Philosophy

Full Terms & Conditions of access and use can be found at https://www.tandfonline.com/action/journalInformation?journalCode=uhcm20

Journal of Health Communication International Perspectives

ISSN: (Print) (Online) Journal homepage: https://www.tandfonline.com/loi/uhcm20

“I Felt Completely Turned off by the Message”: The Effects of Controlling Language, Fear, and Disgust Appeals on Responses to COVID-19 Vaccination Messages

Haijing Ma & Claude Miller

To cite this article: Haijing Ma & Claude Miller (2022) “I Felt Completely Turned off by the Message”: The Effects of Controlling Language, Fear, and Disgust Appeals on Responses to COVID-19 Vaccination Messages, Journal of Health Communication, 27:6, 427-438, DOI: 10.1080/10810730.2022.2119311

To link to this article: https://doi.org/10.1080/10810730.2022.2119311

Published online: 12 Sep 2022.

Submit your article to this journal

Article views: 529

View related articles

View Crossmark data

Citing articles: 1 View citing articles

“I Felt Completely Turned off by the Message”: The Effects of Controlling Language, Fear, and Disgust Appeals on Responses to COVID-19 Vaccination Messages HAIJING MA and CLAUDE MILLER

Department of Communication, The University of Oklahoma, Burton Hall, Norman, Oklahoma, USA

Psychological reactance theory (PRT) posits that when individuals’ perceived freedoms are threatened or restricted, they become aversively aroused and are motivated to reestablish those freedoms, leading to a state of psychological reactance. Applying PRT, this study examined the effects of controlling language, fear, and disgust appeals on responses to COVID-19 vaccination promotion messages. Participants were randomly assigned to one of eight conditions across controlling language (high/low), fear appeals (high/low), and disgust appeals (high/low), wherein they viewed two messages, with responses measured after each message. Results showed persuasion was diminished when the levels of any of these three variables were elevated, as in conditions of either high controlling language, high fear appeals, or high disgust appeals. Relative to low levels of these variables, high levels resulted in greater freedom threat perceptions, reactance, source derogation, and less positive attitudes toward the message. A 2-way interaction between fear and disgust appeals on source derogation and message attitudes in the low controlling language condition was significant—participants reported the least source derogation and most positive attitudes toward the message in response to the low controlling language, low fear, and low disgust appeals.

Since the first reported case on December 31, 2019, the COVID-19 pandemic caused by the severe acute respiratory syndrome coronavirus 2 has killed over 4,770,000 people worldwide by October 2021 (Johns Hopkins Coronavirus Resource Center, 2021). Moreover, variants of the virus with greater virulence have been detected (Doucleff, 2021). Given how vaccines offer up to 95% protection against COVID-19 (Cohen, 2020), vaccination represents a vital instrument for alleviating the health threats posed by the pandemic. Unfortunately, COVID-19 vaccination remains low in several sectors of the US population (Berkeley, 2021). This study seeks to examine the effectiveness of three message strategies to motivate COVID-19 vaccination.

Controlling language, fear, and disgust within health-risk mes- sages have been widely used to promote healthy behaviors (Morales, Wu, & Fitzsimons, 2012; Tannenbaum et al., 2015). However, they hold the potential to backfire if they arouse psycho- logical reactance by posing a threat to perceived freedoms, thereby sabotaging the persuasiveness of health messages (Bahari, 2019; Ma & Miller, 2021). It is therefore important to examine how these strategies may be used to more effectively communicate COVID- 19 risks, and ultimately motivate people to get vaccinated.

Because high fear and disgust appeals may trigger reactance in their attempts to influence behaviors through forceful threats

of dire and repulsive consequences, they may thereby threaten individuals’ freedom to otherwise not engage in the advocated behaviors (Ratcliff, 2021; Yang, 2017). However, under certain conditions where environmental threat levels are particularly high, they may license the use of high controlling language, and thereby increase persuasion (Bessarabova & Massey, 2020; Buller et al., 2000). Therefore, this study focuses on under- standing reactance and responses to messages featuring control- ling language while including fear and disgust content in advocating for COVID-19 vaccination within the menacing context posed by the global pandemic.

Psychology Reactance Theory

Psychological reactance theory (PRT) posits that when one’s freedom is eliminated or threatened with elimination, indivi- duals are motivated to reestablish the threatened freedom, lead- ing to the arousal of reactance (Brehm, 1966; Brehm & Brehm, 1981). A perceived threat to freedom is a prerequisite for reactance (Dillard & Shen, 2005; Quick & Stephenson, 2008), and any communication threatening established freedoms or indicating the possibility of a loss of freedom may potentially generate reactance. Once a threat to perceived freedom is posed, individuals become reactant, and are motivationally aroused to reestablish the threatened freedom (Brehm & Brehm, 1981). Within a health risk context, individuals may engage in unhealthy defensive responses to restore their free- doms (Miller, Lane, Deatrick, Young, & Potts, 2007; Quick, Shen, & Dillard, 2013; Richards, Bessarabova, Banas, &

Address correspondence to Haijing Ma, Department of Communication, the University of Oklahoma, Burton Hall, Norman, 73019, Oklahoma, USA. mail: [email protected]

Journal of Health Communication, 27: 427–438, 2022 Copyright © Taylor & Francis Group, LLC ISSN: 1081-0730 print/1087-0415 online DOI: https://doi.org/10.1080/10810730.2022.2119311

Bernard, 2020), one of which involves source derogation, whereby hostility is expressed toward a freedom-threatening agent (Grandpre, Alvaro, Burgoon, Miller, & Hall, 2003; Miller, 2015). Furthermore, reactance arousal is also associated with less positive attitudes toward the persuasion attempt, lower intentions to act as recommended, and increased attraction to the threatened freedom (Brehm & Brehm, 1981; Dillard & Shen, 2005; Ratcliff, 2021).

