Scientific authority in policy contexts: Public attitudes about environmental scientists,medical researchers,and economists

Abstract

This paper uses data from the US General Social Survey to examine public support for scientists in policy contexts and its link to scientific disciplines. An analysis of attitudes about the amount of influence that environmental scientists, two kinds of medical researchers, and economists should have over policy decisions reveals that in each discipline the extent to which scientists are thought to serve the nation’s best interests is the strongest determinant of attitudes about scientists as policy advisors. Perceptions of scientists’ technical knowledge and the level of consensus in the scientific community also have direct, albeit weaker effects on opinions about scientists’ appropriate roles in policy settings. Whereas previous research has stressed the importance of local variability in understanding the transfer of scientific authority across institutional boundaries, these results point to considerable homogeneity in the social bases of scientific authority in policy contexts.

Keywords

cultural authority of science disciplines group differences scientific experts

1. Introduction

In the United States, organizations such as the Environmental Protection Agency (EPA), the Food and Drug Administration (FDA), and the National Economic Council (NEC) provide scientists a venue to help lawmakers manage uncertainty and risk associated with public issues ranging from climate change to bio-medical research to fiscal policy (Guston and Keniston, 1994; Hilgartner, 2000; Jasanoff, 1994). Despite some evidence of declines in North American and European public confidence and interest in science and technology (European Commission, 2010: 7; Gauchat, 2011; National Science Board, 2008: 7–12; Nowotny, 2005), a large majority of adults support at least some reliance on scientific expertise in political and economic decision-making (European Commission, 2010: 86; Gaskell et al., 2005; National Science Board, 2008: 7–28). Nonetheless, substantial variation exists in the extent to which individuals believe that scientists should influence public policy. In addition to fueling ideological debates about the proper bases for policy, differences in public support for scientists in policy contexts pose an intriguing theoretical question about the nature of scientific authority. Specifically, what makes it possible for scientists to obtain influence in settings that fall beyond science’s historical domain?

Research into public understanding of science offers one set of answers to this question. This survey-based tradition provides a set of well-tested models that emphasize respondents’ scienceliteracy and broader cultural dispositions in predicting their attitudes about scientists and science policies (Evans and Durant, 1995; Miller, 2004). Research into the cultural authority of science provides another set of answers to the question of authority transfer, one that emphasizes scientists’, rather than respondents’, personal characteristics and social contexts (Brown, 2009; Douglas, 2009; Lacey, 2005). The analytic focus of this largely qualitative tradition has been to provide detailed descriptions of historical examples where the boundaries of science were negotiated by competing stakeholders (Fourcade, 2009; Gieryn, 1999; Jasanoff, 2005; Schweber, 2006). Because of their custom of small-sample research, however, few examinations of scientific authority have established general patterns, trends, or “scope conditions” concerning the reach of science in society (Frickel and Moore, 2006: 14). Consequently, this line of study has been criticized as overly idiographic, and generalizable antecedents of scientists’ ability to obtain authority across institutional settings remain elusive.

Recent survey research has begun to bridge these empirical traditions by investigating scientific authority from the perspective of the public (Bauer, Petkova and Boyadijieva, 2000; Gauchat, 2010; Lujan and Todt, 2007). To date, however, this research has focused on survey respondents much more so than on the scientists described by questionnaires. Resultantly, these studies do not take advantage of research into scientific authority that points to ways in which scientists’ individual-level characteristics affect their ability to obtain influence in society. In particular, scientists’ technical knowledge, personal disinterestedness, and the relation of their claims to those of other experts (i.e., consensus) each contribute to scientists’ capacity to harness the cultural authority of science (Brown, 2009; Datson, 1992; Kuhn, [1962] 1996; Lacey, 2005; Latour, 1987; Merton, [1942] 1973). Scientific disciplines are an additional source of authority, a finding that suggests that scientists’ disciplinary associations may mediate their ability to obtain influence in policy contexts (Bourdieu, 1975; Cambrioso and Keating, 1983). Prior research has explored each these factors individually, and each is known to affect the allocation of scientific authority among individuals and organizations (Gieryn 1999). However, the case-based designs of many studies in this area have prevented investigators from analyzing the combined and relative effects of scientists’ knowledgeability, disinterestedness, consensus and disciplines. As such, the distinct impacts of these common explanations for scientific authority, especially as it extends past science’s historical domain, remain unclear.

The General Social Survey (GSS) provides the most comprehensive data available on US adults’ attitudes about scientists in policy settings. Because the survey included questions about scientists’ from several research specialties, the data enable direct comparisons across disciplines of the public sources of scientific authority. The survey used a vignette design, in which respondents were read descriptions of several public policy topics and asked how much influence scientists from several disciplines should have over related policy decisions. Respondents were also asked about the extent to which vignette scientists were knowledgeable, worked in the nation’s best interests, and reflected consensus within the scientific community. I use these data to address the following questions: (1) What effects do perceptions of scientists’ knowledgeability, national interests, and consensus have on public attitudes about the amount of influence scientists should have in policy contexts? (2) Does the strength of the association between desired policy influence and perceived knowledgeability, national interests, or consensus depend on scientists’ discipline of study? Results indicate that each of these variables is uniquely and positively linked to support for scientists, and that the effects of each variable are remarkably similar across disciplines. Overall, this analysis suggests that the public bases of scientific authority may be more general than previously thought. Findings also highlight the importance of accounting for characteristics of scientists in addition to characteristics of publics in examining the social foundations of scientific authority.

