Environmental risk perception among race and ethnic groups in the United States

Transcription

1 Article Environmental risk perception among race and ethnic groups in the United States Ethnicities 2016, Vol. 16(1) ! The Author(s) 2015 Reprints and permissions: sagepub.co.uk/journalspermissions.nav DOI: / etn.sagepub.com Thomas Macias University of Vermont, USA Abstract We examine varying levels of perceived environmental risks among nine U.S. race and ethnic group categories using the 2010 General Social Survey. Three interrelated questions motivate our analysis. First, how does the environmental risk perception of air pollution, water pollution, agricultural chemicals, climate change and nuclear power generation vary among specific categories of race and ethnicity? Second, to what degree are higher levels of perceived risk among immigrant-origin groups subject to change as a result of cultural assimilation over generations? And third, to what degree are race and ethnic differences in environmental risk perception a reflection of other possibly more significant factors such as education, income, gender and conservative political views? Though we find some evidence for attitude assimilation among generations of Mexican-origin people, the dominant pattern is greater perceived risks among non-whites than whites across generations in the U.S. Additionally, not only did race and ethnic groups show greater concern for the threats posed by climate change and nuclear power generation than whites, they also expressed more concern about these two topics than they did for more localized issues such as air pollution emanating from industry and transportation. Despite our statistical controls for socioeconomic status, race and ethnic categories remained strong independent predictors of perceived environmental threats. Keywords Perceived environmental risks, US race and ethnic groups, environmental justice, attitude assimilation, air and water pollution, climate change, nuclear power Corresponding author: Thomas Macias, Department of Sociology, University of Vermont, 31 South Prospect St., Burlington, VT 05405, USA. tmacias@uvm.edu

2 112 Ethnicities 16(1) Environmental justice activists and researchers have begun to shift the nature of inquiry in the area of environmental concern from one focused on environmental protection to one of how people and communities are directly affected, in the places where they live, by environmental threats emanating from industry, transportation, and hazardous waste facilities (Bullard, 1990; Cˇ apek, 1993; Pellow and Brulle, 2005). From this perspective, questions about how often you visit national parks or contribute to environmental organizations are less important than the sense of risk one feels with relation to water and air pollution and exposure to toxic chemicals in the environment. When framed in terms of environmental risks, people of color have demonstrated higher levels of environmental concern than whites (Arp and Kenny, 1996; Bullard et al., 2007; Jones and Rainey, 2006; Mohai and Bryant, 1998; Vaughan and Nordenstam, 1991). Even when more broadly framed, African Americans in particular have been found to have at least as much, if not more, concern for the environment than whites (Jones, 1998; Jones and Carter, 1994; Uyeki and Holland, 2000). Research in this area, however, has encountered some limitations, most notably: (a) a disproportionate focus on dichotomous white versus black or white versus nonwhite outcome differences, neglecting growing diversity in the US population especially of Latino immigrant origin; (b) a relative dearth of national level studies looking beyond the context of specific communities or metro areas; and (c) many studies that do not control for income, education, gender, and political views, all likely predictors of perceived environmental risks which should be included in an analysis of specified race and ethnic correlates. Using nationally representative data culled from the 2010 General Social Survey (GSS), we attempt to address these limitations and explore the nature of greater perceived environmental risks across a spectrum of US race and ethnic groups. The GSS is a national probability sample survey conducted biannually by the National Opinion Research Center (Smith et al., 2011). Along with a standard core of demographic and attitudinal questions, the GSS contains special topic modules of interest. The 2010 GSS is particularly appropriate for this research as it contains a module on environmental attitudes, an expansive list of Hispanic origin categories, and a greater diversity of respondents than was available during earlier iterations of the environmental module. What follows is a review of the literature and its limitations, followed by a test of five hypotheses based on this review using current GSS data. Differing perceptions of environmental risks Social justice: catalyst or competitor for concern? In the early 1970s, social science research on environmental attitudes was trying to explain a relative lack of concern for the natural environment among African Americans, sometimes referred to as black disinterest in environmental issues. Taylor (1989) provides a detailed account of the social, political, and cultural explanations for this divergence. Relative to the early rumblings of the

