Sage Journals: Discover world-class research

Abstract

Researchers studying the influence of strategic science messages have overlooked the fact that people are selective in the content they consume. In this experiment, participants were “forced” to view one of three climate appeals (threat, humor, informational), reflecting the typical between-subjects experiment, or they were permitted to choose between one of these appeals (or topical alternatives), reflecting real-world media choice. Compared with forced exposure, choice exposure resulted in slightly less reactance and slightly greater enjoyment, appreciation, and persuasion outcomes—even after accounting for self-selection biases. These findings underscore the value of treating research participants as selective consumers of science content.

Keywords

persuasion emotional appeals selective exposure psychological reactance enjoyment appreciation

Considering the grave nature of climate change and the immense consequences it poses to human civilization and the planet, science communication scholars have explored a range of strategic messaging strategies in hopes of encouraging the public to think and act in more environmentally friendly ways. Although many questions remain as to which messaging approaches are most persuasive, one matter of great theoretical and practical significance is how the type of exposure audiences have to these climate messages shapes their effectiveness. On some occasions, people are captive audiences to climate messages—for example, looking at poster advertisements while riding public transportation or listening to a public service announcement that cannot be skipped on a music streaming service. Yet, media users play an increasingly active role in selecting and avoiding climate content because the modern media ecosystem provides more opportunities than ever for active consumption of media fare (Feldman & Hart, 2018; Knobloch-Westerwick, 2015). In fact, it has been argued that choice is the predominant way in which media users interact with media content (Knobloch-Westerwick, 2015). Strategic climate appeals must therefore compete for audience attention amid a glut of other media options.

Regrettably, much science communication research has glossed over the role of user choice in theory and method. Message exposure is often taken for granted in the communication literature, and this oversight is reflected in experimental designs that “force” or “require” exposure to the content. By consequence, science communication scholarship has largely failed to capture one way that individuals are regularly exposed to climate messaging—as a function of purposeful content selection. This is a troubling oversight because there is theoretical reason to believe that the type of exposure individuals have to strategic climate content (selected or forced) has implications not only for how they psychologically process the information but also its potential to influence their beliefs, attitudes, and intentions. If this is indeed the case, the available evidence, which relies heavily on forced exposure methods, could be mischaracterizing the effects of climate appeals relative to what is likely to occur outside a tightly controlled experiment.

In this investigation, we consider two theoretical processes that may help elucidate how climate message exposure (forced or choice) affects persuasion: psychological reactance and entertainment responses. We conducted an experiment employing a unique design introduced by political scientists (Arceneaux & Johnson, 2013; Gaines & Kuklinski, 2011) and recently adopted by political communication scholars (e.g., Stroud et al., 2018). We randomly placed individuals into either a forced exposure arm, where people were required to consume the content, or a choice exposure arm, where people were free to select the message that interested them, one of which was the target message about climate change. After weighting for factors known to explain a person’s climate positions and information engagement—namely, their political leaning and their belief in climate change (Feldman & Hart, 2018; Goldberg et al., 2020)—such a design permits stronger causal inference about the relative effectiveness of these two modes of exposure, thereby shedding light on whether the effects of persuasive appeals shift when exposure is not required.

As a starting point for this line of inquiry, we focus our attention on emotional appeals about climate change as a messaging context within which to explore questions about the nature of exposure. Science communication scholars have recognized the limits of deficit model approaches grounded in cognitively oriented, “rational” appeals (Simis et al., 2016). Accordingly, they have increasingly turned their attention to the role of emotions and emotional appeals in motivating pro-science outcomes (Nabi et al., 2018; Yeo & McKasy, 2021). This investigation centers on emotional appeals to threat and humor (in comparison with informational appeals). They represent two of the most widely studied negative and positive emotional appeals in the persuasion literature broadly (Eisend, 2009; Tannenbaum et al., 2015) and in environmental domains more specifically (Ettinger et al., 2021; Kaltenbacher & Drews, 2020).

Exposure Matters in Science Communication Research

Underlying much of the evidence on persuasive appeals about climate issues is a central assumption: Audiences are exposed to the content. Exposure and attention to media content are widely considered necessary for messages to influence audience attitudes and behavior (McGuire, 1985), and communication campaign scholars have long argued that insufficient exposure undermines media campaigns’ effectiveness (Hornik, 2002). In a typical between-subjects experiment, science communication researchers randomly assign participants to one version of a pro-environmental message—in essence, “forcing” or “requiring” exposure to the message—and measuring relevant outcomes afterward, such as attitudes and behavioral intentions, to infer message effects. Forced exposure to the content is desirable from an internal validity standpoint. It increases the likelihood that people are captively exposed to randomly assigned (or randomly ordered) variations of some persuasive message, thereby creating strong conditions to infer causal effects of those messages (Shadish et al., 2002).

Yet, in the modern media landscape, message recipients are often active users who choose to consume the content that appeals to them and pass over the content that does not, as outlined within the uses and gratifications tradition of media use (Rubin, 2009). The recent proliferation of media content via traditional and digital platforms and the fragmentation of media channels make it even more challenging for science messengers to cut through the noise and engage target audiences (Fletcher & Nielsen, 2017). Not to mention, media selection and media effects are reciprocal (if not entangled) processes in which beliefs, emotions, and identities shape message exposure and vice versa (Fisher & Hamilton, 2021; Slater, 2015). Our attention is on the influence of climate messages that media users select, given that the forced exposure experimental designs so often seen in the literature eliminate the element of user choice, thereby creating a somewhat artificial experience. Therefore, there is a great need for further theorizing about the ways in which exposure shapes message processing and effects, and experimental designs that account for selective exposure in assessing message effects.

In this study, we explore the role of exposure in the context of emotional appeals, which have occupied a central role in climate communication work and in persuasion scholarship more broadly. Threat appeals, the most widely studied messaging strategy in this literature, emphasize the dangerous consequences that can occur if a recommended response is not enacted. There is no shortage of theoretical perspectives identifying message features and mechanisms explaining their effects, intended and unintended (e.g., Witte, 1992). Despite theoretical disagreements about when and how threat-based appeals persuade, the accumulated evidence on the whole speaks to their general pattern of effectiveness in changing attitudes and intentions (Tannenbaum et al., 2015). As an upbeat contrast to threat appeals, we also consider the role of exposure in the effects of humor appeals, which have increasingly been used to promote attitudes and behaviors in the context of climate change and related environmental issues (Kaltenbacher & Drews, 2020). Advertising research indicates humor is highly successful at directing audience members’ attention to the content and generating positive emotions (Eisend, 2009). It can also promote favorable attitudes toward the message and source (Eisend, 2009) and, according to a recent meta-analysis, on average has weak but positive effects on persuasion outcomes (Walter et al., 2018).