2. Background

The cultural authority of science

Individual scientists’ ability to obtain scientific authority depends on their ability to position their claims within the cultural boundaries of science (Drori et al., 2003; Gieryn, 1983). Obtaining authority in contexts that fall beyond the traditional realm of science, however, is often complicated by the diverging institutional logics encountered by scientists (Clarke, 1998; Epstein, 1998, 2007; Ezrahi, 2004; Thorpe, 2002; Yankelovich, 2003). For example, public policy decisions in the US have historically been entrusted to political leaders who are accountable to the public through elections. As a result, lawmakers’ authority over policy decisions stems from their positions vis-à-vis constituents. In contrast, science advisors are not immediately subject to electoral politics, and to influence policy decisions they must find other ways to stake their authority onto ground that is traditionally political. Elected leaders’ practice of turning to scientific experts for advice in crafting public policy is one way in which scientists obtain authority in policy settings. However, elected leaders’ decisions to rely on science, and on which science to rely, are ultimately judged by voters, a reality that highlights the importance of understanding public attitudes about scientists in policy arenas. Scientists’ influence over a range of policy decisions is well documented, and so is public support for this arrangement (Gaskell et al., 2005; Hilgartner, 2000; Jasanoff, 1994, 2005; National Science Board, 2008). Nonetheless, the foundations of public support for scientists in policy contexts are not fully understood.

To help explain the movement of scientific authority across institutional boundaries, scholarship on public understanding of science points to certain characteristics of publics that help shape support for science and technology in society. Early research in this tradition focused heavily on the relationship between public knowledge of scientific concepts and general attitudes about science and technology (Miller, 1983, 2004). Recent empirical work has shifted attention away from science literacy and toward a broader view that encompasses other factors that may affect individuals’ attitudes about science, such as their personal experiences and social identities (Barben, 2010; Gauchat, 2011). Despite the analytic shift, these studies continue to emphasize public knowledge of and experience with science in order to understand attitudes about science in society. In this view, scientists’ ability to obtain authority (i.e., public support) in political settings depends on a knowledgeable public that is culturally disposed to accept technocratic authority in a historically democratic context.

Recent calls for more symmetrical analyses of public understanding of science have articulated a need to investigate scientists’ characteristics and backgrounds in addition to those of publics (Barben, 2010). Indeed, relatively little existing research has examined the ways in which scientists themselves influence public attitudes about science, particularly attitudes about the roles scientists play in policy decisions. This omission is surprising given the centrality of scientists’ social contexts and personal attributes in many studies of scientific authority. One of the earliest explanations of scientific authority centers on science’s institutional monopoly over the production of certified knowledge, and underscores scientists’ unique roles as gatekeepers of that information (Merton, [1942] 1973; Nelkin, 1975). This perspective assumes that political leaders’ reliance on scientific expertise results from the instrumental utility of scientists’ technical knowledge. Recent work by Brown (2009) elaborates on this position by distinguishing between scientists’ moral and epistemic authority, arguing that the latter is dependent on scientists’ ability to produce valid and reliable information about the world. Detractors of this position point out that it privileges scientific ways of knowing by devaluing the opinions of other stakeholders, like activists, professionals, and publics, relative to those of scientists (Epstein, 1998; Haraway, 1991; Yearley, 2000). In addition to the controversial question of whether scientific knowledge should be privileged, public support for scientists in policy-influencing roles may depend on scientists’ ability to project an appearance of erudition. One may therefore expect the public to support scientists in policy contexts to the extent that the public believes that scientists are knowledgeable about policy-relevant science.

Also foundational to scientific authority is science’s supposed political and economic neutrality (Brown, 2009; Daston, 1992; Haraway, 1991; Lacey, 2005; Shapin, 2008). The scientific norm of disinterestedness mandates that scientists’ work is unbiased by their personal preferences, ideologies, or social contexts (Merton, [1942] 1973). Although the history of science contains numerous violations of the norm of disinterestedness, the perceived autonomy of science’s practitioners is often cited as an explanation for the cultural authority of science. Recent scholarship, however, has challenged the notion that scientific authority should stem from scientists’ detachment from other social pressures. Douglas (2009), for example, has argued that scientists’ influence in society need not be contingent on the apparent absence of values from the scientific enterprise. Rather, by acknowledging and managing scientists’ personal beliefs in the research process, science can retain its ability to provide authoritative accounts of the world while simultaneously staking certain social and moral positions. Despite her normative argument about the appropriate relationship between science and values, Douglas’s historical analysis, along with many others, suggests that perceptions of scientists’ personal autonomy have been, and remain, central to scientific authority in political arenas. In the context of this article, these findings suggest that the interests that scientists serve—or are believed to serve—directly affect their ability to obtain authority over policy decisions. If scientists appear to be self-serving, then individuals may be less supportive of the use of scientific expertise as a basis for public policy decisions. If, however, scientists appear to be personally disinterested, or if their work appears to be publically beneficial, then individuals may be more supportive of lawmakers’ reliance on scientific knowledge across a range of policy topics.

A third characteristic of scientists, their relationship with the scientific community, may also have generalizable effects on their ability to harness authority in political settings. As well as its instrumental value, consensus within the scientific community provides important symbolic support for the claims of individual researchers (Callon, 1999; Crane, 1972; Kuhn, [1962] 1996; Latour, 1987). Specifically, as scientific opinions become embedded among those of a community of experts, they become more difficult to deconstruct and less susceptible to criticism, especially by non-experts. Several studies of the case of climate change policy have examined the effects of consensus on scientific authority, and found that certainty within the scientific community is crucial to legitimating scientists’ influence over policy decisions (Shackley and Wynne, 1996; Zehr, 2000). These studies indicate that the link between consensus and authority is so strong, in fact, that a primary tactic competing policy interests use to undermine scientific authority is to create a public impression of uncertainty among scientists. In other words, the degree to which scientists are believed to reflect the opinions of other experts matters for scientists’ ability to obtain authority among the public. While consensus science may generate increased support for scientists in policy contexts, uncertain science indicates that individual researchers are not supported by other experts. Scientists whose claims are seen as unsubstantiated by the scientific community may therefore experience decreased public support in policy settings.

