The dark side of close-ties communities: How strong social connections shape health-related risk perceptions

Abstract

This study explores public risk attitudes of environmental hazards with a focus on threats such as air, water and soil pollution or toxicants spillage. We emphasize the social dimension of risk and analyze how strong networks of social connections affect community members’ sensitivity to the health risks from pollution-type hazards. We develop a model that integrates the social dimension using aspects like social capital with a set of individual-level factors like people’s awareness of environmental issues, knowledge of threats, and experience of health problems. We test our models with US national survey data (N=1207) focusing on health hazards and integrate location-based measures of social capital. Our findings highlight the detrimental (albeit relatively small) effect of strong social networks as residents of communities with high social capital downplay the health threats from hazards and report decreased risk attitudes. We also find important role for individual-level factors led by awareness and knowledge of the threats. Our findings offer unique perspective on risk attitudes by demonstrating the potential harmful effects of strong social ties on community members’ sensitivity to health risks from pollution-type hazards. In addition, we offer evidence on factors that shape risk perceptions in the face of less common environmental hazards.

1. Introduction

A 2020 World Health Organization report estimated that environmental risks like air and water pollution cause about a quarter of all deaths and disease burden worldwide, amounting to approximately 13 million annual deaths (WHO 2020). A 2021 report called for additional research on risk perceptions to better understand how people evaluate environmental hazards (WHO 2021). The vast majority of research on risk perceptions focuses on individual-level factors and the evaluation of natural hazards (Akerlof et al. 2013; Bradley et al. 2020; Brody et al. 2008; Han et al. 2022; Kellstedt et al. 2008). Yet, hazards also include toxic pollutions or air and water contaminations. In addition, risk perceptions of such hazards are also influenced by the social environments in which people live.

In this study, we tackle those under-explored aspects of environmental hazards risk perceptions by addressing two questions. First, how do people perceive the health risks from hazards like air pollution or water and soil contamination due to toxicants spillage, waste disposal, etc.? These hazards may result from intentional or unintentional human actions and while not typically viewed as “natural hazards”, they could be a result of extreme weather events such as floods or wildfires, and their consequences may be just as bad. Second, we investigate the social dimension of risk by asking how does the social environment in which people live affects their evaluations of threats like air pollution or water contamination? Testing the role of the social context in conjunction with individual-level factors offers a more comprehensive account of the way individuals assess the health risks posed by these hazards.

We develop a model that describes public health-related risk attitudes by integrating different aspects of the social dimension with individual-level factors such as the awareness of environmental issues, understanding the threats, and multiple socio-economic characteristics. We recruit a national sample of Americans and test the model with a survey instrument that emphasizes the health aspects of environmental hazards.

The results of our empirical tests across a wide range of risk perception measures demonstrate that social conditions can have detrimental effects such that those who reside in communities that are strongly-connected tend to be less sensitive to the health implications of hazards and under-play the associated health risks. We also find evidence for the important role of individual-level factors like awareness of environmental issues, knowledge of the threats and experience of health issues and exposure to environmental threats.

This research extends existing knowledge about threat perceptions of environmental hazards. In particular, we offer evidence on social as well as individual-level drivers of health-related risk assessments facing hazards that are not necessarily natural, are under-investigated, and can be a function of human intervention and action.

2. Risk Perceptions and Environmental Hazards

The concept of risk perception describes a process in which we interpret different signals from various sources, and form subjective judgements of different threats (Slovic et al. 1982). Research on environmental risk perceptions is one of the central areas of study as perceptions are associated with individual behavior and support for mitigation policies (Goldberg et al. 2021; Leiserowitz 2006). The broader area of research explores how climate change relates to individual views of the associated risks (Gilbert and Lachlan 2023; Linden 2015; Siegrist and Árvai 2020; Thaker et al. 2020; Wachinger et al. 2013).

Studies also investigate how risk perceptions emerge due to climate change related hazards like storms and floods (Bradley et al. 2020; Brody et al. 2008; Calculli et al. 2021; Dvir et al. 2022a; Dvir et al. 2022b; Gotham et al. 2018), excessive temperatures (Bergquist and Warshaw 2019; Capstick and Pidgeon 2014; Egan and Mullin 2012), and other extreme weather events (Akerlof et al. 2013; Eck et al. 2020; Gamble et al. 2013; Han et al. 2022; Kellstedt et al. 2008).