3 Macias 113 environmental movement at the time, the Civil Rights Movement had already an established track record of high-profile activism and legislative triumphs (Gale, 1972). That, combined with the sense that pressing social issues tied to racial inequality, education, and unemployment had yet to be fully addressed, contributed to a situation in which, understandably, environmentalism was given a relatively low priority among African Americans (Mohai and Bryant, 1998). By the 1980s, however, a few researchers were beginning to doubt the significance of the concern gap between blacks and whites. The disproportionate focus on environmental group membership, visits to natural areas, and monetary contributions to environmental organizations, for example, caused some to question whether or not standard measures of environmental concern were not, in fact, primarily capturing correlates of affluence and thus underestimating concern among African Americans and other people of color (Arp and Kenny, 1996). Jones and his colleagues produced a series of articles which sought to reframe the way this problem was understood; first, by isolating a single absolute measure of environment concern asked in the GSS over the course of three decades, permitting them to track longitudinally trends in black and white concern for the environment (Jones, 1998; Jones and Carter, 1994). Despite the shortcomings of basing their analysis on a single indicator of support for environmental funding, these two studies went far in dispelling the concern gap hypothesis using national data and a number of demographic controls, including age, gender, education, income, occupational prestige, current residence, residence at age 16, and party affiliation (Jones and Carter, 1994). They demonstrate that, over time and across varying periods of prosperity and economic stagnation, African Americans had at least as much, and sometimes more, concern for the environment than do whites. Moving from the national to a local context specifically Clarksville, Tennessee Jones and Rainey (2006) continued the critique of the concern gap hypothesis including this time multi-indicator scales of environmental concern, perceived environmental health risks, and perceived environmental justice (this last measure referring to the perception of how much effort local officials put into improving neighborhood environmental conditions). Though limited to a survey of one particular environmentally aggrieved community and again limited to an analysis of black/white differences, this study does establish a link between the perception of exposure to environmental risks and concern for local environmental issues, a theoretical rationale the authors refer to as perceived differential exposure to environmental risks (Jones and Rainey, 2006: 480). In this instance, the greater perceived exposure to environmental health risks and environmental injustices are significant factors associated with greater environmental concern among blacks relative to whites. At the national level, Adeola s (2004) discriminate analysis of multiyear data pooled from the GSS support Jones et al. s findings, underscoring the point that the greater perception of environmental risks associated with industrial air pollution, pesticides, and nuclear waste among blacks is shaped by their disproportionate exposure to environmental hazards. Interestingly, this same study found

4 114 Ethnicities 16(1) common ground for blacks and whites with relation to their similar levels of concern for car pollution, water pollution, and the effects of coal and gas usage on the greenhouse effect (Adeola, 2004: 934). This suggests that perceived differential exposure is also reflected in the perception that some environmental threats are distributed more or less equally throughout the population. Latinos, immigration, and a greater diversity of attitudinal differences The past decade has seen a growing interest in capturing broader racial and ethnic diversity in the research on environmental concern. Kalof and her colleagues were the first to indicate greater pro-environmental beliefs among Latinos relative to whites using a pooled national sample, though the authors felt this particular finding needs to be reexamined using a larger sample, especially given the diversity of groups contained within the Hispanic category (Kalof et al., 2002). Since then, more regional studies within the US have produced mixed results. English-speaking Latinos in New Jersey, for example, were found to have similar levels of concern about environmental pollution in this state as whites and blacks, but more concern than Spanish-language-dominant Latinos and Asian Americans (Greenberg, 2005). A sample of attendees at a gun show in Albuquerque, New Mexico found that Latinos expressed both greater concern for the environment and a greater willingness to increase federal funding to address environmental problems relative to whites (Burger et al., 2004). In the sole longitudinal examination to date of Latino environmental concern, Whittaker et al. (2005) pooled survey data across 21 years in the state of California. Placing environmental justice at the center of their analysis, they found growing environmental concern among Latinos for issues directly affecting their communities, though not for more abstract environmental principles. An additional element of diversity suggested by the New Jersey study cited earlier and particularly relevant in a country like the United States that has experienced multiple waves of immigration is cultural assimilation. The question of assimilation s influence on environmental attitudes began to be addressed directly through national surveys such as the GSS in the early 2000s. Though still lacking sufficient diversity among its respondents to test the effect of specific ethnic backgrounds, the GSS did differentiate between US- and foreign-born respondents, as well as between immigrants who had arrived in the US before or after age 16. Using this data, Hunter (2000) found that, on the whole, immigrant environmental attitudes were similar to nonimmigrants, though those who migrated after age 16 expressed higher levels of concern and were more likely to engage in environmentally friendly behavior than their counterparts. Using the cumulative GSS data file, Adeola (2007) also found evidence of cultural assimilation in that native-born respondents as a group had significantly lower perceptions of environmental risks relative to their foreign-born counterparts. This national level research appears consistent with work in social psychology (Schultz et al., 2000) which finds language and cultural assimilation among

