[1] The patterns in the use of data sources do not necessarily mean that people are getting information from less-detailed sources. Newspapers and the Internet include long articles, and the Internet contains links to additional sources of information. In addition, declining reliance on magazines may result from short-term causes, such as a few S&T magazines going out of business without new ones immediately filling the market niche, rather than from a long-term change in information-seeking patterns.

[2] Like most survey data, General Social Survey (GSS) data, used in figure 7-3 and elsewhere in this chapter, are weighted to make them correspond more closely to known parameters in the general population, such as sex and race distributions. In tables and figures that compare different survey years, the data are presented using a weighting formula that can be applied to all years. In tables that present only 2006 survey results, numbers are calculated using a new weighting formula that is designed to produce more accurate figures for that year. As a result, there may be minor discrepancies between the 2006 GSS results that appear in different tables and figures.

[3] A survey that called attention to particular sources of information on the Internet, such as Weblogs, might well produce different results.

[4] Although health news and science and technology news may appear to be closely related categories, the profile of people who follow each type of news closely is different: 63% of Americans who follow health closely are women, whereas 69% of Americans who follow S&T news closely are men (Pew Research Center for the People and the Press 2006a). Many researchers stress that both interest in and knowledge about S&T are often specific to individually defined domains within this broad category and do not generalize to the category as a whole (see sidebar, "Asset-Based Models of Knowledge").

[5] Science, space and technology includes manned and unmanned space flight, astronomy, scientific research, computers, the Internet, and telecommunications media. It excludes forensic science, S&E education, and telecommunications media content. Biotechnology and basic medical research includes stem cells, genetic research, cloning, and agribusiness bioengineering. It excludes clinical research and medical technology. Stories often do not fall neatly into a single category.

[6] People can become involved with S&T through many other non-classroom activities. Participating in government policy processes, going to movies that feature S&T, bird watching, and building computers are a few examples. Data on this sort of involvement with S&T are unavailable.

[7] It is possible that the substantial difference between Pew and IMLS estimates for "zoos or aquariums" is the result of differences in the categories the two surveys offered to respondents. Both surveys asked about zoos and aquariums, but IMLS also asked about nature centers and children's or youth museums, whereas Pew did not. Pew respondents who visited these kinds of museums may have reasoned that "zoo or aquariums" was the most closely comparable category in the survey and classified their visits accordingly.

[8] The NSF surveys asked respondents the number of times in the past year that they have visited an art museum, a natural history museum, a science or technology museum, a zoo or aquarium, or a public library. The Pew survey asks them whether or not they have visited one of the institutions, and includes planetarium in the list. For the S&T-related institutions in the list, the historic NSF numbers are about 4 percentage points higher than the Pew numbers, but the difference may have to do with how the questions were asked. Some research suggests that when surveys ask for the number of times respondents have engaged in an activity, the percentage saying they have engaged in the activity at least once is larger than the percentage who would answer "yes" if asked whether they had engaged in the activity at all, probably because some respondents experience the first type of question as implying that the activity is common or acceptable (Knauper 1998; Sterngold, Warland, and Herrmann 1994). The IMLS survey's institution categories are sufficiently different from the NSF categories to make focused comparisons over time problematic.

[9] The IMLS survey only asked about children's museum visits in households where adults had visited a museum in 2006, either in-person or remotely via the Internet. Because IMLS assumed that children in other households did not visit museums, there is reason to believe that the actual percentages are somewhat higher than the IMLS estimates.

[10] One possible explanation for differences between Europe and the United States in attendance at informal science institutions is that adult leisure patterns reflect patterns that developed in childhood, when, especially for older Europeans, informal science institutions were less readily available than in the United States. The available national data do not permit a test of this explanation.