelfaut et al., 2011) and the resilience bandwagon is beginning to leave its mark on the flood risk perception and management community. Notable recent contributions which incorporate resilience as a formal aspiration for flood threatened communities include work on promoting sustainable flood memory (McEwen et al., 2012) and the cultivation of Learning Alliances (Ashley et al., 2012) as stimulation for shared learning experiences which in turn develop the capacity to accept different perspectives on risk and employ alternative innovatory responses to flood threats. Resilience is often tightly intertwined with concepts of vulnerability and capacity and it is widely acknowledged that resilience remains only a partially useful operational concept (Schelfaut et al., 2011). In broad terms, resilient communities (or systems) are those that can creatively reorganise in the wake of disruptive events (e.g. Folke, 2006; Folke et al., 2010). In the context of managing flood risks (or other natural disasters), a resilience based strategy would shun the objective of returning to a status quo after a flood event, and would instead welcome an element of continual and flexible adaptation to changing circumstances. While the concept remains largely in its infancy in the flood risk management literature, it is the subject of increasing attention (Kuhlicke and Steinführer, 2013). Some of the studies which have worked towards developing the concept have again acknowledged the important role of risk perception without developing it in much depth, often in the context of arguing for better risk communication in flood management strategies (e.g. Schelfaut et al., 2011). As research around resilience and capacity building in relation to flood risk management develops further, it stands to benefit from a more in-depth understanding of flood risk perception. There are currently relatively few empirical studies that have substantively explored flood risk perception in this context. As a result, research around flood risk perception also appears somewhat out of step with theoretical developments in the broader study of risk perception. As previously mentioned, the role of risk perceptions in improving the resilience of people and communities is widely recognised as an important component of the wider field of risk research (Burns and Slovic, 2012), and should, we argue, be used to greater effect in underpinning flood risk research. To this end, the following discussion of key theoretical developments in that wider field is of immediate import.

15

3. Theoretical developments around risk perceptions Approaches to understanding risk perception are generally founded on one of two broad paradigms: a rationalist paradigm or a constructivist paradigm. Here we argue that rationalist approaches to risk perception tend to emphasise individual cognitive processes, and assume that the existence of a hazard will prompt an individual to make an assessment or judgement that feeds into a ‘rational’ decision-making process around the need to adopt protective or precautionary behaviours. Studies rooted in a rationalist paradigm therefore tend to focus on modelling, characterising and predicting behavioural outcomes around different hazards. Constructivist approaches, on the other hand, privilege the concept of risk as a contextual, socially constructed phenomenon, and assume that an individual's judgements and decision-making processes are shaped and constrained by their social environments (Drake, 1992). These social environments (e.g. organisations, institutions, cultures) can also dictate behavioural options as a result of power dynamics and access to resources and knowledge (Kates, 1963; Tierney, 1999; Slovic, 2000). This rationalist/ constructivist distinction is based on similar ones proposed elsewhere in the flood risk literature (e.g. Short, 1984; Johnson et al., 2004; Wachinger and Renn, 2010). Ultimately, this distinction stems from a wider ‘constructivist turn’ in social sciences, which has generated considerable debate around the apparent divide between rationalist and constructivist thinking, most notably in political sciences (e.g. Fearon and Wendt, 2002; Fehl, 2004). It is not within the scope or purpose of this paper to review these wider theoretical developments. We apply a rationalist/constructivist distinction, as defined above, as a tool to highlight a particular discrepancy in approaches to risk perception. We note, however, that such a distinction is not absolute, nor are the two paradigms necessarily mutually exclusive (Fearon and Wendt, 2002). Nonetheless, they provide a useful means of characterising the array of approaches and models that have emerged in relation to risk perception. To help illustrate the analysis, Table 1 provides a breakdown of this characterisation, while each specific approach is discussed in greater detail below. 3.1. Rationalist approaches 3.1.1. Revealed preferences and the psychometric paradigm In 1969 an American engineer attempted a comprehensive study of how people weigh technological risks against their benefits in order to assess ‘how safe is safe enough’ (Starr, 1969). This ‘revealed preference’ approach utilised historical data to determine the riskbenefit trade-offs that societies and authorities choose when faced with uncertainty regarding the benefits and costs associated with technological risks. One of the most significant findings from this research is that people will accept greater risks if they are voluntary as opposed to involuntary (Starr, 1969, 1972). Starr's work largely inspired the research of Slovic and colleagues at the University of Oregon into cognitive processes behind societal-risk taking (Slovic, 2000). A central focus of their research was to show that risk is a cognitive construct that can be psychometrically described so as to reveal quantitative levels of judgement (Fischhoff et al., 1978; De Marchi, 2007). Their research focused initially on gambler uncertainty but, as previously mentioned, it later informed early work on flood risk perception and bounded rationality (Slovic et al., 1974) and decisions around technological risks (Fischhoff et al., 1978). The Oregon School's approach to quantitatively assessing and characterising risk perceptions, usually referred to as the psychometric paradigm, has greatly influenced subsequent research into risk perceptions. One study (Fischhoff et al., 1978) identified nine dimensions or characteristics of risk believed to influence how people perceive risks: 1) whether the risk was voluntary or involuntary; 2) the immediacy of effect; 3) the extent of personal knowledge of the risk; 4) the extent of scientific knowledge of the risk; 5) whether the risk had potential be