Message Strategies and Psychological Reactance

High controlling language uses forceful language, with impera- tives such as “must” and “have to,” to pressure message recei- vers into adopting certain attitudes and actions (Clayton, Leshner, Sanders-Jackson, & Hendrickse, 2020; Miller et al., 2007). In contrast, low controlling language is less forceful, using autonomy-supportive qualifiers (Deci & Ryan, 1987), such as “maybe” and “perhaps.”

High controlling language is often seen as freedom threaten- ing, triggering reactance (Miller, 2015; Miller, Massey, & Ma, 2020) and unfavorable message responses (LaVoie, Quick, Riles, & Lambert, 2017; Reynolds-Tylus, 2019). When indivi- duals realize an influencing agent’s intent to persuade, they will often become defensive, even if an influencing attempt is in the interest of the receiver—merely sensing an attempt to persuade frequently threatens the receiver’s freedom (Miller et al., 2007; Walster & Festinger, 1962). Communication research has iden- tified a range of counterproductive reactance-induced outcomes resulting from the use of controlling language (Clayton et al., 2020; Miller et al., 2020; Quick et al., 2013).

Therefore, high controlling language commanding people to get vaccinated should be seen as more freedom threatening than low controlling, autonomy-supportive language. As a result, individuals should become more reactant, decreasing the effec- tiveness of messages using high controlling language. This forms the basis for the following hypotheses concerning COVID-19 vaccination promotion:

Hypothesis 1: Relative to low controlling language, high con- trolling language induces more a) freedom threat perceptions and b) psychological reactance.

Hypothesis 2: Relative to low controlling language, high con- trolling language leads to a) more source derogation, b) less positive attitudes toward the message, and c) lower COVID-19 vaccination intentions.

Brehm (1966) also noted, not all freedom threat is reactance- arousing; thus, the current study explores circumstances under which controlling language might be effective, as for instance when the potential for a physical threat is high, and high fear and/or high disgust content is present within a persuasive message.

Fear appeals typically describe potential harms people will suffer if they do not follow message recommendations (Dillard, 1994); they are often used to persuade people to forgo harmful behaviors by making relevant health threats salient (Witte, 1992). Relative to low fear appeals, high fear appeals, when accompanied by high perceived efficacy, lead to more positive persuasive out- comes (Tannenbaum et al., 2015; Witte & Allen, 2000).

As a persuasion strategy, however, fear appeals have the potential to threaten freedom and elicit reactance (Ratcliff, 2021; Wechsler et al., 2003). As per the extended parallel process model (Witte, 1992), a fear appeal includes threat and efficacy, wherein threat appraisals evaluate both the severity of threat and individuals’ susceptibility to them. Likewise, efficacy appraisals assess the effectiveness of recommended actions in reducing threats, along with individuals’ ability to perform those recommended actions. Following this conceptualization, high fear appeals generally emphasize both high severity and suscept- ibility while boosting or holding the efficacy components con- stant (Kim & Shin, 2018). In describing the dire consequences of health threats, high fear appeals may both explicitly and implicitly pressure message recipients into performing certain actions to avoid or reduce the risks, while emphasizing how noncompliance will invite negative consequences.

In other words, fear appeals attempt to influence behaviors by pressing upon receivers the dire consequences of noncompliance, frightening them into compliance, and thereby potentially threa- tening perceived freedom to otherwise not engage in the recom- mended behaviors. As studies have observed, the more the intention to influence is apparent, the more even well-intended appeals may be seen as freedom threatening (Grandpre et al., 2003; Miller et al., 2001, 2020). As a result, fear appeals may often lead to perceptions of manipulation, which can elicit reac- tance. It stands to reason that as the level of fear appeal increases and the intention to persuade becomes more explicit, freedom threat perceptions and reactance should increase concomitantly.

On the other hand, fear appeals may not always induce reactance, in situations where freedom is less important, people may deemphasize or forgo freedom concerns (Brehm, 1966; Brehm & Brehm, 1981; Hargrove, 2003). In this light, the potential for more dire consequences should make secondary concerns about freedoms less pressing. Therefore, when a physical or environmental threat is high enough, high fear appeals should be less freedom-threatening than they might otherwise be under more benign circumstances.

Despite half of a century of research, fear appeals have only recently been examined from a reactance perspective (Peng, Shen, Vanderbilt, Kim, & Foley, 2020). For instance, Quick and Stephenson (2008) found, vivid language indicating high threat generated more reactance than pallid language suggesting low threat. Likewise, Dillard, Kim, and Li (2018) showed how anti-sugar-sweetened beverage messages can produce more reactance than control messages, and Clayton, Lang, Leshner, and Quick (2019) showed how high relative to low health threat anti-smoking appeals can produce more anger and counterarguing.

In line with these recent studies showing the potential for fear appeals to arouse reactance, the following hypothesis con- cerning COVID-19 vaccination messages is posited:

Hypothesis 3: Relative to low fear appeals, high fear appeals induce more a) freedom threat perceptions and b) psychological reactance.

As predicted by PRT and found in several recent fear appeal studies, when reactance is aroused, fear appeals tend to be less

428 H. Ma and C. Miller

persuasive (Peng et al., 2020; Quick, LaVoie, Reynolds-Tylus, Martinez-Gonzalez, & Skurka, 2018). As a result, a high fear COVID-19 vaccination appeal may be expected to lead to less positive responses. However, meta-analyses show high relative to low fear appeals should increase persuasion (Tannenbaum et al., 2015; Witte & Allen, 2000). Considering the seeming inconsistencies this entails, the following research question concerning COVID-19 vaccination messages is offered:

RQ1: How do high vs. low fear appeals influence a) source derogation, b) attitudes toward the message, and c) COVID-19 vaccination intentions?