5 Macias 115 foreign-born Latinos to be negatively correlated with a key aspect of environmental concern: a scale measure of the New Ecological Paradigm (NEP), a worldview which sees humans as being closely tied to the natural world and governed by its rules (Dunlap and Van Liere, 1978; Dunlap et al., 2000). Another study using national data found that, though there were significant NEP differences among whites and foreign-born Latinos and blacks (whites scored higher than the other two groups) this outcome had no significant effect on the relative likelihood of different race and ethnic groups engaging in environmentally beneficial behaviors (Johnson et al., 2004). Overall research in this area is somewhat mixed, though the dominant trend among immigrant groups and their descendants is toward assimilation in environmental concern. This is consistent with the broader work on assimilation in the US which finds clear trends among immigrant groups and their descendants toward the dominant norm, especially as regards language, consumer tastes, residential patterns, and intermarriage (Alba and Nee, 1997; South et al., 2005; Waters and Jimenez, 2005). That said, the segmented assimilation hypothesis espoused by some researchers recognizes the structural constraints of socioeconomic class and race placed on many contemporary immigrant-origin groups which make their lived experience of social mobility and opportunity distinctive from that of predominately European-origin groups of an earlier era (Portes and Zhou, 1993; Rumbaut, 1997; Zhou et al., 2008). We expect that, to the extent that immigrant groups and their descendants are integrated on a differential basis into American society, there should be evidence of perceived differential exposure to environmental risks among them, as well. Other socioeconomic factors linked to risk perception and the environment Studies on the whole have found that age correlates negatively with environmental concern, though there is growing evidence that this is a cohort effect wherein more recent generations tend to be more informed and concerned about the environment than previous ones (Barr, 2007; Evans and Jacobs, 1981; Jones and Dunlap, 1992). Women are shown to have higher levels of environmental concern than men (Barr, 2007; Bickerstaff, 2004; Finucane et al., 2000) as are individuals with higher levels of educational attainment, though this latter association has changed over the years as the significance of liberal versus conservative political views has grown stronger leading a few researchers to stress the interactive nature of these variables. That is, though higher levels of educational attainment are positively associated with perceived environmental risks among liberals, the inverse tends to be true among those with extremely conservative views (Barr, 2007; Dietz et al., 2007; Hamilton, 2008). Relatedly, the last decade has seen growing research interest in the role of conservative white males in opposing efforts to address global climate change, specifically, and being relatively unconcerned about environmental risks more broadly (Finucane et al., 2000; McCright and Dunlap, 2011). Even if most conservative white males do not themselves wield very much in the way of economic or political

6 116 Ethnicities 16(1) influence, they tend to identify with conservative leaders who do and hence express opposition to attempts at environmental regulation and deny global environmental risks (McCright and Dunlap, 2013). Relative to the rest of the population, white males in the US have been shown, on average, to perceive as less risky environmental threats such as ozone depletion, climate change, and nuclear power plants (Finucane et al., 2000; Flynn et al., 1994), and the perceived benefits they receive from status quo social arrangements often mesh well with conservative politics (Kahan et al., 2007). With regard to household income, higher earners may be able to focus more energy and time on environmental issues than those who are less affluent (Inglehart, 1995; Jones and Dunlap, 1992). Another body of research, however, challenges this assertion arguing that affluent individuals with high self-perceptions of agency and power are more likely to dismiss environmental concerns and risks because they have more control in their daily lives (Bickerstaff, 2004; Kahan et al., 2005). Work in rural sociology has suggested that rural residents may be less supportive of environmental protection principally because of the economic dependency of rural communities on extractive industries (Theodori et al., 1998). Other research has found that the growth in outdoor recreation and the draw of urban denizens to rural areas has lessened the rural/urban split vis-à-vis environmental concern (Freudenburg, 1991; Lyson, 2004). Hypotheses Base on the above review of the literature, we propose the following hypotheses giving special attention to what we know about the perceived differential exposure to environmental risks among US race and ethnic groups, attitudinal changes associated with cultural assimilation, and the association of other socioeconomic factors with environmental concern: Hypothesis 1: Controlling for key socioeconomic variables, African Americans and Latinos will perceive greater risks than whites from air pollution produced by automobiles and industry given these groups concentrations in counties with poor air quality and high rates of asthma among African Americans. (Centers for Disease Control, 2011; Jones and Rainey, 2006) Hypothesis 2: Controlling for key socioeconomic variables, Mexican and Latino immigrants will perceive greater risks than whites from pesticides and chemicals used in agriculture as a result of these groups disproportionate concentrations in the agricultural sector of the economy and greater exposure to agricultural chemicals and pesticides in urban and rural areas. (McCauley et al., 2001; Mehta et al., 2000). Hypothesis 3: Controlling for key socioeconomic variables, people of color will perceive similar levels of risk to whites regarding the pollution of rivers, lakes and streams

