Notes

[17] Methodological issues make fine-grained comparisons of data from different survey years suspect. Although the question content and interviewer instructions were identical in 2004 and 2006, for example, the percentage of respondents who volunteered "about equal" was substantially different. This difference may have been produced by the change from telephone interviews in 2004 to in-person interviews in 2006 (though telephone interviews in 2001 produced results that are similar to those in 2006). More likely, customary interviewing practices in the three different organizations that administered the surveys affected their interviewers' willingness to accept responses other than those that were specifically offered on the interview form, including "don't know" responses.

[18] The English version of the European question reads, "The benefits of science are greater than any harmful effects it may have." Respondents can strongly agree, tend to agree, neither agree nor disagree, tend to disagree, strongly disagree, or say that they do not know. The U.S. question is prefaced by the statement that "People have frequently noted that scientific research has produced benefits and harmful results" and asks the respondent, "Would you say that, on balance, the benefits of scientific research have outweighed the harmful results, or have the harmful results of scientific research been greater than the benefits." Respondents who say that the benefits are greater are then asked whether "the balance has been strongly in favor of the benefits, or only slightly." Respondents who say the harmful results are greater are asked a parallel question to distinguish strongly from slightly. Some respondents are recorded as saying that the benefits and harmful results are "about equal" when they volunteer this response.

Although these questions differ in their references to "science" and "scientific research," "effects" and "results," and in the exact wording of the response categories, they are similar in their overall thrust and in the availability of a middle category ("neither agree nor disagree," "about equal"). For other questions that are worded similarly in the 2005 Eurobarometer and either the 2004 or 2006 NSF surveys, the presence of a middle category in Europe and the absence of one in the United States makes direct comparison problematic. This lengthy, though incomplete, comparison regarding a single question pair should provide some indication of why international attitude comparisons should be treated with caution.

[19] Unlike the U.S. question, the European question joins two logically independent ideas—more spending on science and less spending on other priorities. In addition, because nations begin from different levels of spending, survey responses cannot be read as indicating different views about the proper level of spending in this area, nor do they indicate the strength of sentiment in different countries. Differences in the connotations of questions posed in different languages add further complexities. Perhaps for some or all of these reasons, variations among European countries in responses to this question are large, with about two-thirds of respondents agreeing in Italy, Spain, and France, but less than one-third in Finland and the Netherlands.

[20] Some Americans may think that science can resolve differences over what to value or settle policy questions without requiring value judgments. This view accords science a kind of influence that goes beyond what the scientific community thinks it can properly exercise. There are no survey data that indicate how many Americans accord science too much influence in this regard.

[21] Although these questions treat economists as scientists and compare them to other categories of scientists, data reported later in this chapter indicate that many Americans do not consider economics to be very scientific. To understand public perceptions of the role of science and scientists in dealing with contested public issues, it helps to have indicators both for disciplines that the public almost universally sees as scientific and for disciplines whose scientific status is less secure in the public's eyes. Many social scientists (e.g., Gieryn 1999) believe that much can be learned from research on how institutional boundaries are defined and maintained. Universities overwhelmingly categorize economics as a social science.

[22] These question batteries were designed as indicators of public views regarding the appropriate influence of science on public issues generally. Questions were posed concerning specific issues both because (1) this is likely to increase the degree to which respondents think of similar situations when they make judgments and (2) because views about the appropriate role of science are likely to depend heavily on context. A study of any one of the specific issues would likely make somewhat different distinctions and ask more and different questions about the topic.

Three other issues are worthy of mention: (1) Because survey respondents are variably familiar with the issues posed in these questions, certain categories are characterized with significant imprecision. For example, "medical researchers" is not an optimal characterization of the kind of researchers who are experts on the health effects of genetically modified foods. (2) Judgments that affect trust in leaders may be difficult to capture in survey questions. A concept such as disinterestedness, for example, (in the sense of a judgment made and expressed in light of appropriate collective interests and independent of personal interests that are not supposed to be given any weight) likely cannot be stated in language that can be used in a survey. (3) Comparable data on other issues is lacking, which makes generalizing observed patterns to other issues hazardous. Just as it is uncertain how attitudes that are highly general shape concrete judgments, it is uncertain how more specific judgments generalize beyond the terms in which they are posed. Because different attitude indicators have different limitations, it can be valuable to have indicators with complementary strengths and flaws. In all cases, it is worth keeping the actual question wording in mind when interpreting the significance of patterns in the data.

[23] The questions were worded as follows:

-"How much influence should each of the following groups have in deciding what to do about global warming? a. Environmental scientists. Would you say a great deal of influence, a fair amount, a little influence, or none at all?" This wording was then repeated in the next two questions, except that "elected officials" and "business leaders" were substituted for environmental scientists.

-"How much influence should each of the following groups have in deciding about government funding for stem cell research? a. Medical researchers. Would you say a great deal of influence, a fair amount, a little influence, or none at all?" This wording was then repeated in the next two questions, except that "religious leaders" and "elected officials" were substituted for medical researchers.

-"How much influence should each of the following groups have in deciding whether to reduce federal income taxes? a. Economists. Would you say a great deal of influence, a fair amount, a little influence, or none at all?" This wording was then repeated in the next two questions, except that "business leaders" and "elected officials" were substituted for economists.

-"Some say that the government should restrict the sale of genetically modified foods. Others say there is no need for such restrictions. How much influence should each of the following groups have in deciding whether to restrict the sale of genetically modified foods? a. Medical researchers. Would you say a great deal of influence, a fair amount, a little influence, or none at all?" This wording was then repeated in the next two questions, except that "elected officials" and "business leaders" were substituted for medical researchers.

[24] The questions were worded as follows: "On a scale of 1 to 5, where 1 means "very well" and 5 means "not at all," how well do the following groups understand" each of four public issues: "the causes of global warming," "stem cell research," "the likely effects of reducing federal taxes," and "the risks posed by genetically modified foods." For global warming, respondents were asked about environmental scientists, elected officials, and business leaders. For stem cell research, respondents were asked about medical researchers, religious leaders, and elected officials. For federal taxes, respondents were asked about economists, business leaders, and elected officials. For genetically modified foods, respondents were asked about medical researchers, elected officials, and business leaders.

[25] The questions were worded as follows: "When making policy recommendations about" each of four public issues "on a scale of 1 to 5, to what extent do you think the following groups would support what is best for the country as a whole versus what serves their own narrow interests?" The issues were "global warming," "stem cell research," "federal income taxes," and "genetically modified foods." If asked about what narrow interests meant, interviewers were instructed to respond "Well, someone might gain financially if a certain policy were adopted or it might advance his or her career."

[26] Three of the four questions were worded as follows: "On a scale of 1 to 5, where 1 means "near complete agreement" and 5 means "no agreement at all," to what extent do" groups of scientists "agree on" an issue. The groups and issues were "environmental scientists/the existence and causes of global warming," "medical researchers/the importance of stem cells for research," "economists/the effects of reducing federal income taxes" and "medical researchers/the risks and benefits of genetically modified foods." The global warming question read "agree among themselves about" instead of "agree on."

[27] Among the considerations that might be considered relevant are the role of ordinary citizens whose interests are specially affected by a decision and the institutional context for a decision (e.g., public versus private, different branches or levels of government). There is an extensive literature, analyzing mostly qualitative and nonnational data, that explores the complexities in when and why the public treats scientists and others as having the authority to influence or make decisions. Although attempts to synthesize that literature and clarify its relationship to what can be learned from national surveys would be welcome, this kind of multivariate analysis and interpretation goes well beyond the scope of this document.