Fear appeals commonly include potentially repulsive content many individuals will find disgusting (Allred & Amos, 2018). Disgust is a discrete, primary emotion functioning to protect organisms from environmental threats (Haidt, McCauley, & Rozin, 1994). As two distinct emotions, fear and disgust are featured by different physiology and action tendencies (Russell, Lewicka, & Niit, 1989; Woody & Teachman, 2000). Considering how threat minimization goals associated with fear coping moti- vate individuals to seek information useful in reducing threats, whereas contamination avoidance goals associated with disgust motivate distancing oneself from repulsive stimuli (Susskind et al., 2008), fear and disgust may work at odds in affecting attention to certain emotional appeal messages, thus adversely influence appraisal of message content and risk assessment (Van Hooff, Devue, Vieweg, & Theeuwes, 2013). Therefore, if fear and disgust co-present, disgust-inducing and fear-inducing stimuli may be sensed in distinct opposition, thus differentially influen- cing message responses relative to when only fear- or only dis- gust-inducing stimuli are present.

However, instead of being studied independently, disgust appeals have primarily been used to strengthen fear appeals, which has confounded and left unanswered the question of whether and how disgust appeals uniquely impact message responses (Nabi, 2002). Moreover, although the inclusion of disgust suggests fear content alone may not fully account for the freedom threat perceptions in response to fear appeals, few studies have examined the role of disgust in arousing reactance (see Yang, 2017). Researchers have only recently begun to explore the effect of disgust appeals on message responses, showing inconsistent findings. For instance, Halkjelsvik and Rise (2015) found high vs. moderate disgust content about smoking appeared to have no significant influence on smoking intentions and beliefs. However, Morales et al. (2012) found advertisements with disgust content to be more persuasive than those with no disgust content in terms of attitudes and beha- vioral intentions. Leshner, Bolls, and Thomas (2009), (2011) studies found high disgust anti-smoking messages produced more cardiac deceleration than low disgust messages, indicating high disgust content may facilitate perceptual decisions (Ribeiro & Castelo-Branco, 2019) and increase cognitive pro- cessing. Leshner and colleagues’ work are relevant to persua- sion since cognition and memory mediate the relationship between message exposure and persuasion (Braun-Latour & Zaltman, 2006).

Given the above evidence indicating the potential for increased persuasiveness following high vs. low disgust

appeals, the following hypothesis is proposed concerning responses to COVID-19 vaccination messages:

Hypothesis 4: Relative to low disgust appeals, high disgust appeals lead to a) less source derogation, b) more positive attitudes toward the message, and c) higher COVID-19 vacci- nation intentions.

Regarding disgust appeals and reactance, aside from Yang’s (2017) study finding disgust in response to human papilloma- virus (HPV) vaccination messages to be positively associated with anger and negative cognitions, no other studies have directly examined the relationship between disgust appeals and reactance. Given this scarcity of empirical findings, the following research question is offered:

RQ2: How do high vs. low disgust appeals influence a) freedom threat perceptions and b) psychological reactance?

As mentioned above, in situations where the importance of freedom is deemphasized, less or no reactance should be expected. Therefore, high controlling language that would otherwise be highly freedom-threatening might not trigger reac- tance. The ostensibly freedom-threatening nature of high con- trolling language may be “licensed” in situations where the relevant choice of freedom is washed out when and if perceived physical threat is deemed severe enough. We explore the pos- sibility of high fear and/or high disgust appeals potentially cushioning the reactance aroused by high controlling language.

Mortality Salience and Psychological Reactance

The terror management health model (TMHM; Goldenberg & Arndt, 2008) posits that, when mortality is salient and within focal awareness, individuals will engage in health-protective behaviors if they believe they can do so. Bessarabova and Massey (2020) integrated the TMHM with PRT to predict the effect of mortality salience (MS) on responses to sexually transmitted disease prevention messages featuring high vs. low freedom threat. They argued that, although reactance moti- vates defensiveness toward freedom-threatening communica- tion, people try to inhibit death thoughts, and may not feel much reactance since death is “a more pressing concern than threats to autonomy” (Bessarabova & Massey, 2020, p. 28). By combining MS with freedom-threatening communication, mes- sage receivers engaging in proximal defense will attempt to remove death thoughts from focal awareness, thereby washing out freedom threat perceptions. Thus, MS should result in reduced reactance. As hypothesized, Bessarabova and Massey (2020) found MS decreased freedom threat perceptions in mes- sages featuring a high threat to freedom during proximal defense, and concluded that in the case of death threatening events requiring immediate compliance, the dire ramifications of noncompliance will overwhelm reactance.

Similarly, when freedoms of low importance are threatened, relatively less reactance should be aroused (Erceg-Hurn & Steed, 2011). Therefore, when people are shown demanding messages in situations involving serious consequences, they should feel fewer freedom threats since freedom of choice

“I Felt Completely Turned off by the Message” 429

should be far less essential. Therefore, they should be more likely to accept high controlling language in response to high vs. low fear appeals (where their freedom to choose should be more pressing). This suggests the possibility that low control- ling language may be less appreciated when levels of fear are high, since receivers may be uncertain about what to do to protect themselves.

