Survey Data Sources

Media Effects

Citizens of economically advanced societies live in a world that is permeated by mass media of communication. A large social science literature probes how these media operate, what kinds of messages they send, how they tailor their messages to reach different audience segments, and how those messages relate to public opinion. Mass media messages interact with the opinions of the American public in complex ways, and teasing out the reciprocal effects is complicated (Perse 2001).

Providers of media content are not free to supply whatever content they prefer. In making content decisions, the people who own and manage media organizations take into account the views of the segments of the public that purchase their products and are well aware that their audiences can select other content providers. Likewise, the journalists who gather information and report stories for mass media transmission, whatever their personal views, are guided by the standards of the organizations for which they work and the professions in which they are trained. In addition, they are typically motivated by the desire to make an impact on a large audience by presenting stories in compelling and dramatic ways.

At the same time, the mass media do not simply reflect the public they serve. Members of the public are dependent on mass media for much of their information about public issues, either through direct exposure or second hand from friends and relatives. Because Americans tend to rely on sources of information that typically adopt a perspective akin to their own, the ways trusted mass media pose new or less familiar issues can assume great importance. Moreover, even for members of the public who search out multiple points of view on an issue, the shared terms and assumptions in the media shape how they think about issues. Interested parties, including newsmakers, are increasingly sophisticated in crafting messages to capture media attention and appeal to the public.

Studies that seek to isolate the effects of mass media face numerous challenges:

• Laboratory research, which can control for factors other than media exposure that influence people’s opinions, has an uncertain relationship to real world situations, in which people choose media programs, interpret media messages through conversations with others, and pay varying degrees of attention to what is said in the mass media. It is difficult to recreate these conditions in laboratories.
• Even when research can demonstrate short-term effects of media exposure, it is hard to know how much these persist over time or affect behavior in natural settings.
• People interpret a media message differently depending on the beliefs they bring to it and the attention they give it. Thus, media messages may affect individuals, but, because the effects are not uniform and can run in opposite directions, aggregate opinion may be left almost unchanged.
• Media messages may have more to do with motivating people who already hold an opinion to become more active in civic and political contexts than with persuading people to adopt new opinions. Surveys may have difficulty capturing this kind of effect.
• In a society with multiple sources of media content, even highly influential media sources, such as programs on major television networks, reach only a fraction of the population and, at any given time, change the perspectives of only a fraction of the people they reach.
• Mass media messages are significantly shaped by events—what is actually taking place in the situations about which they are reporting. Although facts are open to various interpretations and presentations, they set limits beyond which media organizations cannot go without losing credibility.
• It is easy to demonstrate correlations between media content and shifts in public opinion, but hard to demonstrate causation. Sometimes, changes in media content reflect changes in elite opinion or actual circumstances, rather than changes initiated or caused by the mass media themselves.
• Media exposure may have threshold effects, in which a certain amount of repetition is necessary for a message to get through, but beyond that amount further repetition has little or no impact. Compared with effects that work incrementally, threshold effects are harder to isolate.

Recent research in communications has stressed the role of the mass media in shaping the agenda for public debate and political action (agenda-setting) and the terms in which the public sees an issue (framing) (Scheufele and Tewksbury 2007). Agenda-setting works largely through making a topic more salient and accessible to memory by frequent or more prominent mention of it, thereby increasing the public’s sense that the topic is important. Framing refers to ways that mass media construct stories to make a topic comprehensible and relevant to the public. Frames stress some aspects of a topic and minimize others. Some kind of framing is necessary to reduce complexity and provide a focus to make sense of what would otherwise be undigested facts. Interested parties vie to get the mass media to present topics in their preferred frames. Research on how S&T are discussed in the mass media has identified competing frames that have been used to present contested issues (Gamson and Modigliani 1989; Nisbet and Lewenstein 2002). Recognizing that most members of the public pay limited attention to S&T information, some researchers have argued that representatives of the scientific community need to do more to influence how the mass media frame issues (Nisbet and Mooney 2007; Scheufele 2006). In their view, when it comes to influencing public opinion, influencing the frames through which the public processes and understands science-related issues may be more important than increasing the scientific and technical content of news coverage.

What Are Science and Technology?

