value priorities and risk perception observed at the individual level will vary across societies with different cultural emphases and development levels. Results are presented in Table 10.
Table 10
Block 2: Personal values
Incremental R2 (%) 1 4.4*** 1 0.3*** 5.7***
Block 3: Risk perception
Incremental R2 (%) --- -- 3.1***
(1) All interaction terms were entered in a single step as the last block in the model.
(2) The numerals presented here refer to before-entry beta.
(3) ST refers to self-transcendence values. SE refers to self-enhancement values. “Culture” refers to the degree to which a given society promote other-oriented values over self-directed concerns based on Schwartz’s nation scores (2005), with higher scores representing increased emphases on other-orientation. “Development” refers to a given society’s level of development based on the Human Development Index (2005).
Significant cross-level interactions were observed for all interaction terms except for two circumstances. The nature of interactions was further demonstrated
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
60
graphically through plotting techniques. As the results suggested, cultural orientations significantly moderate the effect of self-transcendence and self-enhancement values on perceived seriousness of environmental problems, but this interaction was observed only for issues broad in scope. For local risks that are directly experienced and bring immediate threats, both self-transcendence and self-enhancement values consistently raise concerns across cultural contexts (see Figure 5).
Based on the findings, the positive effect of self-transcendence was more pronounced in countries emphasizing others’ need and welfare (β=.06, p<.001). In other words, in countries where people believe in equal distribution of power and desire to maintain a harmonious relationship with their environment,
self-transcendence values appear to have a stronger effect on global risk perception, making people more likely to be concerned about broader environmental problems.
On the other hand, self-enhancement values were found to have a negative effect on people’s environmental risk perception only in cultures emphasizing other-oriented values (β=.02, p<.001). In cultures promoting self-interests,
environmental risk perception remains low, not significantly influenced by individual differences in value priorities. In other words, in countries where people are socialized to care for others, self-enhancement values distract people’s attention from risks that seem distant and broad. However, in countries where personal power and wealth are emphasized, global environmental concerns remain low. Such lower concerns for distant dangers seem largely determined by cultural forces and not significantly influenced by personal values (see Figure 5).
‧
Figure 5 Self-transcendence and self-enhancement across cultural contexts
Note.
(1) The figure explains how the slope for the relations between personal values (i.e. self-transcendence and self-enhancement) and risk perception varies across cultural contexts. Findings obtained from the individual-level analysis were confirmed, with self-transcendence being positively related to perceived environmental risks. Interaction analysis further pointed out that, when the risks at stake are global in scale, an other-oriented cultural emphasis increases the positive influence of
self-transcendence. In addition, the negative influence of self-transcendence reported at the individual level holds only in other-oriented cultures. In contrast, for cultures emphasizing self-interests, risk perception of global environmental problems remains low, regardless of individual differences in their commitment to self-enhancement values. However, for local environmental issues, cultural orientation does not have a significant moderating effect on value-perception relations.
(2) “Other” refers to the top five countries that emphasize other-oriented cultures (as measured by Schwartz’s Cultural Value Index), while “Self” refers to the bottom five countries on the same index.
(3) “Low ST” refers to the bottom 50 percent of the respondents who rate self-transcendence values (i.e. universalism and benevolence) as their primary life goals on a 10-point scale, while “High ST” refers to the top 50 percent on the same scale. Likewise, “Low SE” refers to the bottom 50 percent of the respondents who rate self-enhancement values (i.e. power, achievement, and hedonism) as their primary life goals on a 10-point scale, while “High SE” refers to the top 50 percent s on the same scale.
Likewise, self-transcendence and self-enhancement values also have different effects in countries at different stages of development. Plotting techniques reveal systematic patterns of cross-level moderation for all interaction terms, as shown in Figure 6.
‧
Figure 6 Self-transcendence and self-enhancement across development levels
Note.
(1) The figure explains how the slope for the relations between personal values and risk perception varies with levels of development and for risks located at different geographical scales. Results from moderation analysis illustrated that the effect of self-transcendence and self-enhancement greatly depends on each society’s development level. Also, patterns of interaction vary for environmental issues located at different geographical scales.
(2) “Low HDI” refers to the bottom five countries on development levels as measured by the 2005 Human Development Index (HDI), while “High HDI” refers to the top five countries on the same index.