Cultural authority and scientific disciplines

Previous studies have linked individual scientists’ knowledgeability, interests, and consensus to their ability to acquire scientific authority, but it remains unclear the extent to which these relationships are similar across scientific disciplines. Scholars have demonstrated that disciplines are themselves sources of scientific authority, a finding that implies that authority is distributed unequally across fields (Bourdieu, 1975; Cambrosio and Keating, 1983; Clarke, 1998; Gieryn, 1999). Studies have yet to clarify, however, how disciplines relate to other sources of scientific authority. On one hand, disciplinary differences in authority may be differences of quantity, resulting from field differences in the attribute (or attributes) that provide individual scientists the ability to harness the cultural authority of science. This account of disciplines aligns with recent conceptualizations of science that argue that science is converging into a unified global form, and that despite continued intellectual differences among disciplines, socio-cultural differences between fields of inquiry are eroding (Drori et al., 2003). On the other hand, field differences in authority may be differences of quality, reflecting fundamental differences across fields in the sources of individual scientists’ authority. This account of disciplines accords with scholarship that argues that scientific authority is differentially constructed across national and historical contexts, and by implication, across disciplines (Jasanoff, 2005; Kagan, 2009; Knorr-Cetina, 1999). This suggests that the foundations of scientists’ authority in policy settings may be mediated by scientists’ fields of study.

Disciplines of natural science are generally more authoritative than those of social science. Differences in public and private funding for natural and social sciences are telling of disparities in authority, and so are differences in public attitudes about the extent to which various fields can be classified as scientific (National Science Board, 2008). Historical accounts indicate that differences in the cultural authority of natural and social sciences have persisted since at least the mid-twentieth century (Gieryn, 1999). Thus, it may not be surprising that compared to social scientists, life and physical scientists have greater overall support in policy contexts (National Science Board, 2008). Because of these field differences in authority, support for natural and physical scientists in policy settings may be more stable than support for social scientists. If so, then the strength of the antecedents of scientific authority may vary across disciplines. In particular, the anticipated positive effects of scientists’ knowledgeability, interests, and consensus on their ability to acquire scientific authority may be relatively weaker for scientists who work in highly authoritative disciplines, such as medicine. Alternatively, the effects of these elements may be relatively stronger in less authoritative disciplines, such as economics.

3. Data, methods, and measures

Data

I address the research questions of this paper using data from a national survey of US residents. Conducted biennially by the National Opinion Research Center, the General Social Survey (GSS) contains data gathered from face-to-face interviews with a full probability sample of non-institutionalized, English-speaking adults in the US. This analysis uses data collected as part of the survey’s 2006 wave. The response rate for this wave of the survey was 71 percent (Smithet al., 2011).

The 2006 GSS contained a module on science and technology that included a series of four vignettes that described public policy topics and asked respondents a set of questions about scientists in the context of each policy issue. Respondents were asked to what extent the vignette scientists:(1) understood policy topics (knowledgeability), (2) served the nation’s best interests, as opposed to their own narrow interests (national interests), and (3) were in agreement among other scientists (consensus). Finally, respondents were asked about the level of influence that scientists should have over public policy decisions (policy influence). The GSS used a split ballot design where a randomly selected half of science module respondents received questions about environmental scientists and medical researchers in the context of global warming and embryonic stem cell research, respectively. The other half of science module respondents received questions about economists and medical researchers in the context of federal income taxes and genetically modified food. Accounting for this design feature and for cases with missing values on variables of interest, the sample size was 694 for the global warming/stem cell research ballot and 658 for the income tax/genetically modified food ballot. These data are unique in that they target attitudes about scientists from multiple disciplines whose research relates to contemporary public issues. As such, the GSS provides a new opportunity for a comparative analysis of the determinants of scientific authority from the perspective of a representative sample of US adults.

Analytical technique

The goals of this analysis are to determine (1) the relative and combined effects of scientists’ perceived knowledgeability, national interests, and consensus on public attitudes about scientists’ role in policy settings, and (2) whether these relationships vary across disciplines. To do so, I first examine responses to each of the four vignettes separately and estimate a set of binary logistic regression models that use the independent variables to predict variation in the level of influence that the public thinks scientists should have over policy decisions related to science. I then pool responses from the four vignettes and use interaction terms to estimate a series of models that tests whether the effects of independent variables on desired policy influence vary across disciplines.¹ Excluding cases with missing information on key variables, the pooled sample contains 2,799 individuals. To help interpret the results, I graph predicted probabilities based on logistic regressions.

Dependent variable: Desired policy influence

The dependent variable is the amount of influence that individuals believe that scientists should have over public policy decisions. This variable comes from survey questions that asked, “How much influence should [environmental scientists/medical researchers/economists/medical researchers] have in deciding [what to do about global warming?/about federal funding for stem cell research?/whether to reduce federal income taxes?/whether to restrict the sale of genetically modified food?]” Answers were provided on a four-point scale ranging from “none at all” to “a great deal.” The first part of the analysis examines responses to this question separately for each vignette. The second part of the analysis transforms the desired policy influence variables into person-period form and pools responses across vignettes. In person-period form, each respondent is characterized by two records—one for each vignette they received—whose value depends on individuals’ responses to each desired policy influence question (Singer and Willett, 2002). The analysis uses a binary specification of desired policy influence that signifies whether or not respondents reported that scientists should have “a great deal” of influence (i.e., the highest available category). Mean levels of the desired policy influence variables were 0.49 for environmental scientists, 0.40 for stem cell medical researchers, 0.21 for economists, and 0.49 for genetically modified food medical researchers.

Independent variables: Perceived knowledgeability, national interests, and consensus

Independent variables come from survey questions that refer to specific policy topics. These variables measure the extent to which respondents believe scientists: (1) are knowledgeable, (2) serve the national interest, and (3) reflect consensus within the scientific community. Responses to these items are measured on five-point scales where the lowest scores mean “no understanding,” “serves own narrow interests,” and “no consensus,” and the highest scores mean “understands very well,” “serves the national interest,” and “near complete consensus.”² Correlations within these sets of variables are low to moderate, ranging from an average of 0.24 between national interests and consensus variables and 0.40 between knowledgeability and national interests variables. The effects of perceived knowledgeability, national interests, and consensus on desired policy influence are first examined separately for each vignette, and then reshaped into person-period form and pooled for analysis in a single model.