High controlling language may be most appreciated when high disgust appeals are used with high fear appeals because high disgust-eliciting objects indicate contamination and death (Haidt et al., 1994). Thus, the inclusion of both high disgust and fear appeals should provide more justification for the use of controlling language. Consequently, freedom threat, reactance, and defensive reactions should be attenuated in response to high controlling language when levels of fear and disgust are high, whereas these responses should be more pronounced when levels of fear and disgust are low. However, as extant research has shown, fear and disgust appeals can potentially threaten perceived freedoms resulting in reactance (Shen, 2015; Yang, 2017). Therefore, it is possible that high fear and disgust appeals using high controlling language may increase reactance. Therefore, the following question is presented:

RQ3: Do controlling language, fear, and disgust appeals interact to influence a) freedom threat perceptions, b) psycho- logical reactance, c) source derogation, d) attitudes toward the message, and e) COVID-19 vaccination intentions?

Method

A 2 (controlling language: high/low) by 2 (fear appeals: high/ low) by 2 (disgust appeals: high/low) by 2 (message variations) mixed-subjects experiment was conducted, with controlling language, fear, and disgust appeals assessed as between- subjects factors, and message variations measured as a within- subjects factor.

Participants

Amazon Mechanical Turk (MTurk) participants (N = 564) were recruited. Responses from 117 participants were discarded since they a) missed any of six attention verification questions, N = 102; e.g., “For this question, please choose ‘Somewhat agree’; and b) took too long or short a time to complete the study based on being within 3 standard deviations of the mean, N = 15), reducing the sample size to 447. Participants were middle-aged (M = 41.68, SD = 12.87), white (N = 315, 70.5%), with roughly the same proportion of females (N = 228, 51%) as males.

Study Design and Procedures

Participants were randomly assigned to one of eight conditions crossing experimental manipulations where they viewed two messages (in random order) of the same experimental features in the respective condition. After each message, participants filled out scales assessing message responses. Finally, partici- pants reported demographic information and were paid.

Stimuli

Two pilot studies were conducted to test message manipula- tions. Pilot study 1 (N = 240) tested the individual manipula- tions of controlling language, fear, and disgust appeals. High vs. low controlling language was manipulated using forceful vs. autonomy-supporting language. Fear appeals were manipulated by varying the severity of COVID-19 infection, and disgust appeals were manipulated by varying the vivid- ness of descriptions referring to at least one of the key dis- gust-eliciting domains specified by Haidt et al. (1994). Furthermore, images representing the consequences of infec- tion were also examined to reinforce fear and disgust appeals. Participants were randomly shown one of two messages in all three pairs of manipulations (i.e., high vs. low controlling language, fear appeals, and disgust appeals). After manipula- tion effectiveness was assessed, participants viewed 32 images relating to COVID-19 infection and scored each image on perceived levels of fear and disgust. Results showed all mes- sage manipulations succeeded, whereupon eight images were selected to reinforce the combination of fear and disgust at high vs. low levels (two images for each condition). Then, individual message manipulations were combined to create 32 treatment messages across eight conditions. Pilot study 2 (N = 497) then tested the 32 treatment messages, from which the 16 most effective messages (two within each con- dition) were selected for the main study (see Appendix for examples).

Measures

Freedom Threat Perceptions Freedom threat perceptions were measured with four items (e.g., “The message threatened my freedom to choose.”) on a 7-point Likert scale (1 = “Strongly disagree,” 7 = “Strongly agree”; Dillard & Shen, 2005). Cronbach’s α = .89.

Psychological Reactance Psychological reactance was assessed by anger and negative cognitions. Anger was measured with four items (e.g., how angry one felt) on a 7-point Likert scale (1 = “None of this feeling,” 7 = “A great deal of this feeling”; Dillard & Shen, 2005), Cronbach’s α = .95. Since a within-subjects component was included in this study and prior research has observed an advantage for scale measure relative to thought-listing (Quick, Kam, Morgan, Montero Liberona, & Smith, 2015; Reynolds- Tylus et al., 2021), negative cognitions were assessed with three items (Silvia, 2006; e.g., “How much did you criticize the message?”), measured on a 7-point Likert scale (1 = “Not at all,” 7 = “Very much”) Cronbach’s α = . 93.

Source Derogation Twelve items adapted from McCroskey and Teven (1999) mea- sured source derogation. Participants reported their attitudes toward the source of the message on a semantic differential scale (e.g., “The source of the message is: unintelligent/ intel- ligent”). A lower score means higher source derogation. Cronbach’s α = .98.

430 H. Ma and C. Miller

Attitudes toward the Message A seven-item semantic differential scale (Dillard & Shen, 2005) measured message attitudes. Participants indicated their atti- tudes toward the message (e.g., “The message was “desirable/ undesirable”). Cronbach’s α = .97.

Behavioral Intentions Three items adapted from Nan (2012) assessed COVID-19 vaccination intentions (e.g., “How likely would you be to get the COVID-19 vaccination once it is available to you?”) mea- sured on a 7-point Likert scale (1 = “Extremely unlikely”; 7 = “Extremely likely.”) Cronbach’s α = .97.

Covariates Included as covariates were participants’ political affiliation, age, general vaccination attitude, and perceptions of message quality.1 General vaccination attitude was assessed using Martin and Petrie’s (2017) Vaccination Attitudes Examination scale (e.g., “I feel safe after being vaccinated.”) measured on a 7-point Likert scale (1 = “Extremely unlikely” and 7 = “Extremely likely”). Cronbach’s α = .80.

Results

Repeated measures multivariate analyses of covariates were performed with message conditions entered as independent variables, message responses as dependent variables, and mes- sage variations as the within-subject factor, along with general vaccination attitudes, message quality, political affiliations, and age entered as covariates.