As we will discuss, there is reason to believe choice exposure to these kinds of appeals (threat- or humor-based) may lead to more desirable processing mechanisms (less psychological reactance, greater entertainment) than forcing exposure, which could result in climate appeals exerting more positive effects when exposure is selected. This would mean the accumulated evidence on climate appeals, which comes from forced exposure designs, may very well be underestimating the impact these messages have “in the wild” when media users are able to make choices about which messages to consume and which to skip.

Unpacking Message Processing for Forced Versus Choice Exposure

Personal choice can be thought of as the primary basis by which we exercise control, a fundamental need that fosters positive affect, self-efficacy, and overall well-being (Leotti et al., 2010), as outlined by self-determination theory (Ryan & Deci, 2006). In the following sections, we consider two psychological processes that may be operative in explaining why having a choice in the message one consumes could lead to different persuasion consequences than not having choice: one account that focuses on how audiences may respond negatively when choice is withheld or eliminated (psychological reactance) and a second account that considers how audiences may respond positively when choice is bestowed (entertainment responses).

Psychological Reactance

Psychological reactance theory operates on the assumption that individuals know they have (and are able to have) the autonomy to act and choose as they so wish (Brehm & Brehm, 1981). This is particularly true in Western cultures where personal autonomy is highly prized (Tripathi et al., 2018). When individuals believe their freedom to act autonomously is at stake, they react negatively and become motivated to restore the freedom. At the heart of this process is reactance—an aversive, motivational state comprised of anger and negative cognitions—and the more reactance a person experiences while processing a persuasive appeal, the less likely they are to be swayed (Dillard & Shen, 2005; Ratcliff, 2019).

Much reactance scholarship examines reactance in response to variations in message language (e.g., domineering language, Ratcliff, 2019), but we investigated a slightly different matter: whether individuals exhibit varying levels of reactance depending on how they are exposed to climate message content (forced or choice exposure). A core principle of reactance theory is that individuals respond unfavorably when they feel their ability to act is constrained. “In general, any event that increases the perceived difficulty of having or not having a potential outcome threatens the exercise of a freedom” (Brehm & Brehm, 1981, p. 3). Applied to the current investigation, the removal of message choice by requiring exposure might therefore trigger reactance (or, put a different way, the provision of choice might dampen reactance) because the ability to select is constrained (or, enhanced when choice is provided).

Direct evidence for this position comes from an experiment assessing the effects of forced versus selective exposure to political news content (Stroud et al., 2018). The researchers reasoned that outside of tightly controlled experimental settings, individuals are accustomed to selecting the messages that appeal to them, and so the elimination of that freedom via forced exposure should prompt reactance. In Stroud et al.’s (2018) study, participants were randomly assigned to read pro-attitudinal, counter-attitudinal, or balanced news articles about various political issues. Other participants were given the chance to choose between the three types of news coverage. Psychological reactance was indeed greater among those forced to consume news articles, and this was the case for both cognitive and affective indicators of reactance (counterarguing and anger, respectively)—though it is worth noting, the effect for anger was driven by one of the four issues studied. This finding indicates that one’s defensive response to (not) being able to exert message choice carries over and shapes defensive responses toward the subsequent message.

If reactance is also found to be lower in response to persuasive climate messages when individuals opt into the message, this would have important implications for expanding our understanding of when reactance occurs in response to strategic messages. That is, in addition to reactance being operative as a function of message content (e.g., domineering language), it may also be operative as a function of the mode of exposure. We therefore hypothesize that psychological reactance will be greater in response to forced exposure relative to selective exposure (H1) and that reactance will in turn be negatively associated with persuasion (perceived risk and behavioral intentions, H2).

Entertainment Responses

Research on the psychology of choice generally shows that providing choice enhances intrinsic motivation: Western individuals are more interested in and derive more inherent satisfaction from activities they select than activities forced on them (for a review, see the work by Iyengar & Lepper, 2002). In fact, having a choice in and of itself is desirable, as indicated by neuroimaging evidence that the anticipation of choice activates brain regions implicated in reward processing (Leotti & Delgado, 2011). Moreover, “the opportunity to exercise control through choice is desirable for choices that lead to both obtaining rewards and avoiding punishment” (Leotti & Delgado, 2014, p. 603). In this way, choice is not only pleasurable; it appears to be adaptive (Leotti et al., 2010)—at least in Western cultures where choice is prized (Iyengar & Lepper, 2002). In summary, enabling choice is pleasant for those in Western societies and should result in especially positive evaluations of the chosen activity or experience.

Applying this logic to our project, which was conducted in the United States, this literature implies that in addition to audiences being less likely to demonstrate reactance if they are allowed to exercise message choice, they could also have a more positive experience, rating the message experiences more favorably and perhaps even deriving more meaning from the message experience. Consistent with these notions, research has found that consumers have more positive attitudes toward an advertised product when given the chance to select an advertising message than when not given the chance to select the message they receive (Schlosser & Shavitt, 2009). To conceptually flesh out the kinds of beneficial reactions audiences may have when message choice is permitted, we integrate theorizing from media psychology on entertainment responses. This theorizing provides a fruitful lens that allows us to acknowledge not only the positively valenced evaluations audiences may have toward their chosen message (akin to positive attitudes or liking) but also their contemplative, more complex reactions that run deeper than merely positive evaluations (i.e., appreciation).

Media psychologists have distinguished two types of media entertainment in what some refer to as two-factor models of entertainment (Oliver & Bartsch, 2010): hedonic responses concerned with the maximization of pleasure (enjoyment) and eudaimonic responses concerned with contemplation, meaningfulness, and feeling moved (appreciation). Although scholars have primarily studied these concepts in the domain of entertainment-oriented genres, such as films and video games, there has been growing interest in entertainment reactions to persuasive content about social issues. Feldman and Borum Chattoo (2019) introduced the concept of perceived entertainment value, a mixture of enjoyment and appreciation, that they argued should explain the influence of strategic messaging about social issues. In support of this assertion, they found that compared with news coverage, satirical coverage of the Syrian refugee crisis prompted greater perceived entertainment value, which in turn was linked to more favorable attitudes toward refugees. It therefore stands to reason that being entertained by the messaging may have implications for persuasion and, as suggested by the psychological literature on choice, could be affected by the nature of exposure. Namely, climate appeals could be more entertaining if the recipient is allowed to choose (than if they are not) because the act of choosing is intrinsically desirable, translating to a more enjoyable, even more meaningful, experience.

We are not aware of much research explicitly linking message choice to entertainment responses, but in one study that manipulated whether individuals were able to choose a narrative or non-narrative message in a learning task (Moore & Green, 2021), the ability to choose resulted in greater learning (compared with no choice) and greater transportation, regardless of whether the message was in a narrative or non-narrative format. These findings indicate having a choice in the media content one consumes fosters greater engagement with the message. If we assume, then, that enjoyment and appreciation also reflect a kind of message engagement, Moore and Green’s findings would lead us to expect that choice could similarly facilitate message enjoyment, and possibly even appreciation. These entertainment responses, in turn, could both then contribute to more positive persuasive outcomes. As argued by Feldman and Borum Chattoo (2019), the enjoyment and appreciation induced by entertaining messaging about serious issues should prompt some recognition of the value of the ideas presented in the message, translating to more favorable impressions of the message’s arguments and ultimately resulting in a greater willingness to accept the message’s recommendations. Considering the novelty of these propositions, we ask: How will entertainment responses (enjoyment, appreciation) differ for forced versus choice exposure (RQ1), and how will these entertainment responses then predict persuasive success (RQ2)?¹

Will Exposure Effects Depend on Type of Appeal?