Disciplines

Analyzing the four vignettes individually permits only rough cross-discipline comparisons of the effects of independent variables on support for scientists in policy settings. To perform a more rigorous test of whether the effects of independent variables vary by scientists’ field of study, the second part of the analysis combines responses across vignettes and estimates a set of models that contain discipline-based interaction terms. In person-period form, each independent variable is associated with a binary variable that signals the discipline of the corresponding scientist. Therefore, a series of interaction terms (each discipline times each independent variable) entered into a multiple regression equation can provide a test of whether the effects of perceived knowledgeability, national interests, and consensus on desired policy influence differ across scientific disciplines.

Control variables: Attitudes and knowledge about science, and respondent characteristics

Multivariate analyses control for a number of variables known to affect attitudes about science. First, science literacy is controlled for using a standardized scale of fourteen true/false questions that measured respondents’ knowledge of scientific facts and familiarity with scientific concepts (range = -1.20–0.85; alpha = 0.74). Second, respondents’ confidence in institutions is controlled for using a standardized scale of questions that asked respondents about their trust in ten social institutions (range = -1.53–1.53; alpha = 0.76).³ Third, the cultural meaning that respondents hold for science is controlled for using three standardized scales that come from a series of survey questions that asked respondents about the importance of method (range = -4.24–0.52; alpha = 0.72), location (range = -2.94–0.92; alpha = 0.60), and external validation (range = -1.60–1.49; alpha = 0.55) in making something scientific.⁴ Finally, a range of respondent characteristics are controlled for, including gender (equals one if female; mean = 0.56), race (equals one if white; mean = 0.80), region of residence (equals one if lives in South; mean = 0.23), age (in years; mean = 46.62), family income (natural log of family income category midpoints, in thousands; mean = 10.56), educational attainment (in years; mean = 13.83), political ideology (from 0 if “extremely liberal,” to 6 if “extremely conservative”; mean = 3.11), and religiosity (frequency of religious service attendance categories; mean = 3.49).

4. Results

The association between desired policy influence and perceived knowledgeability, national interests, and consensus

Table 1 presents results from a series of binary logistic regression models that examine bivariate relationships between each independent variable and its corresponding outcome. The effect of each independent variable is highly significant in the direction one would expect, and the relative magnitudes of the independent variables are similar across vignettes. For three of the four vignettes, perceptions of scientists’ national interests are the strongest predictors of respondents’ odds of reporting that scientists should have a great deal of influence over public policy decisions (.725 for environmental scientist; .382 for stem cell researcher; .556 for economist; .582 for genetically modified food researcher; p < .01). More favorable perceptions about scientists’ consensus and knowledgeability are also associated with significantly increased odds of reporting that scientists should have a great deal of policy influence, although the effects of these variables are relatively weaker.

Table 1.

Unstandardized logit coefficients from bivariate regressions of desired policy influence on perceived knowledgeability, national interests, and consensus^a.

	Environmental scientist			Stem cell medical researcher			Economist			Genetically modified food medical researcher
Knowledgeability	0.597**	−	−	0.591**	−	−	0.508**	−	−	0.400**	−	−
	(0.086)			(0.114)			(0.099)			(0.084)
National interests	−	0.725**	−	−	0.382**	−	−	0.556**	−	−	0.582**	−
		(0.093)			(0.072)			(0.109)			(0.089)
Consensus	−	−	0.578**	−	−	0.296**	−	−	0.502**	−	−	0.295**
			(0.085)			(0.080)			(0.097)			(0.082)
Constant	−2.433	−2.919	−1.990	−2.951	−1.458	−1.122	−3.322	−3.426	−2.848	−1.819	−2.530	−1.198
Observations	694	694	694	694	694	694	658	658	658	658	658	658
Pseudo R-squared	0.070	0.101	0.057	0.045	0.034	0.015	0.048	0.056	0.039	0.032	0.068	0.016

Note: Robust standard errors in parentheses.

p < 0.01.

Source: 2006 General Social Survey.

Table 2 examines whether these bivariate relationships persist in multivariate settings. Doing so will help illustrate the distinct impact of each independent variable on the outcome. Two models are presented for each vignette. The first contains the knowledgeability, national interests, and consensus variables, and the second includes control variables.

Table 2.

Unstandardized logit coefficients from multivariate regressions of desired policy influence on perceived knowledgeability, national interests, and consensus^a.

	Environmental scientist		Stem cell medical researcher		Economist		Genetically modified food medical researcher
Knowledgeability	0.320**	0.286**	0.316**	0.321**	0.284*	0.326**	0.220*	0.197*
	(0.092)	(0.094)	(0.112)	(0.116)	(0.111)	(0.116)	(0.092)	(0.095)
National interests	0.554**	0.510**	0.564**	0.494**	0.381**	0.393**	0.491**	0.461**
	(0.098)	(0.099)	(0.090)	(0.094)	(0.121)	(0.118)	(0.091)	(0.093)
Consensus	0.344**	0.327**	0.355**	0.347**	0.351**	0.324**	0.118	0.127
	(0.095)	(0.097)	(0.095)	(0.097)	(0.105)	(0.106)	(0.087)	(0.091)
Respondent’s Knowledge of science	−	0.040	−	−0.281	−	−0.329	−	−0.534*
		(0.262)		(0.254)		(0.308)		(0.246)
Institutional trust	−	−0.053	−	0.118	−	0.172	−	0.144
		(0.159)		(0.167)		(0.197)		(0.165)
Method	−	0.078	−	−0.156	−	0.063	−	0.258*
		(0.148)		(0.145)		(0.171)		(0.142)
Location	−	0.491**	−	0.592**	−	0.354*	−	0.293**
		(0.148)		(0.159)		(0.177)		(0.132)
Accord	−	−0.027	−	−0.029	−	−0.080	−	−0.259**
		(0.129)		(0.134)		(0.162)		(0.124)
Female	−	0.139	−	0.334	−	0.002	−	−0.075
		(0.184)		(0.180)		(0.220)		(0.182)
White	−	−0.165	−	−0.038	−	−0.806**	−	−0.019
		(0.239)		(0.244)		(0.244)		(0.230)
Education	−	0.028	−	0.041	−	0.066	−	0.021
		(0.039)		(0.038)		(0.050)		(0.036)
South	−	−0.045	−	0.258	−	−0.228	−	−0.354*
		(0.206)		(0.214)		(0.250)		(0.195)
Age	−	0.005	−	0.001	−	−0.005	−	−0.007
		(0.006)		(0.006)		(0.007)		(0.006)
Family income	−	−0.049	−	0.072	−	−0.057	−	−0.023
		(0.091)		(0.094)		(0.125)		(0.103)
Political views	−	−0.139*	−	−0.022	−	−0.032	−	0.079
		(0.064)		(0.067)		(0.081)		(0.062)
Church attendance	−	−0.050	−	−0.065	−	0.044	−	−0.027
		(0.035)		(0.037)		(0.042)		(0.033)
Constant	−4.707	−3.822	−5.277	−6.331	−4.923	−4.561	−3.402	−2.899
Observations	694	694	694	694	658	658	658	658
Pseudo R-squared	0.141	0.169	0.125	0.154	0.090	0.130	0.079	0.107