S. Birkholz et al. / Science of the Total Environment 478 (2014) 12–20

16 Table 1 Summary of approaches to risk perception. Paradigm

Approach

Description & attributes

References

Rationalist

Revealed preferences

• Interested in determining how society evaluates benefits versus costs related to technological risks • Utilises historical empirical data on accidents and health, for a variety of public activities, to evaluate the social cost considered acceptable within benefit–cost analysis of technology risk assessment • Holds that risk is a subjective construct that can be psychometrically described so as to reveal quantitative levels of judgement • Focuses on cognitive variables that influence the perception of risk • Focuses on individual assessments of risks, and the differences between ‘lay’ and/or ‘expert’ risk perceptions • Interested in how different types of risks are assessed and ranked by individuals. • Describes how people make decisions under uncertainty, based on their evaluations of potential losses and gains • Argues that people often resort to mental ‘guidelines’ to make judgements in the context of uncertain risks or risk events, which can result in bias/error • Argues that likely risk perceptions can be determined from cultural adherence and social learning • Argues that structures of social organisation endow and reinforce the perceptions an individual holds • Proposes four worldview categories – fatalistic, hierarchic, individualistic, & egalitarian – each of which are associated with particular risk perceptions • Focuses on societal factors influencing risk estimation practises • Sees risk perceptions as a dependent variable of societal contexts in which they are shaped • Argues for critical analysis of the context of major accidents and disasters • Holds that risk is dynamic and often imposed on societies by power structures, and is experienced unequally by marginalised groups • Focuses on the role of organisations and governments in shaping risk perceptions in societies

Starr (1972)

Psychometric paradigm

Heuristics & judgement

Constructivist

Cultural theory of risk

Social construction of risk

chronic or catastrophic; 6) whether the risk was commonplace or dreaded; 7) the severity of consequences; 8) the level of control people have concerning their exposure to the risk; and 9) the novelty of the risk. When a group of respondents was asked to rate the risk of a range of common activities (i.e. smoking, alcoholic beverages, bicycles, commercial aviation, contraceptives, electric power) relative to these nine dimensions, dread and novelty were the two factors found to explain most of the variance observed (Fischhoff et al., 1978). Since this initial study, research into the psychometric analysis of risk perception has included larger and more diverse groups of respondents, as well as greater variation in the types of hazards explored. For more in-depth reviews of research around the psychometric paradigm we would point the interested reader to contributions by Boholm (1998) and Sjöberg (2002). 3.1.2. Prospect Theory — heuristics and judgement The revealed preference approach also influenced work within cognitive psychology into the anomalies around human behaviour and judgements made under uncertainty. This influenced the development of Prospect Theory, which describes how people make decisions under uncertainty based on their evaluations of potential losses and gains (Tversky and Kahneman, 1982). Prospect Theory has had significant influence on economic theories around financial decisions (Kyle et al., 2006). Through an exploration of how people assess ‘the probability of an uncertain event or the value of an uncertain quantity’ (Tversky and Kahneman, 1982) researchers describe how people rely on a set of heuristics to assess the probability or chance of an event unfolding. The research determined that people often resort to mental guidelines to make judgements in the context of uncertain risks or risk events (Tversky and Kahneman, 1982; Boholm, 1998). These judgements rely on the accessibility of available knowledge (personal and external) concerning a risk. However, these ‘guidelines’ could also be erroneous, and result in bias that negates actual probability measurements (Tversky and Kahneman, 1982, 1983).

Fischhoff et al. (1978), Slovic (2000)

Tversky and Kahneman (1982), Tversky and Kahneman (1983), Johnson and Tversky (1983)

Douglas (1978), Douglas and Wildavsky (1982)

Short (1984), Clarke and Short (1993), OliverSmith (1996), Tierney (1999), Johnson et al. (2004)