Until now, we have centered our attention on how message reactions could differ between forced and choice exposure (i.e., main effects of exposure type). However, it is also important to consider whether the effects of exposure type are conditional on the nature of the messaging itself—that is, threat, humor, or informational appeals. In other words, might exposure type interact with the type of emotional appeal to shape persuasion success?

If emotional appeals about climate change do exert different effects when exposure is forced relative to when it is selected, this would have great implications for our understanding of their influence. The most recent meta-analyses on threat and humor appeals indicate these appeals have small-to-modest effects on persuasion (Tannenbaum et al., 2015; Walter et al., 2018), yet these meta-analytic estimates are based on findings from experiments where exposure to emotional appeals is required. Moreover, for science communication practitioners, the value of a campaign or message often lies in how many people engage with it, and this aspect of campaign reach is overlooked in most meta-analyses. As it remains to be seen whether the type of exposure to different emotional appeals (forced or choice) will affect beliefs or intentions, we ask: How will persuasion outcomes differ in response to forced versus selective exposure to emotional appeals (RQ3)?²

We tested our predictions in an experiment employing a design that enables us to test the influence of the type of exposure media users have to climate appeals (forced or selected) on key persuasion outcomes. Until now, this design only been used to study political and news content (e.g., Arendt et al., 2022; Dahlgren, 2022; Stroud et al., 2018), so this investigation also marks an important contribution in that it applies this innovative design to science-oriented persuasive messaging. We preregistered our hypotheses, design, and analysis plans through AsPredicted (https://aspredicted.org/FZ7_FGJ). All data, syntax, and materials are available online through the Open Science Framework (https://osf.io/xkphw/?view_only=ad00099c4b0b4371beac6581adbdddf1).

Method

Procedure, Stimuli, and Sample

In August 2023, we recruited U.S. adults through CloudResearch Connect, an online crowdsourcing platform. Participants first provided demographics then were randomly assigned to a forced exposure arm or a choice exposure arm. Those in the forced exposure arm were randomly assigned to view an informational, threatening, or humorous climate change video and then completed a posttest survey. We elected to use video clips from real content identified online to create an ecologically valid media experience. The informational video (titled “The Facts Behind Climate Science”) featured a male narrator presenting scientific facts about how climate change happens. The threatening video (“The ‘Horrifying’ Reality of Climate Change”) featured author David Wallace-Wells describing the severe consequences expected to occur without action on climate change (e.g., heatwaves, wildfires). The humorous video (“Climate: Changed—Is the Sun God Angry at Us?”) was a segment from The Late Show with Stephen Colbert in which Colbert shared similar facts about climate change’s impacts, accompanied by good-natured, punchline-oriented jokes (e.g., “Summers in Washington and Atlanta are hotter than summers in Tampa. Tampa, of course, remains the worst possible version of itself”). We edited the videos to be of similar length (~2.5 minutes), which are available in the Supplemental Appendix.

We sought to create a realistic selection experience for the choice exposure arm, so we designed the selection page of the survey to mimic the YouTube home page, including the header, search bar, and four video previews. The previews appeared in random order and were formatted to closely resemble YouTube design aesthetics (see Figure S1 in the Supplemental Appendix). Survey instructions invited participants to select the video they would be most interested in watching. We reasoned that it would be more realistic to incidentally encounter a single message about the target issue (climate change) rather than multiple messages about the topic side-by-side. Therefore, one of the four previews was the target video about climate change, and the survey randomly assigned participants in the choice arm to have their target video be the informational, threat, or humor video. Thus, a third of participants in the choice arm had the chance to choose the informational appeal, a third had the chance to choose the threat appeal, and a third had the chance to choose the humor appeal. The other three previews were filler videos, randomly drawn from a pool of 26 previews about a range of topics (sports, leisure, politics, social issues).

We added “viewer tags” to each preview to further enhance the strength of our emotional appeal manipulations. Although YouTube does not have such tags at present, the streaming service Netflix does currently include tags for its media content, conveying what the media experience may be like for viewers (Koblin, 2024). Each preview had one tag for the substantive topic of the video (“science” for the target videos) and a second tag for the video’s emotional tone (“informative,” “scary,” and “funny” for the target videos). The video titles and viewer tags used in the choice exposure arm were also described to participants in the forced exposure arm before watching their assigned video to hold those context cues constant. The video durations depicted in the previews were all around 2.5 minutes, and we designed the preview images for the target videos to be as similar as possible (see Figure S2 in Supplemental Appendix). Participants who chose the target video proceeded to watch that video and complete the posttest survey, but those who selected a filler video were directed to the end of the survey and received a nominal payment. We conducted two pilot tests. One pilot verified that the videos had the intended emotional effects (e.g., highest evoked fear for the threat video; N = 144) and received similar ratings of argument strength. A second pilot verified emotion expectations for the previews (e.g., highest anticipated fear for the threat video; N = 184). Please see the Supplemental Appendix for pilot details.

Thus, our primary comparison of interest was between participants who were forced to watch the climate change videos (all participants in the forced arm) and participants in the choice exposure arm who elected to watch a climate change video. This comparison raises two important methodological considerations, one related to statistical power and one related to internal validity. First, because we only retained participants in the choice arm if they selected the target videos (to reduce financial costs), a 50%–50% randomization ratio to the two arms could have resulted in a small number of usable responses in the choice arm for analysis, which would compromise statistical power to detect meaningful effects. We therefore used a 20%–80% randomization scheme to assign participants to the forced and selective exposure arms, respectively (Figure 1).

Figure 1.

Randomization Scheme.

Second, focusing only on participants who choose the target videos about climate change means that any differences between forced and choice exposure could be due to participant imbalances between those assigned to the message and those who chose the message—a threat to internal validity others have termed dilution (Arceneaux & Johnson, 2013; Stroud et al., 2018). For example, people should be more likely to select a climate message if they believe in the risks of climate change. If so, comparing the selective and forced exposure arms might give the illusion that choice exposure leads to greater risk perceptions or behavioral intentions when, in reality, the choice arm simply has a greater portion of climate-concerned participants. One way to help address such self-selection biases is to use survey weighting to weight the sample distributions in the forced exposure arm to match those seen in the choice exposure arm. In their study of news exposure effects, Stroud and colleagues (2018) weighted based on participants’ prior preferences for balanced, pro-attitudinal, or counter-attitudinal news expressed in an earlier wave of data collection. Rather than collect multiple waves of data, we opted for a more streamlined approach: statistically weighting based on psychographics known to predict information behavior for the message topic. Although this approach cannot rule out all possible confounding variables, weighting for known differences on plausible predictors of information selection helps reduce the potential influence of self-selection bias. In this way, climate change is an ideal message context because research has demonstrated one’s beliefs about climate change and one’s political orientation are strong predictors of climate change attitudes and information exposure (Feldman & Hart, 2018; Goldberg et al., 2020) and are therefore worth measuring and weighting for.