(3) “Global risk perception” refers to the perceived seriousness of three global environmental issues on a 4-point scale (global warming, loss of biodiversity, and pollution of oceans), while “local risk perception” refers to individuals’ risk ratings of three local environmental issues (poor water quality, air quality, and poor sanitation).
2.8
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Specifically, results suggested that self-transcendence values have stronger positive effects in less developed countries, especially when the risks at stake are broad in scale, bringing no direct effects on individuals (β=.08, p<.001). For local environmental problems, risk perception remains high in poorer countries but keeps low in wealthier ones, independent of individual variation in the degree to which self-transcendence values are emphasized (β=.05, p<.001).
On the other hand, self-enhancement values decrease risk concerns only in highly developed countries and for global risks affecting the world as a whole (β=.05, p<.001). For local risks specific to one’s community, self-enhancement values
increase proximate concerns across societies at different stages of development (β=.03, p<.001). This positive impact is even stronger in less developed countries, possibly because local environmental hazards in these newly emerging economics tend to be more serious, posing direct threats to citizens’ basic survival needs (e.g., health, security), while the public in developed countries, with more advanced technology and higher budget, tend to enjoy a better living environment and are more capable of dealing with potential threats (See Figure 6).
Cross-level Moderation: Perception-Support Relations
While previous studies have empirically demonstrated that increased risk perception raises public support for environmental protection (e.g., Story & Forsyth, 2008), no studies have specifically probed into the conditions under which perceived risks most readily translate into actual behaviors aimed at making the situation less threatening. To address this gap, RQ6 explored whether cultural orientations and development levels may alter the slope for relations between risk perception and policy support. Results are presented in Table 10.
As the results suggested, global risk perception consistently increases public
‧
support across cultures (β=.08, p<.001). However, for local environmental issues, risk perception increases policy support only in cultures emphasizing power distance and mastery over nature (β=.05, p<.001). In cultures that stress others’ need and welfare, local risk perception has limited effects on policy support, which seems to be largely determined by cultural forces (see Figure 7).
Global environmental risks Local environmental risks
Figure 7 The effect of risk perception across cultural contexts
Note.
(1) The figure explains how the slope for the relations between risk perception and policy support varies across cultural contexts. Interaction analysis illustrated that risk perception does not lead to equal amount of support. The effect size depends on larger social contexts such as cultural emphases. Also, the patterns of interaction are different for hazards located at different geographical scales.
(2) “Low risk” for global environmental issues refers to the bottom half of the respondents rating the seriousness of three global environmental issues on a 4-point scale (global warming, loss of biodiversity, and pollution of oceans), while “high risk” refers to the top half of the respondents on the same index. In a similar way, “low risk” for local environmental issues refers to the bottom half of the respondents according to their perceived seriousness of three local environmental issues (poor water quality, air quality, and poor sanitation), while “high risk” refers to the top half of the participants on the same index.
In addition, perception of global environmental problems posing serious threats to the world increases willingness to support environmental policies especially in highly developed countries where citizens are comparatively better off and less
2.00
‧
struggling with basic survival needs (β=.07, p<.001). On the other hand, perception of local environmental problems posing threats to one’s own community was found to increase support in poorer countries but decrease support in more advanced ones (β=.10, p<.001). That is, in countries high on development, the more citizens perceive local environmental problems as serious, the less likely they are willing to support environmental protection in the form of paying more tax. However, in less developed countries, perceiving higher local threats makes the public more likely to support environmental policies at their own cost.
Global environmental risks Local environmental risks
Figure 8 The effect of risk perception across development levels
Note.
(1) The figure explains how the slope for the relations between risk perception and policy support varies with levels of development. Moderation analysis illustrated that risk perception does not lead to equal amount of policy support. Rather, the effect size greatly depends on each society’s level development. Also, the patterns of interaction are different for hazards located at different geographical scales.
(2) “Low HDI” refers to the bottom 5 countries on development levels as measured by the 2005 Human Development Index (HDI), while “high HDI” refers to the top 5 countries on the same index.
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
66
Discussion
The purpose of this study is to explain people’s environmental risk perception and policy support across nations. Several extensions are made to contribute to the literature. First, previous studies have shown that although a majority of people claim they support environmental protection in the abstract, the number drops to half when personal implications are made clear (Dunlap & Scarce, 1991; Gifford et al., 2009).