7 Macias 117 whose impact is not seen as being concentrated among particular race and ethnic groups. (Gallup Organization, 1989; Vaughan and Nordenstam, 1991). Hypothesis 4: The perception of risk from climate change and nuclear power plants will be greater among people of color as a whole relative to whites due in part to the antithetical implications of perceiving these risks for conservative white male identity. (Finucane et al., 2000; Flynn et al., 1994; McCright and Dunlap, 2011). Hypothesis 5: Cultural assimilation over generations will be associated with a change in perceived environmental risks among immigrant-origin groups toward mean levels in the dominant white population. (Adeola, 2007; Hunter, 2000) Data and methods This study takes advantage of the 2010 GSS environmental module whose questions address established areas of interest within environmental sociology, including perceptions of environmental risks (Dietz et al., 1998; Stern et al., 1999). In total, 1430 individuals completed the environmental module; sample characteristics are shown in Table 1. Odds ratios were calculated using ordered-logit regressions of perception of environmental risk dependent variables. The six environmental risk topics are as follows: air pollution caused by cars ; air pollution caused by industry ; pesticides and chemicals used in farming ; pollution of America s rivers, lakes and streams ; a rise in the world s temperature caused by the climate change ; and nuclear power stations. Respondents were asked if they thought each of these risks was not dangerous at all, not very dangerous, somewhat dangerous, very dangerous, or extremely dangerous. Race and ethnic categories Though an expanded list of Hispanic categories has been in use on the GSS since the year 2000, it was not until the year 2010 that both the environmental module and a sufficient diversity of respondents permitted the present analysis which uses race and ethnic categories as predictors of the perception of environmental risks. The combination of race, Hispanic, and place of birth variables allowed us to generate the nine race and ethnic categories described in this section. Referred to simply as US-born white in the analysis, our reference category is non-hispanic native-born whites who make up nearly two-thirds of the sample. The second largest category is US-born Blacks also non-hispanic who make up 14% of the sample. It should be noted that cell counts for foreign-born Blacks were too small for meaningful analysis and they have thus been included in the foreign-born other race/ethnicity category. Likewise, though the data permit us to subdivide foreign-born categories into individuals who, at age 16, either resided in the US or abroad, cell counts for the four US residence categories were prohibitively low

8 118 Ethnicities 16(1) Table 1. Sample characteristics (N ¼ 1430). Variable Percentage Variable Percentage Race, Ethnicity and Place of Birth Annual Household Income US-born White 65.5 Under $20, US-born Black 14.0 $20,000 $34, US-born Mexican American 4.1 $35,000 $59, US-born Latino 2.0 $60,000 $89, US-born other race/ethnicity 2.9 $90,000 $149, Foreign-born White 3.0 $150,000 and over 5.7 Foreign-born Mexican origin 3.2 Political Views Foreign-born Latino 2.2 Extremely liberal 4.2 Foreign-born other race/ethnicity 3.2 Liberal 12.2 Age of Respondent Slightly liberal Moderate Slightly conservative Conservative Extremely conservative and over 26.1 Urban/rural Residence Gender City of 12 largest SMSAs 8.9 Male 42.5 Central city of the remaining 100 Female 57.6 largest SMSAs 12.4 Education Suburbs of the 12 largest 10.6 SMSAs Less than h.s. diploma 15.4 Suburbs of the remaining 100 High school diploma 48.7 largest SMSAs 16.2 Some college 7.6 Counties having towns Bachelor s degree 18.6 of 10,000 or more 40.8 Graduate degree 9.7 Counties having no towns of 10,000 or more 11.2 SMSA ¼ standard metropolitan statistical area. (that of Mexican immigrants in the US at age 16 was nine, for example) and this strategy was not taken. US-born Mexican Americans and Mexican immigrants make up 4 and 3% of the sample, respectively. Both US- and foreign-born Latino categories, each of which constitutes 2% of the sample, represent a vast cultural and geographic

9 Macias 119 diversity of non-mexican, Latin American-origin groups whose categories have been collapsed only because of small numbers in the sample. Likewise, the same could be said of the US- and foreign-born (non-hispanic) other race and ethnic categories which include people with origins in Asia, Africa, and other parts of the non-european world. And finally, as a point of contrast we include the category of foreign-born whites to see if migration itself is a determinant of risk perception apart from differences of race and ethnic origins. Control variables include age, gender, household income, urban residence, education, and conservative political views. Additionally, we include the interaction term conservative x education to capture the degree to which the association of education with perceived environmental risks is moderated by one s political views (Hamilton, 2008). Three of our control variables had the following percentages of missing values: Age of respondent, 0.21; household income, 12.24; conservative political views, We replaced missing values for these three variables through multiple imputations using chained equations in Stata 13. As is common practice, we did not impute scores for the dependent variable so as to reduce bias (von Hippel, 2007; White et al., 2010). Differences in N across columns in Table 2 reflect nonmissing values in the dependent variable only. Results To better specify the predictive value of race and ethnic categories in Table 2, we produce two ordered-logit regression models for each of our six dependent variables; first using only the suite of nine race and ethnic categories, and then using relevant control variables identified earlier that have been associated with environmental concern in the past. African Americans, Mexican immigrants, and other Latin American immigrants all indicate perceiving significantly greater risks from air pollution caused by cars than native-born whites. Though this link is somewhat attenuated with the inclusion of control variables in Model 2, the odds coefficients for these groups (African Americans, 1.776; Mexican immigrants, 1.868; Latin American immigrants, 2.803) remain statistically significant. Among the controls variables, higher household income predicts significantly lower odds of perceiving risk and, in a theme that persists across all but the last dependent variable, higher educational attainment is associated with a greater perception of risk. As indicated by the interaction variable, however, the strength of this last relationship is significantly moderated by the respondents political views. The models for perceived risks from air pollution produced by industry show patterns similar to those of the previous two columns, though with the addition of control variables in Model 2, the odds coefficient for African Americans is no longer statistically significant. Interestingly, while foreign-born Latinos show greater sensitivity to industrial air pollution than their US-born counterparts in Model 2, Mexican immigrants in the same column show a lower perception of risk than US-born Mexican Americans.