We focus on a narrower type of environmental hazards including air pollution, water and soil contamination or toxicants release and spillage. This area of research is more limited and mostly emphasizes the clinical angle, general policy views and pro-environmental behavior (Chen et al. 2023; Chen et al. 2017; Del Ponte et al. 2022; Deng et al. 2020; Johnson 2012; Jung et al. 2022; Mao et al. 2020; Wu et al. 2023). Also, few risk perceptions studies, and especially of pollution-type hazards, account for the social dimension when explaining variations in public evaluations of such hazards. We extend the analysis of risk perceptions by emphasizing the role of the social context in shaping individuals’ assessments of the health-related threats from hazards like air and water pollution, soil contamination and toxicants release.

2.1. Pollution, toxicants and health-oriented risk perceptions

Considering the complex and multi-dimensional nature of risk perceptions (Slovic et al. 1982), scholars introduced a wide range of factors, mostly at the individual-level, that explain variations in hazards’ risk perceptions.

Awareness reflects individuals’ beliefs or general concerns about the environment and climate change. Those who display greater concerns of these issues are more aware of the threats imposed, and are likely to be more sensitive to the implications of air pollution, spillage of toxicants, etc. Within the scope of natural hazards, research has shown that a strong belief in climate change threats increase risk perceptions (Calculli et al. 2021; Lee et al. 2015; Poortinga et al. 2019; Wong-Parodi and Rubin 2022). Akerlof et al. (2015) find that awareness (strong belief in climate change and its implications) is a strong predictor of health-related risk perceptions of environmental hazards. Studies exploring the health threats of microplastics argue that awareness is critical in the formation of risk perceptions (Catarino et al. 2021; Kramm et al. 2022; Wu et al. 2023). Researchers of marine and coastal pollution view awareness as essential to risk attitudes, and increased willingness to adopt mitigating behavior (Begum et al. 2022).

Another individual-level predictor, knowledge, refers to the degree of one’s understanding of the issue and its implications, and is important in shaping her estimation of the probability and severity of risks. Knowledge has a persistent effect in models of climate change risk perceptions (Linden 2015). Recent studies find that perceived knowledge contributes to greater risk perceptions of different hazards (Dvir et al. 2022b; Thaker et al. 2023; Wong-Parodi and Rubin 2022). Yet, there is also evidence of null or negative association between knowledge and risk orientations (Kellstedt et al. 2008; Liu et a. 2014; Stoutenborough and Vedlitz 2014). In that context, scholars suggest that higher degree of knowledge reflects improved sense of control facing hazards and therefore lower reported risk perceptions. Deng et al. (2020) find such effects for risk perceptions of microsplastics in China. Despite the mixed empirical evidence, the logic of this factor is straightforward: “The more one knows about the mechanisms underlying a particular hazard, the more predictable and highly correlated their risk perceptions tend to be” (Siegrist and Árvai 2020, 2195).

Individuals’ experience of hazard events is a central factor driving risk attitudes of natural hazards such as hurricanes or temperature spikes (Bergquist and Warshaw 2019; Brody et al. 2008; Dvir et al. 2022b; Gotham et al. 2018). These effects are a function of psychological mechanisms like the availability heuristic (Gigerenzer 2008; Kahneman et al. 1982) in which individuals recall (available) memories of direct or indirect experience of extreme events like floods or tornadoes (Keller et al. 2006; Tanner and Árvai 2018).

We build on these findings and argue that individuals who experienced (or currently experiencing) health problems are likely to display heightened risk perceptions from environmental conditions. Our logic is that suffering from health issues triggers heuristics like availability with regard to one’s health, and that leads individuals to become increasingly worried about environmental hazards that may worsen their condition. Relevant work show how experience of health risks during outbreaks like the SARS (2003) or Ebola (2014) have substantially heightened concerns (Tucci et al. 2017; Wu et al. 2009). Akerlof et al. (2015) find that vulnerability to health issues increase reported risk perceptions of environmental hazards.

The last set of individual-level factors includes demographic characteristics. Research on risk attitudes of hazards like microplastics or coastal pollution show a role for age and gender (Deng et al. 2020; Heidbreder et al. 2019; Gamble et al. 2013). Race and income levels are essential factors in studies of the health aspect of environmental hazards as evidence point to higher concerns among those who are more socially vulnerable (Akerlof et al. 2015; Mao et al. 2020). Lastly, political orientation is important as studies show that liberals are more sensitive to the threats of climate change hazards (Gregersen et al. 2020).

2.2. The Social dimension of environmental hazard risks

The vast majority of research on risk perceptions facing environmental hazards discuss individual-level factors. Yet, hazards such as floods or air/water pollution usually impact whole communities. As such, the social dimension of risk is an essential (and under-explored) part of the equation. In the context of risk perceptions, the social dimension refers to the effects of the social structure in which individuals live or the role that interactions with other people play in their evaluations of risks (Joffe 2003; Kasperson et al. 1988).

