Skip to main content
NCBI home page
Springer Nature - PMC COVID-19 Collection logoLink to Springer Nature - PMC COVID-19 Collection
. 2023 Mar 4:1–11. Online ahead of print. doi: 10.1007/s00267-023-01805-0

Health and Environmental Protective Behavioral Intentions for Reducing Harm from Water Pollutants

Grace M Little 1,✉, Patrice A Kohl 2, Chloe B Wardropper 1,3
PMCID: PMC9984752  PMID: 36869914

Abstract

Understanding what motivates people to adopt protective behaviors is important in developing effective risk messaging. Motivations may vary depending on the nature of the risk and whether it poses a personal or impersonal threat. Water pollution creates both personal (human health) and impersonal (environmental) threats, yet few studies have examined people’s motivations to protect both personal health and environmental health. Protection motivation theory (PMT) uses four key variables to predict what motivates individuals to protect themselves in relation to a perceived threat. Using data from an online survey (n = 621), we investigated the relationships between PMT variables related to health and environmental protective behavioral intentions related to toxic water pollutants among residents in Oregon, Idaho, and Washington, USA. Among PMT variables, high self-efficacy (belief in one’s own capacity to carry out certain behaviors) significantly predicted both health and environmental protective behavioral intentions for water pollutants, while perceived severity of the threat was only significant in the environmental behavioral intentions model. Perceived vulnerability and response efficacy (belief that a specific behavior will effectively mitigate the threat) were significant in both models. Education level, political affiliation, and subjective knowledge of pollutants were significant predictors of environmental protective behavioral intentions, but not health protective behavioral intentions. The results of this study suggest that when communicating environmental risks of water pollution, highlighting self-efficacy in messaging is particularly important to promote protective environmental and personal health behavior.

Keywords: Environmental protective behavior, Health protective behavior, Water pollution, Protection motivation theory, Threat appraisal, Coping appraisal

Introduction

There is growing recognition that environmental degradation and human health concerns are connected. The accumulation of pollutants in lakes, streams and groundwater, for example, not only undermines the health of our environment, but also poses a serious threat to human health (EPA 2022). Fertilizers, wastewater, atmospheric deposition, urban runoff, and many other contaminate sources pose growing contamination issues. Many water resources are no longer safe to use for drinking or recreation (e.g., fishing, swimming) (Denchak 2022). In 2015, 21 million Americans relied on community water systems that violated the United States (US) Environmental Protection Agency’s (EPA) water quality standards (Odetola et al. 2021) and, globally, an estimated 1.8 million people died due to water pollutants in 2015 (World Health Organization 2015). Institutions have limited regulatory authority and cannot eliminate the environmental and human health threats posed by water pollution. The US Clean Water Act, for example, only has authority over point source water pollution (EPA 2021). Thus, individual behavior change is needed to close the gap where policy and other institutional solutions fall short in protecting environmental and public health. In this survey-based study, we examine how well individuals’ subjective knowledge about water pollution, and their threat and coping appraisals, predict two types of behaviors: behaviors to protect personal health and behaviors aimed at protecting the environment.

Health protective behaviors refer to any behavior performed by a person to protect, promote, or maintain their health (Harris and Guten 1979). An example of a water-related health protective behavior includes flushing pipes in the morning to sweep out contaminants that may have built up overnight (Katner et al. 2018). Environmental protective behaviors include behaviors that reduce adverse impacts of human activities on the natural environment (Paço et al. 2019), such as reducing one’s use of fertilizers and plastics. To date, these two types of behaviors have largely been explored independently, but there are increasing calls to bring health and environmental perspectives closer together (Bentley 2014; Graham and White 2016). In this study we draw on protection motivation theory (PMT)—a theoretical framework widely used to understand protective behaviors—to do a side-by-side examination of what predicts behaviors aimed at protecting the environment and behaviors aimed at protecting personal health. PMT (Maddux and Rogers 1983) uses four key variables to predict protective behaviors—perceived severity and vulnerability (together, threat appraisal), and self-efficacy and response efficacy (together, coping appraisal) (Floyd et al. 2000; Maddux and Rogers 1983).

Threat appraisal variables (perceived severity and vulnerability), have been linked with protective behavior and behavioral intentions in both experimental and survey research. Perceived severity refers to the negative consequences an individual associates with a risk (Miles 2020), while perceived vulnerability refers to the likelihood the individual thinks they will be harmed by a risk (Floyd et al. 2000). With respect to the link between perceived severity and behavior, a meta-analysis of risk messaging studies found that increases in perceived severity resulted in increased acceptance of proposed adaptive behavior or intention in certain situations, such as messaging on the risks of tobacco smoke (Floyd et al. 2000). Similarly, in a survey asking respondents about lead contamination, researchers found higher perceptions of the severity of contamination exposure were associated with greater health protective behavioral intentions (Cooper et al. 2020). In a study examining motivations to engage in behaviors to prevent nearsightedness in children, perceived vulnerability was found to correlate with parents’ preventative behavioral intentions (Lwin and Saw 2007). Lwin et al. found that when parents had higher perceptions of their child’s vulnerability to becoming nearsighted, they also indicated a greater intention to engage in behaviors to prevent nearsightedness in their children. Perceived vulnerability to risks posed by environmental problems has also been found to influence actual behavior. In a study on pro-environmental behaviors in the workplace, for example, Janmaimool (2017) found that when people perceived a potential negative impact of environmental pollutants on their personal health and well-being, they were more likely to engage in recycling.

The relationship between perceptions of risk and behavior are complex. Even if you believe a risk is severe and you feel vulnerable, you might not adopt protective behaviors if you feel you have a limited ability to carry out those behaviors (perceived self-efficacy) or that those behaviors are unlikely to be effective (perceived response efficacy) (Bandura 2000; Floyd et al. 2000). PMT therefore also includes these two coping appraisal variables to better account for individuals’ decisions about protective behaviors. Many previous studies have found protective behaviors and behavioral intentions, can be predicted by perceived self-efficacy (Westcott et al. 2017; Verkoeyen and Nepal 2019) and response efficacy (Flynn et al. 1995; Scholz and Freund 2021). For example, in a study of preventative behaviors toward muscular dystrophy, response efficacy was a significant predictor in parents’ intention to comply with recommended medical treatment (Flynn et al. 1995). Self-efficacy may be particularly important in predicting preventative behaviors. Maddux and Rogers found self-efficacy to be the most powerful predictor of behavioral intentions that precede actual behavior (Maddux and Rogers 1983).