Hypothesis 1 and H2 examined the effects of controlling lan- guage, with multivariate tests showing a significant main effect, Wilk’s Λ = 0.86, F(6, 431) = 11.31, p < .001, η2

p = .14, for freedom threat, F(1, 436) = 61.04, p < .001, η2

p = .12, source derogation, F(1, 436) = 5.28, p = .02, η2

p = .01, and attitudes toward the message, F(1, 436) = 5.43, p = .02, η2

p = .01. Participants in the high controlling language condition reported more freedom threat (M = 4.24, SE = .09), source derogation (M = 5.21, SE = .06), and less positive attitudes (M = 4.86, SE = .08) than those in the low-controlling language condition (M = 3.21, SE = .09; M = 5.41, SE = .06; M = 5.13, SE = .08 respectively). The effect was non-significant on anger, F(1, 436) = 1.79, p = .18, negative cognitions, F(1, 436) = 2.83, p = .09, or vaccination intention, F(1, 436) = .62, p = .43. Thus, both H1 and H2 received partial supported2.

Hypothesis 3 and RQ1 addressed the impacts of fear appeals, with multivariate tests revealing a significant main effect, Wilk’s Λ = 0.95, F(6, 431) = 3.87, p = .001, η2

p = .05, on freedom threat, F(1, 436) = 9.67, p = .002, η2

p = .02, anger, F(1, 436) = 13.26, p < .001, η2

p = .03, negative cognitions, F(1, 436) = 10.52, p = .001, η2

p = .02, source derogation, F(1, 436) = 12.35, p < .001, η2

p = .03, and message attitudes, F(1, 436) = 15.38, p < .001, η2

p = .03. H3 was fully supported: Participants in the high fear appeal condition reported more freedom threat (M = 3.93, SE = .09), anger (M = 2.66, SE = .09), negative cognitions (M = 2.89, SE = .09), and source derogation (M = 5.16, SE = .06), and less positive attitudes

(M = 4.77, SE = .08), than those in the low fear appeal condi- tion (M = 3.52, SE = .09; M = 2.19, SE = .09; M = 2.50, SE = .09; M = 5.46, SE = .06; M = 5.21, SE = .08 respectively). The influence on vaccination intentions was non-significant, F (1, 436) = .001, p = .98.

Hypothesis 4 and RQ2 examined the impacts of disgust appeals, with multivariate tests revealing a significant main effect, Wilk’s Λ = 0.96, F(6, 431) = 3.35, p = .003, η2

p = .05, on source derogation, F(1, 436) = 4.69, p = .03, η2

p = .01, message attitudes, F(1, 436) = 12.05, p = .001, η2

p = .03, freedom threat, F(1, 436) = 11.03, p = .001, η2

p = .03, and anger, F(1, 436) = 4.36, p = .037, η2

p = .01. H4 was not supported since means were in the opposite direction predicted: Participants in the high disgust appeal condition indicated more source dero- gation (M = 5.21, SE = .06), freedom threat (M = 3.94, SE = .09), and anger (M = 2.56, SE = .09), and less positive attitudes (M = 4.80, SE = .08), than those in the low disgust appeal condition (M = 5.41, SE = .06; M = 5.19, SE = .08; M = 3.51, SE = .09; M = 2.29, SE = .09 respectively). The influence was non-significant for COVID-19 vaccination inten- tions, F(1, 436) = 1.60, p = .21, and negative cognitions, F(1, 436) = 0.71, p = .40.

RQ3 explored the interaction between controlling language, fear, and disgust appeals. Multivariate tests showed no signifi- cant three-way interaction, Wilk’s Λ = 0.99, F(6, 431) = 0.74, p = .62. A further examination splitting controlling language into high vs. low conditions, revealed that with low controlling language, the interaction was significant, Wilk’s Λ = 0.93, F(6, 207) = 2.82, p = .012, on source derogation, F(1, 212) = 8.39, p = .004, η2

p = .04, and message attitudes, F(1, 212) = 6.92, p = .009, η2

p = .03 (Figure 1 and 2), however with high controlling language, the 2-way interaction was non- significant, Wilk’s Λ = 0.98, F(6, 216) = 2.82, p = .70. Participants in the low disgust-low fear appeal-low controlling language condition reported the lowest source derogation and the most positive attitudes relative to all other conditions.

Discussion

This study examined the effects of controlling language, fear, and disgust appeals on COVID-19 vaccination messages responses. The findings show, high vs. low levels of controlling language, fear appeals, and disgust appeals produced more freedom threat perceptions, reactance, and source derogation, and less positive attitudes toward the message. Moreover, when low controlling language was used, low fear and disgust appeals generated the least source derogation and the most positive message attitudes.

As seen in previous research (e.g., Clayton et al., 2020), relative to low controlling language, high controlling language appears to produce more freedom threat perceptions, more source derogation, and less positive attitudes toward the mes- sage. There was no significant difference in reactance though freedom threat was associated with anger (r = .61) and negative cognitions (r = .61). These findings reinforce the view that high controlling language should be avoided when promoting

“I Felt Completely Turned off by the Message” 431

COVID-19 vaccination. In support of previous research exam- ining the effects of fear appeals on reactance (Quick et al., 2018), this study found high vs. low fear appeals generate more freedom threat, greater reactance, more source derogation, and less positive attitudes toward the message.

