5

Science Literacy for Individuals

The nature of an individual’s science literacy is tied to the social organizations in which individuals function. System-level factors, such as the differential distribution of knowledge and resources, affect the circumstances that individuals or communities confront, which demand or promote the application of science literacy. These social structures (described in Chapter 3) include a society’s or a community’s education system, health care system, justice system, and governance structure. Results from aggregated individual-based science-knowledge assessments in the United States show differences by race, ethnicity, age, gender, and educational attainment, which are in no small part based on the differing constraints and opportunities that face these subpopulations (see Chapter 3).

This chapter focuses on the value of science literacy and health literacy for individuals. We assess the current empirical evidence regarding the outcomes of science literacy for individuals, including evidence on the association between science literacy and attitudes towards, perceptions of, and support for science and the relationship between science literacy, health literacy, and behaviors (particularly behaviors related to health).

THE RELATIONSHIP BETWEEN SCIENCE LITERACY AND ATTITUDES

As discussed in Chapter 2, science literacy at the individual level has largely been measured by assessing an individual’s knowledge level using content knowledge assessments and measures of understanding of scientific principles.

The items measured are not assumed to be exhaustive of the concept in question; rather, they form a sample from a larger set of potential items that could have been chosen to represent the unobserved characteristic, in this case, science literacy. Individuals who know the answers to particular textbook-style questions about biology, physics, and chemistry have a high probability of knowing other facts in the same domains. In the same way that the questions that graduate school candidates face in the GRE General Test are not the things that the test takers directly need to know for their graduate work but are indicative of necessary and related abilities and knowledge, content knowledge assessments (particularly the Oxford scale science literacy items) are designed to distinguish generally among people who have relatively more or less science knowledge.

The extent to which these knowledge measures have validity in measuring levels of scientific literacy is discussed in Chapter 2, and many scholars, as well as this committee, recognize that they may capture only a narrow aspect of science literacy. That is, they do not specifically assess the many other features of science literacy (detailed in Chapter 2). However, these knowledge measures are the ones most commonly used in studies exploring the potential relationship between science literacy and attitudes toward, public perceptions of, and support for science, as measured by the public’s evaluation of the social impact of science and technology.¹

This section discusses the major scholarly research that considers the potential link between science knowledge and attitudes toward and public perceptions of science, with special attention to the association between science knowledge and support for scientific funding; see Box 5-1. It explores research examining the relationship between science knowledge and a set of broad attitudes toward science that reflect an individual’s assessment of the scientific research enterprise generally, as well as the relationship between knowledge and a more focused set of attitudes toward specific scientific controversies (such as nuclear power, climate change, stem cell research, and genetically modified foods). In addition, it discusses potential moderators and mediators to this relationship and recent experiments aimed at testing interventions to increase science knowledge and their impact on an individual’s attitudes toward science.

The committee recognizes that this may not be the only important empirical question. There is a large body of literature on the influences of attitude

___________________

¹The wordings of many attitudinal items, which are typically worded positively, may be subject to acquiescent response bias (Bauer et al., 2000). Given these measurement concerns, Bauer and colleagues (2000) proposed an alternative, multi-item measure of public science attitudes that focused on understanding the nature of science (factual knowledge, methodological knowledge, and knowledge of the scientific institution). This measure of science attitudes is primarily concerned with understanding people’s views on the controversy over the nature of modern science. Despite cross-validating their measure with multiple samples, the assessment is not currently widely used in research; it presents an opportunity for future research.

formation, such as value predispositions, media use, and perceptions of risks and benefits, among others; see Box 5-2. Though this research literature is informative, the statement of task guided the committee’s narrow exploration of the relationship between science knowledge and attitudes.

Science Knowledge and Attitudes Toward Science

Among the most heavily cited analyses assessing the direct link between science knowledge and attitudes toward and public perceptions of science is a meta-analysis that analyzed publicly available survey data from 193 surveys conducted across 40 countries between 1989 and 2004 (Allum et al., 2008). The

meta-analysis only examined factual knowledge items because items relating to the scientific method and understandings of science were not found in all of the studies considered. The measured attitude items can be classified into five areas: general science attitudes and specific attitudes toward nuclear power, genetic medicine, genetically modified foods, and environmental science.

Controlling for measures of age, gender, and education that were common to all 193 datasets, the meta-analysis found that there was a small, positive overall relationship between science knowledge and attitudes. Equally important, however, the study found that the size of this relationship varied substantially by whether the measure of attitudes was focused on general science or a specific topic and whether the knowledge measure was a general science measure or one

focused on a specific type of knowledge. Specifically, the correlation between general scientific knowledge and a range of specific science attitudes was generally weaker than the correlation between general scientific knowledge and general scientific attitudes. For example, the data suggested almost no relationship between general science knowledge and attitudes about genetically modified food, a potentially negative relationship between biology-specific knowledge and attitudes about genetically modified food, and a small, but negative relationship between that same general science knowledge measure and attitudes toward environmental science (see also Gaskell et al., 2004; Priest et al., 2003). The results further suggested that the basic relationship between general science knowledge and general attitudes was slightly larger than initially estimated

but still small, with little cross-national variation.² Ultimately, the authors characterized the relationship in the following way (Allum et al., 2008, p. 51):

Other scholars have reinforced these findings and shown that this relationship becomes more complicated when assessing specific science knowledge and attitudes. Individuals may have broadly positive (or negative) attitudes toward science and may hold a set of attitudes toward specific scientific issues or disputes that do not align with their general attitudes toward science. For example, O’Connor and colleagues (1999) found a positive relationship between specific attitudes (willingness to take voluntary actions to address climate change) and specific knowledge (understanding of climate change). The study found that willingness was positively related to knowledge of the causes of climate change, although the relationship was weaker once measures of overall environmental attitudes were included in the analysis. In contrast, Bauer and colleagues (1997) analyzed three separate years of Eurobarometer data and found that, while specific knowledge of biology increased across the three surveys, optimism about both biotechnology and genetic engineering actually decreased during that time. The authors also found that scientific knowledge was only weakly correlated with a host of application areas for either biotechnology or genetic engineering. Other studies have concluded that higher levels of scientific knowledge were correlated with negative perceptions of biotechnology (e.g., Midden et al., 2002), pointing to the inconsistent results across studies trying to assess a direct relationship between knowledge and attitudes.

Priest and colleagues (2003) explored public perceptions toward biotechnology in both the United States and Europe. While the authors found science knowledge and educational levels to correlate differently with several different application areas of biotechnology (from a strong, positive correlation of 0.6 for medical applications to weak correlations of 0.05 in areas of food applications and animal cloning), they also concluded that a “knowledge gap” failed to completely explain the much higher European opposition to biotechnology. In this case, “trust gaps” (i.e., the size of the difference in trust in different stakeholders involved in the technology) emerged as a more reliable predictor of biotechnology attitudes than knowledge levels.

The importance of trust in explaining attitudes, relative to science knowl-

___________________

²The small amounts of variation that were present were explained by the proportion of the population in the countries that went on to attend higher education (Allum et al., 2008).

edge levels, has been stressed elsewhere. Priest (2001), using a path analysis, explored the competing roles of awareness, food safety concerns, genetics knowledge, and trust in key scientific institutions on encouragement for biotechnology, including the genetic engineering of crops, cloning, and engineering bacteria to produce pharmaceuticals, among others. This study revealed a moderate positive relationship between specific knowledge of genetics and encouragement of biotechnology applications. However, the strength of the knowledge-attitude link was much less pronounced than the one found between institutional trust and biotechnology encouragement. Brossard and Nisbet (2007) also found a small but positive relationship between factual knowledge of agricultural biotechnology and support for the technology after controlling for a large number of variables, including sociodemographic variables, media use, levels of trust, and reservations about the effects of science. However, they found that the main determinant of public support for agricultural biotechnology was the level of deference toward scientific authority and not knowledge levels or trust in information providers.

Other scholars have observed negative relationships between various measures of scientific knowledge and public attitudes, particularly for issues characterized by ethical debates (Knight, 2009). For example, Cacciatore and colleagues (2012a) found that, even after controlling for a host of factors, increased knowledge about biofuels was associated with a greater tendency to perceive increased risks relative to benefits from the alternative fuel. Similar patterns have been noted for the issue of nanotechnology, where Lee and colleagues (2005) found that general science knowledge negatively predicted people’s perceptions of benefits of the science relative to risks. However, as discussed below, the researchers stressed that processes were complicated and that knowledge had a weaker effect on attitudes for people who showed strong emotional reactions to the topic. In another study, Kahan and colleagues (2012) explored the effects of science literacy and numeracy on climate change attitudes and found that both scientific knowledge and numeracy were associated with decreased risk perceptions regarding the dangers of climate change.

Mediators and Moderators

As illustrated above, there is increasing evidence that the direct link between science knowledge and attitudes toward scientific issues is weak and is mediated or moderated by other factors.³ Acknowledging that the psychology of attitudes is complex and that cognitive and affective factors have to be taken into account, scholars have explored what factors might shape the connections between knowledge and attitudes. Much of this recent work shows that the

___________________

³A moderator variable is one that influences the strength of a relationship between two other variables, and a mediator variable is one that explains the relationship between the two other variables.

relationship between knowledge and attitudes often weakens or disappears as additional variables are controlled for in the analysis (e.g., Cacciatore et al., 2011; Hart and Nisbet, 2012; Ho et al., 2008, 2010, 2011; Scheufele et al., 2009). In other words, individuals may make judgments on specific applications of science not based on their knowledge levels, but based on such other factors as people’s values (political ideology, religiosity, or deference to science), their level of trust in information providers, or other important variables.

Lee and colleagues (2005) found a moderating effect of negative emotion on the relationship between nanotechnology knowledge and perceptions of the risks relative to benefits of nanotechnology. Specifically, nanotechnology knowledge had a significantly stronger effect on perceptions of risks versus benefits among individuals who reported low levels of negative emotion toward the issue. People without strong negative emotions toward nanotechnology were much less concerned about the risks of nanotechnology as their knowledge levels increased, while those with strong negative emotions were relatively unmoved in their perceptions of risks regardless of their knowledge level. This pattern was also found when general support for nanotechnology was the dependent variable of interest.

Ho and colleagues (2008) noted similar patterns in their work investigating public attitudes toward stem cell research. As with the results for nanotechnology, the authors found that the positive effects of knowledge on stem cell support did not persist once a host of demographic and media use variables were controlled for in the regression model. Consistent with the results noted above, they found that the influence of knowledge on support for embryonic stem cell research was significantly stronger for people low in religiosity in comparison with people high in religiosity. They also found that knowledge had a much stronger relationship with support for stem cell research among liberals than conservatives. Finally, a similar pattern was observed for deference to scientific authority, with knowledge having the strongest effect on support among those reporting high levels of scientific deference.

Kahan and colleagues (2012) investigated two competing hypotheses—what they call the science comprehension hypothesis (that increases in scientific knowledge will lead to greater scientific support) and the cultural cognition theory (that people form their perceptions of risks based on the risk perceptions of those groups with whom they identify)—to explain public attitudes toward climate change. Although they found a negative effect of science literacy and numeracy on climate change concern, they also found that general science knowledge interacted with worldviews in predicting such attitudes. Specifically, knowledge served to polarize the viewpoints of egalitarian communitarians and hierarchical individualists, with increased literacy elevating concern about climate change for the communitarians and decreasing the concern of the indi-

vidualists.⁴Guy and colleagues (2014) found that knowledge specific to the issue of climate change was associated with an increased tendency to accept the evidence for climate change (a pro-science attitude) among those with a hierarchical worldview, but not among those with an individualistic worldview. These approaches are consistent with “motivated reasoning,” the idea that individuals tend to select information that is consistent with their views or beliefs and, alternatively, avoid information that is inconsistent with their views or beliefs (see, e.g., Yeo et al., 2015).

As the work discussed above suggests, the path from scientific knowledge to positive attitudes toward science or support for science is not always clear. Knowledge affects different subgroups in a population differently depending on a host of factors, including levels of religiosity, political predispositions and worldviews, and deference to scientific authority. These patterns seem to vary depending on the specific scientific issue being explored and the culture in which the data is collected. More research is needed to understand this phenomenon.

Effects of Interventions to Increase Knowledge on Attitudes

Experiments related to science knowledge typically seek to assess the effect of providing individuals with new information and comparing the views of those individuals to groups who received either no new information, different information, or some other intervention. However, such experiments can be challenging to interpret. Although effects may emerge, simply learning new facts on their own may not be an adequate representation of the effects of science knowledge or literacy. Specific circumstances may place more or less literacy demands on individuals or allow or disallow for opportunities to apply that literacy. Given this, evidence from interventions that do not take into account context may be limited in its general applicability.

Consistent with similar research on nonscience topics, studies on science-related deliberation clearly show that it is possible to increase basic knowledge through various short-term interventions (Sturgis et al., 2010; Doble, 1995; Einsiedel and Eastlick, 2000; Setälä et al., 2010; Delli Carpini et al., 2004; Bauer and Bonfadelli, 2002). However, the studies also typically show that such learning often has little relationship to attitude change. Gastil and Dillard (1999) found these types of outcomes along with differences by ideology for a range of issues, including energy and health topics. More recently, Kronberger and

___________________

⁴Kahan and colleagues (2012) define hierarchical individualists as individuals who tie authority to conspicuous social rankings and eschew collective “interference” with the decisions of individuals possessing such authority. Egalitarian communitarians are defined as individuals who favor less regimented forms of social organization and greater collective attention to individual needs (Kahan et al., 2012, p. 732).

colleagues (2012) found that a reading task followed by focus-group discussions around synthetic biology resulted in increased “opinion certainty,” especially for groups whose members were highly interested in the topic.

Science information-related experiments not involving discussion similarly found that providing participants with information has a limited effect or an effect that is contingent on predispositions, such as ideology or worldviews. Druckman and Bolsen (2011) and Bolsen and colleagues (2014), for example, showed that “framing” various technologies in certain ways can equal or overwhelm the effect of providing someone with basic information related to emerging technologies. Furthermore, these studies show that people tend to interpret any new information in a way that fits with their worldviews (i.e., they engage in “motivated reasoning”) (see, e.g., Ahern et al., 2016). Kahan and colleagues (2009) have further shown that the process of giving people additional information actually helps people figure out how to use their worldviews to judge a new technology. For example, one study found that views about nanotechnology were initially relatively similar across cultural worldviews prior to receiving risk and benefit information but, once such information was provided, people reacted divergently, in a manner consistent with their different cultural predispositions toward technological risk generally (Kahan et al., 2009, p. 88). Similar effects for issues such as climate change have also been found (Braman et al., 2012).

Brossard and colleagues (2005) used a citizen science project (scientific research conducted, in whole or in part, by amateur or nonprofessional scientists) to explore methods of improving understanding and knowledge of the process of science and bird biology. The study found a significant increase in knowledge of bird biology among those in the treatment condition following the completion of the project. However, corresponding increases in participants’ understanding of the scientific process did not occur, and the project did not affect participants’ attitudes toward science or the environment.

To summarize, the available evidence suggests that providing people with an opportunity to learn about a topic may result in some learning, but it is unlikely to substantially affect attitudes on scientific issues. The reason that individuals likely do not change their attitudes in response to new information is similar to the reason that variables such as ideology or worldview moderate the relationship between existing science knowledge and attitudes about science. Individuals use both existing and new information to reinforce existing attitudes rather than to change their attitudes. It is also possible that people typically use new information to figure out what their cultural group likely believes. Once they have made this determination, the tendency is to conform to their group’s beliefs (Gunther and Liebhart, 2006; Mercier and Sperber, 2011; Hart and Nisbet, 2012; 1, 2015).

Because of these phenomena, determining causality—knowing whether knowledge is driving attitudes, whether attitudes are driving people to become

more informed, or whether the relationship is reciprocal—can be difficult. Importantly, experimental work, which can shed some light on this issue, has limits; it can be difficult to increase knowledge levels without framing the information in such a way that makes it difficult to pinpoint knowledge, and not the introduction of a specific viewpoint, as the cause of a shift in attitude. In addition, the nature of experiments is such that analyses of knowledge acquisition and subsequent long-term effects on attitudes are difficult to make.

THE RELATIONSHIP BETWEEN SCIENCE LITERACY, HEALTH LITERACY, AND BEHAVIORS

In this section we turn from the effect of knowledge on attitudes to the effects of science knowledge and health knowledge on actions and behaviors. As discussed throughout this report, science literacy and health literacy can operate at many different levels of society, from the actions and decisions of individuals to the collective actions and decisions of a community or even a society. The benefits of science literacy and health literacy can also accrue at each level of society: the boundaries between individual, community, and societal benefits are fluid, and actions or behaviors at different levels may contribute to effects at other levels. For example, the decision to vaccinate one’s children is made at the individual level, but it has implications at the level of communities and societies, such as increased life expectancies following the eradication or reduction of infectious diseases.

This section discusses the current evidence on the application of science and health knowledge. We first present frameworks for understanding the relationship between knowledge and action and then analyze the evidence on the relationship between science knowledge, health knowledge, and health-related behaviors. Most of the evidence as to the application of science literacy and health literacy focuses on health-related behavior and does not include a wider set of behavioral outcomes. In reviewing the existing evidence, the committee recognizes that there may be individuals (or communities or societies) who are deeply knowledgeable and engaged, yet nevertheless do not act or take actions that may be at variance with the consensus view of scientists on scientific issues (e.g., individuals objecting to vaccinations or individuals who do not pursue preventative care because they lack access to health care).

Framework of Science Literacy and Health Literacy and Action

Identifying the effects of science literacy or health literacy can be challenging. Often, their effects on any sort of action, decision, or behavior are imagined to be linear, unidirectional, and deterministic: science literacy or health literacy causes desirable outcome X. However, science literacy or health literacy alone is rarely entirely necessary or wholly sufficient for producing a particular desir-

able outcome. Although a connection between knowledge and actions exists, other factors influence choices—factors ranging from cultural norms and self-efficacy to an individual’s ability to access services. Therefore, science literacy and health literacy should be seen as only a probabilistic and partial influence on actions, decisions, and behaviors.⁵

In addition, claims that construe science literacy and health literacy as necessary for particular actions fail to acknowledge the plurality of human motives and life circumstances that can lead to the same outcome. For example, if a person makes healthy dietary choices, it does not necessarily mean that she or he did so because of new nutritional knowledge. Personal habits, social norms, and cultural affiliation can all play a role in shaping behaviors, as can a wide range of beliefs that may or may not reflect or derive from scientific knowledge. The assumption that only scientific knowledge and understanding underlies a particular conclusion, action, or behavior is contradicted by the evidence.⁶ Social factors (such as norms, expectations, and regulations) can also shape the resources available to individuals, thus limiting or constraining behavior. For example, in health care systems, the complexity of medical texts, the communication skills of those providing information, and the attributes of institutions that support or impede patients and caregivers and health professionals all shape the behaviors of actors engaged in the system (Pleasant et al., 2016). Furthermore, living in areas with a shortage of health professionals, lacking of health insurance, and facing problems accessing health services may influence health-related behaviors (see e.g., Gore et al., 1999).

Scholars in the social sciences have long worked to conceptualize this problem of human behavior. In general, social scientists argue that behavior may be determined by a range of factors that include: knowledge and skills, perceived risk, attitudes and beliefs, perceived consequences, self-efficacy, social norms, intentions, and demographics. Other social-psychological determinants (e.g., self-concept and self-esteem, occupational stress, religiosity, recreation and leisure, social support networks, and media habits) may also influence a person’s actions (see e.g., Ajzen, 1985; Glanz and Rimer, 1995; Prochaska et al., 1992; Bandura, 1971).

The field of health literacy has developed useful frameworks for under-

___________________

⁵It is always possible to formulate post hoc arguments about why science literacy and health literacy could not operate in a particular circumstance—even if they were present. It is not the committee’s intent to make claims about science literacy and health literacy unfalsifiable. Rather, the committee believes science literacy and health literacy require an assessment of their value on the basis of a range and type of circumstances in which they do appear to shape thoughts, actions, and behaviors in a positive way.

⁶See, for example, Kempton et al. (1995), who found the average knowledge about environmental issues to be low. However, the lack of knowledge was equally strong among environmentalists and nonenvironmentalists, implying that environmental knowledge per se is not a prerequisite for pro- or anti-environmental behavior.

standing individuals’ health behaviors and the relationship between a person’s health knowledge and their health-related behaviors. In general, these frameworks incorporate internal and external factors that contribute to behavior. For example, variables such as knowledge of opportunities for screening and treatment or an individual’s risk perception may spur intentions about a health action and ultimately result in a decision about whether to attempt to perform that health action. Concepts, such as self-efficacy, and practical barriers, like the financial costs associated with the action, influence the translation of intentions into action. Social factors may also influence motivational and volitional processes. In addition to these social cognitive processes, demographic determinants (e.g., gender, employment status, and personal wealth) can influence the likelihood of a health action (see e.g., Paasche-Orlow and Wolf, 2007; von Wagner et al., 2009).

These behavioral theories and frameworks provide a useful guide for thinking about the relationship between science literacy, health literacy, and behaviors or actions. The constraints and, more broadly, the social structures in which individuals live limit their ability to take action on the basis of science literacy or health literacy. Science literacy or health literacy can be a powerful tool, but it can more easily be used by some individuals and in certain circumstances than others. That is, one’s science literacy or health literacy may be adequate in certain situations and may be deficient in others. Understanding this relationship allows for a more realistic understanding of the benefits of science literacy and health literacy, even when, at times, that benefit may be limited by other factors.

Science Literacy, Health Literacy, and Behaviors

As discussed above, the connection between an individual’s science and health knowledge (as assessed by an individual’s knowledge of specific facts) and action is limited. There is particularly strong evidence that knowledge has a mediated relationship to action (e.g., Bord et al., 2000; Kollmuss and Agyeman, 2002).

Science Literacy and Behaviors

Though most of the evidence on use of health literacy examines effects on behaviors related to health, literature in the field of environmental science explores the relationship between knowledge and a nonhealth science-related behavior. In the environmental domain, the available evidence suggests that there is a weak correlation between science knowledge and behavior. For example, Hines and colleagues (1986) published an important meta-analysis that examined 128 pro-environmental behavior research studies. The analysis found that the correlations between knowledge and attitudes, attitudes and intentions, and intentions and actual behavior were weak.

Finger (1994) showed that environmental information and knowledge predicted little of the variability in most forms of environmental behavior. However, the study did find that information and knowledge acquisition appeared to foster protest actions. Hsu (2004) assessed the effects of an environmental education course on students’ environmental behavior. The students’ perceived knowledge of environmental issues and intentions to act did increase after completing the course, but this study examined only perceived knowledge and intentions to act, not completed behaviors.

There is evidence to suggest that understanding the institutional nature of knowledge production (e.g., peer review, conflicts of interest, research funding) can be useful across a wide range of circumstances (see, e.g., Ryder, 2001). This aspect of science literacy may enable individuals to engage critically with science by helping them to frame meaningful questions, even in contexts in which relevant science concepts are highly technical. Critical understanding of the nature of scientific evidence, a grasp of the way that wider issues influence debates about science, and the value of formal scientific evidence may enable individuals to productively engage with scientific controversies (Tytler et al., 2001). In doing so, individuals may create greater accountability—such as demanding caution in environmental assessments or calling for bioethical guidelines in relation to new genetic technologies—in the production and use of scientific knowledge (Jasanoff, 2003; Irwin and Wynne, 1996).

There are numerous factors beyond scientific knowledge that influence behaviors. Kollmuss and Agyeman (2002) found an individual’s ability to undertake environmentally conscious actions was influenced by demographic factors, external factors (e.g., institutional, economic, social, and cultural factors) and internal factors (e.g., motivation, values, attitudes, emotion, locus of control, responsibilities, and priorities). Differences in willingness to act are also mediated by trust and acceptance (Rabinovich et al., 2012). When individuals lack trust in existing political leaders or institutions to respond effectively to climate change, for example, their personal motivation to engage with the issue is dampened, since feelings of reciprocal sacrifice are important motivators of participation (Feldman and Hart, 2016).

Health Literacy and Behaviors Related to Health

The relationship between health literacy, behaviors related to health, and ultimately health outcomes is complex. Much of the research on the relationship between health literacy and behaviors related to health at the individual level focuses on compliance behaviors and the use of health care services. In the field of health literacy, this research tends to use cross-sectional surveys to assess health literacy of a sample of respondents and measure various outcomes. It is common for surveys to use “fundamental” literacy as a proxy for health literacy or instruments such as the Test of Functional Health Literacy in Adults,

the Rapid Estimate of Adult Literacy in Medicine, or the Wide Range Achievement Test to assess an individual’s health literacy (see Chapter 2). The outcomes that are assessed range from compliance behaviors, such as adherence to oral contraceptive pills or rate of breastfeeding, to self-efficacy and health outcomes, such as healthy weight, diabetes control, or management of asthma (Ross et al., 2001; DeWalt et al., 2007; Hawthorne, 1997; Davis et al., 2006; Campbell et al., 2004; DeWalt and Hink, 2009).

It is well accepted that health literacy is an important factor in patients’ abilities to obtain and use health-related information to make decisions about health care as it relates to health services utilization, self-care behaviors, and risks for disease-related morbidities and mortalities, especially when tied to racial and ethnic disparities (DeWalt et al., 2004; Berkman et al., 2011a, 2011b; Osborn et al., 2007, 2011; Paasche-Orlow et al., 2010). For example, in a systematic review the Agency for Healthcare Research and Quality (2011) found evidence that lower levels of health literacy were associated with increased hospitalization, greater emergency care use, lower use of mammography, and lower receipt of the influenza vaccine (see also Berkman et al., 2011a, 2011b; Sheridan et al., 2011).

Although much of this research demonstrates a relationship between knowledge and action, it is typically weak (e.g., Al Sayah et al., 2012). Studies often find that other factors, such as reasoning skills, intelligence, trust, and values, contribute to an individual’s actions (see, e.g., Arnold et al., 2001; Keller et al., 2008). Other mediating factors may be external to the individual, such as access to health care, patient-provider interactions, cost, and management of health and illness (von Wagner et al., 2009). For example, in a meta-analysis reviewing the relationship between health literacy and medication adherence, Zhang and colleagues (2014) found higher health literacy levels were associated with better medication adherence across 6 diseases and 35 samples. However, the effect size of the relationship was weak when compared with other predictors of medication adherence, such as type of disease, medication beliefs, and cost restraints (Zhang et al., 2014).

Some studies have examined interventions intended to improve health behaviors and use of health services. For example, Yin and colleagues (2008) measured parent-reported medication dosing and observed parents preparing a medication dose. Parents in the intervention group received pictogram-based medication instruction sheets with teach-back counseling; parents in the control group received standard care. The parents in the intervention group were more likely to use the correct dose and had greater self-reported adherence to the prescribed medication regimen. In another study, Robinson and colleagues (2008) studied children with asthma who were enrolled in a reading skills and asthma education program. Of the children enrolled, 63 percent visited an emergency room before the intervention while only 33 percent made an emergency visit during the intervention, and the children whose reading improved

the most were least likely to have repeated emergency visits. In contrast, a recent study of parents who received one of four interventions—information explaining the lack of evidence linking vaccinations to autism; textual information about the dangers of diseases prevented by vaccination; images of children who have diseases prevented by the vaccine; and a dramatic narrative about an infant who almost died of a disease prevented by the vaccine—found that none of the interventions increased parental intent to vaccinate a future child. In addition, the results indicated that showing images of sick children increased expressed belief in a vaccine-autism link and the dramatic narrative increased self-reported belief in serious vaccine side effects (Nyhan et al., 2014).

The field of health literacy has acknowledged that many factors—at both the individual level and the system level—may affect an individual’s application of health literacy; see Box 5-3 for a discussion of the relationship between health literacy and health disparities. System factors shape the resources available to and the behaviors of visitors and users, as well as the behaviors and expectations of the broad array of individuals engaged in providing health information, care, and services (Pleasant et al., 2016, p. 3). Studies of interventions demonstrate that, at least in the short term, an individual’s health literacy may be enhanced by an outside intervention, typically a change in the health system. These interventions contribute to literacy and may spur particular behaviors if they can overcome adverse individual and external influences and potential context-based hurdles.

The Effects of Science Literacy on Behaviors Related to Health

As described above, the relationship between health literacy, behaviors related to health, and ultimately health outcomes is complex. Although the effect of health literacy on health outcomes is widely accepted, it is unlikely that health literacy, any more than science literacy, has direct effects on most health outcomes, as this relationship is mediated by a range of factors.

Beyond this literature, there has been little scholarly work analyzing the relationship between science literacy and health-related behaviors. However, studies do show that science knowledge can lead to increased perceived self-efficacy, and it is well established that perceived self-efficacy and response efficacy are important in explaining preventive health care behaviors (e.g., Jayanti and Burns, 1998; Bandura, 2010). Thus, in this way, science literacy may indirectly affect preventive health behaviors. Despite this finding, the available evidence does not demonstrate that science literacy is directly related to specific health behaviors. If enhanced science literacy can be empirically linked to enhanced health literacy, it could indirectly be associated with some behaviors related to health. There is a need for further research to determine if such a relationship exists.

One such promising area is the potential relationship between science

literacy and information-seeking and interpretation of online health-related information. Individuals now routinely seek health-related information online. When doing so, individuals need to possess the ability to determine which information is trustworthy and process the appropriate information in an efficient manner. There is some evidence to suggest that science literate individuals may be more equipped to efficiently look for and process health-related information (Ellis et al., 2012).

CONCLUSIONS

The existing empirical evidence at the individual level on the value of science literacy is drawn largely from two separate research fields: science literacy and health literacy. Studies on the effect of health literacy have largely examined the relationship between knowledge (measured as health-related knowledge or foundational literacy) and behaviors related to health. In contrast, most of the literature on science literacy assesses the relationship between science knowledge and attitudes toward, perceptions of, and support for science.

CONCLUSION 10 Research examining the application of science literacy and health literacy has focused on different things: studies on the impact of health literacy have looked for impact on health-related behaviors and actions (e.g., compliance with medical advice, shared decision making, etc.), whereas studies on the impact of science literacy have mostly examined its relationship to individual attitudes toward science and support for scientific research.

Findings from regression-based analyses—in which the effects of knowledge on attitudes were often weakened or eliminated completely when demographic, value predisposition, media use, and trust variables were included in the analysis—demonstrate that context matters when looking at the relationship between knowledge (as assessed by currently used measures of scientific content knowledge) and perceptions of and support for science. Though there appears to be a small, positive relationship between general science knowledge and general attitudes toward science, the relationship is weak and is moderated or mediated by other factors. Therefore, given the state of current evidence and measures, increasing science literacy should not be seen as the foremost means for improving support for science.

CONCLUSION 11 Available research does not support the claim that increasing science literacy will lead to appreciably greater support for science in general.

Scholars have shown that the relationship between knowledge and attitudes becomes complicated when assessing science knowledge and attitudes toward specific science issues. Knowledge affects diverse subgroups in the population differently depending on a host of factors, including levels of religiosity, political predispositions and worldviews, and deference to scientific authority. These patterns seem to vary depending on the specific scientific issue being explored and the culture in which the data are collected. In fact, there is often an interaction between knowledge and worldviews such that enhanced knowledge has been associated, in cases of controversial issues, with increased polarization, affecting attitudes toward those specific science issues.

CONCLUSION 12 Measures of science literacy in adult populations have focused on a very limited set of content and procedural knowledge questions that have been asked within the constraints of large population surveys. Though available measures are limited in scope, evidence suggests they are reasonable indicators of one aspect of science literacy, science knowledge. Studies using these measures observe a small, positive relationship between science literacy and attitudes toward and support for science in general.

CONCLUSION 12a An individual’s general attitude toward science does not always predict that same individual’s attitude toward a specific science topic, such as genetic engineering or vaccines.

CONCLUSION 12b Some specific science issues evoke reactions based on worldviews (e.g., ideology, religion, deference to scientific authority) rather than on knowledge of the science alone.

The research on the relationship between science literacy, health literacy, and behaviors related to health is limited, but the examples that exist highlight the weak correlation between science literacy and health literacy and behaviors. Like the relationship between science knowledge and attitudes toward science, the causal pathway between science literacy and health literacy and behaviors is complex and mediated or moderated by personal and external factors.

CONCLUSION 13 The commonly used measures of science and health literacy, along with other measures of scientific knowledge, are only weakly correlated with action and behavior across a variety of contexts.

The weak relationships among science literacy, health literacy, attitudes, and behaviors suggest that efforts to simply promote knowledge and understanding to change behavior or attitudes may have limited results. Efforts that focus on increasing knowledge also need to include removing impediments to actions and lowering the literacy demands of particular situations.

This page intentionally left blank.