Social capital refers to norms of reciprocity and trustworthiness that result from social networks and extended coordination and cooperation between individuals (Putnam 2001). It is a latent resource that individuals obtain from the extent of social relations they form within their community. Social capital consists of a cognitive dimension that is central to perceptions of support and civic engagement, trust, and social cohesion. The structural dimension refers to social networks, their strength, and how people are connected within a social environment (Harpham et al. 2002). Jointly, the cognitive aspect promotes cooperation and support among the structure of network members even without the need for rewards or outside enforcement (Nowak 2006).

In the face of community-level threats, social capital has been shown to be beneficial in emphasizing the risks of climate change-related hazards (Hao et al. 2019; Norris et al. 2008). These positive effects are also evident when studying the intentions or willingness of residents in at-risk communities to adopt pro-environmental actions (Cho and Kang 2017; Hao et al. 2020; Macias and Williams 2016).

While multiple studies that explore social capital and environmental hazards point to a positive effect, there is evidence that in some cases, strong community network may have negative effects. The “downsides” of strong community ties relate to the prevalence of narratives of resilience and inner-group uniqueness that may be communicated within close-tied communities. When this type of information is shared among community members, it fosters increased risk tolerance and lower concerns of threats which can be detrimental for vulnerable populations. Facing risks from natural hazards, Wolf et al. (2010) find lower risk perceptions of extreme heat waves among individuals who have rich and extended social networks. Similarly, Babcicky and Seebauer (2017) show that households who view their social environment as supportive and strong tend to downplay the risks from floods.

Social capital also plays a key role in the way individuals perceive their health status. While most evidence points to positive association of social capital with self-reported health status (Poortinga 2012, 2006; Subramanian 2002), strong communities may also spread health damaging information (Villalonga-Olives and Kawachi 2017).

One pathway in which social capital inhibits risk attitudes is that the provision of extended social support allows a person to believe that he or she are cared for, loved, and belongs to a network of mutual obligation (Berkman et al. 2014). This belief generates an ‘informal insurance’ for individuals (Kawachi et al. 2013) who draw upon members of their strong social networks for help in case of hazards like floods (Moore et al. 2004). Similarly, individuals may be less sensitive to the adverse health ramifications of environmental hazards when those risks are ‘shared’ by the entire social group, and a strong community would help and support each other if needed. Previous work on health behavior shows that strong community ties are associated with less preventive care for children in India (Story 2014).

Some research on the social dimension of environmental hazards combines the role of social capital with another closely-related concept, place attachment, to study risk perceptions (Wakefield et al. 2001). Place attachment describes how people are connected to a locality due to both social construction and extended physical presence (Stedman 2003). Similar to social capital, most work on environmental threats show that place attachment is associated with increased risk perception and adopting mitigating behavior (Daryanto and Song 2021). However, some evidence suggest that strong place attachment has a negative effect (Bonaiuto et al. 2016) and is associated with individuals under-estimating local environmental hazards like pollutions and contamination threats (Bickerstaff and Walker 2001; Junot et al. 2018; Venables et al. 2012). Building on the latter rational and empirical evidence, we expect strong social ties to foster these ‘informal insurance’ perspectives, and reduce concerns of environmental hazards like pollution, toxicants spillage and more.

3. Method

We test our model with survey data of a national representative sample collected between October 29^th - November 18^th, 2021. The sample is drawn from Ipsos’ web-enabled KnowlegePanel^®, a probability-based panel designed to be representative of the U.S. population. The sampling strategy uses random selection of telephone numbers and residential address (ABS method) to invite respondents to the opt-in online panel. Participants also complete demographic survey that allows for efficient sampling and weighting representativeness. Our survey median completion time was 14.58 minutes, and the completion rate of 60% yielded 1,207 responses from across the US. The survey instrument was designed with a focus on issues of environmental health hazards and thus offers appropriate data to assess questions about public attitudes in this context.³

3.1. Measures

3.1.1. Health-related risk perceptions

Our dependent variable is respondents’ attitudes of health-related risks from environmental hazards (Davison et al. 2021). Risk perceptions scholars established it as a multi-dimensional concept (Slovic et al. 1982) that involves several aspects including views about the vulnerability and severity of the threats as well as the likelihood of exposure or being affected by the relevant threat (Walpole and Wilson 2021; Wilson et al. 2019).

In order to capture the multi-dimensionality of the risk perceptions, we create three measures.⁴ The main indicator accounts for the vulnerability and severity aspects with an index based on responses to questions about the degree of concern for one’s health from exposure to different hazard sources, with responses fit to a 1–4 scale (“Not concerned at all” to “Very concerned”). The survey includes a total of 11 items, and we use seven items based on factor analysis and reliability tests (Cronbach’s $\alpha =0.933$). The items include sources such as drinking water, air, food and more specific sources like treated wastewater or consumer products.