Historically, PMT was developed to understand protective health behaviors (Maddux and Rogers 1983), but more recently the theory has been used to predict non-health behaviors in environmental contexts (Bubeck et al. 2012; Floyd et al. 2000; Martin et al. 2007). But even in the context of an environmental issue, these studies tend to focus on behaviors that protect against personal risk. Martin et al. (2017), for example, found that perceived vulnerability to personal property damage due to wildfire predicted risk-mitigating behaviors among homeowners, such as removing yard debris. Many of the risk issues that threaten our personal well-being (personal risk) also pose threats that do not directly affect us (impersonal risks) (Kahlor et al. 2006). This includes many contemporary environmental issues. Water pollution, for example, can directly threaten personal health through drinking water, fish consumption and recreational actives, such as swimming. At the same time, water pollutants also harm aquatic environments and the wildlife that rely on them. In the present study, we compare the role of PMT variables in predicting protective behavioral intentions in response to a personal risk (a threat to one’s personal health) and in response to an impersonal risk (threat to the environment) in the context of water pollution.

Finally, in addition to PMT model variables, we also explore whether knowledge may play any role in predicting health and environmental protective behaviors. While there is strong scholarly interest in the relationship between knowledge and risk-related attitudes and behaviors, research results have been mixed (Allum et al. 2008; Light et al. 2022; Malka et al. 2009). Furthermore, science communication efforts predicated on the assumption that attitudes and behaviors can be changed with higher levels of knowledge have been criticized as overly simplistic (Suldovsky 2017). Researchers conducting a meta-analysis found a small positive correlation between general science attitudes and science knowledge (Allum et al. 2008), but results varied considerably when they examined the relationship between knowledge and attitudes by specific domain of science or technology. With respect to the relationship between knowledge and protective behavior, some previous research has found a positive relationship between knowledge and health protective behaviors in the context of communicable disease issues (Faasse and Newby 2020; Petrie et al. 2016). In the present study, we explore whether a relationship between knowledge and protective behaviors might also be found in the context of risks posed by water pollution.

Hypotheses and Research Question

To examine the role of PMT variables and subjective knowledge in intentions to engage in both health and environmental protective behaviors in the context of an environmental issue, we test the following two hypotheses and research question: Hypothesis 1: Health and environmental protective behavioral intentions are positively associated with a high perceived self- and response efficacy. Hypothesis 2: Health and environmental protective behavioral intentions are positively associated with a high perceived severity and vulnerability. Research Question 1: What is the relationship between health and environmental protective behavioral intentions and subjective knowledge of water pollutants?

Methods and Materials

Data Collection

The data used in this study were collected as part of a survey of residents in the Columbia River Basin of the US Pacific Northwest, where water pollutants such as mercury, DDT, PCBs, and PBDEs pose serious risks to both human health and the environment (EPA 2009). Humans and wildlife exposed to water contaminated with these toxins can suffer nervous system, kidney, liver, immune system and reproductive disorders, and cancer (SCDHEC 2019; Harada et al. 2016; EPA 2022). We chose to survey Washington, Oregon, and Idaho because they are home to the Columbia River Basin, a key area of interest to the study’s funder—the EPA’s Columbia River Basin Restoration Program. In developing the survey instrument, we adapted measures from previous research (Bockarjova and Steg 2014), and developed several new measures drawing on an EPA’s Columbia River Basin Report outlining effective behaviors that could protect both human and environmental health (EPA 2009). Three experts with survey and environmental risk background and a group of seven non-experts pretested the initial survey. Based on their feedback we revised to improve clarity and reduce measurement error. Next, we pilot tested the survey with a sample of 50 participants to assess study procedures, including sampling, recruitment, data collection, and analysis. No major changes were made based on results from the 50 pilot study respondents and these 50 initial respondents were included in the final sample (pilot demographics were: 72% female; 82% white; 72% were 25 or older; 46% were from Washington, 38% were from Oregon, and 16% from Idaho; 36% held a bachelor’s degree or higher; 36% leaned liberal; and 6% were vegetarians).

We distributed the final survey instrument online from December 2021 to January 2022. Respondents for both the pilot test and the final survey were sampled from a Qualtrics opt-in panel, a pool of respondents who voluntarily sign up to be solicited for survey participation. Eligible respondents were those at least 18 years of age and residing in Idaho, Oregon, and Washington. In order to create a sample similar to the demographics of the region, we employed quotas for age (18–24 (32%), 25–54 (34%), 55+ (35%), gender (male 50%, female 50%), and state (Idaho (20%), Washington (50%), and Oregon (30%) based on regional proportions. Study procedures were approved and certified exempt by the University of Idaho Institutional Review Board (IRB Protocol #20-186), which assesses human subjects protocols for compliance with ethics and informed consent rules. We had an incidence rate of 31%, meaning that out of all the people who entered the survey, 31% of them were eligible respondents who were able to complete it. Our final number of respondents was 621, so we can estimate ~2003 entrants to the survey in total, the majority of which were terminated. It is standard not to calculate a response rate for opt-in panels like ours (Callegro and DiSogra 2009).

Measures

The primary study variables were measured using a five-point unipolar scale with response labels tailored to each item. Using unipolar scales avoids forcing respondents to choose between contrasting concepts (i.e., agree and disagree) (Alwin et al. 2018). We used a five-point scale because studies suggest that it can result in higher response quality than seven- or eleven- point scales, can minimize respondent burden, and is most appropriate for use with unipolar response categories (Krosnick 2018). This analysis included the survey variables described below.

Perceived self-efficacy

Our study included two perceived self-efficacy variables adapted from measures developed and tested by Bockarjova and Steg (2014)—one focused on environment and a second focused on personal health. Participants responded to the items used to measure these two variables on a five-point scale (1 = not at all confident; 5 = extremely confident). With respect to the environment, we measured perceived response efficacy by asking participants “How confident do you feel in your ability to take any kind of individual action to prevent toxic water pollutants from entering waterways?” (M = 3.9, SD = 0.99). With respect to personal health, we measured perceived response efficacy by asking participants “How confident do you feel in your ability to take any kind of individual action to protect your health from toxic water pollutants?” (M = 3.9, SD = 1.03).

