THE INFLUENCE OF RELIGION AND MESSAGES FROM RELIGIOUS AUTHORITIES: ATTITUDES ABOUT SCIENCE AND ENVIRONMENTAL CONCERN. John M.

THE INFLUENCE OF RELIGION AND MESSAGES FROM RELIGIOUS AUTHORITIES: ATTITUDES ABOUT SCIENCE AND ENVIRONMENTAL CONCERN By John M. Clements A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Sociology Doctor of Philosophy 2013

ABSTRACT THE INFLUENCE OF RELIGION AND MESSAGES FROM RELIGIOUS AUTHORITIES: ATTITUDES ABOUT SCIENCE AND ENVIRONMENTAL CONCERN By John M. Clements This dissertation explores relationships among religious beliefs and practices, attitudes about science, and environmental concern. Discussions about conflict, separation, or ambivalence between science and religion are as old as science itself, with mixed evidence about the relationships between religion and science. Although there is a general lack of understanding of scientific concepts by the public, correct scientific knowledge of environmental problems is often a key predictor of intentions to behave pro-environmentally. At the same time, religion plays an important part in the daily lives of millions of Americans. If religion plays a role in guiding the lived experiences of adherents, even in the face of increasing secularization, it is important to examine the effects of religion, as well as religious messages, on scientific and environmental matters. The main objectives of this dissertation are to: 1) analyze the ways in which religious affiliation, belief, and commitment characteristics affect attitudes about science in contemporary U.S. society, and 2) analyze how scientifically and religiously framed messages affect specific facets of environmental concern. This dissertation is based on theoretically driven hypotheses suggested by the current literature, and is organized into three chapters. In the first empirical chapter, I analyze secondary data from the General Social Survey to test hypotheses about the influence of religious affiliation, beliefs, and behavior on attitudes about science. The results of this study provide

some evidence that Christians in general have more negative attitudes about science than do non- Christians and non-religious people, but claims about major conflict between science and religion may not accurately describe the U.S. general public. In the second and third essays, I conduct two experiments that test the influence of religious and scientific messages on two specific types of environmental concern in subjects recruited using Amazon Mechanical Turk. In the second essay, I use structural equation modeling to analyze data from an experiment that examines how scientifically framed and religiously framed messages about water conservation influence attitudes about a proposed policy to restrict water use. Results indicate that Christians are just as likely as non-christians and non-religious people to agree with a policy calling for water use restrictions. However, among all respondents and also among Christians a Christian religious message reduces agreement with the proposed water use restriction policy. These results suggest that religiously framed messages may not significantly increase environmental concern. In the third essay, I use zero-inflated Poisson regression to analyze data from an experiment that tests the influence of scientifically framed and religiously framed messages about biodiversity loss on the decision to make a donation to an environmental organization that protects against biodiversity loss. Although Christianity does not influence the decision to donate or not, it does affect the amount that donators give; Christians donate less than non- Christians and non-religious subjects. Receiving a religious message has no effect on making a donation or donation amount. Similar to the first experiment, these results suggest that a religiously framed message may not influence this type of donation behavior.

Copyright by JOHN M. CLEMENTS 2013

ACKNOWLEDGEMENTS I could not have completed this dissertation without the participation of some dedicated, supportive, and caring people. There are not enough words to express my gratitude to my dissertation committee, Aaron McCright (Chair), Tom Dietz, Sandy Marquart-Pyatt, and Sean Valles. Thanks to Aaron who guided, cajoled, listened, let me wander off track and pulled me back, and made me realize the importance of every word I use. I thank you from the bottom of my heart for helping me attain this lifelong goal. Even when I had doubts, your well-timed encouragement kept me going. Thanks to Tom and Sandy for your interest in my work and valuable feedback and support during the entire process. Thanks to Sean for your timely assistance and encouragement. I also owe thanks to Soma Chaudhuri and Kyle Whyte for taking an early interest in my work. I thank the Graduate School and the Environmental Science and Policy Program at Michigan State University for providing financial assistance throughout my graduate studies. Finally, I thank my wife Jennifer and my sons Joshua and Jordan for their support and love over the last 18 years, and especially during the last five. I ve ignored you, bored you with my research, and spent too many hours in front of a computer. You ve kept me grounded and sane. Thank you for believing in me, loving me and supporting me. I only hope that I can give back what you have given to me. v

TABLE OF CONTENTS LIST OF TABLES LIST OF FIGURES viii ix CHAPTER 1 INTRODUCTION 1 CHAPTER 2 RELIGIOUS AFFILIATION, COMMITMENT AND BELIEF: IMPLICATONS FOR PUBLIC ATTITUDES ABOUT SCIENCE 8 INTRODUCTION 8 RELIGION AND ATTITUDES ABOUT SCIENCE 12 HYPOTHESES 18 DATA AND METHODS 20 Dependent Variables: Attitudes about Science Indicators 20 Independent Variables 21 Statistical Analyses 22 RESULTS AND DISCUSSION 23 CONCLUSION 27 CHAPTER 3 THE INFLUENCE OF RELIGIOUSLY AND SCIENTIFICALLY FRAMED MESSAGES ON AGREEMENT WITH WATER USE RESTRICTIONS 31 INTRODUCTION 31 RELIGION, SCIENCE AND ENVIRONMENTAL CONCERN 32 Religion, Scientific Knowledge and Environmental Concern 34 Message Framing 36 Water Conservation 37 METHODS 39 Experimental Design 39 Sample 39 Dependent Variable 40 Experimental Conditions 41 Religion Predictors 41 Political, Socio-Demographic and Biophysical Predictors 42 Environmental Concern Indicators 42 Statistical Analysis 45 RESULTS 46 CONCLUSION 49 vi

