Public opinion on S&T includes beliefs about the general promise and benefits of scientific research for society as well as awareness and perceptions of specific scientific topics, including those of recent interest like research on severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) and COVID-19; citizen science; and artificial intelligence (AI), robotics, and automation technology. Social science research also highlights new insights about public perception of long-standing concerns such as climate change and the state of science education in the United States.

Americans’ support for S&T as a general enterprise has been consistently quite positive for at least four decades. For nearly five decades, the General Social Survey (GSS)—a nationally representative survey of adults in the United States—has assessed Americans’ perceptions of S&T (Smith et al. 2012–18). From 1979 to 2018, the GSS found a clear majority of American adults agreed that the benefits of scientific research strongly or slightly outweigh the harmful results (Figure PPS-1). From 1992 to 2018, the GSS also found that most Americans surveyed believed that there would be more opportunities “for the next generation” because of S&T (Figure PPS-2) and that they supported federal funding for basic scientific research, even when they did not expect that research to produce immediate benefits.

One exception to Americans’ tendency to support S&T has been the perception that science makes life change too fast. In the last decade, Americans have been almost evenly split about the view that science has such a downside (Figure PPS-2). From 2010 to 2018, the GSS found that roughly half of respondents agreed or strongly agreed that “science makes our way of life change too fast,” moving up from an average of 38% from 1995 to 1999 to an average of 50% from 2014 to 2018.

Americans also have tended to report that they trust in science, and that stance is similar to residents of the other countries that spend the most on S&T R&D compared to the rest of the world. According to the 2018 Wellcome Global Monitor survey (Gallup 2019b)—the world’s largest study on how people around the world think and feel about science and major health challenges—a majority of Americans surveyed reported that they trust science “some” or “a lot.” This stance was consistent with citizens in the top 16 countries with the largest gross domestic expenditure on R&D as a percentage of gross domestic product (GDP) as of 2017 (Figure PPS-3).

Despite Americans’ general endorsement of science and the stability of their general perceptions of science over time, there are some notable differences in confidence in S&T between some groups. One source of those variations is the extent to which people understand how scientists conduct research and use the logic of science to generate evidence. This issue will be explored later in this report; see section Public Familiarity with Science and Technology Research Processes.

In the decades spanning the late 20th and early 21st centuries, Americans’ confidence in scientists has been high relative to their confidence in other professionals (Krause et al. 2019). A high level of Americans’ confidence in scientists to act in the best interests of society is evident in GSS data spanning from 1985 through 2018. Over that period, several surveys, including the GSS, asked respondents the extent to which they agreed that scientists are dedicated people who work for the good of humanity, help to solve challenging problems, and work to make life better for the average person (Figure PPS-4). A consistently high percentage of Americans agreed with those statements in every survey, although there has been some fluctuation. For example, the percentage of Americans who believe scientists work to make life better for the average person ranged from 80% in 1985 to 89% in 2018. The 2021 GSS cross-section study fielded a similar question from December 2020 to May 2021 asking about the extent to which participants have confidence in the “scientific community,” and results also suggested widespread confidence.^{​Estimates from the 2021 GSS Cross-section Study are not directly comparable to previous GSS estimates of related concepts due to differences in data collection methodology implemented during the COVID-19 pandemic as well as differences in question phrasing.} According to 2021 GSS results, half of Americans had a “great deal of confidence” in the “scientific community,” and another 43% expressed “only some” confidence, whereas a minority of adults—approximately 7%—expressed “hardly any confidence at all” (Davern et al. 2021). The tendency of the majority of Americans to express confidence in scientists and scientific institutions is notable, given that some recent headlines have implied a decline in Americans’ levels of trust or even their widespread mistrust in scientists without accompanying evidence (Fearnow 2021; Piccone 2020).

Following the onset of the COVID-19 pandemic, in April and May 2020 Americans’ confidence in scientists to act in the best interests of the public also appeared to have increased somewhat beyond historical trends (Funk, Kennedy, and Johnson 2020), suggesting that the value of some types of scientific research became more apparent to Americans in the first year of the pandemic compared to the time before the pandemic. In April and May 2020, the Pew Research Center conducted a survey with American adults and repeated questions it had asked in 2016 and 2019 about their confidence in medical scientists and scientists in general to act in the best interests of the public (Funk, Kennedy, and Johnson 2020). In 2016, 24% reported a “great deal” of confidence that medical scientists would act in the public interest. But in spring 2020, this percentage increased to 43% of respondents reporting that level of confidence in medical scientists. For scientists generally, 21% of respondents expressed a great deal of confidence in 2016, compared to 39% in spring 2020.