Note: Robust standard errors in parentheses.

p < 0.01; * p < 0.05.

Source: 2006 General Social Survey.

Table 2 provides additional evidence that favorable perceptions about scientists’ knowledgeability, national interests, and consensus have unique and positive effects on desired policy influence. Furthermore, with the exception of the perceived consensus of genetically modified food medical researchers, the effect of independent variables is similar across vignettes, even after accounting for control variables. The relative magnitudes of independent variables in Table 2 nearly mirror those contained in Table 1: the strongest predictor of support for scientists across policy contexts is the degree to which scientists are believed to serve the nation’s best interests (.510 for environmental scientist; .494 for stem cell researcher; .393 for economist; .461 for genetically modified food researcher; p < .01). Perceptions about the level of consensus in the scientific community and the extent to which scientists understand policy issues are also linked to significantly increased odds of observing the outcome. Although the coefficient for the perceived consensus of genetically modified food medical researchers is not statistically significant, its direction fits the broader pattern.

Table 2 reveals that several control variables have significant effects on desired policy influence. Most notably, the belief that location is important to making something scientific is positively associated with one’s odds of reporting that scientists should have a great deal of policy influence, a finding that holds across vignettes. This pattern aligns with prior research that finds that the cultural meaning that individuals hold for science is particularly important for understanding individuals’ attitudes about science (Gauchat, 2011). Several other significant coefficients are scattered through Table 2, but no other consistent pattern emerges across disciplines.

Scientific disciplines and the association between desired policy influence and perceived knowledgeability, national interests, and consensus

So far, the analysis has shown that perceptions of scientists’ knowledgeability, national interests, and consensus are associated with public opinions about the level of influence scientists should have over policy decisions related to global warming, stem cell research, federal income taxes and genetically modified food. Results presented in Tables 1 and 2 are suggestive of cross-discipline similarities, but comparing the relative magnitudes of regression coefficients across models is an imprecise method of determining whether the links between the independent and outcome variables are mediated by scientists’ disciplinary affiliations. To conduct a more rigorous test, I reshape the vignette variables into person-period form and combine the two halves of the GSS science module ballot. I then estimate a series of regression models that incorporates responses to each of the four vignettes (Table 3). Interaction terms included in the models provide cross-discipline comparisons of the effects of independent variables on the outcome. Because each respondent is characterized by two records in person-period form, standard errors in Table 3 are clustered by respondent (Wooldridge, 2002).

Table 3.

Unstandardized logit coefficients from multivariate regressions of desired policy influence on perceived knowledgeability, national interests, consensus, and interactions^a.

		Model 1	Model 2	Model 3	Model 4
Independent variables	Knowledgeability	0.268**	0.283**	0.328**	0.316**
		(0.049)	(0.050)	(0.091)	(0.091)
	National interests	0.533**	0.507**	0.554**	0.530**
		(0.049)	(0.050)	(0.097)	(0.098)
	Consensus	0.310**	0.279**	0.329**	0.305**
		(0.044)	(0.047)	(0.092)	(0.093)
Disciplines ^b	Stem cell medical researcher	−	−0.461**	−0.600	−0.519
			(0.103)	(0.728)	(0.742)
	Economist	−	−1.044**	−0.275	−0.545
			(0.130)	(0.761)	(0.778)
	Genetically modified food medical researcher	−	−0.060	1.370	1.352*
			(0.116)	(0.722)	(0.736)
Interaction terms	Stem cell medical researcher × knowledgeability	−	−	−0.004	−0.001
				(0.138)	(0.135)
	Economist × knowledgeability	−	−	−0.052	−0.024
				(0.142)	(0.139)
	Genetically modified food medical researcher × knowledgeability	−	−	−0.102	−0.097
				(0.126)	(0.127)
	Stem cell medical researcher × national interests	−	−	0.006	−0.032
				(0.123)	(0.124)
	Economist × national interests	−	−	−0.150	−0.127
				(0.154)	(0.151)
	Genetically modified food medical researcher × national interests	−	−	−0.083	−0.083
				(0.130)	(0.131)
	Stem cell medical researcher × consensus	−	−	0.029	0.039
				(0.125)	(0.127)
	Economist × consensus	−	−	0.020	0.035
				(0.138)	(0.138)
	Genetically modified food medical researcher × consensus	−	−	−0.212	−0.210
				(0.124)	(0.127)
Control variables	Respondent’s Knowledge of science	−	−	−	−0.252
					(0.142)
	Institutional trust	−	−	−	0.030
					(0.091)
	Method	−	−	−	0.068
					(0.084)
	Location	−	−	−	0.452**
					(0.080)
	Accord	−	−	−	−0.125
					(0.072)
	Female	−	−	−	0.119
					(0.100)
	White	−			−0.204
		−	−	−	(0.128)
	Education				0.037
		−	−	−	(0.021)
	South				−0.060
		−	−	−	(0.113)
	Age				−0.001
		−	−	−	(0.003)
	Family income				−0.019
		−	−	−	(0.052)
	Political views				−0.021
		−	−	−	(0.036)
	Church attendance				−0.027
					(0.020)
	Constant	−4.647	−4.150	−4.696	−4.480
	Pseudo R-squared	0.119	0.142	0.145	0.164

Note: Robust standard errors in parentheses.

p < 0.01; * p < 0.05.