CHAPTER 4 ACTUAL PAYMENTS FOR BIODIVERSITY PROTECTION: THE INFLUENCE OF RELIGIOUSLY AND SCIENTIFICALLY FRAMED MESSAGES 51 INTRODUCTION 51 RELIGION, SCIENCE AND ENVIRONMENTAL CONCERN 52 Scientific Knowledge and Religion 53 Message Framing 55 Biodiversity Loss 56 METHODS 58 Experimental Design 58 Sample 58 Dependent Variable 59 Experimental Conditions 61 Manipulation Checks 62 Religion Predictors 62 Values Orientation and Environmental Concern Indicators 62 Political and Socio-Demographic Predictors 64 Statistical Analysis 65 RESULTS 65 CONCLUSION 70 CHAPTER 5 CONCLUSION 73 APPENDICES 78 APPENDIX A: TABLES 79 APPENDIX B: FIGURE 93 REFERENCES 95 vii

LIST OF TABLES Table 1: General Social Survey Variables Used in the Study 80 Table 2: Attitudes about Science among Different Religious Subgroups in 2010 82 Table 3: Standardized Coefficients from OLS Regression Models Predicting Attitudes about Science: Entire Sample 83 Table 4: Standardized Coefficients from OLS Regressions Predicting Attitudes about Science: Christian Subsample 84 Table 5: Description, Coding, Mean, and Standard Deviation of the Variables Used in the Study 86 Table 6: Standardized Coefficients from Structural Equation Model Predicting Agreement with Proposed Water Policy, Entire Sample (N=608) 87 Table 7: Standardized Coefficients from Structural Equation Model Predicting Agreement with Proposed Water Policy, Christian Subsample (N=206) 88 Table 8: Description, Coding, Mean, and Standard Deviation of Variables Used in the Study 89 Table 9: Zero-Inflated Poisson Regression Predicting Donation to Protect against Biodiversity Loss, Entire Sample 91 Table 10: Zero-Inflated Poisson Regression Predicting Donation to Protect against Biodiversity Loss, Christian Subsample 92 viii

LIST OF FIGURES Figure 1: Analytical Model 94 ix

CHAPTER 1 INTRODUCTION This dissertation explores relationships among religious belief and practice patterns, attitudes about science, and environmental concern. Secularization theory posits that as the influence of religious authority on everyday life decreases, society becomes more secularized (Chaves 1994; Christiano, Swatos, and Kivisto 2008; Yamane 1997). Smith (1985) characterizes secularization as the decrease of religious authority on public life, while arguing that religion still influences private life. Apart from trying to define secularization, others argue that a decline in church membership and religious activity is evidence of an increasingly secularized society (Berger 1967, Stark 1999). However, Peter Berger (2008:23), one of the key observers of secularization, recently has pointed out, Religion has not been declining. On the contrary, in much of the world there has been a veritable explosion of religious faith. Over 80% of the U.S. public reports some religious affiliation (Pew Research Center 2013), suggesting that religion plays at least some role in the lives of millions of Americans. Among industrialized nations, the US is unique in this regard, thus making it an interesting case to explore. Therefore, it is important to understand how religion influences attitudes about different social issues, including science and the environment. This dissertation contributes to the literature on religious belief and practice, environmental concern, and attitudes about science in two ways. First, this research explores the influence of religious belief and commitment patterns on attitudes about science. Second, the research studies the influence of religiously and scientifically framed messages on two aspects of environmental concern: support for a proposed policy addressing water rights and actual donations to protect against biodiversity loss. 1

Most of the existing literature examining religion as a predictor of attitudes about science focuses narrowly on Christian religious affiliation (mostly Evangelical Protestantism) and typically uses only single measures of religious behaviors (i.e., attendance at services or frequency of prayer) to examine the relationship between religion and attitudes about science. Investigators assume that those who attend church more often or pray with greater frequency, for example, are more religious than those who perform these activities with lesser frequency. A few patterns do emerge. In general, Christians are less supportive of science than non-christians and non-religious people. Evangelical Protestants are generally less supportive of science and have more negative attitudes about science than do the non-religious, Mainstream Protestants, and Catholics. In addition, people who are more religious or fundamentalist in their belief systems appear to be less supportive of science. Understanding the relationship between religion and science becomes even more important in light of the need to solve problems that require an understanding of science, including environmental problems. Most people tend to get information about science and environmental problems from the media, which obtains information from original scientific sources and interprets this information for the general public. However, religion can also interpret information about science and the environment for adherents of a specific faith. Since the 1990s there has been a supposed greening of Christianity in the United States, with organizations such as the Evangelical Environmental Network, the Southern Baptist Environment, and the Evangelical Climate Initiative, advocating for care of creation. However, recent research indicates that advocacy for creation care at the organizational and elite levels of the green Christian movement has not yet been effective in increasing concern for the environment within rank-and-file Christians (Clements, McCright, and Xiao 2014). 2