10 120 Ethnicities 16(1) Looking at Model 1 for the perception of risks from pesticides and other agricultural chemicals, we see that the two groups with the greatest odds ratios are Mexican immigrants (1.716) and foreign-born Latinos (2.123). With the addition of control variables in Model 2, however, the strength of these coefficients, as well as that of African Americans, is diminished and rendered statistically insignificant. The most significant effects in this column are among the control variables, especially the odds for educational attainment (1.579), gender (1.513), and the interaction variable (0.883). Model 2 for the perceived risk of pollution in the nation s rivers, lakes, and streams also shows no significantly greater or lower odds among minority groups relative to native-born whites, though the odds ratio for foreignborn Latinos (2.071) in Model 1 was statistically significant. Education and the interaction variable stand out as statistically significant variables in predicting this outcome. The most unexpected outcomes are shown in the results for the last two dependent variables of Table 2. With the exception of US-born Latinos whose sample size was relatively low, all non-white categories show significantly greater concern for climate change than native-born whites, with or without the inclusion of background control variables. The odds for perceiving climate change risk among Latin American immigrants were especially high 3.5 to 1 relative to the reference group in Model 2. The odds ratio of the interaction term (0.861) is particularly important in this model, indicating that though the odds ratio for education (1.879) is greater here than in any of the other five control models, this association is dependent on education s relationship with individual political views, i.e. the more conservative someone s views, the slighter the positive association of education with the perception of risk. The pattern of perceived risks associated with nuclear power plants is similar to that of climate change, though in Model 2 the odds ratios for foreign-born Latinos and US-born Mexican Americans are insignificant. Mexican immigrants indicate the highest levels of concern: 3.8 times more likely to perceive risks from nuclear power plants than US-born whites. Five of our seven control variables are also significant in this column. Women are more than twice as likely to perceive risks from nuclear power plants as men, and both age (0.986) and higher household income (0.876) are inversely associated with the outcome category. Conservative views (1.057) and the interaction variable (0.938) were also statistically significant in this column. Discussion We now return to the five hypotheses motivating this analysis. Hypotheses 1 and 2 reflect the perceived differential exposure argument within the environmental justice literature that people of color are more likely to be concerned about environment risks perceived as directly affecting their health and local communities (Arp and Kenny, 1996; Jones and Rainey, 2006). African

11 Macias 121 Table 2. Ordered-logit regression odds ratios for perceived environmental risks. Air pollution caused by automobiles Air pollution produced by industry Pesticides and chemicals used in agriculture Variable Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 US-born White (reference category) US-born Black 2.064*** 1.776*** 1.465** * US-born Mexican American ** 1.700* US-born Latino US-born other race/ethnicity Foreign-born Whitez Foreign-born Mexican origin 2.224** 1.868* * Foreign-born Latino 3.885*** 2.803** 3.443*** 2.674** 2.123* Foreign-born other Controls Age of respondent * Gender (1¼female) * 1.513*** Household income 0.903** 0.890** Education 1.437* 1.572*** 1.579*** Conservative Conservative X education 0.903*** 0.894*** 0.883*** Urban Chi-square 44.52*** *** 27.11*** *** 16.34* *** N (continued)

12 122 Ethnicities 16(1) Table 2. Continued. Pollution of America s rivers, lakes, and streams Temperature rise caused by climate change Nuclear power stations Variable Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 US-born White (reference category) US-born Black *** 1.490* 3.245*** 2.447*** US-born Mexican American ** 1.756* 2.147*** US-born Latino US-born other race/ethnicity *** 2.447** 2.796*** 2.167* Foreign-born White Foreign-born Mexican origin ** 2.013* 4.885*** 3.280*** Foreign-born Latino 2.071* *** 3.546*** 3.006** Foreign-born other *** 2.375** 2.023* 2.468*** Controls Age of respondent ** 0.986*** Gender (1¼female) *** Household income * 0.876*** Education 1.513*** 1.879*** Conservative *** Conservative X education 0.908*** 0.861*** 0.938* Urban Chi-square *** 58.84*** *** *** *** N *p < 0.05; **p < 0.01; ***p <