When Americans refer to science and technology (S&T), they rarely define their terms. Ordinary language rests on the assumption that terms such as these, even if their precise meanings are not quite the same for everyone, invoke a bundle of associations that are similar enough to enable people to communicate. Survey research gathers attitude data about how people respond to the ill-defined linguistic bundles, such as S&T, that people use in ordinary conversation.

For purposes of analysis and comparison, research studies usually classify topics in the news in a way that makes space, environment, and health and medicine separate from S&T. The meaning respondents ascribe to a topic category, such as S&T, is affected by the context in which it appears and the other categories listed with it.

In interpreting survey data that use these terms, it is important to take into account the uncertainties surrounding the meaning of S&T. For example, it is not clear how often survey respondents who are asked about "science and technology" think they are being asked about two separate entities about which they might have different interests or attitudes, rather than about a single complex whole. Likewise, although engineers often think of technology as a broad category of devices and systems that humans construct to solve problems and interact with their environments, there is some evidence that for many people the term technology refers more narrowly to electronic information technology, especially computers (Cunningham, Lachapelle, and Lindgren-Streicher 2005; Rose and Dugger 2002)

Asset-Based Models of Knowledge

Many researchers and educators interested in the public’s understanding of science advocate studying the assets people bring to bear on scientific issues that they deal with in their daily lives. Because individuals encounter S&T in different ways, they acquire different S&T knowledge "assets," which they then can use to make sense of unfamiliar issues. For researchers and educators who favor an asset-based model of scientific literacy, public understanding of science is less a "generalized body of knowledge and skills that every citizen should have by a certain age" than "a series of specific sets of only moderately overlapping knowledge and abilities that individuals construct over their lifetimes" (Falk, Storksdieck, and Dierking forthcoming). In education, asset-based perspectives on knowledge have been useful in helping teachers build on children’s existing strengths to improve their performance.

Generalized assessments of S&T knowledge, by asking questions on topics that may be of little interest to many respondents, may underestimate the assets available to individuals when they deal with S&T matters of greater interest and consequence to them. In contrast, a knowledge assessment that is tailored to an S&T domain with which an individual is familiar might yield very different results. In addition, because people often use their knowledge assets in group interactions, such as a nature outing, some researchers question the value of individual assessments in a test or survey (Roth and Lee 2002).

National indicators that evaluate domain-specific knowledge or group problem-solving are not practical. Surveys cannot use different measures to enable gardeners, auto mechanics, and amateur astronomers to demonstrate their different S&T-related assets and then reliably aggregate the results from different S&T domains. Nonetheless, a perspective on scientific literacy that stresses domain-specific or group assets is useful in that it points to a significant limitation of generalized indicators of individual scientific literacy

Science Knowledge and Civic Knowledge

Political scientists have collected data on how much Americans know about U.S. civic institutions, politics, and history. In an exhaustive review of 50 years of research on civic knowledge, Delli Carpini and Keeter (1996) find patterns that are very similar to those in the distribution of scientific knowledge. More recent data give no indication that these patterns have changed (Pew Research Center for the People and the Press 2004, 2007b).

The following survey results, culled from a long list of knowledge questions about civic institutions and processes, give a flavor of what Americans do and do not know (Delli Carpini and Keeter 1996:70–1):

• Can correctly define Presidential veto (89% in 1989).
• Know that the First Amendment protects free press/speech (75% in 1985).
• Know that English is not the official national language (64% in 1986).
• Can state the substance of the Brown v. Board of Education decision (55% in 1986).
• Know that Congress declares war (45% in 1987).
• Know the length of a term of office in the U.S. House of Representatives (30% in 1978).

These data suggest that limited public mastery of fundamental factual information is not a problem that is unique to S&T.

Patterns in civic knowledge closely parallel those for science knowledge. Thus, much as individuals who demonstrate knowledge of the scientific process (see "Understanding the Scientific Process") also tend to score well on factual knowledge questions, people who are more familiar with the rules that govern civic institutions also tend to be more knowledgeable about political figures, parties, and the substance of public policy. The data on civic knowledge also parallel the data on science knowledge in other respects: political knowledge is strongly associated with formal education, women and minority group members tend to score somewhat less well on knowledge measures, more knowledgeable Americans tend to express more interest in political and civic matters and rely more on longer written sources of information, and political knowledge is associated with higher income.