We excluded data from participants who were duplicates (n = 5), failed an attention check (n = 32), had technical issues hearing or seeing the video (n = 1), or straight-lined survey responses (n = 8).³ Of the remaining 3,341 respondents, n = 694 (21%) were assigned to the forced arm, and n = 2,647 (79%) were assigned to the choice arm, of whom n = 693 (26% of choice arm) selected the target video and provided usable data (Figure 1). Thus, our final analytic sample was N = 1,387. The sample had an average age of M = 40 (SD = 13). The sample was roughly half women (50%) and was mostly non-Hispanic (89%) and white (78%). The majority had a college degree (56%). A majority identified as Democrat (54%) with 19% identifying as Republican and 27% as Independent, another party, or no party affiliation. On a scale of extremely liberal (1) to moderate (4) to extremely conservative (7), participants leaned somewhat liberal (M = 3.16, SD = 1.70). Most expressed belief that climate change is happening (92%) and that it is human-caused (71%).

Measures

We assessed perceived risk of climate change with the question (Kahan et al., 2015), “Assuming it’s happening, how much risk do you think climate change poses to human health, safety, or prosperity?,” ranging from no risk at all (1) to extreme risk (7) (M = 5.95, SD = 1.39).

We measured two types of behavioral intentions. Participants were asked to indicate how likely they would be to engage in various activities in the next few months on a scale of very unlikely (1) to very likely (7). First, 10 items captured intentions to perform climate activism behaviors (e.g., “Contact elected officials to urge them to take action to reduce climate change,” “Join or volunteer with an organization working to reduce climate change”) taken from previous work (Feldman & Hart, 2016), which we averaged into a scale (Cronbach’s α = .93, M = 3.67, SD = 1.45). Second, 10 items captured intentions to engage in personal mitigation behaviors (e.g., “When shopping, use your own bags,” “Avoid meat consumption”) taken from previous work (Hart, 2010), which we averaged into a scale (α = .83, M = 5.06, SD = 1.07). Complete survey items are available in the Supplemental Appendix.

Psychological reactance is a process that encompasses three components: perceived freedom threat, negative cognitions, and anger (Dillard & Shen, 2005). We used a four-item measure of perceived freedom threat (Dillard & Shen, 2005) with responses assessed on a seven-point Likert-type scale in response to statements like “The video tried to make a decision for me” (α = .92, M = 2.11, SD = 1.36). We operationalized negative cognitions as counterarguing, measured with three items borrowed from previous work (Niederdeppe et al., 2015) on a scale of strongly disagree (1) to strongly agree (7), such as “I found myself disagreeing with the video’s points” (α = .81, M = 2.26, SD = 1.36). Our anger measure specified the target of the anger felt (“How much did you feel each of these emotions toward the people who made the video?”) and included three anger-based items (angry, irritated, and annoyed) measured on a five-point scale of not at all (1) to extremely (5) (α = .91, M = 1.38, SD = .80).

We assessed enjoyment with four items (e.g., “I enjoyed watching the video,” “I liked the video”; α = .84, M = 4.93, SD = 1.38) and assessed appreciation with three items (e.g., “I found this video to be very meaningful,” “I was moved by the video”; α = .86, M = 5.10, SD = 1.46), both measured on seven-point scales borrowed from prior work (Oliver & Bartsch, 2010). Zero-order correlations between all variables are available in Figure S5 of the Supplemental Appendix.

Results

Effects of Exposure on Reactance (H1) and Entertainment (RQ1)

We ran independent samples t-tests on reactance (H1), entertainment (RQ1), and all other outcomes to test for differences between the exposure arms (see Table 1 for full results). Compared with forced exposure, choice exposure was associated with greater risk perception (M_choice = 6.28 vs. M_forced = 5.63), activism intentions (M_choice = 3.92 vs. M_forced = 3.41), mitigation intentions (M_choice = 5.23 vs. M_forced = 4.88), enjoyment (M_choice = 5.24 vs. M_forced = 4.61), and appreciation (M_choice = 5.38 vs. M_forced = 4.83). Choice exposure was also associated with lower perceived freedom threat (M_choice = 1.83 vs. M_forced = 2.40), counterarguing (M_choice = 2.02 vs. M_forced = 2.49), and anger (M_choice = 1.30 vs. M_forced = 1.46).

Table 1.

Overall Differences Between Forced and Choice Exposure Arms.

	Raw			Weighted
Variables	Forced (n = 694)	Choice (n = 693)	Test results	Forced (n = 694)	Test results
Persuasion variables
Perceived risk	5.63 (1.57)	6.28 (1.09)	t = 8.90, p < .001, d = 0.48	5.99 (0.98)	t = 5.21, p < .001, d = 0.28
Activism intentions	3.41 (1.51)	3.92 (1.34)	t = 6.68, p < .001, d = 0.36	3.68 (1.38)	t = 3.29, p = .001, d = 0.18
Mitigation intentions	4.88 (1.15)	5.23 (0.95)	t = 6.24, p < .001, d = 0.34	5.04 (1.04)	t = 3.55, p < .001, d = 0.19
Process variables
Freedom threat	2.40 (1.55)	1.83 (1.08)	t = 8.00, p < .001, d = 0.43	2.13 (1.19)	t = 4.92, p < .001, d = 0.26
Counterarguing	2.49 (1.50)	2.02 (1.15)	t = 6.53, p < .001, d = 0.35	2.19 (1.12)	t = 2.79, p = .005, d = 0.15
Anger	1.46 (0.87)	1.30 (0.72)	t = 3.80, p < .001, d = 0.20	1.37 (0.75)	t = 1.77, p = .076, d = 0.10
Enjoyment	4.61 (1.49)	5.24 (1.18)	t = 8.77, p < .001, d = 0.47	4.81 (1.34)	t = 6.34, p < .001, d = 0.34
Appreciation	4.83 (1.60)	5.38 (1.25)	t = 7.08, p < .001, d = 0.38	5.11 (1.29)	t = 3.96, p < .001, d = 0.21

Note. Values represent means with standard deviations in parentheses. We performed weighted t-test calculations with the help of an online t-test calculator.