Perceived response efficacy

Our study included two perceived response efficacy variables, also adapted from Bockarjova and Steg (2014)—one focused on environment and one focused on human health. Participants responded to the items used to measure these two variables on a five-point scale (1 = not at all effective; 5 = extremely effective). With respect to the environment, we measured perceived response efficacy by asking participants “How effective do you think taking individual action is at reducing toxic water pollutants from entering waterways?” (M = 3.6, SD = 0.99). With respect to personal health, we measured perceived response efficacy by asking participants “How effective do you think taking individual action is at protecting your physical health from exposure to toxic water pollutants?” (M = 3.87, SD = 0.93).

Perceived severity

Our study included two measures of the perceived severity of the risks posed by toxic water pollutants adapted from (Bockarjova and Steg 2014)—one focused on environment and one focused on human health. Participants responded to items used to measure these two variables using a five-point scale (1 = Not at all severe; 5 = Extremely severe). With respect to the environment, we measured perceived severity by asking participants “How severe do you think the negative consequences of toxic water pollutants are for the health of the environment?” (M = 4.17, SD = 0.93). With respect to human health, we asked participants “How severe do you think the negative consequences of toxic water pollutants are for human health?” (M = 4.19, SD = 0.92).

Perceived vulnerability

Our study included two perceived vulnerability variables adapted from Bockarjova and Steg (2014)—one focused on environment and a second focused on personal health. Participants responded to the items used to measure these two variables using a five-point scale (1 = not at all vulnerable; 5 = extremely vulnerable). With respect to the environment, we measured perceived vulnerability as the averaged response to two items, asking participants “In your opinion, how likely is it that each condition is vulnerable to toxic water pollutants: wildlife on land, wildlife in water (M = 4.05, SD = 0.93, r2 = 0.4). With respect to personal health, we measured perceived responsibility as the response to one item, asking participants “When you consider the possibility of toxic water pollutants affecting your physical health, how vulnerable do you feel?” (M = 3.56, SD = 1.08).

Subjective knowledge

Respondents were asked to self-assess their water pollutant-related knowledge adapted from Frewer et al. (1994), Glanz et al. (1997) and Liu and Jiao (2018). We created two knowledge scores, one that measured human health knowledge related to water pollutants score and one that measured environmental knowledge related to water pollutants. The first set of items considered the effects of water pollutants on the human body, how water pollutants enter the human body, and how to prevent water pollutant exposure. Respondents answered all items on a response scale from 1 = “not at all knowledgeable” to 5 = “extremely knowledgeable” (M = 2.48, SD = 1.16, α = 0.9). An “I don’t know” option was not offered to respondents. The three items were averaged to create an individual subjective knowledge on human health score used in the health protective behavior model of the regression. The remaining three items asked specifically about several issues related to water pollutants and the environment: effects on the environment, effects on wildlife, sources of water pollutants. Respondents answered all items on a response scale from 1 = “not at all knowledgeable” to 5 = “extremely knowledgeable” (M = 2.6, SD = 1.14, α = 0.93). The three items were averaged to create an individual subjective knowledge on the environment score used in the environmental protective behavior model.

Behavioral intention

Respondents were asked to rank their likelihood of performing behaviors that protect the environment and protect their physical health. For environmental protective behaviors, respondents were asked “Consider the actions listed below related to preventing toxic water pollutants from entering waterways, how likely is it that you will perform these behaviors in the next year? (1) Minimize my use of plastic, (2) Dispose of cleaning products properly based on the label, (3) Encourage people I know to reduce using fertilizer on their lawn, (4) Dispose of medicine properly by taking them to a Drug Take Back Program.” Respondents answered all items on a response scale from 1 = “not at all likely”, 5 = “extremely likely” (M = 3.79, SD = 0.93, α = 0.8). The environmental protective behavior median score was used in the environmental protective behavior regression model. For health protective behaviors, respondents were asked “Consider the actions listed below related to protecting your physical health toxic water pollutants, how likely is it that you will perform these behaviors in the next year? (1) Install a water filter in your household, (2) Eat fish less frequently, (3) Flush pipes with cold water in the morning and (4) Cook with cool tap water rather than hot. Respondents answered all items on a response scale from 1 = “not at all likely”, 5 = “extremely likely” (M = 3.71, SD = 0.94, α = 0.7). The health protective behavior median score was used in the health protective behavior regression model. These behaviors were drawn from behavioral recommendations in EPA’s Columbia River Basin: State of the River Report for Toxics (2009) and from Katner et al. (2018).

Sociodemographic characteristics

Six sociodemographic items were included in the final analysis as controls due to their possible influence on perceived threat and coping appraisal (Table 1). Survey respondents reported their gender (0 = “male”, 1 = “female”), age, race and ethnicity (White, Black or African American, American Indian or Alaskan Native, Asian, Native Hawaiian or Pacific Islander, Hispanic or Latino), level of education (less than high school degree, high school graduate, some college but no degree, college degree, and advanced degree), income level (from less than $20,000 to greater than $120,000), and political ideology (ranged from −3 = “strongly liberal” to 3 = “strongly conservative”). We also asked if participants ate meat (0 = no, 1 = yes) because one of our health protective behaviors was “eat fish less frequently” so we wanted to exclude respondents for whom this behavior change was not possible. Race and ethnicity were recoded such that 0 = “White” and 1 = “non-White”. Education was recoded as high and low education, split at the median such that 0 = low education and 1 = high education. Age was measured as a continuous variable.

Table 1.