High vs. low disgust appeals were associated with more freedom threat perceptions, anger, and source derogation, and less positive attitudes toward the message. This finding is inconsistent with some extant research. For instance, Halkjelsvik and Rise (2015) found high vs. modest disgust content did not influence smoking intentions and beliefs. Unfortunately, studies examining disgust appeals across con- texts are both scarce and inconsistent; thus, more studies are needed to explore the effects of disgust appeals on message responses. The lack of a significant difference in negative cognitions between high vs. low disgust appeals accords with research showing that disgust is more of a gut-level

response, featuring immediate rejection of sensory informa- tion (Rozin & Fallon, 1987). Therefore, the gut-level proces- sing of disgust appeals does not necessarily lead to a difference in negative cognitions, which is what we found.

Meta-analyses have shown how high fear appeals may increase persuasion (Tannenbaum et al., 2015; Witte & Allen, 2000), however, in line with recent studies by Peng et al. (2020) and Quick et al. (2018), the findings reported here indicate messages with high fear and high disgust appraisals tend to be reactance-eliciting, suggesting the positive effects fear and dis- gust appeals have on attitude change will be diminished or extinguished to the extent such appeals also threaten perceived freedoms, induce reactance, and increase source derogation. More research is needed to explore the conditions under which fear and disgust appeals lead to reactance.

Though some extant research suggests a potential positive effect of controlling language associated with fear appeals in

Figure 1. Fear and disgust appeal interaction on source derogation in the low controlling language condition (higher scores indicate less source derogation).

Figure 2. Fear and disgust appeal interaction on message attitudes in the low controlling language condition (higher scores indicate more favorable attitudes).

432 H. Ma and C. Miller

increasing persuasion without arousing reactance under certain circumstances (Brehm, 1966; Buller et al., 2000; Hargrove, 2003), as mentioned, it should be noted that they did not directly measure reactance and its antecedent—freedom threat, since a valid measure of reactance and perceptions of freedom threat was not established until Dillard and Shen’s (2005) work.

Nonetheless, despite significant differences in reactance, source derogation, and attitudes toward the message, these out- comes did not appear to influence COVID-19 vaccination inten- tions. Although no conclusions may be drawn from null results, this non-significant finding is consistent with previous research (e.g., Clayton et al., 2020), and the attitude-intention inconsis- tency following a brief message exposure aligns with the notion that decision-making incorporates various motivation-based factors moderating the influence of message content on inten- tions (Crano & Prislin, 1995). For example, misinformation and conspiracy theories about the pandemic may moderate the impact of attitudes on vaccination intentions (Stein et al., 2021).

There were significant interactions between fear and disgust appeals in the low controlling language condition. When both fear and disgust appeals were low, the least source derogation and the most positive attitudes were reported. This finding suggests that when high controlling language is used, levels of fear and disgust do not matter—persuasion is going to be diminished. Whereas, when low controlling language is used, the use of low fear and disgust eliciting content is likely to prompt the most positive message responses.

High fear and disgust appeals were ineffective in mitigating the effects of high controlling language, which appears to be inconsistent with Bessarabova and Massey’s (2020) findings. Their results likely stemmed from the unique effects of MS on death thought awareness that does not appear to have been elicited by the high fear and disgust appeal stimuli used in the present study. Though the high fear and high disgust appeals used emphasized the life-threatening consequences of infection, they likely did not make death thoughts salient, particularly if message recipients were not inclined to feel susceptible to COVID-19 infection. If the high fear and high disgust stimuli used failed to elicit high enough level of threat to make MS, the messages used were likely insufficient to justify the use of high controlling language. Bessarabova and Massey (2020) manipu- lated MS through a separate manipulation prior to presenting their high controlling language messages rather than by the content of those messages. Further research is needed to exam- ine the nature of message stimuli necessary for inducing MS in the absence of a separate MS manipulation. We believe MS may be elicited by fear and disgust stimuli that are high enough to make relevant threats both plausible and salient.

Limitations and Implications

This research has several limitations. Aside from the high fear and disgust conditions being insufficient to make actual physi- cal threats salient, the sample primarily consisted of low- to middle-income, middle-aged whites; thus, the findings may not be generalizable to other groups, such as certain racial mino- rities and the elderly (Sandoiu, 2020). Another limitation con- cerns the length of the treatment messages. To more fully

present the consequences of COVID-19 infection, the messages used in this study were relatively long, thus they may not be ideal for real-world campaigns advocating vaccination. Perhaps shorter, more intense messages would be more effective. We recommend that future research designs include MS measures3

to gauge the level of death thought necessary for the use of high controlling language within high threat, health risk prevention messages.

Nevertheless, the findings offer important implications from a reactance perspective and insights into message designs. The study provides empirical support for recognizing the potentially reactance-eliciting features of fear and disgust appeals. For public health messaging promoting COVID-19 vaccination, these findings suggest avoiding high fear and disgust inducing descriptions of COVID-19 infection, particularly in conjunction with forceful language. Finally, although this study focused on COVID-19 vaccination, these findings are likely to apply to other kinds of vaccinations more generally as well.

Disclosure Statement

No potential conflict of interest was reported by the author(s).

Funding

The author(s) reported there is no funding associated with the work featured in this article.

ORCID Haijing Ma http://orcid.org/0000-0002-6154-4090 Claude Miller http://orcid.org/0000-0002-8514-936X

References Allred, A. T., & Amos, C. (2018). Disgust images and nonprofit children’s

causes. Journal of Social Marketing, 8(1), 120–140. doi:10.1108/ JSOCM-01-2017-0003

Bahari, A. (2019). Psychological reactance management via nonlinear dynamic motivation in classroom and telecollaborative second lan- guage learning contexts. Journal of Educational, Health and Community Psychology, 8(1), 40–58. doi:10.12928/jehcp.v8i1.10657

Berkeley, L.,sJr. (2021). Biden’s next fight: Anti-vaxxers jeopardize plans to protect U.S. against Covid. CNBC. https://www.cnbc.com/2021/02/ 10/biden-covid-vaccine-anti-vaxxers-us.html

Bessarabova, E., & Massey, Z. B. (2020). Testing terror management health model and integrating its predictions with the theory of psycho- logical reactance. Communication Monographs, 87(1), 25–46. doi:10. 1080/03637751.2019.1626992

Braun-Latour, K. A., & Zaltman, G. (2006). Memory change: An intimate measure of persuasion. Journal of Advertising Research, 46(1), 57–72. doi:10.2501/S0021849906060077

Brehm, J. W. (1966). A theory of psychological reactance. New York: Academic Press.