However, as noted previously, the choice exposure arm was subject to self-selection biases that may interfere with our ability to causally compare the two arms. Indeed, although the two arms were balanced with respect to age, gender identity, ethnicity, race, and educational attainment (Supplemental Table S3), the choice arm was more ideologically liberal than the forced arm (M = 2.91 vs. M = 3.41, p < .001), had a greater proportion of Democrats (60% vs. 47%, p < .001), and had greater proportion of individuals who believe climate change is happening (95% vs. 88%, p < .001) and human-caused (79% vs. 63%, p < .001). We therefore weighted the forced exposure data to match the distributions seen in the choice exposure data with respect to political party and the two climate change beliefs. Table S3 of the Supplemental Appendix shows that applying these weights made the two arms equivalent with respect to the proportions of political partisans and climate change believers, and it did not markedly shift the other demographic variables. Thus, our weighting approach seems to have viably reduced selection bias in the choice arm with respect to one’s climate beliefs and political leaning, putting the forced and choice arms on more level ground for between-arm comparisons.

In these post hoc analyses (Table 1), differences between the arms became smaller in size but persisted for perceived risk (d_raw = 0.48 vs. d_weighted = 0.28), activism intentions (d_raw = 0.36 vs. d_weighted = 0.18), mitigation intentions (d_raw = 0.34 vs. d_weighted = 0.19), freedom threat (d_raw = 0.43 vs. d_weighted = 0.26), counterarguing (d_raw = 0.35 vs. d_weighted = 0.15), enjoyment (d_raw = 0.47 vs. d_weighted = 0.34, and appreciation (d_raw = 0.38 vs. d_weighted = 0.21). Exposure differences became non-significant for anger (d_raw = 0.20 vs. d_weighted = 0.10).⁴

To assess the robustness of the findings from this set of weighted analyses, we also performed linear regressions that adjusted for a range of covariates (demographics, political ideology, political party, belief that climate change is happening, and belief that climate change is human-caused). These regressions retrieved the same results as the weighted analyses: Exposure effects remained for all outcome variables except for anger, which suggests the two analytic approaches tell a similar story (results available in the data and syntax online). To summarize, these results partially support our reactance hypothesis (H1) and also suggest that choice exposure was linked to greater entertainment responses (RQ1) and more favorable persuasion outcomes.

Reactance (H2) and Entertainment Responses (RQ2) Predicting Persuasion

We used structural equation modeling to construct a parallel mediation model in which exposure arm indirectly affected the three dependent variables through (a) reactance variables (b) enjoyment, and (c) appreciation (see Figure 2). For the reactance variables, we constructed a serial mediation path whereby freedom threat preceded reactance, a second-order latent variable comprising counterarguing and anger (Dillard & Shen, 2005). We modeled all variables as latent except for perceived risk (a single-item measure) and the two intention scales, which, as aggregations of various behaviors, may not necessarily reflect latent variables that determine their indicators (Bollen & Lennox, 1991).

Figure 2.

Visualizing the Indirect Effects of Exposure Type on Persuasion Outcomes.

Although it would have been ideal to run a weighted structural equation model to parallel the weighted analyses reported earlier, to our knowledge, it is not possible to weight only the forced exposure data with available analytic tools for structural equation modeling. Instead, we adjusted for all covariates to account for self-selection differences between the forced and choice arms. We are reasonably confident that this approach provides similar estimates as a weighted analysis would, given that the two analytic approaches yielded the same substantive results in our earlier tests of H1 and RQ1. In addition, we included dummy variables for the video watched, setting the informational video as the reference level (see video-specific results in Table S1 of the Supplemental Appendix). We set the covariates as predictors of each mediator and dependent variable.

An initial measurement model demonstrated acceptable fit to the data, χ²(110) = 1,164.32, CFI = .94, RMSEA = .08 [90% CI = .08, .09], SRMR = .07, as did the final structural model, χ²(365) = 2.772.69, CFI = .90, RMSEA = .07 [90% CI = .067, .071], SRMR = .08. Choice exposure was associated with less freedom threat (B = −.13) but greater enjoyment (B = .18) and appreciation (B = .11), consistent with the weighted t-tests presented previously. Freedom threat positively predicted reactance (B = .60), which in turn predicted lower perceived risk (B = −.28) and greater activism intentions (B = .15) but did not predict mitigation intentions (B = −.01), offering mixed support for H2. Thus, the data are consistent with a model in which freedom threat and reactance serially mediated the influence of exposure on perceived risk and activism intentions but not mitigation intentions (see Table 2 for all indirect effects). Enjoyment predicted lower perceived risk (B = −.09) but did not predict activism intentions (B = .07) or mitigation intentions (B = −.06), meaning enjoyment negatively mediated the relationship between exposure and perceived risk only. Answering RQ2, appreciation positively predicted all three outcomes (B_risk = .36, B_activism = .50, B_mitigation = .45). Therefore, a causal interpretation of the model would indicate that choice exposure, by generating more appreciation than forced exposure, indirectly promoted perceived risk and behavioral intentions.

Table 2.

Indirect Effect Estimates of Exposure Type on Persuasion Outcomes.

	B	SE	p
Choice (vs. forced) → enjoyment → _____
Perceived risk	−.02	.02	.033
Activism intentions	.01	.03	.171
Mitigation intentions	−.01	.02	.257
Choice (vs. forced) → appreciation → _____
Perceived risk	.04	.03	<.001
Activism intentions	.06	.04	<.001
Mitigation intentions	.05	.03	<.001
Choice (vs. forced) → freedom threat → reactance → _____
Perceived risk	.02	.01	<.001
Activism intentions	−.01	.01	.001
Mitigation intentions	.001	.01	.642

Note. B-coefficients are standardized. SE = standard error.

Of note, the direct paths from exposure to perceived risk and mitigation intentions were positive and significant (B_risk = .06, B_mitigation = .05), whereas the direct path to activism intentions was not significant (B_activism = .04). Thus, these mechanisms partially mediated the association that exposure type had with perceived risk and mitigation intentions but fully mediated the association between exposure type and activism intentions. We do wish to point out, however, that all three direct paths were of similar magnitude.

Effects of Type of Exposure to Each Emotional Appeal on Persuasion (RQ3)

To understand the relative persuasiveness of exposure type for the different emotional appeals (RQ3), we ran three multivariable linear regressions that interacted exposure type with emotional appeal and adjusted for the same covariates as our structural equation model (Table 3). None of the interaction terms were significant for models of perceived risk, activism intentions, or mitigation intentions. Therefore, there is no evidence to suggest the effect of each emotional appeal depended on the mode of exposure.

Table 3.

Linear regressions with interactions between emotional appeal and exposure type.