Johnson and Tversky (1983) expanded on these findings by investigating the role of affect (emotion) in influencing a person's estimate of the frequency of a hazardous event. Their study involved asking a group of respondents to estimate the frequency of fatalities from a list of 18 risks (i.e. tornadoes, floods, lightning, fire, electrocution, accidental falls, traffic accidents, airplane accidents). They determined that in general people tend to make judgements that are compatible with their current mood (affect), even if the subject matter is unrelated to the cause of that mood. This means that when interpreting public perception of a risk, the potentially manipulative role mood has on lay judgements must not be overlooked. The study also outlines two observations of bias or error in people's estimates of hazards: they overestimate infrequent causes of death while underestimating more frequent causes; and they tend to overestimate causes of death that are dramatic and sensational, and underestimate causes of death from unspectacular events that generally claim one victim at a time. The authors speculated that these biases could be attributed to an ‘availability heuristic’, in that ease of recall was influenced by media attention to the overestimated risks (Johnson and Tversky, 1983). A more recent study explored the judgements people make when faced with images of hazards and events, arguing that the mental representations people have can carry an ‘affect tag’ – i.e. they are ‘tagged’ with faint negative or positive feelings – and this tag can then influence the risk/benefit perceptions that people have of a thing, event or hazard (Finucane et al., 2000). Additionally, the influence of the affect heuristic and associated factors, such as mood, on risk perceptions has been more comprehensively examined by Slovic (2010) and Slovic et al. (2004, p. 311) who describe rational behaviour in relation to risk as requiring a ‘complex interplay between emotion and reason’. The rationalist paradigm has allowed the development of business and epidemiological risk assessments by nourishing the assumption that individual risk preferences and behavioural outcomes are precipitated via a logical assessment of likely outcomes and associated costs

S. Birkholz et al. / Science of the Total Environment 478 (2014) 12–20

and benefits. In this context the rationalist paradigm has been both successful and useful and informs many algorithms which shape decision making in the financial, insurance and management sectors. However, contemporary critiques of the limitations of a rationalist perspective point out that human action is personal, contextual, and socially constructed, making reliance on universal rationality as a foundational concept rather limiting. Work on the affect heuristic has demonstrated that our affective reactions can be manipulated (e.g. through advertising) and this can exert a powerful influence on our ‘rational systems’ of thinking, sometimes leading to seemingly irrational behavioural outcomes (Slovic et al., 2004). Perhaps the most articulate critique of the rationalist paradigm came from Jaeger et al. (2001) who, in addition to providing a nuanced distinction between three levels of interpretation for the rational actor paradigm, pointed out that it erroneously assumes that all human behaviour can be modelled as variants of optimisation problems. This thinking is influencing others to develop innovative approaches to risk management which privilege both relativist understandings of risk and poly-centric governance (see, for example, Heazle et al., 2013). 3.2. Constructivist approaches 3.2.1. Political ecology and the social construction of risk During the 1970s and 1980s an increasing number of sociologists and anthropologists questioned the assumptions and approaches of the rationalist paradigms around assessing risk and risk perception (Short, 1984; Douglas, 1985; Johnson et al., 2004). They argued that such rationalist approaches did not effectively take into account the social structures and organisations within which perceptions of risk developed (Tierney, 1999). One response to this critique was the application of political ecology theory to risk-related research. Political ecology research seeks to develop knowledge around a society's relationship with hazards by examining contextual factors such as structural constraints and the political/economic forces that result in varying access to resources and varying degrees of vulnerability amongst marginalised groups. One effect of this line of thinking was that it helped shift the focus of risk research towards exploring the role of contextual factors in shaping risk and perceptions of risk (Mustafa, 2002). These ideas were closely related to the development of constructivist thinking around risk, which largely rejects the idea that hazards are objective phenomena external to the social system. Hazards and risks are seen as socially constructed and therefore intricately linked with the dynamics of the social system — culture, institutions, organisations, values, beliefs, etc. (Drake, 1992; Oliver-Smith, 1996; Tierney, 1999; Weichselgartner, 2001; Johnson et al., 2004). Moreover, the overall understanding of risk is characterised as highly context-dependent — i.e. what may be considered a hazard in one social context may be seen as a resource in another (Weichselgartner, 2001). One early example of a constructivist understanding of risk highlighted the role of the media in contributing to social perceptions of risk (Short, 1984). This work added a new perspective to earlier rationalist studies around the role of affect in risk perceptions, which had pointed out that media coverage could potentially contribute to biased judgement of risks (Johnson and Tversky, 1983). The constructivist stance (Short, 1984) argued that more dramatic and disastrous events receive more media coverage than everyday hazards, and that this preference for catastrophe (which helps popularise news stories) results in everyday hazards being perceived as less dangerous, less frightening, and more easily solved, thereby leading to underestimation of the risks associated with more frequent events. This illustrates how the media can play a key part in the social construction of risk perceptions. As constructivist thinking developed further across different disciplines, subsequent authors began expanding on how it could be more comprehensively applied within risk-related research. Some