To address the conflicts between immediate self-interests and long-term collective good, the main cause for low public engagement as suggested by many, this study goes beyond low-cost behavioral change, focusing instead on willingness to support environmental policies in the form of paying more taxes. This particular form of engagement is important, as it involves explicit personal costs but less observable environmental impact.
Second, past scholarship has recognized that a general measure of
environmental risk perception may miss out important qualitative distinctions among risks located at different geographical scales (Dietz et al., 2005; Gökşen, et al., 2002;
Gifford et al., 2009; Lima & Castro, 2005). Fairbrother (2012), for example, noted that local and global risk perception measure different aspects of broad environmental concerns and are better treated as different cognitive constructs driven by different causal factors. Given this, however, few studies examined perception of environment risks located at different geographical scales. To address this gap, environmental issues are distinguished according to their geographical scales.
Third, most environmental studies focused on the individual as the primary level of analysis (Boeve-de Pauw & Van Petegem, 2010). However, as individuals are always born into and grow up in specific cultural contexts, a more appropriate level of analysis should be the individual within its context (Cordano et al., 2010; Deng,
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Walker, & Swinnerton, 2006; Dunlap & York, 2008; Franzen & Meyer, 2010;
Gelissen, 2007). Hence, beyond personal attributes, country-level characteristics are also incorporated as contextual influences shaping public perception and responses to environmental risks. The following discussion will highlight some important findings and practical implications.
Demographics
A number of prior studies conducted at the individual level focused their attention on socio-demographics (Boeve-de Pauw & Van Petegem, 2010; Shen &
Saijo, 2008). In general, more consistent effects were found for age and education, yet weaker and mixed results for income and gender (Kemmelmeier et al., 2002;
Marquart-Pyatt, 2012; Xiao & Dunlap, 2007). Results from this study offered additional evidence that age, education and gender differentially influence the perceived seriousness of environmental problems located at different geographical scales. Furthermore, it demonstrated that public risk perception and policy support have different social bases, driven by different sets of factors.
Age
A number of past studies suggested that age was negatively related to environmental concerns and behavior (Hersch & Viscusi, 2006; Mertig & Dunlap 2001; Zelezny et al., 2000; Theodori & Luloff, 2002; Jain & Kaur, 2006). Inglehart (1997) explained that the low environmental concerns and commitment frequently associated with older people is due to age differences in value priorities. That is, older people tend to give priority to materialist values over post-materialist ones, and it is this value difference that makes older people less environmentally concerned. Jones and Dunlap (1992) also noted that younger people are more environmentally
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
68
concerned because they are less committed to traditional and materialist values, less integrated into the dominant social order, and thereby less affected by conflicts between environmental concerns and economic interests.
Partially supporting these findings, evidence presented here shows that older people tend to perceive lower environmental risks, both for global and local
environmental problems. However, no significant relations have been found between age and support for environmental policies. The reason might be that although
younger people may judge environmental problems to be more serious than their older counterparts, they might be more constrained by their limited budget such that higher concerns do not correspond to increased support, especially when substantial material costs are required. On the contrary, older people might be more economically
better-off such that their concerns for the environment are more likely to find expressions in behavioral forms.
Supporting this, several studies found that people with higher ability to pay are more willing to sacrifice for the environment (e.g., Shen & Saijo, 2007; Kemmelmeier et al., 2002; Lazo, Kinnell, & Fisher, 2000). Brechin and Kempton (1994) also
stressed that the observed reluctance to pay for environmental protection by
socio-economically disadvantaged individuals is not due to a lack of environmental concern, but the monetary bias inherent in the tradeoff measures. Since past studies focused more on low-cost behavioral change (e.g. recycling), individual difference in the ability to pay was not taken into account as an important factor determining support.
Education
In general, most relevant studies reported a positive relationship between levels of education and environmental engagement (e.g., Fransson & Gärling, 1999;
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Olofsson & Ö hman, 2006; Theodori & Luloff, 2002; Guber, 2003; Mertig & Dunlap 2001; Barr, 2007). In support, this study cross-culturally confirmed that people with higher education tend to perceive higher environmental risk, which in turn makes them more willing to pay for the environment. However, adding new evidence to the literature, the positive impact of education holds only when the issues under
discussion were located at broader geographical scales. For local environmental threats that individuals were directly exposed to, education was not related to risk concerns.