There are some minor differences, too. Older Americans tend to be better informed about civic matters but not about science. Unlike science knowledge, Americans’ civic knowledge shows no signs of increasing over time and appears to be slightly weaker than that in other developed countries.

Divisions among scholars over the implications of data on Americans’ civic and science knowledge follow similar lines (Delli Carpini and Keeter 1996; Lupia 2006; Nisbet 2003; Toumey 2006). Some stress that by trusting knowledgeable people, Americans can adequately perform necessary tasks without acquiring much civic or scientific knowledge. Others stress that considerable knowledge is required as context for deciding whom and what to trust. Similarly, for some scholars, singling out civic or scientific knowledge as distinctively valuable amounts to imposing elite preferences on people who would rather not spend time learning about either science or politics. To others, however, knowledge of these domains seems central to active problem-solving and participation in the shared cultural life of a modern society.

Evolution and the Schools

The American Association for the Advancement of Science (AAAS) gave its annual Award for Scientific Freedom and Responsibility for 2006 to 10 people "who have been on the front lines of the battle to prevent introduction of ‘intelligent design’ into science classrooms as an alternative to evolution" (AAAS 2007). According to Dr. John Marburger, the head of the White House Office of Science and Technology Policy, the theory of evolution is "the cornerstone of modern biology" (Bumiller 2005). In a March 4, 2005, letter to National Academy of Sciences (NAS) members, Dr. Bruce Alberts, then president of NAS, characterized the theory of evolution as "one of the foundations of modern science," urged America’s leading scientists to help in their states and localities "to confront the increasing challenges to the teaching of evolution in the public schools," and cited the succession of NAS efforts devoted to ensuring that evolution is taught appropriately (Alberts 2005).

Yet, despite endorsements of evolution from these and other representatives of the scientific and political establishment, controversy over how evolution should be taught in public schools remains a perennial feature of American life and shows no sign of disappearing. Instead, the controversy is evolving.

Eight of the AAAS awardees were science teachers in the Dover, Pennsylvania, school district who fought their school board’s decision to require that they read a disclaimer about the theory of evolution to their ninth grade biology students. After stating that evolution was a theory, not a fact, and had "gaps," the disclaimer directed students’ attention to intelligent design, "an explanation of the origin of life that differs from Darwin’s view."*

The Dover disclaimer was successfully challenged in court (Kitzmiller v. Dover 2005). The case turned on whether the disclaimer violated the Establishment Clause of the First Amendment to the U.S. Constitution, which deals with the relationship between government and religion. The court concluded that intelligent design was a religious view and not a scientific theory and that, because the school board’s policy was animated by a religious purpose and had a religious effect, neither the policy nor the disclaimer that implemented it was constitutional.

In reaching this conclusion, the court reviewed the history of efforts to have biblical views of the origins of life taught in the public schools, the legal decisions that posed obstacles to these efforts, and the subsequent efforts to exclude the teaching of evolution from the schools or undermine the scientific status of the theory in the eyes of high school students. It traced a succession of legal conflicts in which the teaching of creationism and creation science had been found to violate the Establishment Clause and that had led to the development of intelligent design.

The Discovery Institute (2007), a Seattle policy and research organization, is the leading proponent of intelligent design. The Discovery Institute characterizes itself as a secular institution and maintains that intelligent design is not based on the Bible and is not the same as creationism. It does not advocate requiring that intelligent design be taught in schools. Rather, it "recommends that states and school districts focus on teaching students more about evolutionary theory, including telling them about some of the theory’s problems." At the same time, it believes "there is nothing unconstitutional about discussing the scientific theory of design in the classroom." Framed in this way, intelligent design may appear to be more distant from religion and less vulnerable to legal challenge than doctrines such as creationism and creation science, which have failed to pass constitutional muster (for a discussion of framing, see sidebar, "Media Effects").

Even where, as in Dover, legal controversies over the teaching of evolution are resolved with affirmations of scientific evidence and criteria, thorough and substantive presentation of the theory of evolution in the schools is by no means guaranteed. The possibility that parents and students may object to the teaching of evolution, let alone evidence of organized efforts to resist it, may discourage some teachers from covering the topic in depth (Dean 2005). In addition, not all high school biology teachers subscribe to the accepted view of evolution or are well versed in the topic (Monastersky 2006).