A second measure accounts for the likelihood aspect of risk perceptions. We use responses to a series of questions that ask how likely it is that environmental health hazards cause or complicate different health conditions (the flu, cancer, respiratory diseases, pregnancy complications and more). We create an index measure (with responses fit to a 1–4 scale) for the likelihood that environmental hazards cause multiple health problems (all 14 items load on a single factor, Cronbach’s $\alpha =0.955$).

Third, we include a broader measure of views about health risks from environmental hazards using survey item that asks: “How concerned are you about the public issue of local environmental health threats?”. Responses fit to a 1–4 scale (“Not concerned at all” to “Very concerned”).

3.1.2. Social Dimension

We use two measures for the main independent variable of social context. First, we integrate to our survey data county-level social capital measures from the Social Capital Project, a US government initiative to create updated state and county-level indicators of social capital with multiple subcomponents that capture the many aspects of social capital.⁵ We use this data source for two reasons: first, it is relatively updated (2018). Second, it includes a specific county-level health-focused sub-index which offers a good fit with aspects of community-level social capital and how strong connections are fostered among members.

Our Social capital measure is the community health sub-index that combines information about the number of non-profit organizations, religious congregations and informal civil society at the county level.⁶ Higher index values describe strong community ties; thus, we expect it to display a negative association with concerns from environmental threats. Second, Community duration is a survey indicator of how many months respondents have lived in their communities (mean duration = 260). We expect those that have longer duration to form stronger social connections and thus higher place attachment.

3.1.3. Awareness of hazards

Awareness captures respondents’ beliefs or general concerns about the environment. The survey item asks: “How concerned are you about the following public issue: The environment?” Responses are measured on a 1–4 scale (“Not concerned at all” to “Very concerned”).

3.1.4. Hazards Knowledge

We measure respondents’ degree of knowledge using both perceived and factual indicators. Perceived knowledge is an index measure of eight items (Cronbach’s $\alpha =0.91$) asking respondents about their degree of knowledge of several environmental threats (air pollution, food and soil contaminants, chemicals in consumer products). Responses range from “Not knowledgeable at all” to “Very knowledgeable”. Factual knowledge is based on four items that assess how familiar are respondents with environmental threats and their consequences. We code each correct response as 1 and incorrect or don’t know responses as 0. Then, we tally the four items to create a score for each respondent.

3.1.5. Health Experience

We use two variables to measure experience. Health Experience is an item asking respondents to rate the health conditions in their household (“1-Excellent” to “5-Poor”). We also use a more direct indicator, Chronic health, asking whether someone in respondents’ household has been diagnosed with a chronic health condition (binary indicator “Yes” or “No”).

3.1.6. Structural environment

Research on risk perceptions suggest that the physical environment represents an important factor that contributes to variations in attitudes. For instance, heightened risk perceptions are associated with living in close proximity to coastal flood zones (Brody et al. 2008) or urban areas with concentrated air pollutants (Li et al. 2021). We account for this issue with relevant EPA information on residents’ potential exposure to different hazards. Waste Proximity is a county-level measure of the number of hazard waste facilities within 5 km, each divided by the distance in kilometers.⁷ Higher values of this measure correspond with increased likelihood of exposure to hazard waste and thus higher (objective) health threat. We use this objective measure to complement the survey items and present a more comprehensive view of respondents’ attitudes about health threats from this type of environmental hazards.

3.1.7. Socio-Economic variables

We add measures for demographic factors including age, gender, race, income and education levels. Also, as issues of climate change and the environment became highly politicized over the last decade (Goldberg et al. 2020), we control for respondents’ political ideology (on a standard 1–7 scale).

4. Results

We begin this section with a descriptive analysis of the social capital factor and how it relates to risk perceptions. In table 1, we present the proportions of reported high-risk perceptions for each of the hazard source items (the main dependent variable).⁸ We group the respondents based on high and low values of the social capital measure and compare the proportions of reported high-risk perceptions for each group.

Table 1:

Social capital levels and high-risk perceptions proportions

	Social capital levels (county)
Source	10th percentile	25th percentile	75th percentile	90th percentile
Drinking Water	71.4%	69.4%	53.5%	55.3%
Air	74.4%	71.7%	61.9%	60.7%
Food	63.7%	62.2%	52.5%	51.2%
Agriculture	57.9%	61.1%	43.7%	37.4%
Consumer Products	67.5%	64.9%	52.5%	50.4%
Treated wastewater	62.9%	63.3%	45.3%	47.5%
Soil	58.5%	56.9%	39.2%	37.3%
Source Index means (1–4 scale)	2.83 (0.057)	2.79 (0.044)	2.5 (0.041)	2.47 (0.067)

The results in table 1 indicate that residents of communities with high social capital report relatively lower risk perceptions of all sources of hazards. The gap between high and low social capital (90^th or 75^th versus 10^th or 25^th percentiles) ranges between 10–21%. The last row in table 1 displays the means of the main risk perceptions index measure, and we also see the decrease in mean values as social capital levels increase (the differences between the relevant social capital groups are significant based on separate t-tests).