13 Macias 123 Americans, in particular, have relatively high rates of asthma in the population, and African Americans and Latinos both have disproportionate population concentrations in US counties with poor air quality (Centers for Disease Control, 2011; Jones and Rainey, 2006). Support for Hypothesis 1 is most clearly evident in outcomes for perceived risks of air pollution from automobiles among African Americans, and Latino and Mexican immigrants, and for perceived risks of air pollution from industry among African Americans, Mexican Americans and Latino immigrants, though the odds ratio of this last perceived risk for African Americans is no longer statistically significant after the inclusion of controls in Model 2. Ambiguous support for Hypothesis 2 is evident in outcomes for risks perceived from pesticides and agricultural chemicals. Here, we would expect a greater sense of risks among Latino and Mexican immigrants relative to other groups given their disproportional representation in the agricultural sector (US Department of Labour, 2005), and we see this in Model 1 of this variable. However, once controlling for background variables in Model 2, this relationship is no longer statistically significant. Thus, through our use of controls in the agricultural chemical model and, to a lesser extent in the industrial air pollution model, we can see that perceived differential exposure can, for certain risks, be largely dependent on socioeconomic factors apart from race and ethnic identity. That said, it is still the case that 90% of agricultural laborers in the US are of Mexican or other Latin American origin and a disproportionate percentage of African Americans live near polluting industries in US metropolitan areas (Bullard et al., 2007; Mehta et al., 2000). It would appear that disentangling socioeconomic characteristics from race and ethnic identity in the face of perceived environmental risk is a more straightforward exercise in ordered-logit regression than in the real world where race and ethnicity often track closely with socioeconomic status. Hypothesis 3 holds that people of color will perceive levels of risk similar to those of whites for those threats believed to be relatively equally distributed among the population. Previous research found this to be true of perceived risks from the pollution of rivers, lakes, and streams (Gallup Organization, 1989) and, once relevant background variables are controlled for, we also found no significant differences for this perceived risk among race and ethnic categories relative to the reference category of US-born whites. However, in addition to gauging the respondent s sense of environmental risks, the wording of this question in the GSS in particular, with reference to America s rivers, lakes and streams also seems to encompass national pride and the desire for pristine nature more commonly associated with preservationism than environmental justice. It would be interesting to see if reframing this topic as simply water pollution would yield different results. Hypothesis 4 considers climate change and nuclear power plants, two perceived risks which, it has been argued, are closely tied to conservative white male identity (Brody, 1984; Finucane et al., 2000; Flynn et al., 1994; Kahan et al., 2007;

14 124 Ethnicities 16(1) McCright and Dunlap, 2011). Both these risks show notable contrasts between the reference category of US-born whites and those of race and ethnic minorities. In the case of climate change, however, though the interactive odds ratio of conservative political views and education is more consequential (in this case, smaller) than that for any other perceived risk in Table 2, the odds ratio for gender is statistically insignificant. That is, for this particular perceived risk, we find evidence pointing toward a conservative white effect, the question of gender being inconsequential. Regarding the perceived risks of nuclear power plants, results are similar but different there are notable contrasts between whites and nonwhites, but now the odds ratio for gender indicates a greater than 2 1 sense of risk among women relative to men on the GSS scale. Though the interactive variable for this perceived risk has a statistically significant odds coefficient, perhaps the more remarkable outcomes are the positive relationship of conservative political views and the inverse relationship of household income with the dependent variable. That is, in the case of nuclear power plants, the data analyzed here points toward an affluent, white, liberal male effect, that being the profile least likely to perceive risk in this area. Hypothesis 5 states that cultural assimilation over generations of an immigrantorigin group in the US will be associated with changes in perceived environmental risks among that group toward the mean of the majority population. When considering cross-generational differences in environmental risk perception, the Mexican-origin population in this sample is particularly important since it is the one ethnic category for which we know all respondents share a national origin. USborn Latinos, for example, will disproportionately have ancestry from Puerto Rico and Cuba, while their foreign-born counterparts will have greater representation from Central and South America. The differences between US- and foreign-born other race and ethnic categories present similar issues. With a focus on Mexicanorigin people in Table 2, then, we see across categories a consistent pattern of assimilation such that the odds for perceived risks are lower among the nativeversus foreign-born segments of this population. The one exception to this is the air pollution caused by industry models wherein Mexican Americans perceive greater risk than Mexican immigrants. A possible explanation for this is that earlier generations of Mexican Americans were more likely to have worked or have had parents who worked in an industrial plant during the postwar era than current immigrants. Beyond this one exception, the dominant pattern of cross-generational assimilation is tempered by the still significantly greater odds of risk perception among US-born Mexican Americans, especially regarding climate change and nuclear power plants. That US-born African American, Mexican American, and other race and ethnic groups all show statistically significant greater odds of risk perception relative to US-born whites suggests that distinctive race and ethnic dispositions toward the environment persist over time and generations a segmented assimilation of perceived environmental risks among groups of non-european origin.