	Perceived risk	Activism intentions	Mitigation intentions
Intercept	3.69***	2.27***	3.73***
Intercept	[3.20, 4.19]	[1.64, 2.91]	[3.25, 4.22]
Emotional appeal (vs. informational)
Threat appeal	0.31***	0.03	−0.01
Threat appeal	[0.13, 0.49]	[−0.20, 0.26]	[−0.19, 0.17]
Humor appeal	0.11	−0.30*	−0.17
Humor appeal	[−0.08, 0.29]	[−0.53, −0.06]	[−0.35, 0.01]
Choice arm (vs. forced)	0.29**	0.19	0.15
Choice arm (vs. forced)	[0.11, 0.47]	[−0.04, 0.42]	[−0.02, 0.32]
Interactions
Threat appeal × arm	−0.03	0.18	0.09
Threat appeal × arm	[−0.29, 0.22]	[−0.15, 0.51]	[−0.16, 0.33]
Humor appeal × arm	0.04	0.10	0.09
Humor appeal × arm	[−0.22, 0.30]	[−0.23, 0.43]	[−0.16, 0.34]
Observations	1,387	1,387	1,387
R ²	.49	.23	.19

Note. Cells represent unstandardized beta coefficients and 95% confidence intervals. Bolded values are statistically significant. Models also controlled for demographics, political ideology, political party, belief that climate change is happening, and belief that climate change is human-caused.

p < .05 ** p < .01 *** p < .001.

Discussion

Facing incredible latitude among media users to choose and avoid content in the media ecosystem, science communicators must be cognizant of the way in which exposure matters for the influence of the messages they study. Unfortunately, science communication scholarship has devoted far less attention to the theoretical underpinnings of message exposure effects than message content effects. This study drew from theories of persuasion, entertainment psychology, and the psychology of choice to derive predictions about how the nature of exposure (forced or selected) shapes responses to messages about climate change.

We find that, compared with those randomized to forced exposure, those randomized to choice exposure who selected a climate change–related video had higher levels of perceived risk, activism intentions, and mitigation intentions—differences that did not depend on whether participants were exposed to a threat appeal, humor appeal, or informational appeal. Mediation analyses suggest these differences in persuasion outcomes were explained by diminished reactance and heightened enjoyment and appreciation. Crucially, adjusting for self-selection biases reduced, but did not eliminate, these estimated differences. We note, however, that these differences were small in size with effect sizes ranging from d = 0.15 for counterarguing (meaning there was 94% overlap between the counterarguing distributions of the forced and choice arms) to d = 0.34 for enjoyment (87% overlap between the enjoyment distributions) (Magnusson, 2022). We see several theoretical and methodological takeaways from these findings.

First, we conceptually replicate prior work (Stroud et al., 2018), showing choice exposure may inhibit reactance in terms of perceived freedom threat and counterarguing but not anger. These reactance patterns are strikingly similar to those reported by Stroud and colleagues, who found uniform exposure type differences in negative cognitions but topic-dependent effects on anger. Thus, even though we used one issue context and one set of stimuli in this study, a clear picture is emerging from the available evidence: Requiring exposure to media content (persuasive or otherwise) may engender slightly greater cognitive but not affective reactance. However, it is worth acknowledging that there were indirect effects of exposure type on reactance, a latent state comprising counterarguing and anger, through perceived freedom threat. Reactance theory has chiefly focused on reactance responses to message content (e.g., domineering language) (Ratcliff, 2019), but our work underscores the need to think beyond message content to develop a more complete understanding of how prosocial persuasion efforts elicit reactance. Moreover, although theorists often hail reactance as an obstacle to message success, our evidence suggests requiring message exposure may somewhat overestimate the level of reactance that individuals experience relative to more naturalistic environments where people can opt into the messaging.

Second, not only can the nature of exposure affect defensive processing, but it may also matter for how entertained users are and, by consequence, how persuaded they are. That choice exposure was linked to greater enjoyment and appreciation supports psychological research that choice is intrinsically desirable in Western cultures, resulting in especially favorable evaluations of the chosen activity (Iyengar & Lepper, 2002; Leotti & Delgado, 2011). Also, whereas previous studies comparing forced with choice exposure have failed to see differences on persuasion-related outcomes (Arendt et al., 2022; Dahlgren, 2022; Stroud et al., 2018), we found that individuals expressed somewhat greater message-consistent belief and intentions when they chose the message. Because prior studies have largely focused on non-strategic communication (e.g., news, discussion comments), our results may imply a boundary condition whereby requiring exposure only inhibits persuasion when the message has persuasive intent (see also Schlosser & Shavitt, 2009).

Third, although the potential effects of exposure on appreciation were smaller than the effects on enjoyment or freedom threat, mediation analyses revealed appreciation was the mechanism that most consistently explained how exposure predicted persuasion. Thus, although message choice made for both a more fun experience and a more profound, moving experience, it was this latter experience of appreciation that carried the most weight in explaining differences in persuasive outcomes. These patterns mirror earlier findings on the importance of the entertainment value of prosocial messaging (Feldman & Borum Chattoo, 2019) and illustrate the benefit of integrating concepts from media entertainment theory into persuasion contexts (Oliver & Bartsch, 2010). Perhaps appreciation predicted greater persuasion outcomes because it encouraged deeper elaboration about the ideas presented, whereas enjoyment reflected a more surface-level response that failed to motivate viewers to want to take action (Feldman & Borum Chattoo, 2019). Reactance inconsistently predicted our persuasion outcomes when considering these other mechanisms of influence. This tells us that even if requiring message exposure does induce greater reactance, reactance may not by itself account for the differences in outcomes, further highlighting the significance of entertainment responses—appreciation in particular.

An important limitation of this work is that we jettisoned participants in the choice exposure arm who chose a non-climate message. Although statistically weighting for key variables known to be among the strongest predictors of climate engagement (i.e., climate change beliefs, political orientation) should account for major differences between the forced and choice arms, we cannot eliminate all possible confounding factors given self-selection biases in the choice exposure arm (e.g., the impact of previous pro-environmental behaviors above and beyond climate beliefs). Our findings may not generalize beyond those individuals who are interested in the topic of climate change; that is, people are who are not particularly favorable toward climate change mitigation might not have favorable message responses regardless of their mode of exposure. In contrast to previous studies of a similar nature that permitted participants to choose only between messages about the same issue (e.g., Arendt et al., 2022; Stroud et al., 2018), we felt it was important to create as ecologically valid a media selection experience as possible, in which media users browse between messages about multiple topics, incidentally encountering content about climate change.

In future work, it would be prudent to adjust for a wider range of covariates—previous behavioral engagement, in particular—as a more stringent test of the differences between selected and required exposure. Indeed, although political affiliation and belief in climate change together serve as reasonable proxies for prior climate change engagement, simulation-based analyses would suggest that the risk of confounding bias increases as the correlation between a measured proxy variable (here, political affiliation and climate belief) and an unmeasured confounder (prior behavior) decreases (Desai et al., 2024). An alternative design approach would be to allow participants to select from multiple climate change messages side by side, similar to the designs used in previous work (e.g., Arendt et al., 2022; Stroud et al., 2018). This approach would virtually eliminate the risk of confounding bias because all of the message options would be about the topic of interest, so none of the participants permitted to choose their message would need to be removed before analysis. In any case, the logistical challenges posed by questions about the nature of message exposure will require researchers to leverage a range of innovative methodological and analytic strategies. Only by triangulating across these strategies will we develop the clearest possible understanding of how message effects change when exposure is not required.