17

argued that a constructivist stance was essential for opening new and fruitful lines of enquiry in risk research, such as understanding the power dynamics that frame the terms of debate about risks in society, as well as the influence of organised interest groups over risk management policy (Clarke and Short, 1993). Such constructivist arguments also arose in the literature around natural hazards and disasters — the popularisation of conceptual frameworks around vulnerability, discussed earlier in Section 2.3, is strongly rooted in constructivist thinking around risk. Some have also argued that social constructivist approaches will benefit the field of risk research further by shedding light on a range of factors which drive, transform, and moderate the dynamics between risk perception and management. Such factors might include: the mechanisms by which social agents create and demarcate what is considered a risk or hazard; how and why certain mechanisms of assessing or measuring risk are legitimised over others; the rhetorical strategies used by decisionmakers, researchers, the private sector, and others when making claims regarding risks; and the organisational, institutional and societal factors that contribute to disastrous events (Tierney, 1999). 3.2.2. Cultural theory of risk Another important theoretical development, which was broadly associated with constructivist thinking, was the cultural theory of risk (Douglas and Wildavsky, 1982). Cultural theory suggests that it should be possible to determine likely risk perceptions from cultural adherence and social understandings. It approaches the concept of risk in society from the perspective that structures of social organisation endow and reinforce the perceptions of risk that an individual holds (Douglas and Wildavsky, 1982). The theory is based on ‘the uniquely human capacity to classify experiences, encode such classifications symbolically, and teach such abstractions to others’ (Oltedal et al., 2004, pp. 17). Douglas (1978) made a distinction between ‘cultural bias’ (shared values and beliefs) and ‘social relations’ (patterns of interpersonal relations), and saw ‘way of life’ being a combination of these two aspects. The enculturation of a younger generation thereby acts to reproduce the established ‘way of life’ (Oltedal et al., 2004). Based on this understanding, Douglas and Wildavsky (1982) propose four broad ‘way of life’ categories or groups, which (in general) can describe an individual's attitudes and actions: individualistic, egalitarian, hierarchical and fatalistic. These categories are also associated with certain assumptions or attitudes towards risks, as summarised in Table 2. 4. Challenges for policy and management at the contested interface between perceptions and intervention The challenges of drawing lessons from risk perception research to inform policy and management have long been recognised and the gaps between rationalist and constructivist interpretations do not make these challenges any easier. A significant confounding factor here is the need for those who are sanctioned to plan and execute flood management strategies to communicate, validate and justify proposed measures. This becomes particularly complex where those strategies are founded on more constructivist tenets (which rely on principle and debate for their validity) and where measures seek to promote more collective actions around flood risk reduction. The following examples both illustrate some of these problem areas and indicate where an improved understanding of the interface between perceptions and management, particularly one that is more comprehensively underpinned by constructivist paradigms, might usefully inform improved flood risk management interventions. One of the more frequent assertions made in the risk management literature is that understandings of risk perception should inform risk communication strategies. For instance, Burns and Slovic (2012, p. 581) argue that the outcomes of risk perception research can help to ‘better prescribe risk management and communication strategies, and thereby lessen the societal costs of major disasters’. In flood risk

18

S. Birkholz et al. / Science of the Total Environment 478 (2014) 12–20

Table 2 Characteristics of worldview categories, with attitudes towards risks. Adapted from Oltedal et al. (2004). Worldview

Characteristics

Individualist

• • • • • • •

Egalitarian

Fatalistic

Hierarchical

• • • •

See nature as self-preserving, with the ability to re-establish its own status quo — hence people don't really need to care for nature See risk as an opportunity as long as it does not limit freedom Tend to be sceptical of expert knowledge, because they are suspicious of misused authority See nature as fragile and vulnerable to human interventions Generally oppose any risk that might inflict irreversible dangers on many people or future generations Have minimal involvement in social life; they tend to feel tied and regulated by social groups they belong to See nature as unpredictable and providing very little reliable feedback as to whether people are doing the right thing or not, therefore problems need to be dealt with as and when they arise Assume risk to be unavoidable High degree of trust and faith in expert knowledge See nature as self-preserving, though within strict and rigid limits — if such limits are crossed nature will no longer be able to heal itself Accept risks that have been justified by the government or experts