Source: 2006 General Social Survey, n =2799; ^b referent is environmental scientist.

Results contained in Table 3 corroborate the patterns suggested by Tables 1 and 2: perceptions of scientists’ national interests are the strongest predictor of attitudes about scientists in policy contexts. Furthermore, perceptions of scientists’ knowledgeability and consensus have distinct, albeit weaker effects on support for scientists in policy settings. Table 3 provides no evidence of disciplinary differences in these relationships. Specifically, Model 1 indicates that each of the independent variables has a significant effect on the odds of observing the outcome (.268 for knowledgeability; .533 for national interests; .310 for consensus; p < .01), and Model 2 shows that the associations persist after controlling for the discipline of the vignette scientist. Significant coefficients for discipline variables reflect discipline differences in mean levels of the outcome variable. Model 3 contains the crucial series of interaction terms that test for disciplinary differences in links between independent variables and desired policy influence (discipline times knowledgeability, discipline times national interests, and discipline times consensus). Upon adding the interactions, the main effects of independent variables remain significant and their magnitudes increase slightly. Importantly, none of the nine interaction terms has a statistically significant effect of the model, meaning that the relationships between independent variables and desired policy influence are not statistically different across disciplines. A final model adds controls for attitudes and knowledge about science and respondent characteristics. Although several control variables have statistically significant effects on the model, they do not alter any of the main conclusions of the analysis.

The meaning of these results is made clearer by examining how the predicted probabilities of observing the outcome vary by discipline. In addition, graphing predicted probabilities across the range of independent variables has been proposed as a solution to an identification problem inherent to the use of interaction terms to examine group differences in non-linear regression models (Allison, 1999; Long, 2009). Figure 1 displays the predicted probabilities of observing the outcome for each discipline at different levels of the independent variables. Note that a similar linear pattern is evident for each independent variable. Although there are clear discipline differences in the chances of observing the outcome, the roughly parallel slopes of the predicted probability curves throughout the figure signify that the effect of each independent variable on desired policy influence is similar across disciplines. For example, moving from the minimum to the maximum value of the national interests variable more than doubles a respondent’s probability of reporting that medical researchers should have a great deal of influence over genetically modified foodpolicy, from about 0.3 to about 0.7. Change across the range of the perceived national interests variable is associated with a similar absolute increase in a respondent’s probability of reporting that environmental scientists should have a great deal of policy influence, from less than 0.1 to nearly 0.5. Similar interpretations can be provided for perceptions about scientists’ knowledgeability and the level of consensus in the scientific community. Overall, the data reveal striking similarities across disciplines in the associations between perceptions of scientists’ knowledgeability, national interests, and consensus and public reports about the level of influence scientists should have over political and economic decision-making.

Figure 1.

Predicted probabilities of reporting that scientists should have “a great deal” of policy influence

5. Discussion

To date, the study of the extension of science’s cultural authority across institutional boundaries has been an area dominated by small-sample, qualitative research. Survey methods cannot match the thick description offered by these in-depth studies, but public opinion research provides a valuable tool for identifying broad patterns in the extent to which scientific authority extends beyond its historical domain. In particular, the 2006 GSS provides a unique opportunity to examine US public opinion about the role of scientific expertise in public policy decisions.

Previous research has identified a gap between public and scientific worldviews, a finding that points to the importance of examining scientists’ social characteristics in addition to their intellectual ones in order to understand public attitudes about the role of science in modern life (Ezrahi, 2004; Yankelovich, 2003). In particular, the relationships between scientists’ knowledgeability, disinterestedness, consensus, and disciplines have been explored by prior research into scientific authority. However, whereas previous studies have typically examined these factors individually, this article is uniquely positioned to evaluate their relative and combined effects on the extension of scientific authority into policy contexts. The most important single factor identified by this analysis for determining public support for scientists in policy settings is the perception that scientists serve the nation’s best interests. This finding is consistent with research that suggests that the authority of science stems from its supposed autonomy from other institutions and the related neutrality of its practitioners (Brown, 2009; Daston, 1992; Haraway, 1991; Lacey, 2005; Shapin, 2008). Perceptions about the degree to which scientists are knowledgeable, and about the level of consensus that exists within the scientific community are also linked, although to a lesser degree, to public attitudes about scientists’ proper role in policy contexts. Thus, one of the primary contributions of this paper has been to clarify the relative impacts of three common explanations for scientific authority.

It is worth reiterating that these results are based on a survey of US adults. Recent public opinion research suggests that, despite some key differences, adults in the US, Canada, and Europe have similar attitudes about many aspects of science and technology in society (European Commission, 2010; National Science Board, 2008). For example, a majority of adults in the US, Canada, and Europe agree that scientists should play some role in the creation of public policy related to technical issues (Gaskell et al., 2005). Cross-national similarities in attitudes about science in society suggest that the bases of public support for scientists in policy settings may also generalize to other Western nations, a possibility that should be tested by future empirical work. Another promising avenue for future research involves identifying the links between the findings of this study and more general attitudes about science and technology. Investigators have identified an important distinction between general and particular attitudes about science and its application in public policy (Bak, 2001; Gauchat, 2011). This article has examined attitudes about scientists in several specific policy settings, but important questions remain concerning the linkages between individuals’ attitudes about scientists’ role in specific policy decisions and individuals’ general dispositions toward science and technology in society.