Globally, 2018 and 2020 Wellcome Global Monitor studies in 113 countries also have demonstrated an increase in the percentage of those who trust scientists “a lot.” The 2018 and 2020 studies differed in the number of countries included and by interview modality; the 2020 study, conducted primarily between September and December 2020, occurred in fewer countries than the 2018 study and included telephone interviewing rather than face-to-face interviews. Nonetheless, each of the two studies included participants from a common set of 113 countries. Data from those countries point to an increased level of trust in scientists: 43% of people answered “a lot” when asked about how much they trust in “scientists in this country” in late 2020 versus 34% who answered “a lot” in 2018 (Gallup 2021).

Although confidence in scientists has remained high for decades, Americans are not uniform in their expressed confidence, suggesting some variation in trust in scientists. According to November 2020 data from Pew Research Center’s American Trends Panel (ATP),^{​The Pew Research Center provided restricted-use data from November 2020 for this analysis that are presented here and in other sections of this report with the center’s permission. The Pew Research Center’s ATP is a nationally representative survey panel composed of more than 10,000 randomly selected adults in the United States. Respondents are frequently surveyed about a variety of topics over time, allowing researchers to build a more detailed understanding of public opinion than is possible with a single survey. New panelists are added each year to replace panelists who have dropped out or been retired from the panel. One concern with the ATP is that panelists may become conditioned to the questions, spurring them to change their behavior in response to a survey or become more skilled at answering the questions. Still, many ATP surveys interview only a subset of the panelists, which prevents survey fatigue and helps ensure that the surveys are nationally representative with less weighting. For more information about the ATP, see https://www.pewresearch.org/our-methods/u-s-surveys/the-american-trends-panel/.} 84% of U.S. adults expressed “a fair amount” or “a great deal” of confidence in scientists to act in the best interests of the public (Table PPS-1). Confidence in scientists differed by education and income. For example, 54% of U.S. adults with a postgraduate degree expressed a great deal of confidence in scientists, whereas 30% of U.S. adults with a high school degree or less did. Half of U.S. adults in the highest of three family income tiers in the survey expressed a great deal of confidence, while 32% of U.S. adults in the lowest family income tier expressed that same level of confidence. What accounts for the differences in confidence in scientists between adults with different education and income levels is an important empirical question. The 2020 American Trends Panel data demonstrate limited differences in confidence in scientists as a function of respondent race and no differences as a function of respondent sex (Table PPS-1). Later, this report will assess one factor that predicts confidence—namely, the extent to which people understand how scientific inquiry ideally occurs. (See section Public Familiarity with Science and Technology Research Processes.)

Social science researchers have limited evidence of the extent to which Americans draw fine distinctions between the categories of scientists and engineers. Some experimental evidence comparing survey respondents’ answers to questions about scientists and engineers suggests that Americans tend not to differentiate between scientists and engineers in terms of their value to society (see NSB Indicators 2020 report "Science and Technology: Public Attitudes, Knowledge, and Interest"). Notably, for example, 2012 GSS survey respondents who were asked general questions about scientists (as to whether a respondent would be happy if their child became a scientist or whether scientists work for the good of humanity) responded similarly to those who were asked questions about engineers (regarding being happy if their child became an engineer and whether engineers work for the good of humanity).

In 2013, Pew Research Center surveyed Americans regarding their perceptions of the contributions of various occupational groups to society’s well-being. That work suggested a majority of American adults hold medical doctors, scientists, and engineers in roughly equal regard (Pew Research Center 2013). Among American adults, 63% believed engineers contribute a lot to societal well-being, 65% believed scientists contribute a lot to societal well-being, and 66% believed medical doctors do so. Moreover, only a small percentage of adults believed medical doctors (8%), scientists (8%), or engineers (7%) contribute nothing or not very much to societal well-being. Those positive perceptions of engineering generally align with earlier survey research commissioned for the National Academy of Engineering (NAE 2008).

In addition, older evidence on perceptions of engineers and engineering (and on comparative perceptions of engineers and scientists) in the United States has come primarily from student populations at the elementary and middle school levels. Some evidence suggests that elementary school students tend to perceive engineers as men engaged in the activities of “building” or “making,” while other evidence suggests middle school students tend to have no clear perception of engineers or their work activities (Capobianco et al. 2011; Reeping and Reid 2014; Fralick et al. 2009). These same middle school students had more substantial mental models of science, reporting that they mostly work indoors conducting experiments (Fralick et al. 2009).