management research, the need for rethinking risk communication strategies is often made clear. For instance, one study raised concerns about the U.K. Environment Agency's Indicative Floodplain Maps – online tools that highlighted potential exposure to flooding – arguing that the maps might give people a false impression of their likelihood of being flooded, because of a lack of information around how to interpret the maps appropriately (Brown and Damery, 2002). Similarly, a more recent study argued that some California residents may have been given a false sense of security because their land is protected by levees and therefore not ‘officially’ designated as a floodplain, despite being below sea level (Ludy and Kondolf, 2012). Others have also argued for a re-design of flood insurance tools in the U.S., in order to provide more transparent assessments of homeowners' exposure to flood risk, so as to encourage them to adopt long-term outlooks in their decisions to seek insurance coverage (Michel-Kerjan and Kunreuther, 2011). Such studies illustrate how some risk communication mechanisms can have unintended consequences in shaping risk perceptions, and that understanding these influences can help inform the design of more effective communication strategies. However, such studies generally maintain a relatively narrow focus — i.e. on the need for policy-makers to improve their (oneway) communication with those at risk of flooding, in order to shape their perceptions of flood risk so that they are encouraged to adopt more protective measures. This focus risks ignoring some of the wider lessons from the more constructivist approaches to risk, which emphasise that all facets of society, not just those ‘at risk’, have a role in shaping how risk is understood and ultimately dealt with. Furthermore, these different facets of society may seek to handle risks very differently. For instance, the interest of ‘attracting additional inhabitants or enterprises into a floodplain region in order to strengthen the regional economic development’ (Messner and Meyer, 2006) might result in attempts to tone down predictions of exposure to flooding, or objections to unattractive protection measures. Other research has demonstrated that different stakeholders can support different management interventions in relation to flooding, or support the same interventions but for very different reasons (Haasnoot et al., 2013). Therefore, a research agenda for flood risk perceptions that is comprehensively underpinned by constructivist understandings (as we argue for in the next section) could inform the development of flood management policies and strategies by helping them to consider the interactions of these different (and potentially competing) interests. Finally, flood risk perception research has also underscored the importance of perceived responsibility for risk management. In other words, the extent to which people feel responsible for taking protective actions is closely related to their belief in the effectiveness of ‘public’ protective measures. This is perhaps especially true in regards to flood risks, for which there has long been a tendency (in developed countries at least) to rely primarily on large-scale, publicly funded, structural protection. In the desire to move away from absolute reliance on structural measures, policy makers in many countries are

seeking to divest themselves of sole responsibility for flood protection, and encourage more household and community-level actions. This requires much more than a one-way risk communication strategy — indeed, many authors now argue for long-term engagement based on dialogue between those at risk, policy makers, and other stakeholders (e.g. Baan and Klijn, 2004; Terpstra and Gutteling, 2008; Kuhlicke et al., 2011; Burns and Slovic, 2012). Ideally, this process should not be seen as just a way of transmitting ‘correct’ flood risk information, or a way of convincing certain stakeholders to adopt more protective measures, but as (to some degree) a negotiation of shared responsibility for flood protection. Undertaking such long-term engagement requires some appreciation of how societies (not only those at risk) understand and value personal protection vs. public protective measures, as well as how public measures might be valued against other societal objectives (e.g. economic growth). Again, a renewed research agenda around flood risk perceptions can help develop this understanding and provide a stronger basis for long-term dialogue. 5. A renewed agenda for flood risk perception research In Section 2 we discussed two substantive approaches to examining flood risk perceptions — bounded rationality and protection motivation theory. These can be seen as rooted largely in a rationalist paradigm, as they are closely related to the cognitive psychology work discussed in Section 3.1. They reveal the complex interrelatedness of the factors that can influence decision-making around protection from flood risk. However, they also reveal that those influential factors can be contextual — for instance, the existence of ‘public’ flood defence measures can significantly affect willingness to adopt ‘personal’ protective measures (Grothmann and Reusswig, 2006; Terpstra and Gutteling, 2008). Contrastingly, constructivist thinking advocates a more comprehensive exploration of how socio-cultural context shapes wider understandings of risk and influences the outcomes of disastrous events. As discussed in Section 2.3, contemporary flood risk management research is increasingly informed by more holistic, socio-cultural concepts such as vulnerability, capacity and resilience, and these concepts share close ties with constructivist perspectives. However, within those three concepts, the nature and role of flood risk perception are still under-developed. As a result, there is a need for a re-invigoration of flood risk perception research, in order to deliver a more comprehensive understanding of how risk perceptions influence the vulnerability, capacity and resilience of individuals and communities in the face of flooding. This renewed impetus should be more substantively underpinned by constructivist thinking around broader risk perceptions (e.g. Nelkin, 1989; Vaughan and Seifert, 1992). This is not to argue that rationalist approaches should be disregarded. Much of the recent research around flood risk perceptions has concentrated on the cognitive perceptions of those at risk, and the links between these perceptions and willingness to take protective actions. Resulting insights have generated new