Although I have focused mostly on the positive effects of favorable perceptions of scientists’ knowledgeability, national interests, and consensus on scientific authority, one of the most interesting findings of this analysis is the absence of disciplinary differences in the ways that the independent variables translate into desired policy influence. Variation across disciplines in research practices, epistemic standards, and relationships with the public suggests that researchers’ ability to acquire scientific authority in political and economic spheres may vary at the level of disciplines (Hargens, 1975; Kagan, 2009; Knorr, 1981; Knorr-Cetina, 1999). This paper, however, suggests otherwise. Almost without exception, perceptions of scientists’ knowledgeability, national interests, and consensus have similar effects across fields on reports about the level of influence science advisors should have over public policy decisions. Thus, a second contribution of this paper has been to provide evidence that disciplinary differences in scientific authority result at least partly from disciplinary differences in the extent to which researchers are thought to be knowledgeable, serve the nation’s interests, and represent the opinions of the scientific community.

Discipline variables used in this analysis vary across vignettes, and therefore also vary by public issue. Consequently, one may argue that survey responses reflect attitudes toward public policies rather than attitudes toward scientists. If so, responses to vignette-specific questions should vary across vignettes more so than within them. For example, if responses reflect attitudes about public issues, then we would expect relatively higher scores on vignette variables for scientists whose work relates to popular policies, like reducing federal income taxes, and relatively lower scores for scientists whose work relates to contentious topics, like climate change. Again, however, the data indicate otherwise. Although policy issues may have some impact on responses to vignette questions, the stability of vignette-specific variables suggests that these questions primarily capture attitudes about scientists. This conclusion is consistent with other analyses of GSS science module data that find that vignette variables tap attitudes about scientists more so than attitudes about policy topics (Kennedy et al., 2011).

Given the centrality of scientific knowledge in contemporary political and economic decisions, this research relates to more than simply academic issues. Science advisors and regulators are typically appointed by political leaders, and therefore members of government’s “fifth branch” are not publically accountable in the same ways as elected officials. By voting for political leaders who are more or less receptive to considering scientific evidence as a part of policy deliberations, the public can begin to express its beliefs about the proper place of science in society. Survey research such as this provides an additional, more direct way of gauging public support for allowing scientists or other unelected leaders to influence public decisions. If, as scholars have argued, public opinion helps shape legislative action (Brooks and Manza, 2007; Burstein, 2003), then the results of this paper speak to questions about how scientists can be more effectively brought into or excluded from policy discussions. For example, this analysis suggests that change in the ways that science advisors affect political processes may be precipitated by convincing the public of scientists’ knowledgeability or lack thereof, the extent to which scientists serve national or private interests, and the degree to which consensus exists or is absent from the scientific community. In other words, this research suggests that public perceptions of scientists’ knowledgeability, national interests, and consensus can directly affect the role that science plays in public policy decisions.

Footnotes

Acknowledgements

Thanks go to Tom Gieryn, Brian Powell, the editor, and two anonymous reviewers for helpful feedback on earlier versions of this article.

Funding

Funding for this research was provided by grant number SRS 0935815 from the National Science Foundation.

Notes

Author biography

Timothy L. O’Brien is a PhD candidate in the Department of Sociology at Indiana University. His research focuses on scientific careers as well as the ways in which scientific authority is negotiated outside of science’s traditional domain.

References

Allison

(1999) Comparing logit and probit coefficients across groups. Sociological Methods and Research 28: 186–208.

Bak

H-J

(2001) Education and public attitudes toward science: Implications for the “deficit model” of education and support for science and technology. Social Science Quarterly 82: 779–795.

Barben

(2010) Analyzing acceptance politics: Towards an epistemological shift in the public understanding of science and technology. Public Understanding of Science 19: 274–292.

Bauer

Petkova

Boyadijieva

(2000) Public knowledge of and attitudes to science: Alternative measures that may end the “science war”. Science, Technology and Human Values 25: 30–51.

Bourdieu

(1975) The specificity of the scientific field and the social conditions of the progress of reason. Social Science Information 14: 19–47.

Brooks

Manza

(2007) Why Welfare States Persist: The Importance of Public Opinion and Democracies. Chicago: University of Chicago Press.

Brown

(2009) Imperfect Oracle: The Epistemic and Moral Authority of Science. University Park, PA: The Pennsylvania State University Press.

Burstein

(2003) The impact of public opinion on public policy: A review and an agenda. Political Research Quarterly 56: 29–40.

Callon

(1999) Some elements of a sociology of translation: Domestication of the scallops and the fishermen of St. Brieuc Bay. In: Biagioli

(ed.) The Science Studies Reader. London: Routledge, pp. 67–83.

10.

Cambrosio

Keating

(1983) The disciplinary stake: The case of chronobiology. Social Studies of Science 13: 323–353.

11.

Clarke

(1998) Disciplining Reproduction: Modernity, American Life Science and the Problems of Sex. Berkeley, CA: University of California Press.

12.

Crane

(1972) Invisible Colleges: Diffusion of Knowledge in Scientific Communities. Chicago: University of Chicago Press.

13.

Daston

(1992) Objectivity and the escape from perspective. Social Studies of Science 22: 597–618.

14.

Douglas

(2009) Science, Policy, and the Value-Free Ideal. Pittsburg, PA: University of Pittsburg Press.

15.

Drori

Meyer

Ramirez

Schofer

(2003) Science in the Modern World Polity: Institutionalization and Globalization. Palo Alto, CA: Stanford University Press.

16.

Epstein

(1998) Impure Science: Aids, Activism and the Politics of Knowledge. Berkeley, CA: University of California Press.

17.

Epstein

(2007) Inclusion: The Politics of Difference in Medical Science. Chicago: University of University Press.

18.

European Commission (2010) Science and Technology Report. Special Eurobarometer 340 / Wave 73.1-TNS Opinion and Social.

19.

Evans

Durant

(1995) The relationship between knowledge and attitudes in the public understanding of science in Britain. Public Understanding of Science 4: 57–74.