Demographics of full sample of respondents

Characteristic Sample
Mean (SD) (% (frequency))
Agea
 18–34 31.2% (194)
 35–54 33.1% (206)
 55+ 35.5% (221)
Gender
 Female 48.7% (303)
 Male 51.1% (318)
Race/ethnicity
 White 74.6% (464)
 Black or African American 5.1% (32)
 American Indian or Alaskan Native 1.8% (11)
 Hispanic or Latino 13.2% (82)
 Asian 3.2% (20)
 Native Hawaiian or Pacific Islander 1.9% (12)
Highest education
 Advanced degree 11.4% (71)
 College degree (2 or 4 year) 32.8% (204)
 Some college but no degree 29.6% (184)
 High school graduate 22.3% (139)
 Less than high school degree 3.7% (23)
Occupational status
 Working full-time 35% (218)
 Working part-time 11.3% (70)
 Student 4.5% (28)
 Unemployed 11.4% (41)
 Retired 24.1% (150)
 Homemaker 7.6% (47)
 Other 5.9% (37)
Approximate household income
 Less than $20,000 18.3% (114)
 $20,000–$49,999 34.9% (217)
 $50,000–$79,999 22.8% (142)
 $80,000–$99,999 8.5% (53)
 $100,000–$119,999 5.9% (37)
 $120,000 or more 9.3% (58)
Region
 Idaho 19.65% (122)
 Washington 50.24% (312)
 Oregon 30.11% (187)
Open in a new tab

We removed 37 respondents in our final analysis because they indicated that they were vegetarian

aWe show age in categories for descriptive statistics, but we treat age as continuous in analysis

Statistical Analysis

All analyses were conducted with RStudio (2022.07.1). We analyzed data with two regression models testing our hypotheses and research question. In the first regression model, we regressed five sociodemographic variables (age, gender, education, political party, and income), the four health focused PMT variables (perceived severity, perceived vulnerability, perceived self-efficacy, and perceived response efficacy), and human health subjective knowledge of water toxics on intentions to engage in health protective behaviors. For the second model, we regressed five sociodemographic variables (age, gender, education, political party, and income), the four environmental focused PMT variables (perceived severity, perceived vulnerability, perceived self-efficacy, and perceived response efficacy), and environmental subjective knowledge of water, on intentions to engage in environmental protective behaviors.

We used the lmsupport package to analyze our regression models and retrieve effect sizes (Curtin 2018). Bivariate correlations for all predictor variables were lower than 0.70, indicating that multicollinearity is likely not a major concern for subsequent model testing (Dormann et al. 2013). For each regression model, we calculated partial eta-squared (ηp2) to quantify predictor-variable effect sizes. Effect sizes with ηp2 are considered small at 0.01, medium at 0.09, and large at 0.25 (Tabachnick and Fidell 2007; Watson 2017).

Results

Regression Analysis

Results for both the health protective behavior regression model (Model 1) and environmental protective behavior intention regression model (Model 2) are reported in Table 2. Both models explain a similar levels of variance in our two dependent variables, behavioral intentions to protect one’s health (R2 = 0.358) and behavioral intentions to protect the environment (R2 = 0.377). Effect sizes were calculated for significant variables.

Table 2.

Summary of multiple regression models for variables predicting health and environmental protective behavior intention (n = 584)

Model Model 1
Health protective behavior intention		 Model 2
Environmental protective behavior intention
Independent variable B (SE) β ηp2 B (SE) β ηp2
Age 0.05 0.00 0.05 0.00
Gender 0.15* 0.08* 0.01 0.06 0.03
Education −0.02 −0.01 −0.09** −0.09** 0.01
Political Party 0.00 −0.01 −0.10*** −0.1*** 0.01
Income 0.04 0.02 0.00 0.00
Threat appraisal
Vulnerability 0.21*** 0.24*** 0.06 0.12** 0.11** 0.01
Severity 0.07 0.07 0.00 0.12** 0.12** 0.01
Coping appraisal
Self-efficacy 0.27*** 0.30*** 0.09 0.29*** 0.31*** 0.10
Response efficacy 0.20*** 0.20*** 0.04 0.19*** 0.20*** 0.04
Subjective knowledge 0.02 0.03 0.10** 0.12** 0.01
Adjusted R2 0.36 0.37
F for ΔR2 33.44*** 33.93**
Open in a new tab

* < 0.05; ** < 0.01; *** < 0.001

The results support Hypothesis 1, which predicted that health and environmental protective behavior intentions would be positively associated with high perceived coping appraisal variables. Higher perceived self-efficacy (β = 0.30, p < 0.001) and response efficacy (β = 0.20, p < 0.001) were both associated with health protective behavior intention. Similarly, a high perceived self-efficacy (β = 0.31, p < 0.001) and response efficacy (β = 0.20, p < 0.001) were both associated with environmental protective behavior intention. Moreover, effect size was considered medium for self-efficacy in both the health (ηp2 = 0.09) and environmental (ηp2 = 0.1) model and small for response efficacy.

The results mostly support Hypothesis 2, which predicted that high threat appraisal would be associated with environmental protective behavior intention but reject that it is associated with health protective behavior intention. Higher health protective behavioral intentions was associated higher perceived vulnerability (β = 0.24, p < 0.001), but not perceived severity. Higher environmental protective behavior intention was associated with both perceived vulnerability (β = 0.11, p < 0.01) and severity (β = 0.12 p < 0.01).

Finally, our analyses testing whether subjective knowledge would be positively associated with protective behavioral intentions (Research Question 1) revealed one significant result. In Model 2, higher subjective knowledge was associated with environmental protective behavior (β = 0.12, p < 0.001, ηp2 = 0.01). We did not find the same link between subjective knowledge and health protective behavior in Model 1.

Discussion

In this study we assessed the strength of the PMT model in predicting behavioral intentions aimed at both protecting personal health and protecting environmental health in the context of the issue of water toxics pollution. Our analyses revealed similar results for PMT variables in predicting behavioral intentions both for protecting health and for protecting the environment. Overall, we found more consistent results and greater effect sizes for the coping appraisal variables (perceived self-efficacy and response efficacy) in predicting behavioral intentions, than for the threat appraisal variables (perceived vulnerability and severity). In particular, perceived self-efficacy was the strongest predictor of behavioral intentions both for protecting health and protecting the environment. This result is consistent with previous scholarship suggesting self-efficacy is an especially important variable in predicting behavior within the PMT model (Floyd et al. 2000; Westcott et al. 2017; Shafiei and Maleksaeidi 2020). In an environmental behavior context, Shafiei and Maleksaeidi (2020) applied PMT to understand the motivators behind environmental behaviors and found self-efficacy to be the strongest predictor of behavior. Furthermore, in a study that assessed factors influencing farmers’ environmental behavior with respect to non-point water pollutants, self-efficacy was significant in predicting behavior intention (Wang et al. 2019). This finding is consistent with previous research (Westcott et al. 2017; Verkoeyen and Nepal 2019). Fear appeal research suggests that a greater sense of threat could increase persuasiveness to engage in risk-reducing behaviors, but only if the recipient feels capable of avoiding the threat by executing the recommended behavior (Ruiter et al. 2001; Carey and Sarma 2016). Our research and these previous studies suggest that people need to feel confident in their ability to perform behaviors; thus, instilling confidence to take action may be critical when developing communication regarding environmental threats.