Although Americans have tended to broadly support S&T, they sometimes express concerns about specific issues that arise with the publication of new research and the introduction of new technologies. As described in this section, recent peer-reviewed literature highlights evidence on public perceptions of a variety of topics, including research related to COVID-19; understanding of AI, robotics, and automation technology; perceptions of climate change and climate change research; and beliefs about STEM education. These issues have been prominent in recent public discussions or may be relevant to evaluating Americans’ trust in scientific institutions, understanding of scientific processes, or exposure to scientific activities.

Data from 3M’s State of Science Index Survey suggest some uncertainty among Americans over the definition of artificial intelligence (AI).^{​3M shared data for this analysis that are presented here and used with 3M’s permission. The 3M State of Science Index Survey is an independent, nationally representative research study commissioned by 3M to track global attitudes toward science. It has been conducted annually since 2018, but due to the coronavirus pandemic, two waves of data were released in 2020 after an additional survey was fielded during summer 2020. The 2020 Pre-Pandemic Survey was conducted in 14 countries, while the 2020 Pandemic Pulse Survey was conducted in 11 countries; the United States was included in both surveys. For more information about the survey methodology, see https://www.3m.com/3M/en_US/state-of-science-index-survey/2020-summary/.} When Americans were asked how much they know about AI, over a fifth reported knowing “nothing” about AI (22%), a minority (17%) reported that they know “a lot,” and 62% reported knowing “some” (3M 2020). Recent evidence also suggests that public understanding of AI, robotics, and automation technology may change in coming years. Evidence suggests, for example, that popular conceptions of automation technology and robotics change as more people have opportunities for direct experience with various automated applications. Tenhundfeld and colleagues (2019, 2020) found that participants’ willingness to rely on an automatic parking feature in an electric car varied as a function of how much experience they had with the technology. Over time, as they gained more experience with the feature, participants’ tendency to allow automation to control the car increased (measured as the lack of behavioral intervention to stop the automated system from operating) (Tenhundfeld et al. 2020).

In a different example, Sanders and colleagues (2017, 2019) investigated human perceptions of robots in terms of perceived trust and willingness to allow a robot to perform various tasks. One of these studies (Sanders et al. 2017) found that prior interaction with robots was positively associated with trust in them. Another study (Sanders et al. 2019) found participants were more likely to choose a robot for a task that was relatively dangerous and was likely to result in death. Respondents were also more likely to choose humans to do mundane warehouse tasks, noting job and income considerations for human workers and the implications of robots replacing human workers.

Popular imagination regarding AI beyond automated mechanical tasks and robotics is potentially fertile ground for future investigation, but currently much about human perceptions of AI remains undocumented. Available evidence suggests that AI has become an increasingly relevant topic in public discussions about science. Fast and Horvitz (2017) studied 30 years of New York Times references to AI—between 1986 and 2016—and found that mentions of AI, including references both to optimism and concerns about ethics and loss of control, began increasing in 2009.

Recent survey research shows some ambivalence in public opinion about AI R&D. On the positive side, analysis by Zhang and Dafoe (2019) of a public opinion poll of 2,000 adults (age 18 and older) found that a substantial number (nearly half) of Americans support further development of AI, defined in the survey as “computer systems that perform tasks or make decisions that usually require human intelligence” (Zhang and Dafoe 2019:5). This study is consistent with results from a Pew Research Center report (Johnson and Tyson 2020) in which roughly half of U.S. respondents said that the development of AI “has mostly been a good thing for society.”

Recent survey research on AI, however, suggests a substantial proportion of the United States is currently uncertain about AI R&D. More than a third of participants in the Zhang and Dafoe (2019) analysis neither supported nor opposed AI development (28%) or were unsure about what they thought of AI development (10%). What support currently exists for research on AI appears to be conditional. The vast majority (82%) of those surveyed by Zhang and Dafoe believed robots or AI should be carefully managed. Zhang and Dafoe also note that educational attainment is positively associated with support for AI development and that trust in organizations to develop and manage AI varied as a function of the type of organization, with relatively more support for universities compared with some other types of organizations. Taken together, current public perception research on AI suggests that many Americans lack awareness about AI or feel uncertain about it, yet they feel some conditional optimism about it as well. The vast majority appear to have some concern about future management of the technology.

The COVID-19 pandemic is relevant to this report in several ways. First, evidence suggests that COVID-19 news coverage may have increased public consideration of, and trust in, science generally during spring 2020, as detailed in the Perceptions of Scientists section. (For more, see Funk, Kennedy, and Johnson 2020.) Second, Americans’ understanding of COVID-19 and their perceptions of COVID-19 research itself fall squarely within the subject matter covered here. Finally, the pandemic affected many public opinion data collection efforts beginning at least as early as 2020, constraining some potential sources of information for this report and others. The accompanying sidebar highlights relevant evidence on U.S. perceptions of COVID-19 research specifically and suggests how the pandemic experience offers an example of public perceptions developing over time. (See sidebar U.S. Public Perceptions of COVID-19 Research.)

There has been an increase in recent decades in the percentage of Americans who have expressed concern about the rise in the Earth’s average temperature over time (see Indicators 2020 report Science and Technology: Public Attitudes, Knowledge, and Interest). In 1994, 35% of Americans who participated in the GSS believed that a “rise in the world’s temperature caused by the greenhouse effect” is “extremely” or “very” dangerous. By 2018, the GSS reported that 58% of Americans surveyed believed this. This pattern of increased concern is consistent with other recent studies (Leiserowitz et al. 2019; Gallup 2019a). Although concern has increased generally, younger Americans appear to be relatively more concerned than their older counterparts. In the 2018 GSS sample, concern varied as a function of education and age; younger respondents and those with relatively more education were more concerned than their counterparts who were older and had relatively fewer years of education (see Indicators 2020 report Science and Technology: Public Attitudes, Knowledge, and Interest).

The extent to which Americans are concerned about climate change, however, is not necessarily related to their perceptions of climate change research. Research on Americans’ beliefs about climate change science suggests both relatively broad support for including climate scientists in government policy deliberation as well as a common perception that climate scientists do not yet understand climate change well. A majority of U.S. adult respondents in a 2016 Pew Research Center study agreed that climate scientists should have a major role in making decisions about policy issues related to global climate change (Pew Research Center 2016). At the same time, 33% agreed that climate scientists understand the occurrence of climate change “very well.” An even smaller percentage (28%) agreed that climate scientists understand “very well” the causes of climate change, and only 19% agreed that “climate scientists understand very well the best ways to address climate change.” An April 2021 Pew Research Center survey found similar results on perceptions of scientists’ understanding of climate change causes and remedies (Funk 2021). These results suggest that Americans generally acknowledge the relevance of climate science research to societal decision-making and that they also focus on what they believe is not yet empirically known to climate researchers.

Recent research on public understanding of climate change offers insight on factors that can shape and influence perceptions. Exposure to news stories can directly affect public opinion about climate change—both in terms of the general importance of the issue as well as issue-framing effects (Newman, Nisbet, and Nisbet 2018). News references to the credibility of science and scientific institutions can indirectly affect beliefs about the credibility of climate change research (Hmielowski et al. 2014). The extent to which a person has thought about climate change previously also appears to limit possibilities for media content to affect beliefs about climate change (Wonneberger, Meijers, and Schuck 2020). Research indicates that perceptions of climate change and climate change research are functions of both existing beliefs and patterns in the information environment—suggesting potential for change but also relative stability as consistent news coverage and online information accumulate over time.

Public perception of STEM education in K–12 U.S. public schools comprises a mix of fond recollection for STEM classes, concern about present investment in K–12 schools, and widespread judgment that STEM education offered to elementary, middle, and high school students in the United States is worse than that offered in at least some other countries. A Pew Research Center survey (Funk and Parker 2018) found that 75% of adult respondents reported that they liked science courses during their time as K–12 students. When asked to choose whether they liked those courses because of the subject matter itself or because of the way the subject matter was taught, 68% of those who liked their science courses said the subject matter was the main reason they enjoyed those classes. Despite their fondness for their own STEM experiences, only a quarter of respondents considered K–12 STEM education in the United States to be at least above average “compared with other developed nations” (Funk and Parker 2018:86). They held similar perceptions of both undergraduate and graduate STEM education in the United States, as fewer than half of respondents thought either undergraduate or graduate STEM education in the United States outranked what is available in other countries. Future inquiry could explore the basis for such perceptions.

Respondents in the 2018 Pew Research Center report (Funk and Parker 2018) saw opportunities for improvement in U.S. STEM education. Approximately 3 in 10 respondents believed that K–12 public schools should emphasize at least one STEM subject more than was currently the case in schools at the time of the survey. (For the Pew Research Center survey, STEM subjects included mathematics and statistics, science and engineering [S&E], and computers and computer science.) In terms of future opportunities for improvement, a majority of respondents cited lack of parental involvement in public STEM education as a concern, and approximately half believed that teaching methods should further emphasize critical thinking and applying STEM subject matter to everyday life. Together, these beliefs suggest that most Americans see value in STEM education but also believe that STEM instruction in U.S. public schools could be improved.