Brehm, S. S., & Brehm, J. W. (1981). Psychological reactance: A theory of freedom and control. New York: Academic Press.

Buller, D. B., Burgoon, M., Hall, J. R., Levine, N., Taylor, A. M., Beach, B., & Melcher, C. (2000). Long-term effects of language inten- sity in preventive messages on planned family solar protection. Health Communication, 12(3), 261–275. doi:10.1207/S15327027HC1203_03

Clayton, R. B., Lang, A., Leshner, G., & Quick, B. L. (2019). Who fights, who flees? An integration of the LC4MP and psychological reactance

“I Felt Completely Turned off by the Message” 433

theory. Media Psychology, 22(4), 545–571. doi:10.1080/15213269. 2018.1476157

Clayton, R. B., Leshner, G., Sanders-Jackson, A., & Hendrickse, J. (2020). When counterarguing becomes the primary task: Examination of dog- matic anti-vaping messages on psychological reactance, available cog- nitive resources, and memory. Journal of Communication, jqaa010. doi:10.1093/joc/jqaa010

Cohen, E. (2020). Head of operation warp speed says he expects corona- virus vaccine to be highly effective, “in the 90%” range. CNN. https:// www.cnn.com/world/live-news/coronavirus-pandemic-07-30-20-intl /h_5f70b27a0d4dd245b1317ea2aafcd198

Crano, W. D., & Prislin, R. (1995). Components of vested interest and attitude-behavior consistency. Basic and Applied Social Psychology, 17 (1 & 2), 1–21. doi:10.1080/01973533.1995.9646129

Deci, E. L., & Ryan, R. M. (1987). The support of autonomy and the control of behavior. Journal of Personality and Social Psychology, 53, 1024–1037. doi:10.1037/0022-3514.53.6.1024

Dillard, J. P. (1994). Rethinking the study of fear appeals: An emotional perspective. Communication Theory, 4(4), 295–323. doi:10.1111/j. 1468-2885.1994.tb00094.x

Dillard, J. P., Kim, J., & Li, S. S. (2018). Anti-sugar-sweetened beverage messages elicit reactance: Effects on attitudes and policy preferences. Journal of Health Communication, 23(8), 703–711. doi:10.1080/ 10810730.2018.1511012

Dillard, J. P., & Shen, L. (2005). On the nature of reactance and its role in persuasive health communication. Communication Monographs, 72(2), 144–168. doi:10.1080/03637750500111815

Doucleff, M. (2021). Why scientists are very worried about the variant from Brazil. NPR. https://www.npr.org/sections/goatsandsoda/2021/01/ 27/961108577/why-scientists-are-very-worried-about-the-variant-from- brazil

Erceg-Hurn, D. M., & Steed, L. G. (2011). Does exposure to cigarette health warnings elicit psychological reactance in smokers? Journal of Applied Social Psychology, 41(1), 219–237. doi:10.1111/j.1559-1816. 2010.00710.x

Goldenberg, J. L., & Arndt, J. (2008). The implications of death for health: A terror management health model for behavioral health promotion. Psychological Review, 115(4), 1032–1053. doi:10.1037/a0013326

Grandpre, J., Alvaro, E. M., Burgoon, M., Miller, C. H., & Hall, J. R. (2003). Adolescent reactance and anti-smoking campaigns: A theoretical approach. Health Communication, 15(3), 349–366. doi:10.1207/S15327027HC1503_6

Haidt, J., McCauley, C., & Rozin, P. (1994). Individual differences in sensitivity to disgust: A scale sampling seven domains of disgust elicitors. Personality and Individual Differences, 16(5), 701–713. doi:10.1016/0191-8869(94)90212-7

Halkjelsvik, T., & Rise, J. (2015). Disgust in fear appeal anti-smoking advertisements: The effects on attitudes and abstinence motivation. Drugs: Education, Prevention and Policy, 22(4), 362–369. doi:10. 3109/09687637.2015.1015491

Hargrove, L. M. (2003). Psychological reactance as a personality char- acteristic: Relationships to attachment and autonomy. [Doctoral Dzissertation]. Louisiana Tech University.

Johns Hopkins Coronavirus Resource Center (2021). COVID-19 vaccina- tion: What you need to know. https://www.hopkinsmedicine.org/health/ conditions-and-diseases/coronavirus/covid-19-vaccine-what-you-need- to-know

Kim, H. J., & Shin, W. (2018). The effects of message source and fear appeal on young adults’ response to Sexually Transmitted Disease (STD) messages in Singapore. Asian Journal of Communication, 28 (2), 185–204. doi:10.1080/01292986.2017.1384030

LaVoie, N. R., Quick, B. L., Riles, J. M., & Lambert, N. J. (2017). Are graphic cigarette warning labels an effective message strategy? A test of psychological reactance theory and source appraisal. Communication Research, 44(3), 416–436. doi:10.1177/0093650215609669

Leshner, G., Bolls, P., & Thomas, E. (2009). Scare’em or disgust’em: The effects of graphic health promotion messages. Health Communication, 24(5), 447–458. doi:10.1080/10410230903023493