S. Birkholz et al. / Science of the Total Environment 478 (2014) 12–20

decision support frameworks for policy makers (e.g. Raaijmakers et al., 2008) and over time have enabled many countries to improve the ways in which they visualise, plan for and address the risks from flooding. However, a greater engagement with constructivist perspectives might broaden and enrich this field of research by drawing attention to a wider range of flood risk perceptions (such as those of policymakers, or those of tax-payers who live outside flood affected areas) and their links with larger-scale protective measures (such as statesupported flood insurance schemes). As previously mentioned, research has already highlighted that different stakeholders can have different and competing priorities around flood risk management. Constructivist-oriented flood risk perception research could deepen our understanding of these differences and deliver a more nuanced recognition of how flood risk cultures are fashioned by individual and collective experiences, histories, beliefs, relationships and understandings. In addition, the manner in which such socio-cultural forces interact with and shape the affect heuristics of individuals (the study of which is still largely in its infancy — Slovic et al., 2004) presents an important nexus between rationalist and constructivist perspectives, worth of much further exploration. Two contributions provide particular promise as starting points for a renewed research agenda around flood risk perception. The constructivist position argued by Tierney (1999), and discussed above, draws attention to the mechanisms through which particular perceptions of risk might be disseminated and become legitimised at a wider, societal level. Tierney's short but intellectually potent manuscript offers a penetrating view of how institutions dictate public understanding of risk and of how social inequalities compound risk exposure. Published around the same time, Ortwin Renn's (1998) contribution argues that both constructivist and realist perspectives are insufficiently authoritative to be able to deliver sound guidance in isolation. However, he goes on to argue that scientific expertise, rational decision making and public values are not at odds with each other and can be reconciled through well managed discourse. The insights offered by both contributions offer a point of departure that has yet to be fully engaged with and embellished by academia or responded to by practitioners. Both are integrative and progressive, provide intellectual stimulus and transform debate beyond sterile concerns about disciplinary or paradigmatic hegemony.

Acknowledgements The research which informed this contribution was financed by the European Community's Seventh Framework Programme through the CORFU (Collaborative Research on Flood Resilience in Urban Area) project — EC Contract #244047. References Adger WN. Vulnerability. Glob Environ Chang 2006;16(3):268–81. Adikari Y, Yoshitani J. Global trend in water-related disasters — an insight for policymakers. Paris: UNESCO; 2009. Ashley RM, Blanskby J, Newman R, Gersonius B, Poole A, Lindley G, et al. Learning and Action Alliances to build capacity for flood resilience. J Flood Risk Manag 2012;5(1): 14–22. Baan PJA, Klijn F. Flood risk perception and implications for flood risk management in the Netherlands. J River Basin Manag 2004;2(1):1–10. Birkmann J, editor. Measuring vulnerability to natural hazards—towards disaster-resilient societies. Tokyo, New York: UNU Press; 2006. Blaikie P, Cannon T, Davis I, Wisner B, editors. At risk: natural hazards, people's vulnerability, and disasters. London: Routledge; 1994. Boholm Å. Comparative studies of risk perception: a review of twenty years of research. J Risk Res 1998;1(2):135–63. Brown JD, Damery SL. Managing flood risk in the UK: towards an integration of social and technical perspectives. Trans Inst Br Geogr 2002;27:412–26. Bubeck P, Botzen WJW, Aerts JCJH. A review of risk perceptions and other factors that influence flood mitigation behavior. Risk Anal 2012;32(9):1481–95. Burns WJ, Slovic P. Risk perception and behaviors: anticipating and responding to crises. Risk Anal 2012;32:579–82.