20.

Ezrahi

(2004) Science and political imagination in contemporary democracies. In: Jasanoff

(ed.) States of Knowledge: The Co-Production of Science and Social Order. London: Routledge, pp. 254–273.

21.

Fourcade

(2009) Economists and Societies: Discipline and Profession in the United States, Britain, and France, 1890s to 1990s. Princeton, NJ: Princeton University Press.

22.

Frickel

Moore

(2006) Prospects and challenges for a new political sociology of science. In: Frickel

Moore

(eds) The New Political Sociology of Science: Institutions, Networks, and Power. Madison, WI: University of Wisconsin Press, pp. 3–31.

23.

Gaskell

Einsiedel

Hallman

Priest

Jackson

Olsthoorn

(2005) Social values and the governance of science. Science 310: 1908–1909.

24.

Gauchat

(2011) The cultural authority of science: Public trust and acceptance of organized science. Public Understanding of Science 20: 751–770.

25.

Gieryn

(1983) Boundary-work and the demarcation of science from non-science: Strains and interests in the professional ideologies of scientists. American Sociological Review 48: 781–795.

26.

Gieryn

(1999) Cultural Boundaries of Science: Credibility on the Line. Chicago: University of Chicago Press.

27.

Guston

Keniston

(eds) (1994) The Fragile Contract: University Scientists and the Federal Government. Cambridge, MA: MIT Press.

28.

Haraway

(1991) Simians, Cyborgs, and Women. London: Free Association Press.

29.

Hargens

(1975) Patterns of Scientific Research: A Comparative Analysis of Research Work in Three Scientific Fields. ASA Rose Monograph Series, Washington, DC.

30.

Hilgartner

(2000) Science on Stage: Expert Advice as Public Drama. Palo Alto, CA: Stanford University Press.

31.

Jasanoff

(1994) The Fifth Branch: Scientists as Policy Makers. Cambridge, MA: Harvard University Press.

32.

Jasanoff

(2005) Designs on Nature: Science and Democracy in Europe and the United States. Princeton, NJ: Princeton University Press.

33.

Kagan

(2009) The Three Cultures: Natural Sciences, Social Sciences, and the Humanities in the 21st Century. Cambridge: Cambridge University Press.

34.

Kennedy

Powell

Long

O’Brien

Meanwell

Dodson

Noy

(2011) Examining the Science and Technology Items in the General Social Survey. Unpublished report to the National Science Foundation.

35.

Knorr

(1981) Social and scientific method, or what do we make of the distinction between natural and social sciences? The Philosophy of Science 11: 335–360.

36.

Knorr-Cetina

(1999) Epistemic Cultures: How the Sciences Make Knowledge. Cambridge, MA: Harvard University Press.

37.

Kuhn

([1962] 1996) The Structure of Scientific Revolutions, 3rd edition. Chicago: University of Chicago Press.

38.

Lacey

(2005) Values and Objectivity in Science: The Current Controversy about Transgenic Crops. London: Routledge.

39.

Latour

(1987) Science in Action: How to Follow Scientists through Society. London: SAGE.

40.

Long

(2009) Group comparisons in logit and probit using predicted probabilities. Working paper. Available at: http://www.indiana.edu/~jslsoc/files_research/groupdif/groupwithprobabilities/groups-with-prob-2009-06-25.pdf (accessed December 2011).

41.

Lujan

Todt

(2007) Precaution in public: The social perception of the role of science and values in policy making. Public Understanding of Science 16: 97–109.

42.

Merton

([1942] 1973) The normative structure of science. In: Storer

(ed.) The Sociology of Science. Chicago: University of Chicago Press, pp. 267–280.

43.

Miller

(1983) Scientific literacy: A conceptual and empirical review. Daedalus 112: 29–48.

44.

Miller

(2004) Public understanding of and attitudes toward scientific research: What we know and what we need to know. Public Understanding of Science 13: 273–294.

45.

National Science Board (2008) Science and Engineering Indicators, 2008. Arlington, VA: National Science Foundation.

46.

Nelkin

(1975) The political impact of technical expertise. Social Studies of Science 5: 35–54.

47.

Nowotny

(2005) High- and low-cost realities for science and society. Science 308: 1117–1118.

48.

Schweber

(2006) Disciplining Statistics: Demography and Vital Statistics in France and England, 1830– 1885. Durham, NC: Duke University Press.

49.

Shackley

Wynne

(1996) Representing uncertainty in global climate change science and policy: Boundary-ordering devices and authority. Science, Technology and Human Values 21: 275–302.

50.

Shapin

(2008) Science in the modern world. In: Hackett

Amersterdamska

Lynch

Wajcman

(eds) The Handbook of Science and Technology Studies. Thousand Oaks, CA: SAGE, pp. 433–448.

51.

Singer

Willett

(2002) Applied Longitudinal Data Analysis: Modeling Change and Event Occurrences. Oxford: Oxford University Press.

52.

Smith

Marsden

Hout

Kim

(2011) General social surveys, 1972–2010 [machine-readable data file]/Principal Investigator, Smith TW; Co-Principal Investigator, Marsden PV; Co-Principal Investigator, Hout M; Sponsored by National Science Foundation. NORC ed. Chicago: National Opinion Research Center [producer]; Storrs, CT: The Roper Center for Public Opinion Research, University of Connecticut [distributor].

53.

Thorpe

(2002) Disciplining experts: Scientific authority and liberal democracy in the Oppenheimer case. Social Studies of Science 32: 525–562.

54.

Wooldridge

(2002) Econometric Analysis of Cross Section and Panel Data. Cambridge, MA: MIT Press.

55.

Yankelovich

(2003) Winning greater influence for science. Issues in Science and Technology Summer: 7–11.

56.

Yearley

(2000) Making systematic sense of public discontents with expert knowledge: Two analytical approaches and a case study. Public Understanding of Science 9: 105–122.

57.

Zehr

(2000) Public representations of scientific uncertainty about global climate change. Public Understanding of Science 9: 85–103.