Response efficacy, the second component of coping appraisal, was significant in both the environmental protective model and the health protective model. Response efficacy has been positively associated with behaviors in previous literature (Floyd et al. 2000; Janmaimool 2017). For example, in an environmental context, Westcott et al. (2017) found response efficacy as an important indicator for families to protect their homes from wildland fires. Westcott et al. (2017) advised that response efficacy could be best achieved by on-the-ground fire fighters offering advice to homeowners to instill confidence in their actions. In personal health situations, Ling et al. found that response efficacy was the strongest predictor of intention to vaccinate for seasonal influenza, noting that response efficacy may be particularly important to target because it can stop the threat from impacting the individual (Ling et al. 2019). Our findings suggest that in the case of protection from water pollutants, it may be beneficial for communications to clearly explain the mechanism of how a behavior protects a person or the environment from pollution exposure.

Threat appraisal findings shed light on the relative importance of vulnerability and perceived severity variables in responses to environmental and health threats. Our finding related to perceived vulnerability, which was significant in both models but had a higher coefficient in our health protective model, could indicate that some environmental issues need to be communicated with specific tailoring about their impact on human health for perceived vulnerability to be stimulated. In fact, many communication campaigns emphasize the public health risks that are associated with environmental issues to incite behavior change (Myers et al. 2012; Sauerborn et al. 2009). Past studies from both environmental and public health literatures have found higher perceived vulnerability explains engagement in protective behavior (Saylor et al. 2011; Janmaimool 2017; Lwin and Saw 2007). For example, in a recent study on COVID-19, researchers found that perceived vulnerability to COVID-19 was related to the use of more protective behaviors (González-Castro et al. 2021). While we did not find a significant relationship between perceived severity and protective health behaviors, our results did find a significant relationship between perceived severity and environmental protective behaviors, which is consistent with previous research findings. For instance, a study on green consumerism in response to species decline due to invasive lionfish found increasing severity messaging was effective in changing consumption behavior (Huth et al. 2018). In the absence of a communication prompt, on the topic of farmers protection of their land during drought, individuals had higher intentions to practice environmental protective behaviors when they perceived the risk of drought as severe (Keshavarz and Karami 2016). As we discussed in the literature review, threat appeals alone can be unreliable and a balanced message emphasizing both threat severity and self-efficacy may be more effective to motivate behavior.

Political affiliation and education affected environmental protective behavioral intentions, while gender affected health protective behavioral intentions. We found the more liberal the participant, the higher their intention was to participate in environmental behaviors. Studies consistently find that liberalism is positively and significantly related to environmental concern (e.g., Cruz 2017). More conservative individuals also tend to be less sensitive to diffuse threats (or threats spread out over a large area), such as the environmental threat of water pollution (Choma et al. 2013). We found a negative relationship between education and intent to participate in environmental protective behavior. These findings deviate from previous literature that finds individuals with higher education levels tend to be more environmentally friendly (Meyer 2015; Wang et al. 2022). Additionally, some researchers find that education effects are mostly driven by different levels of knowledge on environmental issues; for example, having greater knowledge about climate change, perceiving its risks to be higher, and being aware of its causes positively impact green behavior among respondents (Hoffmann and Muttarak 2020). Because we found that high subjective knowledge on water pollutants correlated with environmental protective behavior, this finding could explain why we see a deviation in education level and behavior intention. To perform environmental protective behavior, an individual’s subjective knowledge on the topic may be more important than formal education level. Gender significantly affected health protective behavior in our study. Women were more likely than men to intend to perform health protective behaviors. This finding is consistent with previous health protection literature; women are often more likely to practice behaviors intended to reduce the consequences of environmental risks that impact their health (Davidson and Freudenburg 1996; Dömötör et al. 2019). This phenomenon could be related to risk perception and the gender bias that women are socialized to be more concerned about health issues than men (Deeks et al. 2009; Siegrist et al. 2005).

Our study has a few limitations. First, despite efforts to maximize sample representativeness with demographic quotas, the survey did not utilize a random sample, and thus the findings here should not be generalized to the national population. Second, risk researchers have developed an extended PMT model that considers maladaptive response costs (meaning the time, effort, or financial costs of taking protective behaviors) (Floyd et al. 2000). We did not implement the extended PMT model for two reasons. The first was because we were assessing a diverse range of behaviors and the response costs would change dramatically from behavior to behavior. Further, if we were to ask questions for response costs and maladaptive response rewards for each behavior, that would have significantly lengthened the survey. We wanted to avoid respondent fatigue which could deteriorate the quality of data (Ben-Nun 2008). Additionally, in our Qualtrics panel we only offered male and female options to avoid small response categories, but this is a limitation because we did not encapsulate gender fluidity. Finally, in this study, we tested behavioral intention rather than actual behavior. Many scientists research the intention-behavior gap and have found one variable that influences behavioral intention is actual behavioral control (Nguyen et al. 2019; Wang and Mangmeechai 2021). We addressed the intention-behavior gap by asking self-efficacy questions to understand the individual’s confidence to perform the behavior. Further research should explore actual behavior or past behavior to determine if there are changes between behavioral intention and include the PMT extended model.

Conclusion

The PMT was developed primarily in the context of public health behavior; in our study, we compared behavioral intention concerning both behavioral intentions for health (personal) and environmental (impersonal) risks in the context of an environmental issue. Our findings contribute to both health and environmental behavior literature as we found that self-efficacy had the most substantial relationship with behavioral intention across both types of behavioral intentions. These conclusions suggest that instilling confidence in the person performing a suggested behavior is essential when developing communication tools. These findings may help policy and decision makers because they can ensure that they communicate the severity of environmental threats to garner more support—as we see that severity when discussing environmental issues is important for people to preform environmental protective behaviors. Further, policy makers can lean on the importance of self-efficacy and ensure that they instill confidence in voters to encourage them to participate in environmental protective behaviors surrounding policies, for example supporting stricter environmental regulations.