We also wish to acknowledge some additional limitations. First, related to the causal inference issues discussed above, our mediation analyses rely on cross-sectional data, and so the temporal relationships between our proposed mediators and outcomes could be called into question. As an example, it is possible that changes in appreciation occur at the same time as changes in perceived risk, which would mean appreciation does not in fact function as a mediator. Second, our findings are limited by the use of a single topic (climate change) and a single set of messages. Although we found no evidence that the nature of the messaging moderated the persuasive influence of exposure type, we look forward to work assessing the effects of selective exposure to other commonly studied science communication tactics (e.g., gain-loss framing, narratives) to test whether the current findings may be conceptually replicated across message contexts. Third, even in the choice arm of our experiment, participants had a limited number of message options from which to select. Because too many options can hamper consumer satisfaction (Chernev et al., 2015), we may have observed less of a benefit for selected over forced exposure had we provided a larger assortment of options in the choice arm. Finally, with regard to constraints on the generality of our findings (Simons et al., 2017), non-Western cultures place less value on personal choice than Western cultures, and they may even respond more favorably when no choice is permitted (Iyengar & Lepper, 2002). We therefore suspect that the findings reported here with our (largely white, educated) sample may only hold for Western populations.

Conclusion

Randomized experiments are often heralded as the gold standard for causal inferences, including the impact of science communication efforts. Yet as they are normally executed, messaging experiments neglect the fact that media users are often purposeful in choosing the types of science content they watch, listen to, and read. We view our findings as somewhat heartening for extant research on strategic science communication. Although requiring exposure generally resulted in less favorable message responses than choice exposure, these differences were quite small in magnitude and may very well disappear should additional confounding factors be taken into account. Thus, it remains possible that message effects seen in required exposure contexts may not be far off from effects that occur “in the wild” where message choice is the norm. If differences do occur, it appears science communication experiments slightly underestimate favorable message reactions. We believe the field would benefit from additional investigations that compare forced to selective exposure, but we also acknowledge that such designs may not be logistically or financially feasible for all researchers. At the very least, we hope these findings encourage science communication scholars to shift away from viewing message recipients as passive audiences and instead regard them as active users of science media content.

Supplemental Material

sj-docx-1-scx-10.1177_10755470251316777 – Supplemental material for How Forced Versus Selective Exposure Matters for the Influence of Emotional Appeals About Climate Change in a Sample of Americans

Supplemental material, sj-docx-1-scx-10.1177_10755470251316777 for How Forced Versus Selective Exposure Matters for the Influence of Emotional Appeals About Climate Change in a Sample of Americans by Chris Skurka, Rainer Romero-Canyas, Helen H. Joo and Jeff Niederdeppe in Science Communication

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iDs

Chris Skurka

Helen H. Joo

Supplemental Material

Supplemental material for this article is available online at

Notes

Author Biographies

Chris Skurka (PhD, Cornell University) is an assistant professor in the Department of Film Production and Media Studies at Penn State University. (corresponding)

Rainer Romero-Canyas (PhD, Columbia University) is a lead senior social scientist at the Environmental Defense Fund.

Helen H. Joo (MA, New York University) is a senior social science research analyst at the Environmental Defense Fund.

Jeff Niederdeppe (PhD, University of Pennsylvania) is a professor in the Department of Communication at Cornell University.

References

Arceneaux

Johnson

(2013). Changing minds or changing channels? Partisan news in an age of choice. University of Chicago Press.

Arendt

Forrai

Mestas

(2022). News framing and preference-based reinforcement: Evidence from a real framing environment during the COVID-19 pandemic. Communication Research, 50(2), 179–204. https://doi.org/10.1177/00936502221102104

Bollen

Lennox

(1991). Conventional wisdom on measurement: A structural equation perspective. Psychological Bulletin, 110, 305–314. https://doi.org/10.1037/0033-2909.110.2.305

Brehm

S. S.

Brehm

J. W.

(1981). Psychological reactance: A theory of freedom and control. Academic Press.

Chernev

Böckenholt

Goodman

(2015). Choice overload: A conceptual review and meta-analysis. Journal of Consumer Psychology, 25(2), 333–358. https://doi.org/10.1016/j.jcps.2014.08.002

Dahlgren

P. M.

(2022). Forced versus selective exposure: Threatening messages lead to anger but not dislike of political opponents. Journal of Media Psychology: Theories, Methods, and Applications, 34(3), 150–164. https://doi.org/10.1027/1864-1105/a000302

Desai

R. J.

Bradley

M. C.

Lee

Eworuke

Weberpals

Wyss

Schneeweiss

Ball

(2024). A simulation-based bias analysis to assess the impact of unmeasured confounding when designing nonrandomized database studies. American Journal of Epidemiology, 193(11), 1600–1608. https://doi.org/10.1093/aje/kwae102

Dillard

J. P.

Shen

(2005). On the nature of reactance and its role in persuasive health communication. Communication Monographs, 72, 144–168. https://doi.org/10.1080/03637750500111815

Eisend

(2009). A meta-analysis of humor in advertising. Journal of the Academy of Marketing Science, 37, 191–203. https://doi.org/10.1007/s11747-008-0096-y

10.

Ettinger

Walton

Painter

DiBlasi

(2021). Climate of hope or doom and gloom? Testing the climate change hope vs. Fear communications debate through online videos. Climatic Change, 164(1), 19. https://doi.org/10.1007/s10584-021-02975-8

11.

Feldman

Borum Chattoo

(2019). Comedy as a route to social change: The effects of satire and news on persuasion about Syrian refugees. Mass Communication and Society, 22(3), 277–300. https://doi.org/10.1080/15205436.2018.1545035

12.

Feldman

Hart

P. S.

(2016). Using political efficacy messages to increase climate activism: The mediating role of emotions. Science Communication, 38, 99–127. https://doi.org/10.1177/1075547015617941

13.

Feldman

Hart

P. S.

(2018). Broadening exposure to climate change news? How framing and political orientation interact to influence selective exposure. Journal of Communication, 68, 503–524. https://doi.org/10.1093/joc/jqy011

14.

Fisher

J. T.

Hamilton

K. A.

(2021). Integrating media selection and media effects using decision theory. Journal of Media Psychology, 33(4), 215–225. https://doi.org/10.1027/1864-1105/a000315

15.

Fletcher

Nielsen

R. K.

(2017). Are news audiences increasingly fragmented? A cross-national comparative analysis of cross-platform news audience fragmentation and duplication. Journal of Communication, 67(4), 476–498. https://doi.org/10.1111/jcom.12315

16.

Gaines

B. J.

Kuklinski

J. H.

(2011). Experimental estimation of heterogeneous treatment effects related to self-selection. American Journal of Political Science, 55(3), 724–736. https://doi.org/10.1111/j.1540-5907.2011.00518.x

17.