The results in table 1 point to the potential negative role of social capital in shaping risk perceptions. Across all hazard sources, residents of communities with strong inter-connections reported substantially lower proportions of high concerns compared to residents of counties with lower levels of social capital. We build on these results and conduct a more extensive test of the social capital factor and the rest of our conceptual framework using the national survey data.⁷ Table 2 presents the results of three regression models that test the relationships between our explanatory factors, led by measures of social capital, and the three measures of health-related risk perceptions. ⁸

Table 2:

Environmental Hazards Risk perception – OLS regression models

	Hazard Risk: Source Model 1 (β)	Hazard Risk: Likelihood Model 2 (β)	Hazard Risk: General Model 3 (β)
Social Capital	−0.065** (0.01)	−0.044** (0.01)	−0.04** (0.01)
Community Duration	−0.009 (0.02)	−0.003 (0.02)	−0.02 (0.02)
Awareness	0.311*** (0.02)	0.231*** (0.02)	0.487*** (0.02)
Perceived Knowledge	0.168*** (0.01)	0.115*** (0.01)	0.119*** (0.01)
Factual Knowledge	−0.003 (0.01)	0.04** (0.02)	−0.021 (0.02)
Health Experience	0.053** (0.02)	0.074** (0.02)	0.025 (0.02)
Chronic Health	0.015 (0.02)	0.029 (0.02)	0.032 (0.02)
Waste Proximity	0.008** (0.001)	0.012** (0.004)	0.017** (0.001)
Age	0.001* (0.001)	−0.004** (0.001)	0.002* (0.001)
Gender	0.096** (0.036)	0.084** (0.036)	0.083** (0.037)
Race (White)	−0.305*** (0.04)	−0.173*** (0.04)	−0.089** (0.04)
Income	−0.031** (0.012)	−0.026** (0.01)	−0.012 (0.012)
Education	−0.031 (0.02)	−0.019 (0.02)	−0.019 (0.02)
Political Ideology	−0.001 (0.01)	−0.005 (0.01)	−0.052*** (0.01)
Constant	2.87*** (0.11)	2.87*** (0.12)	2.93*** (0.12)
Observations	1120	1100	1140
R 2	0.399	0.281	0.49

Notes: Standardized coefficients reported Standard errors in parenthesis;

^*p<0.1

^**p<0.05

^***p<0.001

Model 1 focuses on risk perceptions from several sources of hazards (air, soil, wastewater, etc.), and we find that residents of counties where social connections are high report lower risk perceptions. In other words, the social capital coefficient suggests that living in a community with strong social ties has a negative effect ($\beta =-0.065,p<.001$) that reduces locals’ concerns about the potential health risks from pollution and contamination hazards. The coefficient of the second measure of social context, place attachment (duration of residence in the region/community) is negative but fails to reach statistical significance.

Model 2 explores a different angle of risk perception by measuring the question of likelihood or the degree to which residents view environmental hazards as a central cause to various health problems. The results of this model mirror model 1 with respect to the social dimension variables. In particular, the social capital indicator is negative and significant indicating that respondents who reside in communities that share many close ties assign lower likelihood to the possibility that hazards are exacerbating or causing different health problems. The coefficient of place attachment is negative but insignificant.

In model 3, we use a measure of concern with “local environmental health threats”. This variable does not describe a specific type of hazard but instead captures broader risk perceptions of the issue at-stake. The results of the social capital measure are consistent showing that living in counties (communities) where residents share more connections via community meetings and frequent interactions is associated with lower evaluations of health risks. At the same time, the duration measure does not have a significant effect.

Our results of the main explanatory factor of social dimension demonstrate that strong social connections have detrimental effects on health-related risk perceptions of environmental hazards. For the social capital factor, we find negative and significant coefficients across all three models. At the same time, the magnitude of these effects is relatively small as the predicted change in reported risk increases by 3–6% when shifting from low to high social capital levels (25^th to 75^th and 10^th to 90^th percentiles) across all measures of risk perceptions, controlling for all other factors. While these results indicate that the magnitude of the effect of social capital is limited, the consistent results of this factor across all models (and extended robustness tests) suggest that close-ties within a community can be problematic and can decrease the sensitivity of local residents to the potential risks that pollution-type hazards may represent to residents’ health.