Gelissen (2007) explained that education socializes individuals to espouse values that evoke greater commitment to the common good. This positive effect is more salient when the issues are broad in scope and abstract in nature. Franzen and Meyer (2010) noted that it is the knowledge acquired through education that enables individuals to be aware of global environmental problems that bring no tangible consequences at the present. On the other hand, for local environmental issues that are posing perceivable threats to people’s real lives, education might have limited effect on their risk judgments (Olofsson & Ö hman, 2006; Park, Russell, & Lee, 2007;
Grendstad, Selle, Strømsnes, & Bortne, 2006).
Gender
Several studies suggested that women are more environmentally concerned even after controlling for their income or education (Zelezny et al., 2000; Xiao &
McCright, 2012; Teal & Loomis, 2000). These gender differences are often attributed to a traditional socialization process (Blocker & Eckberg, 1997; Hunter, et al., 2004).
According to this perspective, women’s social roles as caregivers and nurturers make them more likely to embrace an altruistic worldview based on concern for the
maintenance of life and relationships (Hunter et al., 2004). This ‘‘motherhood
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
70
mentality’’ induces them to define themselves as embedded within a larger world and to hold a more encompassing protective attitude toward nature (Blocker & Eckberg, 1997; Davidson & Freudenburg, 1996; Hunter et al., 2004).
On the contrary, the role of men as “bread-winners’’ encourages them to be more controlling and separatist in their pursuit of economic success (Blocker &
Eckberg, 1997). This “marketplace mentality” makes them more likely to objectify the non-self and hold strong desire to dominate the environment (Teal & Loomis, 2000; Hunter et al., 2004; Gelissen, 2007). Eagly (1987) summarized this gender role difference, suggesting that women as care-givers are socialized to be more expressive, interdependent, compassionate, and cooperative, whereas men as economic providers are socialized to be more independent, competitive, and accumulative.
Contrary to this, however, this study found that men are only slightly more sensitive to global environmental risks than women, and score no significantly lower in local risk perception and policy support (also see, Hersch & Viscusi, 2006). These inconsistent findings could be explained in at least two ways. First, studies have suggested that the effect of gender was largely mediated by other socio-psychological variables, such as value priorities or affective connection with nature (Dietz, et al., 2002; Hinds & Sparks, 2008). Dietz et al. (2002), for example, explained gender differences in environmental concerns by means of women’s putting more emphasis on altruistic values than men. As a consequence, when personal values were entered into the regression model, the influence of gender was reduced to insignificance.
Contradictory findings could also be attributed to different forms of environmental behaviors that men and women typically prefer and engage in. As noted by several researchers (e.g., Blocker & Eckberg, 1997; Tindall, Davies, &
Mauboules, 2003; Hunter et al., 2004), women are more likely to engage in private (household-oriented) pro-environment behaviors (e.g., recycling, driving less, eating
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
or buying organic food, conserving energy, and reusing objects) than men, but not significantly more willing to participate in public (community/society-oriented) pro-environment behaviors (e.g., join protests, sign a petition, take part in environmentalist groups, and pay more taxes). Since the particular form of
engagement examined in this study was more public oriented, the previously observed negative relationship might not hold in this study.
Self-transcendence and Self-enhancement Values
Built on previous studies conducted at the national level (Thøgersen &
Beckmann, 1997; Nordlund & Garvill, 2002, 2003; Slimak & Dietz, 2006; Whitfield, 2009), this study cross-culturally confirmed the role of values in shaping people’s environmental risk perception and policy support. Results further demonstrated that values might have different impact across risks located at different geographical scales.
Overall, personal values are more effective in predicting risk perception of global environmental problems than that of local ones. The reasons behind might be that for environmental issues that are abstract in nature and broad in scope, potential negative impact might be delayed in time and not readily observable at the present. As
Overall, personal values are more effective in predicting risk perception of global environmental problems than that of local ones. The reasons behind might be that for environmental issues that are abstract in nature and broad in scope, potential negative impact might be delayed in time and not readily observable at the present. As