Numerous efforts are under way in the scientific community to make materials available to middle and high school teachers that will help them do a better job presenting the scientific evidence about evolution (Holden 2006; Monastersky 2006). Niles Eldredge, a prominent researcher in evolutionary biology, has announced plans to initiate a new journal, Evolution: Education and Outreach, to serve as a resource for teachers at all levels who wish to improve their treatment of the topic. The journal is scheduled to begin publication in March 2008 (Monastersky 2007).

Attitudes and Question Wording

In the first paragraph of a May 16, 2006, press release, the Coalition for the Advancement of Medical Research (CAMR) reported that "nearly three-quarters of Americans support embryonic stem cell research." Two weeks later, the United States Conference of Catholic Bishops (USCCB) issued a press release. The first paragraph of that press release indicated that "48% of Americans oppose federal funding of stem cell research that requires destroying human embryos, while only 39% support such funding" (CAMR 2006; Levin 2006; Nisbet 2004; USCCB 2006).

How could two surveys, conducted by telephone 2 weeks apart and using similar methodologies, arrive at such different results?

The answer lies in wording and context (Schuman and Presser 1996). To their credit, later in their press releases, both organizations provided the wording of the actual questions respondents were asked:

CAMR question: I’m going to read you a brief description of embryonic stem cell research, and then get your reaction. Embryonic stem cells are special cells that can develop into every type of cell in the human body. The stem cells are extracted from embryonic cells produced in fertility clinics and then frozen days after fertilization. If a couple decides that the fertilized eggs are no longer needed, they can choose to donate the embryos for research or the clinic will throw the embryos away. Scientists have had success in initial research with embryonic stem cells and believe that they can be developed into cures for diseases such as cancer, Parkinson’s, heart disease, juvenile diabetes, and spinal cord injuries. Having heard this description, do you strongly favor, somewhat favor, somewhat oppose, or strongly oppose medical research that uses stem cells from human embryos?

USCCB question: Stem cells are the basic cells from which all of a person’s tissues and organs develop. Congress is considering the question of federal funding for experiments using stem cells from human embryos. The live embryos would be destroyed in their first week of development to obtain these cells. Do you support or oppose using your federal tax dollars for such experiments?

These two questions provide very different contextual information about stem cell research. To the organizations that sponsored the two surveys, the questions doubtless present the most relevant information for informed decisions. Most members of the public do not follow issues such as stem cell research very closely (Pew Research Center for the People and the Press 2006b), and the way questions are framed can influence their views.

Even neutral survey organizations ask questions in different ways and produce different results. Their questions, although generally more useful for scientific research on public attitudes, neither present a "correct" context, create a situation in which context plays no role in how people respond, nor establish a context that closely approximates the one in which most citizens make decisions. Because survey responses are affected by subtle differences in wording and context, thoughtful researchers pay attention to precisely how questions are asked, give more weight to patterns and trends in survey results than to the percentage of people who choose a particular response, and examine the degree to which responses are stable across different surveys on the same topic.

How Knowledge Relates to Attitudes

In an analysis of data from almost 200 nationally representative surveys conducted in 40 countries between 1989 and 2003, Allum et al. (2008) examined how knowledge of science relates to attitudes toward S&T. Data are mostly from Europe and North America, but suitable surveys from countries in other regions were also included; these tended to be economically developed countries, such as Japan and New Zealand.

The analysis divided knowledge indicators into two groups depending on whether they involved general knowledge of scientific facts and processes or knowledge of a relatively specific scientific domain such as biology or genetics. It grouped attitude indicators by topic, distinguishing among science in general, nuclear power, genetic medicine, genetically modified food, and environmental science.