Goldberg

M. H.

Gustafson

Ballew

M. T.

Rosenthal

S. A.

Leiserowitz

(2020). Identifying the most important predictors of support for climate policy in the United States. Behavioural Public Policy, 5, 480–502. https://doi.org/10.1017/bpp.2020.39

18.

Hart

P. S.

(2010). One or many? The influence of episodic and thematic climate change frames on policy preferences and individual behavior change. Science Communication, 33, 28–51. https://doi.org/10.1177/1075547010366400

19.

Hornik

R. C.

(2002). Exposure: Theory and evidence about all the ways it matters. Social Marketing Quarterly, 8(3), 31–37. https://doi.org/10.1080/15245000214135

20.

Iyengar

S. S.

Lepper

M. R.

(2002). Choice and its consequences: On the costs and benefits of self-determination. In Stapel

Tesser

Wood

J. V.

(Eds.), Self and motivation: Emerging psychological perspectives (pp. 71–96). American Psychological Association. https://doi.org/10.1037/10448-003

21.

Kahan

D. M.

Jenkins-Smith

Tarantola

Silva

C. L.

Braman

(2015). Geoengineering and climate change polarization: Testing a two-channel model of science communication. The ANNALS of the American Academy of Political and Social Science, 658, 192–222. https://doi.org/10.1177/0002716214559002

22.

Kaltenbacher

Drews

(2020). An inconvenient joke? A review of humor in climate change communication. Environmental Communication, 14(6), 717–729. https://doi.org/10.1080/17524032.2020.1756888

23.

Knobloch-Westerwick

(2015). Choice and preference in media use: Advances in selective exposure theory and research. Routledge.

24.

Koblin

(2024, January 14). A few words about Netflix’s success: Vivid. Snappy. Tags. The New York Times. https://www.nytimes.com/2024/01/14/business/media/netflix-streaming-movies-ratings.html

25.

Leotti

L. A.

Delgado

M. R.

(2011). The inherent reward of choice. Psychological Science, 22(10), 1310–1318. https://doi.org/10.1177/0956797611417005

26.

Leotti

L. A.

Delgado

M. R.

(2014). The value of exercising control over monetary gains and losses. Psychological Science, 25(2), 596–604. https://doi.org/10.1177/0956797613514589

27.

Leotti

L. A.

Iyengar

S. S.

Ochsner

K. N.

(2010). Born to choose: The origins and value of the need for control. Trends in Cognitive Sciences, 14(10), 457–463. https://doi.org/10.1016/j.tics.2010.08.001

28.

Magnusson

(2022). Interpreting Cohen’s d. https://rpsychologist.com/cohend/

29.

McGuire

W. J.

(1985). Attitudes and attitude change. In Lindzey

W. G.

Aronson

(Eds.), Handbook of social psychology (pp. 233–346). Random House.

30.

Moore

M. M.

Green

M. C.

(2021). Narrative and choice effects on learning outcomes. Communication Quarterly, 69(4), 410–430. https://doi.org/10.1080/01463373.2021.1951795

31.

Nabi

R. L.

Gustafson

Jensen

(2018). Framing climate change: Exploring the role of emotion in generating advocacy behavior. Science Communication, 44(4), 442–468. https://doi.org/10.1177/1075547018776019

32.

Niederdeppe

Heley

Barry

C. L.

(2015). Inoculation and narrative strategies in competitive framing of three health policy issues. Journal of Communication, 65, 838–862. https://doi.org/10.1111/jcom.12162

33.

Oliver

M. B.

Bartsch

(2010). Appreciation as audience response: Exploring entertainment gratifications beyond hedonism. Human Communication Research, 36(1), 53–81. https://doi.org/10.1111/j.1468-2958.2009.01368.x

34.

Ratcliff

C. L.

(2019). Characterizing reactance in communication research: A review of conceptual and operational approaches. Communication Research, 48(7), 1033–1058. https://doi.org/10.1177/0093650219872126

35.

Rubin

A. M.

(2009). Uses and gratifications: An evolving perspective of media effects. In Nabi

R. L.

Oliver

M. B.

(Eds.), The SAGE handbook of media processes and effects (3rd ed., pp. 147–160). Sage.

36.

Ryan

R. M.

Deci

E. L.

(2006). Self-regulation and the problem of human autonomy: Does psychology need choice, self-determination, and will? Journal of Personality, 74(6), 1557–1586. https://doi.org/10.1111/j.1467-6494.2006.00420.x

37.

Schlosser

A. E.

Shavitt

(2009). The effect of perceived message choice on persuasion. Journal of Consumer Psychology, 19(3), 290–301. https://doi.org/10.1016/j.jcps.2009.03.006

38.

Shadish

W. R.

Cook

T. D.

Campbell

D. T.

(2002). Experimental and quasi-experimental designs for generalized causal inference. Houghton, Mifflin and Company.

39.

Simis

M. J.

Madden

Cacciatore

M. A.

Yeo

S. K.

(2016). The lure of rationality: Why does the deficit model persist in science communication? Public Understanding of Science, 25(4), 400–414. https://doi.org/10.1177/0963662516629749

40.

Simons

D. J.

Shoda

Lindsay

D. S.

(2017). Constraints on generality (COG): A proposed addition to all empirical papers. Perspectives on Psychological Science, 12(6), 1123–1128. https://doi.org/10.1177/1745691617708630

41.

Slater

M. D.

(2015). Reinforcing spirals model: Conceptualizing the relationship between media content exposure and the development and maintenance of attitudes. Media Psychology, 18(3), 370–395. https://doi.org/10.1080/15213269.2014.897236

42.

Stroud

N. J.

Feldman

Wojcieszak

Bimber

(2018). The consequences of forced versus selected political media exposure. Human Communication Research, 45(1), 27–51. https://doi.org/10.1093/hcr/hqy012

43.

Tannenbaum

M. B.

Hepler

Zimmerman

R. S.

Saul

Jacobs

Wilson

Albarracín

(2015). Appealing to fear: A meta-analysis of fear appeal effectiveness and theories. Psychological Bulletin, 141, 1178–1204. https://doi.org/10.1037/a0039729

44.

Tripathi

Cervone

Savani

(2018). Are the motivational effects of autonomy-supportive conditions universal? Contrasting results among Indians and Americans. Personality and Social Psychology Bulletin, 44(9), 1287–1301. https://doi.org/10.1177/0146167218764663

45.

Walter

Cody

M. J.

L. Z.

Murphy

S. T.

(2018). A priest, a rabbi, and a minister walk into a bar: A meta-analysis of humor effects on persuasion. Human Communication Research, 44(4), 343–373. https://doi.org/10.1093/hcr/hqy005

46.

Witte

(1992). Putting the fear back into fear appeals: The extended parallel process model. Communications Monographs, 59, 329–349. https://doi.org/10.1080/03637759209376276

47.

Yeo

S. K.

McKasy

(2021). Emotion and humor as misinformation antidotes. Proceedings of the National Academy of Sciences, 118(15), Article e2002484118. https://doi.org/10.1073/pnas.2002484118

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

3.10 MB