Much of the existing literature about the relationship between religion and environmental concern provides evidence of a negative relationship between religion and environmental concern (e.g., Eckberg and Blocker 1989, 1996; Kanagy and Nelson 1995). Other studies suggest a positive relationship (Boyd 1999; Guth et al. 1995; Kanagy and Willits 1993), and some studies find no relationship at all (Hayes and Marangudakis 2000; Woodrum and Hoban 1994). Although the general public generally lacks understanding of scientific concepts (Dunlap 1998; Jenkins 2003), correct scientific knowledge of environmental problems is a key predictor of intentions to participate in effective decision making about environmental issues (Bord, O Connor, and Fisher 2000). Given the general negative relationship between Christian religious behavior and attitudes about science, as well as environmental concern, it is important to determine how religious affiliation, commitment, and belief characteristics influence attitudes about environmental problems that require scientific solutions. The previous discussion leads to the following questions about how religion and messages from religious authorities influence attitudes about science and the environment: 1. How do religious belief and commitment characteristics influence attitudes about science in contemporary U.S. society? 2. Do religiously and scientifically framed messages differently influence specific facets of environmental concern? The dissertation directly addresses these questions in two ways. First, I analyze secondary data from the General Social Survey to test hypotheses relevant to the first question. Second, I conduct two experiments testing the effects of religiously and scientifically framed messages on 1) support for a proposed water use restriction policy, and 2) actual donation behavior to prevent 3

loss of biodiversity. Guided by these questions, I organized the dissertation into three empirical chapters. In the first essay (Chapter 2), I use nationally representative data from the 2010 General Social Survey to test hypotheses about the influence of religious affiliation, beliefs, and behavior on attitudes about science. I compare the influence of Christian and non-christian affiliations, and also compare attitudes about science among Christian affiliations: Mainline Protestants, Evangelical Protestants, Black Protestants, and Catholics. I also use composite, rather than single-item (e.g. church attendance or biblical literalism), indicators of religiosity and fundamentalism to test their influence on attitudes about science. Finally, I test whether Christian religiosity and fundamentalism moderate the relationship between Christian religious affiliation and attitudes about science. Recent research finds no significant greening among Christians in the U.S. approximately 20 years into a supposed greening of Christianity (Clements, McCright, and Xiao 2014). In addition, previous research reports that Christians are less supportive of science than non- Christians and non-religious people (Ellison and Musick 1995; Evans 2002, 2012; Gauchat 2008). Despite these results, it remains an interesting empirical question whether religiously or scientifically framed messages increase environmental concern about different environmental problems among U.S. citizens. The second and third essays report the results of experiments that test the influence of religiously and scientifically framed messages on two specific types of environmental concern: support for a water use restriction policy (Chapter 3) and donations to protect against biodiversity loss (Chapter 4). For each experiment, I recruited subjects using Amazon Mechanical Turk, which allows requesters (such as myself) to solicit workers to complete Human Intelligence Tasks, (HIT) (i.e., the experiments). Amazon Mechanical Turk 4

samples are closer to the U.S. general public than are typical university samples (Mason and Suri 2012; Paolacci, Chandler, and Ipeirotis 2010) and tend to be more diverse than typical Internet samples (Buhrmester, Kwang and Gosling 2011). Amazon Mechanical Turk provides a quick, inexpensive method to collect experimental data from a wide cross-section of the general public. The second essay (Chapter 3) examines how religiously and scientifically framed messages about water conservation influence attitudes about policies that restrict water use. Worldwide, over seven billion people depend on one percent of water on earth that is available for use (Shiklomanov 1993). In the summer of 2012, approximately 81% of the U.S. was classified as abnormally dry by the United States Geological Survey, and 67% of the U.S. was in a moderate to extreme drought (United States Drought Monitor 2013). As drought events increase nationwide, it becomes increasingly important to understand how governments, municipal water suppliers, and water conservation organizations can elicit support for different methods to conserve water to meet our present and future needs. The second essay uses structural equation modeling to test whether religiously or scientifically framed messages mediate the relationship between various socio-demographic, religious, and environmental concern characteristics, and support for a fictional water use restriction policy. The third essay (Chapter 4) explores how religiously and scientifically framed messages about biodiversity influence the choice to make donations to protect against biodiversity loss. Over the last 100 years there have been over 100 well-documented extinctions worldwide, mostly due to habitat loss and ecosystem destruction from human activities. In the United States alone, it costs hundreds of millions of dollars per year to control invasive species and respond to floods and fires that are becoming more extreme as ecosystem changes increase (Millennium Ecosystems Assessment 2005). These costs are covered by taxes collected by governments, as 5