Individual-level factors

We focus in this study on the social dimension and how factors like social capital shape risk perceptions. At the same time, our model also accounts for individual-level factors that contribute to the variations in public risk assessments of environmental hazards.

Awareness reflects individuals’ beliefs or general concerns about the environment. Across our three models, this factor is positive and statistically significant suggesting that individuals who are more concerned about environmental issues report higher health-related risk perceptions of hazards like pollution of contamination. The effects of this factor are substantial. Using model 1 coefficient, the predicted degree of concern is 14.1% higher for those who report high (75^th percentile) versus low (25^th percentile) awareness.

We also test the effects of perceived and factual knowledge. We find that perceived knowledge displays a consistent positive association with risk perceptions indicating that individuals who think they have a better understanding of the implications of pollution hazards are likely to report higher degree of concern from the associated health ramifications. On the other hand, factual knowledge is less consistent and is statistically significant (and positive) only in model 2. Considering that most studies highlight the role of factual knowledge, these results point to the possibility that for more specific issues like health threats from pollution-type hazards, it is less about the facts of what the risks are, and more about how individuals perceive their understanding of the potential risks these hazards represent to their health.

Our experience factor highlights health-related experience using a general household conditions, and specific chronic health variables. We find that respondents who face worse health conditions in their households report higher risk perceptions (models 1 and 2). Using model 1 coefficient, the predicted risk level of those who report health conditions as Excellent (2% of sample) is about 7–10% lower than those whose household health status is Poor (8% of sample) or Fair (36% of sample). In contrast to the household general health status results, direct health issues (chronic health) are not associated with public risk assessments.

We added to our models an objective measure of structural environmental conditions. The Waste Proximity variable offers an indication of potential exposure to hazard waste and thus higher health risks within a community. The coefficients for this variable across the three measures of risk perceptions (models 1–3) are positive and significant suggesting that residents of at-risk counties (due to close proximity to hazard waste facilities) are more concerned about potential health threats.

The last set of factors we test are socio-demographic factors. Across all models, women report higher risk perceptions than men. We find that age is significant and positive for models 1 and 3, indicating increased risk perceptions among older respondents (reasonable since they are more likely to encounter health problems). In model 2, age is negative suggesting older respondents do not necessarily connect the hazards to existing health conditions. Race is another significant factor. Based on model 1, the average predicted risk level among whites is about 12% lower than non-whites. We also find a limited negative effect for income as respondents with higher reported income display lower risk evaluations. Finally, political ideology matters when testing broader conceptions of risk attitudes (model 3) as the coefficient is negative and significant indicating that conservatives (as expected) are less concerned about the health risks of local environmental hazards.

5. Discussion

Our empirical analysis provides extensive evidence about the role of the social dimension in shaping individuals’ health-specific risk attitudes in the context of less common environmental hazards.

While most studies of hazards’ risk perceptions highlight individual-level factors, we emphasize the role of the social context. In studies of natural hazards risk perceptions, this factor captures social connections in the community and to one’s place of residence and is measured with survey-based indicators of social capital or place attachment. The majority of evidence points to positive association with risk attitudes (Cho and Kang 2017; Hao et al. 2020; Macias and Williams 2016). We offer a more controversial viewpoint according to which strong communal ties have a negative effect as individuals in such conditions trust their fellow community members to provide help in case of-need. As a result, they are likely to display lower health risk attitudes in the face of hazards such as air pollution or wastewater contamination. While counter-intuitive, such arguments have psychological roots (Berkman et al. 2014) and limited empirical evidence (Story 2014; Wolf et al. 2010).

Our main indicator, Social Capital, is community-based and is a function of the number of nonprofit and religious groups, civil society and extent of informal community interactions. We find that this location-based indicator has a strong negative association with all health-related risk perception measures. This suggests that individuals who have strong social networks and reside in communities that are rich in social associations, and members interact frequently in various activities, are more likely to report lower health-related concerns. Across multiple models that employ different risk perception measures (as well as alternative social capital measures), the negative association persists and corresponds with the results of studies that highlight the possible “downsides” of strong community ties (Villalonga-Olives and Kawachi 2017) in reducing concerns from environmental hazards (Babcicky and Seebauer 2017; Bonaiuto et al. 2016; Junot et al. 2018).

While we find a consistent negative association between social capital and risk perceptions, the magnitude of these effects is limited. In other words, the explanatory power of the social capital factor is comparably smaller in our models. One way to explain this outcome is that the measures we employ are not individual-based but rather a county-level aggregate or average indicators. As a result, we infer about the extent of social connections each individual respondent may have based on an aggregate measure. This is not an uncommon approach (Dvir et al. 2024; Reames et al. 2021; Zhang et al. 2020) and in fact represent an important contribution to the literature as most studies rely on survey-based self-report measures of social capital (Babcicky and Seebauer 2017; Hao et al. 2020). Nevertheless, adopting this approach does presents costs in the sense of effect size.