19

Burton I, Kates RW, Mather JR, Snead RE. The shores of megalopolis: coastal occupance and human adjustment to flood hazard. Publ Climatol 1965;18(3). [Available at: http://rwkates.org/pdfs/1965.01.pdf Accessed 7 August 2013]. Burton I, Kates RW, White GF. The human ecology of extreme geophysical events. Paper 78. FMHI Publications; 1968 [http://scholarcommons.usf.edu/fmhi_pub/78 Accessed 7 August 2013]. Burton I, Kates RW, White GF. The environment as hazard. London: Guildford Press; 1993. Clarke L, Short JF. Social organisation and risk: some current controversies. Ann Rev Sociol 1993;19:375–99. Crichton D. The risk triangle. In: Ingleton J, editor. Natural disaster management. London: Tudor Rose; 1999. Cutter SL. Vulnerability to environmental hazards. Prog Hum Geogr 1996;20(4):529–39. Cutter SL, Mitchell JT, Scott MS. Revealing the vulnerability of people and places: a case study of Georgetown County, South Carolina. Ann Assoc Am Geogr 2000;90(4): 713–37. Cutter SL, Boruff BJ, Shirley WL. Social vulnerability to environmental hazards. Soc Sci Q 2002;84(2):242–61. Cutter SL, Boruff BJ, Shirley WL. Social Vulnerability to Environmental Hazards. Social Science Quarterly 2003;84(2):242–61. de Bruijn GJ, Kremers SPJ, De Vet E, De Nooijer J, Van Mechelen W, Brug J. Does habit strength moderate the intention–behaviour relationship in the theory of planned behaviour? The case of fruit consumption. Psychol Health 2007;22:899–916. De Marchi B. Flood risk management with the public. Proceedings of the European Symposium on Flood Risk Management Research. Istituto di Sociologia Internazionale di Gorizia; 2007. p. 153–4. [Dresden, 6–7 February 2007]. Douglas M. Cultural bias. Occasional Paper no. 35. Royal Anthropological Institute of Great Britain and Ireland; 1978. Douglas M. Risk acceptability according to the social sciences. Russell Sage Foundation; 1985. Douglas M, Wildavsky AB. Risk and culture: an essay on the selection of technical and environmental dangers. Berkeley: University of California Press; 1982. Drake K. Myths of nature: culture and the social construction of risk. J Soc Issues 1992;48(4):21–37. Faulkner H, Ball D. Environmental hazards and risk communication. Environ Hazards 2007;7(2):71–8. Fearon J, Wendt A. Rationalism v. constructivism: a skeptical view. In: Carlsnaes W, Risse T, Simmons BA, editors. Handbook of International Relations. London.: Sage; 2002 Fehl C. Explaining the international criminal court: a ‘practice test’ for rationalist and constructivist approaches. Eur J Int Relat 2004;10(3):357–94. Few R. Flooding, vulnerability and coping mechanisms: local responses to a global threat. Prog Dev Stud 2003;3(1):43–58. Finucane ML, Alhakami A, Slovic P, Johnson SM. The affect heuristic in judgments of risks and benefits. J Behav Decis Mak 2000;13:1–17. Fischhoff B, Slovic P, Lichtenstein S, Read S, Combs B. How safe is safe enough? A psychometric study of attitudes towards technological risks and benefits. Policy Sci 1978;9: 127–52. Folke C. Resilience: the emergence of a perspective for social–ecological systems analyses. Glob Environ Chang 2006;16(3):253–67. Folke C, Carpenter SR, Walker B, Scheffer M, Chapin T, Rockström J. Resilience thinking: integrating resilience, adaptability and transformability. Ecol Soc 2010;15(4):20. Grothmann T, Reusswig F. People at risk of flooding: why some residents take precautionary action while others do not. Nat Hazards 2006;38:101–20. Haasnoot M, Kwakkel JH, Walker WE, ter Maat J. Dynamic adaptive policy pathways: a method for crafting robust decisions for a deeply uncertain world. Glob Environ Chang 2013;23:485–98. Heazle M, Tangney P, Burton P, Howes M, Grant-Smith D, Reis K, et al. Mainstreaming climate change adaptation: an incremental approach to disaster risk management in Australia. Environ Sci Pol 2013;33:162–70. Jaeger C, Renn O, Rosa E, Webler T. Risk, uncertainty, and rational action. London: Earthscan Publications; 2001. Johnson EJ, Tversky A. Affect, generalization, and the perception of risk. J Pers Soc Psychol 1983;45:20–31. Johnson JG, Wilke A, Weber EU. Beyond a trait view of risk-taking: a domain- specific scale measuring risk perceptions, expected benefits, and perceived-risk attitude in German-speaking populations. Pol Psychol Bull 2004;35:153–72. Kates RW. Perceptual regions and regional perception in flood plain management. Pap Proc Reg Sci Assoc 1963;11:217–27. Kates RW. Hazard and choice perception in floodplain management: Department of Geography Research Paper No. 78. Chicago: University of Chicago Press; 1964. Kates RW. Natural hazard in human ecological perspective: hypotheses and models. Econ Geogr 1971;47:438–51. Kates RW, Kasperson JX. Comparative risk analysis of technological hazards: a review. Proc Natl Acad Sci 1983;80:7027. Kellens W, Terpstra T, De Maeyer P. Perception and communication of flood risks: a systematic review of empirical research. Risk Anal 2013;33(1):24–49. King D, MacGregor C. Using social indicators to measure community vulnerability to natural hazards. Aust J Emerg Manag 2000;15:52–5. Koerth J, Vafeidis AT, Hinkel J, Sterr H. What motivates coastal households to adapt pro-actively to sea-level rise and increasing flood risk? Reg Environ Chang 2013;13(4):897–909. Kuhlicke C, Steinführer A. Searching for resilience or building social capacities for flood risks? Planning Theory Pract 2013;14/1:114–20. Kuhlicke Christian, Steinführer Annett, Begg Chloe, Bianchizza Chiara, Bründl Michael, Buchecker Matthias. Perspectives on social capacity building for natural hazards: outlining an emerging field of research and practice in Europe. Environ Sci Pol 2011;14:804–14.