Future research might explore communication manipulations to understand the effectiveness of self-efficacy messaging to promote behavior with varying levels of threat; experimental work could further elucidate our findings. Additionally, we did not assess PMT under varying levels of threat as our study asked about relatively low threat level pollutants, though they can be of great concern at high concentrations. It would be helpful to understand how PMT acts under more acute levels of pollution threats. Finally, we found that perceived severity and self-efficacy were the most important variables in predicting environmental protective behavior intention; thus, future research should explore threat-self-efficacy messaging further as it may be the most effective way to communicate environmental risks to stimulate environmental protective behavior.

Author Contributions

GML and CBW designed and directed the study. GML processed the data, performed the analysis, drafted the manuscript, and prepared all tables. All authors discussed the results, edited, and commented on the manuscript.

Compliance with Ethical Standards

Conflict of Interest

The authors declare no competing interests.

Footnotes

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

References

  1. Allum N, Sturgis P, Tabourazi D, Brunton-Smith I. Science knowledge and attitudes across cultures: a meta-analysis. Public Underst Sci. 2008;17(1):35–54. doi: 10.1177/0963662506070159. [DOI] [Google Scholar]
  2. Alwin DF, Baumgartner EM, Beattie BA. Number of response categories and reliability in attitude measurement†. J Surv Stat Methodol. 2018;6(2):212–239. doi: 10.1093/jssam/smx025. [DOI] [Google Scholar]
  3. Bandura A (2000) Social Cognitive Theory: An Agentic Perspective. Annu Rev Psychol 52(1), 1–26. 10.1146/annurev.psych.52.1.1 [DOI] [PubMed]
  4. Ben-Nun P (2008) Respondent fatigue. In: Lavrakas PJ (ed) Encyclopedia of survey research methods. Thousand Oaks, Calif.: Sage Publications, Inc., p 743. 10.4135/9781412963947.n480
  5. Bentley M. An ecological public health approach to understanding the relationships between sustainable urban environments, public health and social equity. Health Promot Int. 2014;29(3):528–537. doi: 10.1093/heapro/dat028. [DOI] [PubMed] [Google Scholar]
  6. Bockarjova M, Steg L. Can protection motivation theory predict pro-environmental behavior? Explaining the adoption of electric vehicles in the Netherlands. Glob Environ Change. 2014;28(1):276–288. doi: 10.1016/j.gloenvcha.2014.06.010. [DOI] [Google Scholar]
  7. Bubeck P, Botzen WJW, Aerts JCJH. A review of risk perceptions and other factors that influence flood mitigation behavior. Risk Anal. 2012;32(9):1481–1495. doi: 10.1111/j.1539-6924.2011.01783.x. [DOI] [PubMed] [Google Scholar]
  8. Callegro M, DiSogra C. Computing response metrics for online panels. Public Opin Q. 2009;72(5):1008–1032. doi: 10.1093/poq/nfn065. [DOI] [Google Scholar]
  9. Carey RN, Sarma KM. Threat appeals in health communication: messages that elicit fear and enhance perceived efficacy positively impact on young male drivers. BMC Public Health. 2016;16(1):1–16. doi: 10.1186/s12889-016-3227-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Choma BL, Hanoch Y, Gummerum M, Hodson G. Relations between risk perceptions and socio-political ideology are domain- and ideology-dependent. Pers Individ Differ. 2013;54(1):29–34. doi: 10.1016/j.paid.2012.07.028. [DOI] [Google Scholar]
  11. Cooper CM, Langman JB, Sarathchandra D, Vella CA, Wardropper CB (2020) Perceived Risk and Intentions to Practice Health Protective Behaviors in a Mining-Impacted Region. Int J Environ Res Public Health 17(21):7916. 10.3390/ijerph17217916 [DOI] [PMC free article] [PubMed]
  12. Cruz SM. The relationships of political ideology and party affiliation with environmental concern: a meta-analysis. J Environ Psychol. 2017;53:81–91. doi: 10.1016/j.jenvp.2017.06.010. [DOI] [Google Scholar]
  13. Curtin J (2018) lmSupport: support for linear models. R package version 2.9.13. https://CRAN-R-project.org/package=lmSupport
  14. Davidson D, Freudenburg W (1996) Gender and environmental risk concerns.pdf. Environ Behav 302–339. 10.1177/0013916596283003
  15. Deeks A, Lombard C, Michelmore J, Teede H. The effects of gender and age on health related behaviors. BMC Public Health. 2009;9:1–8. doi: 10.1186/1471-2458-9-213. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Denchack M (2022) Water pollution: everything you need to know. NRDC. https://www.nrdc.org/stories/water-pollution-everything-you-need-know
  17. Dömötör Z, Nordin S, Witthöft M, Köteles F (2019) Modern health worries: a systematic review. J Psychosom Res 124. 10.1016/j.jpsychores.2019.109781 [DOI] [PubMed]
  18. Dormann CF, Elith J, Bacher S, Buchmann C, Carl G, Carré G, Marquéz JRG, Gruber B, Lafourcade B, Leitão PJ, Münkemüller T, McClean C, Osborne PE, Reineking B, Schröder B, Skidmore AK, Zurell D, Lautenbach S (2013) Collinearity: a review of methods to deal with it and a simulation study evaluating their performance. Ecography 36(1):27–46. 10.1111/j.1600-0587.2012.07348.x
  19. EPA (2009) Columbia River Basin: state of the river report for toxics January 2009. Environmental Protection Agency
  20. EPA (2021) Basic information about nonpoint source (NPS) pollution. EPA. https://www.epa.gov/nps/basic-information-about-nonpoint-source-nps-pollution
  21. EPA (2021) Clean Water Act. EPA. https://www.epa.gov/laws-regulations/summary-clean-water-act
  22. EPA (2022) The effects: human health. EPA. https://www.epa.gov/nutrientpollution/effects-human-health
  23. Faasse K, Newby J (2020) Public perceptions of COVID-19 in Australia: perceived risk, knowledge, health-protective behaviors, and vaccine intentions. Front Psychol 11. 10.3389/fpsyg.2020.551004 [DOI] [PMC free article] [PubMed]