To isolate the relationship between knowledge and attitudes, the study used statistical techniques to control for factors that might be expected to influence both knowledge and attitudes, such as the age, sex, and education level of the respondent and the country in which the survey was conducted. Controlling for these influences, it reached several conclusions:

• There is "a small positive correlation between [favorable] general attitudes toward science and general knowledge of scientific facts and processes." Though small, this relationship appears consistently across countries.
• The relationship is stronger in the United States than in any of the other countries studied.
• The strength and nature of the relationship between knowledge and attitudes did not vary systematically over time during the period studied.
• Favorable attitudes about topics in a particular domain are more closely related to knowledge in that domain than to general science knowledge. Attitudes about genetically modified food, for example, show a stronger relationship to knowledge about biology and genetics than to general science knowledge.
• Contrary to findings in some other, less comprehensive studies, the relationship between knowledge and attitudes did not vary depending on differences in the level of economic development of the countries studied.

The study does not establish a causal link between knowledge and attitudes. Indeed, the authors conclude that "scholars have overlooked the need to provide a satisfactory account of how knowledge of science relates to preferences regarding its technological implementation in society," and recommend that researchers address "the social and psychological mechanisms that generate the associations we observe."

Designs on Nature

In Designs on Nature (2005), Sheila Jasanoff analyzed how the United States, Great Britain, and Germany have grappled with recent developments in biotechnology. Her study sought to explain numerous differences among these three leading S&T powers in the kind of political dynamics spawned by biotechnology:

• Agricultural biotechnology generated much more concern in British public life than in either Germany or the United States.
• Embryo research was relatively uncontested in Britain, publicly divisive in the United States, and debated in institutionalized governmental forums in Germany without becoming a salient political issue for a wider public.
• Patenting life forms was seen as an ethical issue in Europe but not in the United States.
• All three nations considered bioethics important, but each understood it very differently.

For Jasanoff, differences in public opinion do not, for the most part, account for these political differences. Rather, differences in political culture and institutions shape when, whether, and how public opinion is mobilized in the political arena and becomes a significant force affecting biotechnology issues. Often, elite deliberations are relatively insulated from public attitudes, and elite politics plays a large role in how the public defines and responds to new scientific issues.

Jasanoff points to differences in how knowledge becomes publicly validated in the three countries, differences that affect "national discourses of risk and safety, naturalness and artificiality, innovation and ownership, constitutional rights, and bioethics" (Jasanoff 2005, pp. 20–1). Differences in public discourse, combined with differences in regulatory approaches, legal institutions, and styles of managing conflict, affect how these countries respond to the new ethical and policy challenges posed by biotechnology. Although countries tend to respond in accordance with long-standing cultural and institutional patterns, Jasanoff observed that countries also alter and adapt these patterns to deal with the novel issues that biotechnology raises.

For each country she studied, Jasanoff identified a dominant cultural and institutional paradigm that characterizes its general approach to issues at the intersection of science and politics:

• United States. In a predominantly "contentious," adversary process, groups with competing interests vie to define relevant knowledge. Courts loom unusually large as arbiters of disputes, and federal administrators are relatively passive. Public optimism about technology creates an environment that is open to experimentation unless there are demonstrated risks or pre-existing regulatory barriers. New technologies often validate themselves only after they are introduced, by not causing unacceptable harms. Skepticism about expertise makes it difficult to resist demands for quantified measures, formal credentials, and transparent decision processes. Science is viewed as a sphere of objective knowledge separate from "the contaminating touch of politics" (Jasanoff 2005, p. 288).
• Great Britain. Biotechnology policy is developed in an atmosphere in which the credibility of state-regulated science has been damaged by the nation’s experience with mad cow disease. Public trust in experts imbued with an ethic of public service and a reputation for character and judgment, although damaged, remains a key resource for validating scientific knowledge. Scientific experts associated with the government are trusted to be able to consult with affected parties, gather relevant information, and reach objective decisions that "discern the public’s needs" (Jasanoff 2005, p. 268). Transparency is more an option than a requirement.
• Germany. Decisionmaking is consensus oriented, with interested parties participating in institutionalized deliberative processes organized by the federal government. Public debate is largely restricted to matters of values, and technical and factual issues are reserved for expert committees whose work is largely removed from public view. Committee members derive their stature from public trust in the institutions they represent. The public assumes that the state can assemble competent expert bodies composed of reasonable individuals who, although they reflect diverse interests and perspectives, can negotiate a shared view of the public interest.

Jasanoff emphasized that these patterns, though resilient, are not rigid, and that actual political processes are more fluid than the central tendencies she described.