15 Macias 125 Though the first two dependent variables of Table 2 show clear evidence in support of the perceived differential exposure argument, the even greater disparity between whites and nonwhites in the climate change column was not anticipated. An interesting question in regard to persistent views is whether concern about climate change among people of color has increased over the past decade as high-profile disasters such as Hurricane Katrina have had disproportionate impact on the nonwhite population. We have no way of comparing the 2010 GSS data exactly to an earlier time since questions have changed and earlier samples were not as diverse. However, the year 2000 GSS did include the same question on climate change as 2010 and, though the sample sizes for other groups were too low to make a meaningful comparison, we can still compare African Americans to nativeborn whites. In that regard, we calculated with year 2000 data that 45.4% of US-born whites (n¼841) and 56.4% of African Americans (n¼156) felt that the temperature rise caused by climate change was very or extremely dangerous. These figures are very close to corresponding 2010 percentages 42.3 (n¼885) and 55.1 (n¼178), for whites and blacks, respectively suggesting the direction and spread of racial differences in attitudes toward climate change is not a recent development. Conclusion In the literature on race, ethnicity, and perceived environmental risks, there are now a number of studies that look beyond white/nonwhite dichotomies (Adeola, 2007; Burger et al., 2004; Greenberg, 2005; Kalof et al., 2002; Schultz et al., 2000; Whittaker et al., 2005), others that use national level data (Adeola, 2004, 2007; Dietz et al., 1998; Finucane et al., 2000; Hunter, 2000; Johnson et al., 2004; Jones 1998; Jones and Carter, 1994), and a few which control for an array of background demographic variables (Jones and Carter, 1994; Mohai and Bryant, 1998; Uyeki and Holland, 2000; Whittaker et al., 2005), but this is the only research in this area to our knowledge that does all three. Building from this unique configuration of data, the analysis presented here provides three essential reasons why the environmental concerns of race and ethnic minorities in a nation need to be at the core of effective strategies for socioecological change. First, it appears that environmental values among people of color persist over generations. Though there is some evidence for value assimilation over generations especially between foreign- and US-born people of Mexican origin, the more remarkable finding is persistently high levels of perceived environmental risks among US-born race and ethnic minorities, especially as regards climate change and nuclear power stations. This growing portion of the US population appears open to the message of doing something to address global environmental threats. Second, even with controls for key demographic variables, the statistically significant link between ethnic and racial categories and perceived risks persists in four out of six of our dependent variables. That is, the association of perceived environmental risks from air pollution (both from industry and automobiles), climate change, and nuclear power plants among people of color cuts across the key socioeconomic

16 126 Ethnicities 16(1) characteristics of age, gender, household income, education, political views, and rural/ urban place of residence. There are, of course, other measures of socioeconomic status not measured here such as accumulated wealth, social network advantages, and neighborhood effects, all of which would presumably play an important role in shaping environmental risk perceptions. Nonetheless, the evidence presented here suggests race and ethnic identities may serve as an effective wedge for broadening environmental concern throughout the population. And third, the fact that our statistical controls were not sufficient to explain away race and ethnic differences in perceived risks reaffirms themes environmental justice activists and researchers have asserted all along. Namely, environmental risks in America s cities, towns, and rural areas are unequally distributed on the basis of race and ethnicity and this form of inequality is quite palpable for those who live and/or work near transportation centers and roadways, industrial plants, and hazardous waste facilities. Moreover, even for those risks once thought to be more equally distributed throughout society such as climate change and nuclear power plants, this study supports previous research which finds a greater sense of risk associated with these issues among people of color than among members of the majority population. This research affirms that the movement to reduce high-risk anthropogenic impacts on the ecosystem will do well by bringing to prominence the voices and addressing the needs of the people who themselves perceive the greatest levels of environmental hazard. Funding This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors. References Adeola F (2004) Environmentalism and risk perception: Empirical analysis of black and white differentials and convergence. Society and Natural Resources 17: Adeola F (2007) Nativity and environmental risks: An empirical study of native-born residents of the USA. Research in Human Ecology 14: Alba R and Nee V (1997) Rethinking assimilation theory for a new era of immigration. International Migration Review 31: Arp W and Kenny C (1996) Black environmentalism in the local community context. Environment and Behavior 28: Barr S (2007) Factors influencing environmental attitudes and behaviors. Environment and Behavior 39: Bickerstaff K (2004) Risk perception research: Socio-cultural perspectives on the public experience of air pollution. Environment International 30: Brody CT (1984) Differences by sex in support for nuclear power. Social Forces 63: Bullard RD (1990) Dumping in Dixie: Race, Class, and Environmental Quality. Boulder, CO: Westview. Bullard RD, Mohai P, Saha R, et al. (2007) Toxic Wastes and Race at Twenty : Grassroots Struggles to Dismantle Environmental Racism in the United States. Cleveland, OH: United Church of Christ Justice and Witness Ministry.