  24. Floyd DL, Prentice-Dunn S, Rogers RW (2000) A meta-analysis of research on protection motivation theory. J Appl Soc Psychol 30:407–429
  25. Frewer LJ, Shepherd R, Sparks P. The interrelationship between perceived knowledge, control, and risk associated with a range of food-related hazards targeted at the individual, other people, and society. J Food Saf. 1994;14(1):19–40. doi: 10.1111/j.1745-4565.1994.tb00581.x. [DOI] [Google Scholar]
  26. Flynn MF, Lyman RD, Prentice-Dunn S (1995) Protection motivation theory and adherence to medical treatment regimens for muscular dystrophy. J Soc Clin Psychol 14(1):61–75
  27. Glanz K, Lewis FM, Rimer BK (1997) Health behavior and health education: theory, research, and practice. San Francisco, CA: Jossey-Bass
  28. González-Castro JL, Ubillos-Landa S, Puente-Martínez A, Gracia-Leiva M (2021) Perceived vulnerability and severity predict adherence to COVID-19 protection measures: the mediating role of instrumental coping. Front Psychol 12. 10.3389/fpsyg.2021.674032 [DOI] [PMC free article] [PubMed]
  29. Graham H, White PCL. Social determinants and lifestyles: integrating environmental and public health perspectives. Public Health. 2016;141:270–278. doi: 10.1016/j.puhe.2016.09.019. [DOI] [PubMed] [Google Scholar]
  30. Harada T, Takeda M, Kojima S, Tomiyama N (2016) Toxicity and carcinogenicity of dichlorodiphenyltrichloroethane (DDT). Toxicol Res 32(1):21–33. 10.5487/TR.2016.32.1.021 [DOI] [PMC free article] [PubMed]
  31. Harris DM, Guten S. Health-protective behavior: an exploratory study. J Health Soc Behav. 1979;20(1):17–29. doi: 10.2307/2136475. [DOI] [PubMed] [Google Scholar]
  32. Hoffmann R, Muttarak R (2020) Greening through schooling: Understanding the link between education and pro-environmental behavior in the Philippines. Environ Res Lett 15(1):014009. 10.1088/1748-9326/ab5ea0
  33. Huth WL, McEvoy DM, Morgan OA. Controlling an invasive species through consumption: the case of Lionfish as an impure public good. Ecol Econ. 2018;149(July):74–79. doi: 10.1016/j.ecolecon.2018.02.019. [DOI] [Google Scholar]
  34. Janmaimool P. Application of protection motivation theory to investigate sustainable waste management behaviors. Sustainability. 2017;9(7):1079. doi: 10.3390/su9071079. [DOI] [Google Scholar]
  35. Kahlor L, Dunwoody S, Griffin RJ, Neuwirth K. Seeking and processing information about impersonal risk. Sci Commun. 2006;28(2):163–194. doi: 10.1177/1075547006293916. [DOI] [Google Scholar]
  36. Katner A, Pieper K, Brown K, Lin HY, Parks J, Wang X, Hu CY, Masters S, Mielke H, Edwards M (2018) Effectiveness of prevailing flush guidelines to prevent exposure to lead in tap water. Int J Environ Res Public Health 15:7. 10.3390/ijerph15071537 [DOI] [PMC free article] [PubMed]
  37. Keshavarz M, Karami E. Farmers’ pro-environmental behavior under drought: application of protection motivation theory. J Arid Environ. 2016;127:128–136. doi: 10.1016/j.jaridenv.2015.11.010. [DOI] [Google Scholar]
  38. Krosnick JA (2018) Improving question design to maximize reliability and validity. In: Vannette DL, Krosnick JA (eds) The Palgrave handbook of survey research. London, UK: Springer International Publishing, p 95–101. 10.1007/978-3-319-54395-6_13
  39. Lacroix K, Gifford R. Psychological barriers to energy conservation behavior: the role of worldviews and climate change risk perception. Environ Behav. 2017;50(7):749–780. doi: 10.1177/0013916517715296. [DOI] [Google Scholar]
  40. Lavrakas P (2011) Respondent fatigue. Sage Research Methods. https://methods.sagepub.com/reference/encyclopedia-of-survey-research-methods/n480.xml
  41. Light N, Fernbach PM, Rabb N, Geana MV, Sloman SA. Knowledge overconfidence is associated with anti-consensus views on controversial scientific issues. Sci Adv. 2022;8(29):eabo0038. doi: 10.1126/sciadv.abo0038. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Ling M, Kothe EJ, Mullan BA. Predicting intention to receive a seasonal influenza vaccination using protection motivation theory. Soc Sci Med. 2019;233:87–92. doi: 10.1016/j.socscimed.2019.06.002. [DOI] [PubMed] [Google Scholar]
  43. Liu T, Jiao H. How does information affect fire risk reduction behaviors? Mediating effects of cognitive processes and subjective knowledge. Nat Hazards. 2018;90(3):1461–1483. doi: 10.1007/s11069-017-3111-0. [DOI] [Google Scholar]
  44. Lwin MO, Saw S. Protecting children from Myopia: a PMT perspective for improving health marketing communications. J Health Commun. 2007;12(3):251–268. doi: 10.1080/10810730701266299. [DOI] [PubMed] [Google Scholar]
  45. Maddux JE, Rogers RW. Protection motivation and self-efficacy: a revised theory of fear appeals and attitude change. J Exp Soc Psychol. 1983;19(5):469–479. doi: 10.1016/0022-1031(83)90023-9. [DOI] [Google Scholar]
  46. Malka A, Krosnick JA, Langer G. The association of knowledge with concern about global warming: trusted information sources shape public thinking. Risk Anal Int J. 2009;29(5):633–647. doi: 10.1111/j.1539-6924.2009.01220.x. [DOI] [PubMed] [Google Scholar]
  47. Martin IM, Bender H, Raish C. What motivates individuals to protect themselves from risks: the case of wildland fires. Risk Anal. 2007;27(4):887–900. doi: 10.1111/j.1539-6924.2007.00930.x. [DOI] [PubMed] [Google Scholar]
  48. Meyer A (2015) Does education increase pro-environmental behavior? Evidence from Europe. Ecol Econ 116:108–121. 10.1016/j.ecolecon.2015.04.018
  49. Miles A (2020) Perceived severity. Perceived Severity | Division of Cancer Control and Population Sciences (DCCPS). Retrieved March 2, 2023, from https://cancercontrol.cancer.gov/brp/research/constructs/perceived-severity
  50. Myers TA, Nisbet MC, Maibach EW, Leiserowitz AA. A public health frame arouses hopeful emotions about climate change: a letter. Clim Change. 2012;113(3–4):1105–1112. doi: 10.1007/s10584-012-0513-6. [DOI] [Google Scholar]
  51. Nguyen HV, Nguyen CH, Hoang TTB. Green consumption: closing the intention-behavior gap. Sustain Dev. 2019;27(1):118–129. doi: 10.1002/sd.1875. [DOI] [Google Scholar]
  52. Odetola L, Sills S, Morrison S. A pilot study on the feasibility of testing residential tap water in North Carolina: implications for environmental justice and health. J Expo Sci Environ Epidemiol. 2021;31(6):972–978. doi: 10.1038/s41370-021-00352-2. [DOI] [PubMed] [Google Scholar]
  53. Paço AD, Shiel C, Alves H. A new model for testing green consumer behaviour. J Clean Prod. 2019;207:998–1006. doi: 10.1016/j.jclepro.2018.10.105. [DOI] [Google Scholar]
  54. Petrie K, Faasse K, Thomas M. Public perceptions and knowledge of the Ebola Virus, willingness to vaccinate, and likely behavioral responses to an outbreak. Disaster Med Public Health Prep. 2016;10(4):674–680. doi: 10.1017/dmp.2016.67. [DOI] [PubMed] [Google Scholar]
  55. Qualtrics (2022) Simple random sampling: the ultimate guide. Qualtrics. https://www.qualtrics.com/uk/experience-management/research/simple-random-sampling/#:~:text=Simple%20Random%20Sampling%3A%20The%20Ultimate%20Guide%20%7C%20Qualtrics
  56. R Core Team . R: a language and environment for statistical computing. Vienna, Australia: R Foundation for Statistical Computing; 2021. [Google Scholar]
  57. Revelle W. psych: procedures for personality and psychological research. Evanston, Illinois, USA: Northwestern University; 2021. [Google Scholar]
  58. Ruiter RAC, Abraham C, Kok G. Scary warnings and rational precautions: a review of the psychology of fear appeals. Psychol Health. 2001;16(6):613–630. doi: 10.1080/08870440108405863. [DOI] [Google Scholar]
  59. Sauerborn R, Kjellstrom T, Nilsson M (2009) Health as a crucial driver for climate policy. Glob Health Action, 2(1). 10.3402/gha.v2i0.2104 [DOI] [PMC free article] [PubMed]
  60. Saylor A, Prokopy LS, Amberg S. What’s wrong with the tap? Examining perceptions of tap water and bottled water at Purdue University. Environ Manag. 2011;48(3):588–601. doi: 10.1007/s00267-011-9692-6. [DOI] [PubMed] [Google Scholar]
  61. SCDHEC (2019) Health risks of mercury. https://scdhec.gov/environment/your-home/mercury/health-risks-mercury
  62. Scholz U, Freund AM. Determinants of protective behaviors during a nationwide lockdown in the wake of the COVID-19 pandemic. Br J Health Psychol. 2021;26(3):935–957. doi: 10.1111/bjhp.12513. [DOI] [PMC free article] [PubMed] [Google Scholar]
  63. Shafiei A, Maleksaeidi H. Pro-environmental behavior of university students: application of protection motivation theory. Glob Ecol Conserv. 2020;22:e00908. doi: 10.1016/j.gecco.2020.e00908. [DOI] [Google Scholar]
  64. Siegrist M, Keller C, Kiers HAL. A new look at the psychometric paradigm of perception of hazards. Risk Anal. 2005;25(1):211–222. doi: 10.1111/j.0272-4332.2005.00580.x. [DOI] [PubMed] [Google Scholar]
  65. Slovic P, Fischhoff B, Lichtenstein S (1980) Judgmental biases in risk perceptions. Perceived risk: psychological factors and social implications, 1–13
  66. Suldovsky B (2017) The information deficit model and climate change communication. In: Climate Science (ed) Oxford research encyclopedia of climate science. Oxford, UK: Oxford University Press. 10.1093/acrefore/9780190228620.013.301
  67. Tabachnick BG, Fidell LS. Using multivariate statistics. Boston, Massachusetts: Pearson Education; 2007. [Google Scholar]
  68. Turns A (2022) How ‘forever chemicals’ are using marine life to travel the world by sea. The Guardian. https://www.theguardian.com/environment/2022/feb/14/forever-chemicals-pcb-pfas-use-marine-life-to-travel-world-by-sea
  69. US Census Bureau (2022) Census Bureau releases New Educational Attainment Data. Census.gov. https://www.census.gov/newsroom/press-releases/2022/educational-attainment.html
  70. Verkoeyen S, Nepal SK. Understanding scuba divers’ response to coral bleaching: an application of protection motivation theory. J Environ Manag. 2019;231:869–877. doi: 10.1016/j.jenvman.2018.10.030. [DOI] [PubMed] [Google Scholar]
  71. Wang H, Mangmeechai A. Understanding the gap between environmental intention and pro-environmental behavior towards thewaste sorting and management policy of China. Int J Environ Res Public Health. 2021;18(2):1–16. doi: 10.3390/ijerph18020757. [DOI] [PMC free article] [PubMed] [Google Scholar]
  72. Wang Q, Niu G, Gan X, Cai Q. Green returns to education: does education affect pro-environmental attitudes and behaviors in China? PLoS ONE. 2022;17(2 Feb):1–20. doi: 10.1371/journal.pone.0263383. [DOI] [PMC free article] [PubMed] [Google Scholar]
  73. Wang Y, Liang J, Yang J, Ma X, Li X, Wu J, Yang G, Ren G, Feng Y. Analysis of the environmental behavior of farmers for non-point source pollution control and management: an integration of the theory of planned behavior and the protection motivation theory. J Environ Manag. 2019;237:15–23. doi: 10.1016/j.jenvman.2019.02.070. [DOI] [PubMed] [Google Scholar]
  74. Watson P. Rules of thumb on magnitudes of effect sizes. Cambridge, United Kingdom: University of Cambridge; 2017. [Google Scholar]
  75. Westcott R, Ronan K, Bambrick H, Taylor M (2017) Expanding protection motivation theory: Investigating an application to animal owners and emergency responders in bushfire emergencies. BMC Psychol 5(1). 10.1186/s40359-017-0182-3 [DOI] [PMC free article] [PubMed]
  76. World Health Organization (2015) Drinking-water. World Health Organization. https://www.who.int/news-room/fact-sheets/detail/drinking-water

Articles from Environmental Management are provided here courtesy of Nature Publishing Group

RESOURCES