well as through donations to nongovernmental organizations that work to prevent further biodiversity loss. In order to continue this work, it is important that governments and NGOs understand what influences peoples willingness to pay or sacrifice for environmental protection. Investigations about willingness to pay for environmental protection use indicators that measure a person s self-rated willingness to pay higher taxes, pay higher prices, and change their standard of living to protect the environment. However, these measures typically ask people to rate their hypothetical willingness to pay or sacrifice to deal with an environmental problem. The third essay reports results of an experiment that is unique to the willingness to pay literature in that it measures actual donation behavior to a nonprofit organization that operates to protect against biodiversity loss. Considered together, the three essays provide additional evidence about the role of religion in influencing attitudes about science and environmental concern. While some people perceive a conflict between religion and science, evidence suggests both negative and positive relationships between religion and attitudes about science. However, there is a consistent negative relationship between Christian religious affiliation and environmental concern. In addition, there is a general lack of understanding about science by the US public which is troubling when effective solutions to environmental problems require an understanding of science. Because religion plays an important part in the daily lives of millions of Americans, and because of possible negative relationships between religion and science, and religion and environmental concern, it is important to examine the effects of religion and religious messages on matters related to science and the environment. My research adds to the current literature by updating evidence about the relationship between religion and attitudes about science using data from 2010. In addition, I also test the influence of religiously and scientifically framed messages on 6

environmental concern which allows me to make some observations about the conflicts between religion and science, as well as a possible greening of Christianity. 7

CHAPTER 2 RELIGIOUS AFFILIATION, COMMITMENT AND BELIEF: IMPLICATONS FOR PUBLIC ATTITUDES ABOUT SCIENCE INTRODUCTION Science is an integral part of contemporary U.S. society, as it helps drive our economy, promotes technological developments, and gives us significant power to change our world in noticeable ways. While the U.S. general public lacks knowledge of basic scientific facts (e.g., Dunlap 1998; Jenkins 2003; Miller 2004), it nevertheless does trust scientists who have relevant expertise (e.g., Shapin 2007; VCU Center for Public Policy 2010). In fact, 80% of the general public believes that science is essential to the country s economy, healthcare system, and global reputation (California Academy of Sciences 2009; Miller 2004). In spite of this support, 58% of U.S. adults believe that scientific research does not pay enough attention to the moral values of society, and 50% believe that scientific research has created as many problems for society as it has solutions (VCU Center for Public Policy 2010). Further, some scientific issues (e.g., embryonic stem cell research, vaccines) and technological developments (e.g., genetically modified foods) provoke significant opposition to science by the U.S. public. Public understanding of science (PUS) scholarship aims to explain public understanding of, involvement in, and trust in science. The deficit model posits that a lack of support for science is due to a lack of understanding about science; if scientists can find a way to fill this knowledge deficit, then support for science will increase (Lach and Stanford 2010; Maranta et al. 2003; Powell and Colin 2008; Sturgis and Allum 2004). Others find mixed results as to whether increased knowledge leads to positive attitudes about science in the general public (Allum et al. 8

2008; Bhaduri 2003; Critchly 2008; Miller 2004), among women (Simon 2010), and among political conservatives (Gauchat 2012), specifically. A more contemporary model holds that what is deficient is not the public s knowledge, but rather its trust in science, and in scientific experts specifically (Gauchat 2012; Lewenstein 1992; Solomon 1993; Wynne 1992, 1993). PUS scholarship also provides some evidence that females (Bak 2001; Baker 2012; Gauchat 2012; Hayes and Tariq 2000; Trankina 1993), conservatives (Gauchat 2012), older adults (Bak 2001; Gauchat 2012), and non-whites (Gauchat 2008) report lesser trust and confidence in science (and increased worry about the risks due to science) than their respective counterparts. The relationships between these demographic and political characteristics and attitudes about science are well documented, but less so are the relationships between religion and attitudes about science. Current scholarship that investigates relationships between religion and attitudes about science depends heavily upon single-item indicators of religion too rarely measuring the concept with multidimensional composite indicators. As religion is a multi-faceted experience, a single measure of only one facet of religious practice likely does not represent the full range of religious belief systems. Composite indicators of religious life might better represent the range of religious beliefs and behaviors. Because religion plays a vital role in the everyday life of over a hundred million U.S. citizens and because the content of religious ideas and values may be challenged by the methods and substantive knowledge of science, it is important to determine the extent of the conflict between religion and science in this country. Some studies indicate that religious strength correlates negatively with support for science funding (e.g., Brossard et al. 2009) and with the belief that science does not pay enough attention to the morals of society (VCU Center for Public Policy 2010). Other studies indicate that biblical literalism, theological orthodoxy, and perceived 9

ubiquity of sin are associated with moral criticism of science (e.g., Baker 2012; Ellison and Musick 1995; Nisbet and Goidel 2007). Most of the existing PUS literature examining religion as a predictor of attitudes about science focuses narrowly on Christian religious affiliation (mostly Evangelical Protestantism) and typically uses only single measures of religious behaviors (i.e., attendance at services or frequency of prayer) to examine the relationship between religion and science. Investigators assume that those who attend church more often or pray with greater frequency, for example, are more religious than those who perform these activities with lesser frequency. Unfortunately, we don t know if these single item measures reflect the same characteristic (Evans 2013). For instance, people may pray several times per day, but never attend religious services. By one measure they would be very religious, but by the other, not religious at all. Religiosity and fundamentalism indicators comprised of several individual measures may provide better measures of these concepts. Investigators generally consider affiliation, religiosity, and fundamentalism separately in their analyses. However, they infer that those with a specific affiliation who perform more religious behaviors or hold more fundamentalist beliefs than others are more faithful and will follow the tenets of their faith with greater fervor. This suggests that religiosity and fundamentalism may influence the strength or direction of affiliation on attitudes about science. To my knowledge, no studies investigate if religious behaviors and beliefs moderate denominational affiliation to influence attitudes about science. By addressing the preceding limitations, this study contributes to the scholarship that addresses this question: How do religious belief and commitment characteristics influence attitudes about science in contemporary U.S. society? 10

I use nationally representative data from the 2010 General Social Survey to test hypotheses about the influence of religious affiliation, beliefs, and behavior on attitudes about science. This study makes three contributions to the literature. First, I move away from limited considerations of religious affiliation (e.g. only Evangelical Protestantism) to examine how attitudes about science are influenced by a wider range of religious affiliation. Specifically, I compare the influence of Christian and non-christian affiliations and within Christianity compare the influence of Mainline Protestant, Evangelical Protestant, Black Protestant, and Catholic affiliations. Second, I include more comprehensive measures of religiosity and religious fundamentalism than typically are utilized in the literature. Most studies use single-item measures of religiosity (e.g., attendance at religious services) and fundamentalism (e.g., belief in the Bible as the literal word of God). As is the nature of most large data sets that poll the U.S. public, measures of religious belief and commitment focus on Christian religious beliefs and behaviors. However, I employ composite, rather than single-item, indicators of these characteristics. Third, I test whether Christian religiosity and fundamentalism moderate the relationship between Christian religious affiliation and attitudes about science, by examining the performance of interaction terms in my analytical models. In the next section I review the literature on the relationship between science and religion, focusing first on religious affiliation and then on religious strength, activity, and beliefs. I follow this by developing specific hypotheses for testing and by describing my analytical model. After that I detail my dataset and methods, before presenting and discussing my results and ending with suggestions for future research. 11

RELIGION AND ATTITUDES ABOUT SCIENCE Scholars cite many examples as evidence of conflict between religion and science. Among these are the harsh treatment of Galileo and Copernicus by the Catholic Church, the response to Darwin s The Origin of Species by Samuel Wilberforce (Bishop of Oxford) as contrary to scripture and unsupported by theology (Eister 1978: 352), and Thomas Aquinas s thought of Eternal law as existing in the Mind of God and governing the whole universe (Eister 1978: 350). Over the last 100 years, social scientists have contemplated this conflict. Max Weber (1946 [1918]) alludes to a tension between religion and the world, explaining that science is not useful for explaining the fundamental questions of life such as why we are here and how we should live our lives. Robert Merton (1938) identifies organized religion as an institution inherently in conflict with the organized skepticism of science, as the former demands unqualified faith something that science fundamentally challenges. In recent decades several scholars have investigated the relationship between religion and science in the U.S. (see e.g., Campbell and Curtis 1996; Eister 1978; Evans 2002; Evans and Evans 2008) as well as in other countries (Buckser 1996; Campbell and Curtis 1996). This literature on the relationship between religion and attitudes about science focuses on three main aspects of religion: affiliation (what religion people are), religiosity (how strong their religious belief or commitment is, usually measured as attendance at church services or frequency of prayer) and fundamentalism (adherence to orthodox beliefs, usually represented by belief in a literal interpretation of the bible). Studies examining the relationship between religious affiliation and attitudes about science generally investigate Christian religions and mainly focus on Evangelical Protestant affiliation (Binder 2007; Ellison and Musick 1995; Evans 2002; Gauchat 2008). Others study the influence 12

of faith traditions such as Mainline Protestantism, Black Protestantism, Catholicism, and Judaism, as well as people with no religious affiliation (Evans 2002, 2012; Freeman and Houston 2011; Hayes and Tariq 2000; Hochschild, Crabill, and Sen 2012; Nisbet 2005; Scheitle 2005). Some studies find that Evangelical Protestantism is associated with negative perceptions of science in general (e.g. Ellison and Musick 1995; Gauchat 2008). These studies show that Evangelical Protestant affiliation is related to beliefs that science pries into inappropriate areas and impairs people s ideas of right and wrong (Ellison and Musick 1995), that we believe too much in science and not enough in faith, that changes caused by science make things worse, and that science does more harm than good (Gauchat 2008). Other studies report opposition by Evangelical Protestants to specific types of science-related issues (Binder 2007; Evans 2002). Evans (2002:756) provides evidence that Evangelical Protestants are more opposed to unrestricted research on human cloning than other Protestants, and Catholics, likely due to a view that cloning usurps God s role. In addition, most of the opposition to teaching evolution in public schools, as well as support for including intelligent design as a scientific theory, comes from Evangelical Protestant traditions (Binder 2007). Further, Freeman and Houston (2011) find that compared to Jews and Mainline and Black Protestants, Evangelical Protestants report less support for federal funding for scientific research, especially stem cell research. While the influence of Evangelical Protestantism is fairly consistent across studies, Evans (2013) finds no association between conservative Protestant affiliation and confidence in the scientific institution. However, this ethnographic study does not define which Protestant affiliations are conservative. Another group of studies investigates the influence of other Christian affiliations such as Catholicism (Evans 2002; Hayes and Tariq 2000; Nisbet 2005; Scheitle 2005), Mainline 13

Protestantism (Freeman and Houston 2011), and Black Protestantism (Freeman and Houston 2011) on attitudes about science. A few studies investigate the influence of non-christian affiliation (Freeman and Houston 2011; Scheitle 2005) and no religious affiliation (Evans 2012; Hochschild, Crabill, and Sen 2012), on attitudes about science. Hayes and Tariq (2000) find that Catholics have more positive attitudes about science than do non-catholics in the U.S., but this is not observed in other Anglo countries. Evans (2002) finds that Catholics show greater support for unrestricted research on human cloning than Evangelical Protestants, but Nisbet (2005) reports that Catholics show greater opposition to stem cell research compared to Protestant traditions. Scheitle (2005) reports no difference in optimism about biotechnology among Catholics, Jews, and Protestants, while members of a group of other religions (e.g., Hindu, Buddhist, Mormons, among others) are more optimistic about biotechnology than are Protestants. Freeman and Houston (2011) find that Mainline Protestants and Catholics believe we spend too much on scientific research compared to non-religious people, but there is no difference in support among Black Protestants, Evangelical Protestants, Jews, and non-religious people. In an ethnographic study, Evans (2012) finds no difference in support for science between religious people (regardless of tradition) and non-religious people. Finally, Hochschild, Crabill, and Sen (2012) find that religious affiliation has no influence on optimism about science and technology. The above group of studies has several limitations. First, with the exception of Ellison and Musick (1995) and Evans (2002), these studies only include religion as a statistical control in their analytical models. Thus, they spend minimal time analyzing the relationship between religious affiliation and attitudes about science. Some of these studies use data sets from the late 1980s or 1990s (Ellison and Musick 1995; Gauchat 2008; Scheitle 2005), so they have limited 14

efficacy for explaining more recent patterns in the relationship between science and religion. Other studies use interview and media data (Binder 2007; Evans 2012) that is difficult to quantitatively analyze. Finally, there is a wide diversity in operationalizing attitudes about science. Some studies ask about attitudes towards science funding (Freeman and Houston 2011) or the compatibility of science and faith and harm to society from science (e.g. Ellison and Musick 1995; Gauchat 2008). Still others focus on attitudes about specific types of science such as reproductive cloning (Evans 2002) or embryonic stem cell research (Nisbet 2005). Another group of studies examines the relationship between religiosity and attitudes about science. One limitation here is the use of a few single-item indicators to measure religiosity, including attendance at religious services, the perception that religion is a guide to daily decision making, and self-rated strength of beliefs. These measures suggest that religiosity (especially attendance at religious services) negatively influences attitudes about science (e.g. Brossard et al 2009; Ellison and Musick 1995; Gauchat 2008, 2011, 2012). However, as described earlier, single-item indicators may not effectively measure a multi-faceted construct such as religiosity in the same way (Evans 2013). One set of studies operationalizes religiosity as perceptions that religion is a guide to daily decision making. These studies report a negative relationship between those who report that religion has a great deal of influence on their daily decision making and support for science funding for general nanotechnology, likely due to the influence of strong religious belief systems (Brossard et al. 2009) or concerns that science and technology tamper with nature by playing God (Ho, Scheufele and Corely (2010:2709). Nisbet and Goidel (2007) find that people who report that religion has a great deal of influence on their daily life have negative opinions of embryonic stem cell research and therapeutic cloning. 15

Another group of studies that examines the relationship between religiosity and attitudes about science includes measures such as frequency of prayer and attendance at religious services. These studies find that prayer and attendance at religious services are negatively related to attitudes about science (Gauchat 2008; Sturgis and Allum 2004) and confidence in science (Gauchat 2011, 2012). However, when people understand that scientific activities have well established methods and review processes, then this relationship disappears (Gauchat 2011). Increased attendance at religious services is also related to a lack of confidence in political institutions, suggesting that increased religiosity is related to a general lack of confidence in social institutions (Gauchat 2012). Finally, Freeman and Houston (2011) note a correlation between attendance and decreased support for government funding of stem cell research. In contrast, other studies find that attendance is not related to skepticism about genomic science (Hochschild, Crabill, and Sen 2012), optimism about biotechnology (Scheitle 2005), or support for science funding (Freeman and Houston 2011). There is some evidence to suggest that religiosity characteristics moderate the relationship between religious affiliation and attitudes about science. Among people who frequently attend church, there is a negative relationship between Evangelical Protestantism and opinions about therapeutic cloning, which is not observed in frequent church attending Catholics or liberal Protestants (Evans 2002). In addition, Evans (2013) also suggests that frequent church attendance and biblical literalism together moderate the relationship between Protestantism and confidence in scientific institutions. However, frequent church attendance alone does not moderate the respective relationships between Evangelical Protestantism, Mainline Protestantism, Black Protestantism, and Catholicism, and confidence in scientific institutions. 16

A final group of studies considers the influence of fundamentalist religious beliefs on attitudes about science (e.g. Baker 2012; Ellison and Musick 1995, Evans 2013; Freeman and Houston 2011; Nisbet and Goidel 2007; Scheitle 2005). Most operationalize fundamentalism as the belief that the Bible is the literal word of God. Such studies find that biblical literalism negatively influences attitudes about science (Ellison and Musick 1995, Nisbet and Goidel 2007), confidence in science as an institution (Evans 2013), and support for federal science funding and stem cell research funding more specifically (Freeman and Houston 2011). Further, biblical literalism is positively associated with the belief that there is a conflict between science and religion (Baker 2012). One study suggests that fundamentalist religious beliefs may moderate the relationship between religious affiliation and attitudes about science. Evans (2012) finds that the only participants in his ethnographic study who do not support science are Protestants who report fundamentalist religious beliefs. This suggests that it may be worthwhile to investigate whether fundamentalism moderates the relationship between religious affiliation and attitudes about science. One limitation of existing studies that include measures of religiosity and fundamentalism is that they generally use single item indicators to measure religiosity and fundamentalism when multidimensional composite indicators may better measure these concepts. Religious belief is a multifaceted construct and one measure cannot reliably discern more religious people from less religious people (Evans 2012). Measures of religiosity such as frequency of church attendance and self-rated religious strength do not provide the same information about religiosity. Inclusion of disparate religiosity measures in separate studies makes it difficult to make conclusions about the influence of religiosity on attitudes about science. Similarly, the use of one main measure of 17

fundamentalism ignores the influence of other facets of fundamental beliefs such as being born again, or behaviors such as trying to get others to accept Jesus Christ as their savior. Different denominations within Christianity have different patterns of church attendance and beliefs about the Bible (Evans 2013) that may differently influence attitudes about science. In summary, Evangelical Protestants are generally less supportive of science and have more negative attitudes about science than do the non-religious, Mainstream Protestants, and Catholics. In addition, people who are more religious and fundamentalist in their belief systems appear to be less supportive of science. These results suggest several hypotheses for testing. HYPOTHESES Following previous studies that report negative attitudes about science by adherents of Christian denominations (e.g. Binder 2007; Ellison and Musick 1995; Evans 2002, 2012; Gauchat 2008), I expect that those who report no religious affiliation have more positive attitudes about science than those who report a Christian religious affiliation (Hypothesis 1). Much of the existing literature only examines the relationship between the Christian faith and attitudes about science. Few studies examine the relationship between non-christian religions and attitudes about science, yet this is an important comparison. Because of the negative relationship between Christian affiliation and attitudes about science (e.g. Ellison and Musick 1995, Evans 2002, Freeman and Houston 2011), and because of a recent trend of U.S. Christian opposition to stem cell research, reproductive cloning, climate change, and teaching evolution, I expect that non-christians have more positive attitudes about science than Christians (H2). As described earlier, some evidence suggests that attitudes about science vary across Christian denominations. While Catholics have more positive attitudes about science than non- Catholics in the U.S. (Hayes and Tariq 2000), they may not support science funding (Freeman 18

and Houston 2011) and are opposed to stem cell research (Nisbet 2005). Mainline Protestants are generally as supportive of science as the non-religious, while Black Protestants have similar attitudes as Evangelical Protestants. Therefore, I expect that Mainline Protestants have more positive attitudes about science than Catholics, Evangelical Protestants, and Black Protestants (H3). Many studies find a negative relationship between religiosity (e.g. Brossard et al. 2009; Gauchat 2008 2011, 2012; Nisbet and Goidel 2007; Sturgis and Allum 2004) or fundamentalism (e.g. Baker 2012; Ellison and Musick 1995, Evans 2013; Scheitle 2005) and attitudes about science. Almost all of these studies consider religiosity or fundamentalism in the context of Christianity, generally Evangelical Protestantism. Items that commonly measure religiosity and fundamentalism are based on Christian belief or practice. For instance, attendance at a religious service as a measure of religiosity is inadequate for some non-christians who do not attend services as part of their faith, and non-religious people who generally do not attend at all. Similarly, fundamentalism is often times measured by belief in the Bible as the literal word of God, or reporting that a person has tried to get another to accept Jesus Christ as their savior. These measures are clearly inadequate for non-christians. Thus, I only examine the effects of religiosity and fundamentalism among Christians. Some studies find that prayer and attendance at religious services are negatively related to attitudes about science (Gauchat 2008; Sturgis and Allum 2004) and confidence in science (Gauchat 2011, 2012). Therefore, I expect that religiosity negatively relates to attitudes about science among Christians (H4). Further, Ellison and Musick (1995) and Nisbet and Goidel (2007) report that biblical literalism is negatively associated with attitudes about science. Therefore, I expect that fundamentalism negatively relates to attitudes about science among Christians (H5). 19

Finally, some studies suggest that religiosity and fundamentalism moderate the relationship between Evangelical Protestant affiliation and attitudes toward science (Evans 2012, Evans 2013). Therefore, I expect that religiosity moderates the relationship between Evangelical Protestant religious affiliation and attitudes about science (H6) and that fundamentalism moderates the relationship between Evangelical Protestant religious affiliation and attitudes about science (H7). In the next section, I discuss the dataset and describe the selected variables and the statistical techniques I use to analyze the data. DATA AND METHODS I use data from the 2010 General Social Survey (GSS) (Smith et al. 2011) to test my hypotheses. The GSS is a survey of the American public that is conducted every two years to collect demographic information of U.S. citizens as well as elicit public opinion about government, daily life, economics, and religion, among other things. I use the 2010 edition of the GSS because it includes several variables related to opinions about science, theological orientation, religious affiliation, and socio-demographic information. Table 1 presents information about all variables used for this study. Dependent Variables: Attitudes about Science Indicators I use four indicators of attitudes about science as dependent variables in my models. Each of the four asks respondents the extent to which they disagree or agree with a statement about science. One is an indicator of support for scientific funding: Even if it brings no immediate benefits, scientific research that advances the frontiers of knowledge is necessary and should be supported by the federal government. The other three are indicators of general attitudes toward science: Overall, modern science does more harm than good, We believe too often in science, 20

and not enough in feelings and faith, and One of the troubles with science is that it makes our way of life change too fast. Independent Variables Religious Affiliation: I classified religious affiliation following the recommendations of Steensland et al. (2000) using the following GSS items: DENOM Specific Denomination, OTHER Other Protestant Denomination, and RELIG Respondent s Religious Preference. Following Steensland et al. (2000) I recoded these variables to the following dummy variables: No Religion, Christians, Non-Christian, Catholic, Black Protestant, Evangelical Protestant, and Mainline Protestant to allow for comparisons as suggested by my hypotheses. Catholics and Protestants are the majority of Christian religious people in the U.S. Smaller groups of Christian denominations (e.g., Eastern Orthodox, Latter Day Saints, and Jehovah s Witnesses, among others) are represented in the GSS, but because of their small numbers, I exclude them from this study. Other Religion Indicators: Past research examining the relationship between religion and attitudes about science focuses on the levels of respondents religiosity and fundamentalism. The 2010 GSS has several items that I use to create indicators of religiosity and fundamentalism. For religiosity and fundamentalism, I used exploratory factor analysis and Cronbach s alpha reliability test to form composite indicators of each. In each case, items loaded onto one factor and Cronbach s alpha tests indicate good reliability. I used exploratory factor score weights to form composite indicators of religiosity and fundamentalism. The religiosity indicator is created from seven GSS items (factor loadings between.68 and.82; Cronbach s α =.85): how often you attend religious services, the strength of your belief in God, how often you pray, how often you take part in the activities and organizations of a church 21

or place of worship other than attending service, the strength of your religious group identification, how hard you try to carry your religious beliefs over into all other dealings in life, and how religious you consider yourself. Due to the varying number of response categories for each item (i.e., frequency of attendance is a nine-point scale, while prayer is a six-point scale), I first standardized each variable prior to creating each composite measure. For purposes of this analysis higher scores indicate increased levels of religiosity. The fundamentalism indicator is created from four GSS items (factor loadings between.70 and.80; Cronbach s α =.76): the strength of your belief in the literal interpretation of the Bible, the liberalism or fundamentalism of your religion, whether or not you have been born again, and whether or not you have ever tried to encourage someone to believe in Jesus Christ or accept Jesus Christ as his or her savior. Again, higher scores indicate a respondent who is more fundamentalist in his or her belief system. Political and Socio-Demographic Control Variables: I use the following political indicators in my analyses: political ideology (1= extremely conservative to 7= extremely liberal ) and party identification (1= strong Republican to 7= strong Democrat ). Finally, I use the following socio-demographic variables as controls: gender (female), race (white), age in years, number of years of education, and family income. Statistical Analyses I conducted a series of OLS regression models to predict values on each of the four attitudes about science indicators. First, I ran models to compare the attitudes of non-religious respondents, Christians, and non-christians (H1 and H2). Then, I ran a second set of OLS regressions to test within-christian variation in attitudes about science (H3). I included religiosity and fundamentalism in each of my models to test the influence of each on attitudes 22