17 Macias 127 Burger J, Myers O, Boring CS, et al. (2004) Perceptions of general environmental problems, willingness to expend federal funds on these problems, and concerns regarding the Los Alamos national laboratory: Hispanics are more concerned than whites. Environmental Research 95: Cˇ apek SM (1993) The Environmental Justice frame: A conceptual discussion and an application. Social Problems 40: Centers for Disease Control (2011) Unhealthy air quality United States, Morbidity and Mortality Weekly Report 60: Dietz T, Dan A and Shwom R (2007) Support for climate change policy: Social psychological and social structural influences. Rural Sociology 72: Dietz T, Stern PC and Guagnano GA (1998) Social structural and social psychological bases of environmental concern. Environment and Behavior 30: Dunlap RE and Van Liere KD (1978) The environmental paradigm : A proposed measuring instrument and preliminary results. Journal of Environmental Education 9: Dunlap RE, Van Liere KD, Mertig A, et al. (2000) Measuring endorsement of the new ecological paradigm: A revised NEP Scale. Journal of Social Issues 56: Evans GW and Jacobs SV (1981) Air pollution and human behavior. Journal of Social Issues 37: Finucane ML, Slovic P, Mertiz CK, et al. (2000) Gender, race, and perceived risk: The white male effect. Health, Risk and Society 2: Flynn J, Slovic P and Mertz CK (1994) Gender, race, and perception of environmental health risks. Risk Analysis 14: Freudenburg WR (1991) Rural-urban differences in environmental concern: A closer look. Sociological inquiry 61: Gale R (1972) From sit-in to hike-in: A comparison of the civil rights and environmental movements. In: Burch WR, Cheek NH, Taylor L (eds) Social Behavior, Natural Resources and the Environment. New York: Harper and Row, pp Gallup Organization (1989) General environmental concern. Gallup Report 285: Greenberg M (2005) Concern about environmental pollution: How much difference do race and ethnicity make? A New Jersey case study. Environmental Health Perspectives 113: Hamilton LC (2008) Who cares about polar regions? Results from a survey of U.S. public opinion. Arctic, Antarctic, and Alpine Research 40: Hunter L (2000) A comparison of the environmental attitudes, concern, and behaviors of native-born and foreign-born U.S. residents. Population and Environment 21: Inglehart R (1995) Public support for environmental protection: Objective problems and subjective values in 43 societies. PS: Political Science and Politics 28: Johnson C, Bowker JM and Cordell HK (2004) Ethnic variation in environmental belief and behavior: An examination of the new ecological paradigm in a social psychological context. Environment and Behavior 36: Jones RE (1998) Black concern for the environment: Myth versus reality. Society and Natural Resources 11: Jones RE and Carter LF (1994) Concern for the environment among Black Americans: An assessment of common assumptions. Social Science Quarterly 75: Jones RE and Dunlap RE (1992) The social basis of environmental concern: Have they changed over time? Rural Sociology 57:

18 128 Ethnicities 16(1) Jones RE and Rainey SA (2006) Examining linkages between race, environmental concern, health and justice in a highly polluted community of color. Journal of Black Studies 36: Kahan DM, Braman D, Gastil J, et al. (2005) Gender, Race and Risk Perception: The Influence of Cultural Status Anxiety. New Haven, CT: Yale Law SchoolPublic law and legal theory research paper no. 86. Kahan DM, Braham D, Gastil J, et al. (2007) Culture and identity-protective cognition: Explaining the white-male effect in risk perception. Journal of Empirical Legal Studies 4: Kalof L, Dietz T, Guagnano G, et al. (2002) Gender and environmentalism: The atypical values and beliefs of white men. Race, Gender and Class 9: Lyson TA (2004) Civic Agriculture: Reconnecting Farm, Food, and Community. Medford, MA: Tufts University Press. McCauley LA, Lasarev MR, Higgins G, et al. (2001) Work characteristics and pesticide exposures among migrant agricultural families: A community-based research approach. Environmental Health Perspectives 109: McCright AM and Dunlap RE (2011) Cool dudes: The denial of climate change among conservative white males in the United States. Global Environmental Change 21: McCright AM and Dunlap RE (2013) Bringing ideology in: The conservative white male effect on worry about environmental problems. Journal of Risk Research 16: Mehta K, Gabbard S, Barrat V, et al. (2000) Findings from the National Agricultural Workers Survey (NAWS) : A Demographic and Employment Profile of United States Farmworkers. Washington, DC: U.S. Department of Labor. Mohai P and Bryant B (1998) Is there a race effect on concern for environmental quality? Public Opinion Quarterly 62: Pellow D and Brulle RJ (2005) Power, justice and the environment: Toward critical environmental justice studies. In: Pellow D, Brulle RJ (eds) Power Justice and the Environment: A Critical Appraisal of the Environmental Justice Movement. Cambridge, MA: MIT Press, pp Portes A and Zhou M (1993) The new second generation: Segmented assimilation and its variants. The Annals of the American Academy of Political 530: Rumbaut RG (1997) Assimilation and its discontents: Between rhetoric and reality. International Migration Review 31: Schultz WP, Unipan JB and Gamba RJ (2000) Acculturation and ecological worldview among Latin Americans. Journal of Environmental Education 31: Smith TW, Marsden PV, Hout M, et al. (2011) General Social Surveys, : Cumulative Codebook. Chicago: National Opinion Research Center. South SJ, Crowder K and Chavez E (2005) Migration and spatial assimilation among U.S. Latinos: Classical versus segmented trajectories. Demography 42: Stern PC, Dietz T, Guagnano GA, et al. (1999) A value-belief-norm theory of support for social movements: The case of environmentalism. Human Ecology Review 6: Taylor DE (1989) Blacks and the environment: Towards an explanation of the concern and action gap between blacks and whites. Environment and Behavior 21: Theodori GL, Luloff AE and Willits FK (1998) The association of outdoor recreation and environmental concern: Reexamining the Dunlap-Heffernan thesis. Rural Sociology 63: