Ethical Considerations of Using Narrative to Communicate Science

What Is the Role of Communication Within Science Policy?

According to the public understanding of science (PUS) model, controversies about science are rooted in ignorance caused by a deficit of science literacy, and the role of communication is to rectify this deficit by educating the public and reducing the controversy (Miller & Kimmel, 2001; Miller, Pardo, & Niwa, 1997). Implicit in this model is that controversy is not desirable and is something that can be remedied by moving away from irrational beliefs toward factual knowledge. This model treats the public as passive vessels needing more knowledge that scientists provide and communicators translate.

Contrasting this model is the public engagement in science and technology (PEST) model, in which controversies about science represent a necessary function of the democratic process, and the role of communication is to facilitate discussion about the benefits and risks of policy informed by societal values and technical information (Dickson, 2001; Walker, Simmons, Irwin, & Wynne, 1999). In this model, communication serves a two-way function, with the public actively engaging in deliberations about the benefits and risks of controversial science-related issues. Science communication scholars have shifted support from the PUS model toward the PEST model over the past decade, noting the necessary consideration of personal values and autonomy for appropriate scientific policy making (Besley, Kramer, Yao, & Toumey, 2008; Einsiedel, 2008; Irwin & Michael, 2003; Powell & Kleinman, 2008).

What Is the Role of Scientists Within Political Policy?

Pielke (2007) discusses the possible roles that scientists can personify when contributing information toward policy decisions. A pure scientist avoids commenting on policy options and instead summarizes knowledge from his or her particular field. A science arbiter answers technical questions about a particular policy but avoids prescribing what should be done. An issue advocate does prescribe a particular policy action, aligning with one side and limiting policy options. Finally, an honest broker expands policy options by commenting on existing policy and offering options that may not have appeared yet in the political agenda. Pielke (2007) claims that all four of these roles can be ethically appropriate.

What is inappropriate, according to Pielke (2007), is when a scientist claims that scientific information compels a certain policy action. In such an instance, the scientist is trying, consciously or not, to use the credibility of science to obscure the larger value debates underlying the controversy. Pielke calls this role a stealth issue advocate and uses climate change as an example where society is still arguing over the legitimacy of the science when the underlying value systems are driving much of the controversy. This critique about assuming that science has the power to drive policy has been echoed by others (Nelson & Vucetich, 2009; Nisbet, 2009) and suggests that while there are multiple roles a scientist can play in policy debate, scientists should not use their credibility or the objectivity of science to suppress the deeper value debates intrinsic to controversial science issues.

How Do Specific Communication Techniques Align With the Roles of Communication and of Scientists Within Policy?

The communication literature discusses many techniques used to attract, hold, inform, and persuade audiences and offers empirical tests as to their effectiveness under varied contexts. Yet there is much less discussion as to the ethical considerations of using these communication techniques.

One field that does reflect on the ethics of using communication techniques is advertising. As summarized by Nebenzahl and Jaffe (1998), a communication technique’s ethicality is a function of the degree to which it causes harm to consumers, such as by (a) manipulating and controlling consumers’ behaviors, (b) infringing on consumers’ level of privacy, or (c) violating consumers’ rights to be informed.

Braybrooke (1967) argued that by repeatedly showing advertisements of a certain brand or product, advertising creates a limited set of choices for consumers and manufactures a new set of desires, potentially violating consumers’ autonomy by preventing them from following their rational desire or will (Braybrooke, 1967; Crisp, 1987). Likewise, exposure to advertising is not always voluntary and may happen when consumers are not conscious of their own exposure (Nebenzahl & Jaffe, 1998). Product placement where a product is seamlessly inserted into a film may therefore result in the invasion of consumers’ privacy (Nebenzahl & Jaffe, 1998). Since “listeners are entitled to know by whom they are being persuaded” (Federal Communications Commission, 1963), product placements and press releases where sponsors are not clearly stated could also be argued to be violating consumers’ rights to be informed.

While these ethical considerations have been discussed in an advertising context, the relation of these concerns to science communication remains relatively unexplored. One of the only discussions of ethical considerations regarding a communication technique in a science policy context has focused on the use of frames (Nisbet, 2009). As defined by Entman (1993), “To frame is to select some aspects of a perceived reality and make them more salient in a communicating text, in such a way as to promote a particular problem definition, causal interpretation, moral evaluation, and/or treatment recommendation” (p. 52). Nisbet (2009) has identified seven typologies of frames often used in science-related policy debates, such as framing the issue based on social progress, scientific or technical uncertainty, or public accountability. The choice of how to frame an issue often has covert effects on an audience’s beliefs and attitudes toward an issue (Cobb, 2005; Davis, 1995; Entman, 1993; Scheufele, 1999).

Nisbet (2009) recommended four guiding principles to determine the ethicality of using framing techniques within science policy: (a) frames should be used “to emphasize common ground and promote dialogue” (p. 70) rather than managing information in a top-down fashion, (b) frames should clearly communicate the underlying values guiding a policy choice rather than suggesting science information compels a decision, (c) frames need to remain accurate and not distort or exaggerate the meaning of the content, and (d) frames should not be used to typecast a particular social group or for political leaders to use deliberately for electoral gains.

While Nisbet’s (2009) principles of ethical framing present a useful guide within our fourth ethical domain, considerations for other communication techniques, including narrative, have yet to enter the discussion. In the final section of this article, we extend the ongoing discussion of this ethical domain to the use of narrative in a science policy context.

What Is the Underlying Purpose of Using Narrative: Comprehension or Persuasion?

This consideration requires a reflection on the appropriate role of communication within science policy. Why engage in science communication? The simple answer is to “inform,” but such an answer often distracts from the underlying assumptions about what informing the public is supposed to achieve. Is the purpose of informing the public to reduce controversy about a science-related policy as assumed under the PUS? Or is the purpose of informing the public to facilitate the controversy necessary to reach a policy as assumed under PEST? Should science communication promote personal autonomy to make choices or create disengaged compliance toward a preferred outcome? This partition manifests itself in the narrative literature between the areas of narrative persuasion and narrative comprehension.

Narratives offer benefits of persuasion through their ability to make normative claims without needing to explicitly state or defend them. These benefits are amplified through factors such as a reduced ability to counterargue when processing narrative information and identification with characters designed to exemplify the central persuasive message. A science communicator whose assumptions underlie the PUS model may decide to design a narrative with the goal of reducing a science-related controversy and generating consensus through the persuasive benefits intrinsic to narratives. Creating a narrative for these ends could involve choosing a series of events that provide a causal explanation of the preferred side of the issue, portraying those events through the eyes of a character who either normatively agrees with the preferred side of the issue or learns to do so throughout the narrative and either concealing or undermining the values underlying the opposing side. Such a narrative has the potential to unobtrusively persuade an audience to be more receptive to a particular science-related policy.

Narratives also offer benefits in comprehension through their increased ease of processing and their ability to make information more relevant and contextual. Much of this benefit lies in the fact that narratives mirror daily experience and are therefore easier to understand and put into a human perspective. A science communicator whose assumptions underlie the PEST model may decide to design a narrative with the goal of facilitating informed debate by increasing comprehension of the science-related factors. Creating a narrative for these ends could involve selecting causal events that explain the factors underling the science issue, portraying the events through a character neutral to the issue at hand or through multiple characters in order to represent multiple sides, and personifying the underlying social values that intersect with the issue. Such a narrative has the potential to engage a wider public in the debate, enhance understanding of the science, and create greater connections with existing knowledge. The area of narrative medicine can provide a relevant example of using narratives for these purposes. Formed partially in response to the critique of medical personnel treating diseases instead of people, narrative medicine claims that narratives are necessary to facilitate a two-way dialogue between patient and care provider and to increase comprehension for both (Harter & Bochner, 2009).

As evidenced by recent attacks on the PUS model as inadequate (Einsiedel, 2008; Irwin & Michael, 2003), it can be assumed that science communication scholars would support using narrative to facilitate discussion toward informed policy. In fact, three of Nisbet’s (2009) principles for the ethical use of framing can be addressed by this comprehension function of narrative, namely, promoting dialogue, clearly stating values, and not using frames for personal gain. These principles raise the question of whether it is ethical to use a communication technique that must remain hidden in order to be effective. The persuasive impact of narrative has been demonstrated to decrease markedly once the persuasive intent becomes known (Moyer-Guse & Nabi, 2010). Yet the improved comprehension derived from narrative is not affected if its intent becomes salient. It therefore seems safe to assume that using narrative for increased comprehension is ethically justifiable.

However, it may be naive to assume that using narrative for comprehension in a science context is appropriate while using narratives for persuasion is inappropriate, as there may be instances where manufacturing compliance represents an ethical decision. Persuasion is often the underlying purpose behind health narratives trying to promote healthy attitudes or behaviors, either as veiled entertainment programs or as carefully selected testimonials within a larger nonnarrative message (Vaughan et al., 2000; Zillmann, 2006). In such cases, the social benefits of increased vaccination or environmentally conscious behaviors may justify reduced personal autonomy.

Such a decision may depend in part on the type of science issue. Pielke (2007) differentiates between two types of science issues: tornado politics and abortion politics. Tornado politics represent issues with high consensus where science can justify the best course of action (because everyone agrees they want to escape the tornado and wants to know how). Abortion politics represent issues with low consensus where science can never resolve a conflict of underlying values. Persuasion toward science policy may be more justifiable under tornado politics, where there is a clearly supported outcome, than under abortion politics, where values become more contested. Likewise, the decision may also depend on the normative expectations held by the public about the communicator. Scientists and journalists are often expected to remain objective, and using narrative for persuasion may take advantage of this expectation. Public information officers, on the other hand, are expected to promote their employer, and persuasive narratives may be more justifiable because they are expected.

In sum, narrative can be used in a science context toward either increased persuasion or increased comprehension. The consideration of which underlying purpose should drive the creation of a science narrative will intersect with a consideration of the appropriate role of communication within science policy.

What Are the Appropriate Levels of Accuracy to Maintain Within the Narrative?

Implicit in the previous discussion is that the science being communicated in a narrative should remain accurate. Nisbet’s (2009) third principle for the ethical use of framing states this explicitly, claiming that scientific accuracy in communication needs to be maintained. However, unlike evidence-based communication, where each fact can be individually assessed for accuracy, narratives rely on contexts for their meanings, which introduces differing levels of accuracy.

It may seem that the most obvious measure of accuracy may be whether the narrative describes a real, nonfictional series of events or an imagined, fictional account. While nonfiction narratives may indeed be more accurate, this level of accuracy may not always be necessary. Choosing to construct a fictional narrative may sometimes be more appropriate because hypothetical situations can be created to explain relationships that have yet to occur or to model an “average” or “extraordinary” experience that might not actually occur in the complex interactions of the real world, yet would be instructive in understanding the science. Likewise, many elements within a fictional narrative can still be accurate—such as scientific procedures, cause-and-effect relationships, or probabilities of risk—and much research has demonstrated that individuals learn facts about the world just as well from fiction as from nonfiction narratives (Dahlstrom, 2010; Marsh et al., 2003). To take a concrete example, consider constructing a narrative to explain the future impact of sea level rise due to climate change. Since the future impact has yet to occur, creating a nonfiction narrative would entail focusing on tangential events, such as how a particular scientist predicted the rise in sea levels or how an individual experiences the current impact of rising sea levels. Alternatively, a fictional narrative could describe how a hypothetical individual might experience the world with sea levels at the predicted height to provide a more accurate perception of the future impact than could be provided by either of the nonfiction examples.

Beyond fiction versus nonfiction, the literature on perceived realism can offer vocabulary to describe additional levels of accuracy intrinsic to narrative. Busselle and Bilandzic (2008) define narrative elements that are accurate relative to the real world as high on external realism. The previous examples of potentially accurate elements within a fictional narrative show how fictional narratives may nonetheless be high on external realism. Likewise, a nonfiction narrative may actually be low in external realism, such as in a historically accurate narrative that bears little resemblance to the world of today. It is likely that external realism is being conceptualized when scientists and scholars speak of the importance of maintaining accuracy within science communication, namely, that it accurately represents the real world. One challenge, then, of creating science narratives lies in deciding what elements of the narrative need to maintain high external realism and what elements can be relaxed toward low external realism for the larger purposes of communication. This selective external accuracy is already present within science itself, such as when a scientist uses the impossible assumption of a frictionless surface or infinite plane to more clearly focus on some other aspect of reality. Some of the narrative elements that may or may not need to be high on external realism include types of characters, characters’ motivations and actions, settings, situations, events, cause-and-effect relationships, procedures, chronologies, and time frames.

Again, to offer an example, consider constructing a narrative to explain the process of converting grain to ethanol by personification of the components into characters. Describing yeast as waiting until the proper temperature to eat its lunch of sugar is a cause-and-effect relationship low on external realism, but the inputs and requirements of the procedure can retain high external realism and accurately describe the process in an understandable form.

In contrast to external realism, Busselle and Bilandzic (2008) define elements of a narrative that are accurate relative to the rules set forth within the narrative world as high on narrative realism. Rules are established early in every narrative about how the narrative world operates, such as how a character relates to the world around him or her, the properties of an object, or the importance of certain objectives. If these rules are later broken, processing is hindered (Albrecht & Obrien, 1993; Kaup & Foss, 2005), persuasion is decreased (Green, 2006; Green & Brock, 2000), and the narrative may even be rejected (Busselle & Bilandzic, 2008; Hall, 2003). The influence of narrative realism becomes obvious in the case of fantasy, where a dragon breathing fire is seen as more accurate than a dragon breathing water, even though both are unquestionably externally inaccurate. Narrative realism represents a second level of accuracy that is less often discussed in science communication, but the implications remain. A narrative that maintains accuracy through appropriate external realism may nonetheless be rejected because it is not accurate as perceived through narrative realism. For instance, constructing a persuasive narrative to promote the acceptance of genetically modified foods by showing a character benefiting from their use may maintain appropriate external realism, yet be perceived as narratively inaccurate if the character does not behave as earlier descriptions would suggest or if the narrative does not complete a story arc and feels incomplete.

Yet another level of accuracy within narrative communication lies in a narrative’s representativeness. Narratives intrinsically lead to the abstracting of specific examples to general trends. Unlike the generalizable content of evidence-based communication, narratives provide a single or small number of exemplars relative to an issue. The representativeness of these exemplars will therefore determine the accuracy with which an audience can generalize to other contexts. Exemplification theory clearly shows the power of narrative to affect perceptions of representativeness. Even when base rate information is present claiming a particular risk is low, the presence of exemplars skew perceptions toward the typicality of the specific exemplar used (Gibson & Zillmann, 1994). This fourth level of accuracy raises the possibility of creating a narrative that maintains accuracy through appropriate external and narrative realism but fails to accurately depict a series of events that is representative of the larger science issue. For instance, a narrative of the experience of an individual who decided not to be vaccinated and developed polio may maintain the desired previous levels of accuracy, but it also represents a worst-case scenario that is not generalizable to what is likely to occur in such a situation and is therefore representationally inaccurate. Of course, selecting a nonrepresentative narrative could be beneficial for a science communicator attempting to use narrative to persuade an audience toward a predetermined end.

While accuracy remains crucial for appropriate science communication, the maxim of maintaining accuracy becomes more complex when considering narrative communication. While maintaining accuracy through narrative realism is likely necessary for any effective narrative, the choice in constructing a fiction versus nonfiction narrative, what elements of the narrative should exhibit high external realism, and whether to select a representative example will intersect with the science issue at hand, the nature of what is to be communicated, and the underlying purpose of using narrative in a science context.

Should Narrative Be Used at All?

Whereas some communication techniques are unavoidable, including the frames underlying Nisbet’s (2009) ethical principals referenced throughout this section, narratives do represent a communicative choice—a choice between using evidence-based arguments through paradigmatic processing or a mediated experience through narrative processing. What are the ethical considerations of using one cognitive pathway over the other? It now becomes necessary to examine specific actors within our broad “science communicator” label. Journalists, public information officers, and other roles that specialize in communication are often expected to use narrative. In fact, the Field Guide for Science Writers specifically cites narrative as one of the effective techniques for covering science (Blum, Knudson, & Henig, 2006).

Scientists, however, hold a much different role in society, and their use of narrative raises its own considerations. Do the perceptions of scientists so closely align with evidence-based communication of the paradigmatic pathway that scientists will be perceived as violating normative expectations if they dabble in the narrative pathway? Could such normative violations cause particular scientists to lose their credibility in the public sphere or, worse, cause the science itself to lose credibility? Bruner (1986) expands on this idea, saying that a scientist caught using the presupposition or subjectification common in narrative would become the “butt of jokes” (p. 28), while a novelist could not maintain suspense or reader involvement without them (Bruner, 1986). While successful scientist popularizers, such as Carl Sagan, are often championed as the cure for poor science literacy (Nisbet & Scheufele, 2009), a 2006 survey of scientists’ views of engaging with nonscientists reveals that scientists who communicate with nonscientists are often viewed by their peers as “fluffy” or “not good enough” for an academic career (The Royal Society, 2006). There is yet no research into what normative expectations the public holds regarding the communication of scientists to a nonscientist audience, so these questions cannot be addressed here.

Nonetheless, other communicators within the debate will likely use narratives, and a choice to avoid using narratives completely due to normative expectations may represent a capitulation to those who do. This dilemma may be becoming more salient as exemplified in a 2008 speech where National Public Radio science correspondent Robert Krulwich criticized scientists for not using their own narratives to counter the “beautiful” narratives of creationism and told young scientists that they should tell stories as a way to fight back (Krulwich, 2008). Regardless of any existing normative expectations aligning scientists to the paradigmatic mode of communication, it may, in fact, be unethical for scientists not to use narrative and surrender the benefits of a communication technique to the nonexpert side of an issue.

The question of whether or not to use narrative at all in a science policy context is essentially a philosophical question. While all of the considerations thus far are relevant when evaluating the question on a situational level, it may be illustrative to ground the considerations in contrasting ethical theories. Utilitarianism is one such theory that focuses on the aspect of harm, stating that ethical actions are those that produce the greatest good for the greatest number of people (Rachels & Rachels, 2007). Following utilitarianism, the use of narrative is ethically justified anytime the effects benefit more people than they harm. In contrast, Kant’s (1981) categorical imperative focuses on autonomy and states that ethical actions are those that respect individuals’ rationality and do not treat others as a means to an end. Following this aspect of Kant’s categorical imperative, the use of narrative is ethically justified anytime it does not attempt to restrict an individual’s autonomy to make decisions.

To illustrate the contrasts between these two ethical theories, let us examine two possible questions about the use of narrative: (a) Are narratives always manipulative? (b) When is manipulation appropriate? The first question would be of utmost importance under Kant’s (1981) categorical imperative, as narratives could only be ethically justified if the answer was no. Under utilitarianism, the first question is of little consequence, yet the second question demands careful consideration of the outcome of the manipulation. Returning to Kant’s categorical imperative, the second question has a simple answer: never.

Rather than promote a particular ethical theory on which to ground a proposed ethical conduct of narrative, it is our purpose to articulate the relevant concerns that arise when considering the use of narrative in a science and policy context. Not only is it suggested that reliance on a single ethical theory is not how individuals actually make decisions (Daniels, 1979), but it also oversimplifies the complexities that exist when communicating science in a policy context among conflicting viewpoints. It is more important to emphasize that specific instances of communication are situational, and the decision whether or not to use a narrative should be one of the considerations.