In the literature on the social dimension of environmental hazards risks, there is little consensus whether the effects of social ties are positive (Macias and Williams 2016) or negative (Babcicky and Seebauer 2017). In a recent meta-analysis, Daryanto and Song (2021) point to the importance of context in identifying positive or negative effects of place attachment (another element of the social context). This magnifies the importance of our findings as the social aspect is rarely studied for the type of hazards that we explore, and we point to a possible downside of extended social connections. Therefore, it may be that our results represent another important context in which the effects of the social dimension are not as expected, similar to the arguments of Daryanto and Song (2021), and warrants further research.

In the empirical analysis, we also account for multiple individual-level factors that drive risk perceptions. Across all our tests, we find that Awareness is a critical factor. Studies of climate change and natural hazards risk perceptions demonstrate that those who report general concerns about these issues are more likely to display higher risk perceptions of associated extreme weather events like excessive temperature or floods (Akerlof et al. 2015; Lee et al. 2015; Thaker et al. 2023; Linden 2015). We offer a unique insight with a focus on hazards that are not necessarily natural and can be man-made. In testing such threats, we find that awareness is strongly associated with growing risk attitudes as the projected risk perceptions of those with high awareness levels are about 14% higher than those with lower awareness across different measures of health-related risk perceptions. Our work joins studies that focus on hazards like air or marine pollution and microplastics and highlight that awareness of environmental issues is a critical factor in shaping risk orientations and behavior (Begum et al. 2022; Catarino et al. 2021; Kramm et al. 2022; Li et al. 2021).

Models that study climate change and hazards risk perceptions point to the role of knowledge in understanding the severity, and estimating the probability, of events (Dvir et al. 2022b; Kahlor et al. 2020; Mitter et al. 2019; Linden 2015). In our analysis, Perceived knowledge has a consistent effect indicating that those who believe they understand the threats from environmental hazards also report higher risk perceptions. Scholars of environmental hazards find this type of knowledge to be predictive of increased concerns (Wong-Parodi and Berlin 2022). At the same time, factual, objective knowledge is not associated with hazards risk perceptions, a null effect that corresponds with some arguments about the mixed results in the literature (Liu et al. 2014; Stoutenborough and Vedlitz 2014).

The experience of hazards is also critical in studies of environmental risk attitudes (Bergquist and Warshaw 2019; Brody et al. 2008; Gotham et al. 2018; Wachinger et al. 2013). Our view of experience is slightly different as we focus on the experience of health problems, and how it shapes evaluations of health risks. We find that experience does matter but not in the same capacity. In most of our models, having an already existing health condition increases reported risk perceptions, a result that has been documented in research on natural hazards (Akerlof et al. 2015). This may point to the limited role of experience when exploring a more unique type of hazards. The null effect of our specific measure (chronic health) further highlights this possibility. In other words, it may be that relatively few individuals draw clear connections between their pre-existing health problems and the possible health ramifications from such hazards.

A different assessment of the role of experience in this context may be of an indirect effect. We add to our data measures of relative proximity of respondents’ place of residence to hazards waste facilities based on EPA environmental indicators. The data is measured at the county level (and thus may be less accurate per respondent), yet we find that waste proximity is correlated with risk perceptions. This is a very interesting finding with several implications. First, it may represent an indirect measure of experience of such hazards and thus a better fit with how natural hazards studies conceptualize experience and health. Akerlof et al. (2015) use objective floodplain measures and find it to be positively correlated with health-related risk perceptions. Following that logic, we also find a positive association between indirect-objective experience of hazards and risk perceptions. In the discussion of experience above, we speculate that individuals may not directly link their existing health problems or vulnerability to this type of hazards. It is possible that the objective measure of proximity (at least partially) addresses this gap. A second implication of the objective measure results is that, similar to studies of natural hazards risks that add objective measures of exposure to their models (Akerlof et al. 2015; Brody et al. 2008; Li et al. 2021), our waste proximity measure demonstrates that objective conditions play an essential role when assessing different kind of environmental hazards like toxicants spillage, water contamination and air pollution.

Lastly, we test several socio-economic variables and find that women display higher risk perceptions, a result that has been documented in other studies (Akerlof et al. 2015; Brody et al. 2008; Dvir et al. 2022b; Mao et al. 2020). Race is another factor that has a consistent negative association with health attitudes (income also shows negative association, albeit the results are less consistent). This is an important finding as it suggests that weaker communities (lower income and minorities), that in many cases are more susceptible to face such threats, are also more concerned about the health ramifications of the hazards (Mao et al. 2020). Such findings are prevalent in work on health risks from natural hazards and point to the increased vulnerability of at-risk communities and how socio-economic factors reflect growing concerns from exposure to environmental hazards (Akerlof et al. 2015).