20

S. Birkholz et al. / Science of the Total Environment 478 (2014) 12–20

Kyle AS, Ou-Yang H, Xiong W. Prospect theory and liquidation decisions. J Econ Theory 2006;129(1):273–88. López-Marrero T. An integrative approach to study and promote natural hazards adaptive capacity. Geogr J 2010;176(2):150–63. Ludy J, Kondolf GM. Flood risk perception in lands ‘protected’ by 100-year levees. Nat Hazards 2012;61(2):829–42. McEwen LJ, Krause F, Jones O, Garde Hansen J. Sustainable flood memories, informal knowledge and the development of community resilience to future flood risk. WIT Trans Ecol Environ 2012;159:253–64. Messner F, Meyer V. Flood damage, vulnerability and risk perception — challenges for flood damage research. In: Schanze J, Zeman E, Marsalek J, editors. Flood risk management — hazards, vulnerability and mitigation measures. Springer; 2006. Michel-Kerjan E, Kunreuther H. Redesigning flood insurance. Policy Forum 2011:333. [22nd July]. Mustafa D. Linking access and vulnerability: perceptions of irrigation and flood management in Pakistan. Prof Geogr 2002;54:94–105. Nelkin D. Communicating technological risk: the social construction of risk perception. Annu Rev Public Health 1989;10:95–113. Nott J. Extreme events — physical reconstruction and risk assessment. Cambridge: Cambridge University Press; 2006. Oliver-Smith A. Anthropological research on hazards and disasters. Annu Rev Anthropol 1996;25:303–28. Oltedal S, Moen BE, Klempe H, Rundmo T. Explaining risk perception: an evaluation of cultural theory. Trondheim: Norwegian University of Science and Technology; 2004. Pelling M. What determines vulnerability to floods; a case study in Georgetown, Guyana. Environ Urban 1997;9(1):203–26. Raaijmakers R, Krywkow J, van der Veen A. Flood risk perceptions and spatial multi-criteria analysis: an exploratory research for hazard mitigation. Nat Hazards 2008;46:307–22. Renn O. The role of risk perception for risk management. Reliab Eng Syst Saf 1998;59(1): 49–62. Rogers RW. A protection motivation theory of fear appeals and attitude change. J Psychol 1975;91:93–114. Schanze J. Flood risk management — a basic framework. In: Schanze J, Zeman E, Marsalek J, editors. Flood risk management — hazards, vulnerability and mitigation measures. Springer; 2006. Schelfaut K, Pannemans B, van der Craats I, Krywkow J, Mysiak J, Cools J. Bringing flood resilience into practice: the FREEMAN project. Environ Sci Pol 2011;14: 825–33. Schiermeier Q. Climate and weather: extreme measures. Nature 2011;477(7363):148–9. Schneiderbauer S, Ehrlich D. Risk, hazard and peoples' vulnerability to natural hazards: a review of definitions, concepts and data. Brussels: EC JRC; 2004.

Scott M, White I, Kuhlicke C, Steinführer A, Sultana P, Thompson P, et al. Living with flood risk. Plan Theory Pract 2013;14(1):103–6. Short JF. The social fabric of risk: toward the social transformation of risk analysis. Am Sociol Rev 1984;49:711–25. Sjöberg L. Are received risk perception models alive and well? Risk Anal 2002;22(4): 665–9. Slovic P. Perception of risk. Science 1987;236:280–5. Slovic P. The perception of risk. London: Earthscan; 2000. Slovic P, editor. The feeling of risk: new perspectives on risk perception. London: Earthscan; 2010. Slovic P, Kunreuther H, White GF. Decision processes, rationality, and adjustment to natural hazards. In: White G, editor. Natural hazards: local, national, global. Oxford: Oxford University Press; 1974. Slovic P, Finucane ML, Peters E, MacGregor DG. Risk as analysis and risk as feelings: some thoughts about affect, reason, risk and rationality. Risk Anal 2004;24(2):311–22. Smit B, Wandel J. Adaptation, adaptive capacity and vulnerability. Global Environ Chang 2006;16:282–92. Starr C. Social benefit versus technological risk. In: Chalk R, editor. Science, technology, and society: emerging relationships. Washington: American Association for the Advancement of Science; 1969. Starr C. Benefit cost studies in sociotechnical systems. Perspectives on benefit–cost decision-making, Report of a Colloquium conducted by the Committee on Public Engineering Policy. Washington DC: National Academy of Engineering; 1972. Terpstra T, Gutteling JM. Households' perceived responsibilities in flood risk management in the Netherlands. Int J Water Resour Dev 2008;24:555–65. Tierney KJ. Toward a critical sociology of risk. Sociol Forum 1999;14(2):215–42. Tversky A, Kahneman D. Judgement under uncertainty: heuristics and biases. In: Kahneman D, Slovic P, Tversky A, editors. Judgement under uncertainty: heuristics and biases. New York.: Cambridge University Press; 1982 Tversky A, Kahneman D. Extensional vs. intuitive reasoning: the conjunction fallacy in probability judgment. Psychol Rev 1983;90:293–315. Vaughan E, Seifert M. Variability in the framing of risk issues. J Soc Issues 1992;48(4): 119–35. Wachinger G, Renn O. Risk perception and natural hazards. CapHaz-Net WP3 Report. Stuttgart: DIALOGIK Non-Profit Institute for Communication and Cooperative Research; 2010 (available at: http://caphaz-net.org/outcomes-results/CapHazNet_WP3_Risk-Perception.pd). Weichselgartner J. Disaster mitigation: the concept of vulnerability revisited. Disaster Prev Manag 2001;10(2):85–94. White G. Human adjustment to floods: a geographical approach to the flood problem in the United States (Research Paper 29, Department of Geography). Chicago: University of Chicago; 1945. White R. On measurement of systematic risk [Unpublished Ph.D Dissertation]MIT; 1972.