6. Conclusions

In this study, we investigate public risk perceptions in the context of environmental hazards and focus on two aspects that are under-explored in the literature. First, we explore environmental hazards such as air pollution or water and soil contamination. These may be the result of intentional or unintentional human actions. Also, while those are not natural hazards, they may result from extreme weather events such as floods, excessive heat or wildfires, and their consequences may be as severe. Second, while many studies address the individual-level determinants of risk perceptions, we place an emphasis on the social dimension by studying how the social context of individuals who face such hazards affects their evaluations of the health-related risks. Our model focuses on the social dimension by incorporating measures of social capital and we account for a host of individual-level factors like awareness of environmental issues, understanding of the threats and the experience of health problems. Using a survey instrument that emphasize the health aspect, we explore a national sample of Americans and their risk perceptions of such hazards.

The results of our study offer several contributions to research on environmental hazards and risk evaluations. First, most work on risk perceptions tend to focus on natural hazards and climate change (Gilbert and Lachlan 2023; Siegrist and Árvai 2020; Thaker et al. 2020; Linden 2015; Wachinger et al. 2013). A smaller volume of studies address the health-specific aspects of risk perceptions (Akerlof et al. 2015; Thaker et al. 2023). Our work combines the unique aspect of health-related risk perceptions with an emphasis on less common environmental hazards (air pollution or water and soil contamination). We find that the effects of most individual-level factors correspond with existing research, yet the social context actually represents a “downside” and hinders risk perceptions. These findings expand existing knowledge on risk perceptions facing a different, less-common type of hazards, and ones that may have direct health ramifications.

Second, we offer a contrasting view of the role of social connections in driving risk attitudes. Most work suggest that strong community ties increase individuals’ concerns of hazards (Cho and Kang 2017; Hao et al. 2019; Hao et al. 2020; Macias and Williams 2016). We show that for health-specific purposes, members of tight communities are likely to underestimate the risks from toxicants or pollution sources. This may be a result of residents over-reliance on the support of their social circle in case of danger, and the emergence of ‘informal insurance’ that residents believe would be available if needed (Bonaiuto et al. 2016; Junot et al. 2018; Kawachi et al. 2013). Regardless of the reason, our results highlight a “dark side” of strong social ties, especially for at-risk communities who may face dangerous health threats from air or water pollution hazards. It is important to emphasize to locals that even strong communities where residents have powerful connections with their neighbors or community should not lead them to ignore such hazards and the risks they may pose to residents’ health.

Third, research on public views of hazards like pollution or toxicants is limited (Chen et al. 2023; Chen et al. 2017; Deng et al. 2020; Johnson 2012; Jung et al. 2022). Furthermore, few studies use large-n data for quantitative analysis, and most work highlight a small number of contributing factors (Catarino et al. 2021; Kramm et al. 2022; Wu et al. 2023). We present a comprehensive framework that demonstrates the role of multiple factors at the individual as well as the social level (the community where one resides). We also incorporate objective structural elements (proximity to waste treatment facilities). The results of extended empirical tests highlight the multi-faceted nature of health-related risk perceptions in the face of less common environmental hazards.

Despite the encouraging results, our work is limited in several aspects and warrants further research. First, our main independent variable, social capital, is an aggregate level indicator. We discuss this issue at length above and recognize the limitations of assuming a certain level of social capital to all county residents. At the same time, the consistent results across multiple measures of social capital and risk perceptions are encouraging and point to the problematic role of the social context for risk attitudes. Future work should strengthen this link by integrating individual-level survey measures of social capital with such aggregate measures. Such multi-measure approach would offer a more comprehensive account of social capital and its role in shaping threat perceptions. Second, our study does not include a measure that accounts for direct exposure to hazards like pollutions and their health ramifications. While we do show that general health condition is part of the equation when it comes to risk perceptions, collecting data on direct exposure to this type of hazards can greatly enhance our findings and complement objective measures of collective exposure. Third, we recruit a representative sample of US citizens that provide us some level of general implications. Yet, replicating our work in other national contexts can add more insights on individual risk perceptions and the role of social capital that is likely to be different in other countries or regions around the globe.

Research on concerns of environmental hazards tend to pay relatively less attention to threats like pollution and contamination or the release of toxicants. Our study offers important insights about the health-related risk perceptions in the face of such hazards, and the role of individual factors, and especially the potential “downside” of a strong social networks within close-knit communities, many of which are vulnerable to the health threats that may emanate from various types of environmental hazards.