Leshner, G., Bolls, P., & Wise, K. (2011). Motivated processing of fear appeal and disgust images in televised anti-tobacco ads. Journal of Media Psychology, 23, 77–89. doi:10.1027/1864-1105/a000037

Ma, H., & Miller, C. H. (2021). The effects of agency assignment and reference point on responses to COVID-19 messages. Health Communication, 36(1), 59–73. doi:10.1080/10410236.2020.1848066

McCroskey, J. C., & Teven, J. J. (1999). Goodwill: A reexamination of the construct and its measurement. Communication Monographs, 66(1), 90–103. doi:10.1080/03637759909376464

Miller, C. H. (2015). Persuasion and psychological reactance: The effects of explicit, high-controlling language. In R. Shulze & H. Pishwa (Eds.), The exercise of power in communication: Devices, reception and reac- tion (pp. 269–286). London, UK: Palgrave McMillan.

Miller, C. H., Burgoon, M., Alvaro, E. M., Hall, J. R., Grandpre, J., Broneck, K., & Frank, C. (2001). Creating intrinsically motivating health messages with interactive multimedia: Targeting and testing adolescent populations. The Electronic Journal of Communication, 11.

Miller, C. H., Lane, L. T., Deatrick, L. M., Young, A. M., & Potts, K. A. (2007). Psychological reactance and promotional health messages: The effects of controlling language, lexical concreteness, and the restoration of freedom. Human Communication Research, 33(2), 219–240. doi:10. 1111/j.1468-2958.2007.00297.x

Miller, C. H., Massey, Z. B., & Ma, H. (2020). Psychological reactance and persuasive message design. In H. D. O’Hair & M. J. O’Hair (Eds.), The handbook of applied communication research (pp. 457–483). New York: John Wiley & Sons, Inc.

Morales, A. C., Wu, E. C., & Fitzsimons, G. J. (2012). How disgust enhances the effectiveness of fear appeals. Journal of Marketing Research, 49(3), 383–393. doi:10.1509/jmr.07.0364

Nabi, R. L. (2002). Discrete emotions and persuasion. In J. P. Dillard & M. Pfau (Eds.), The persuasion handbook: Developments in theory and practice (pp. 289–308). Thousand Oaks: SAGE.

Nan, X. (2012). Communicating to young adults about HPV vaccination: Consideration of message framing, motivation, and gender. Health Communication, 27(1), 10–18. doi:10.1080/10410236.2011.567447

Peng, L., Shen, L., Vanderbilt, R. R., Kim, Y., & Foley, K. A. (2020). The impact of fear versus state empathy on persuasion and social stigma. Media Psychology, 23(1), 1–24. doi:10.1080/15213269.2018.1535321

Quick, B. L., Kam, J. A., Morgan, S. E., Montero Liberona, C. A., & Smith, R. A. (2015). Prospect theory, discrete emotions, and freedom threats: An extension of psychological reactance theory. Journal of Communication, 65(1), 40–61. doi:10.1111/jcom.12134

Quick, B. L., LaVoie, N. R., Reynolds-Tylus, T., Martinez-Gonzalez, A., & Skurka, C. (2018). Examining mechanisms underlying fear-control in the extended parallel process model. Health Communication, 33(4), 379–391. doi:10.1080/10410236.2016.1266738

Quick, B. L., Shen, L., & Dillard, J. P. (2013). Reactance theory and persuasion. In J. P. Dillard & L. Shen (Eds.), The SAGE handbook of persuasion: Developments in theory and practice (pp. 167–183). Thousand Oaks: SAGE.

Quick, B. L., & Stephenson, M. T. (2008). Examining the role of trait reactance and sensation seeking on perceived threat, state reactance, and reactance restoration. Human Communication Research, 34(3), 448–476. doi:10.1111/j.1468-2958.2008.00328.x

Ratcliff, C. L. (2021). Characterizing reactance in communication research: A review of conceptual and operational approaches. Communication Research, 48(7), 1033–1058. doi:10.1177/0093650219872126

Reynolds-Tylus, T. (2019). Psychological reactance and persuasive health communication: A review of the literature. Frontiers in Communication, 4(4), 56. doi:10.3389/fcomm.2019.00056

Reynolds-Tylus, T., Bigsby, E., & Quick, B. L. (2021). A comparison of three approaches for measuring negative cognitions for psychological

434 H. Ma and C. Miller

reactance. Communication Methods and Measures, 15(1), 43–59. doi:10.1080/19312458.2020.1810647

Ribeiro, M. J., & Castelo-Branco, M. (2019). Neural correlates of antici- patory cardiac deceleration and its association with the speed of per- ceptual decision-making, in young and older adults. NeuroImage, 199, 521–533. doi:10.1016/j.neuroimage.2019.06.004

Richards, A. S., Bessarabova, E., Banas, J. A., & Bernard, D. R. (2020). Reducing psychological reactance to health promotion messages: Comparing preemptive and postscript mitigation strategies. Health Communication, 1–9. doi:10.1080/10410236.2020.1839203

Rozin, P., & Fallon, A. E. (1987). A perspective on disgust. Psychological Review, 94(1), 23–41. doi:10.1037/0033-295X.94.1.23

Russell, J. A., Lewicka, M., & Niit, T. (1989). A cross-cultural study of a circumplex model of affect. Journal of Personality and Social Psychology, 57(5), 848–856. doi:10.1037/0022-3514.57.5.848

Sandoiu, A. (2020). The impact of the COVID-19 pandemic on older adults. Medical News Today. https://www.medicalnewstoday.com/arti- cles/the-impact-of-the-covid-19-pandemic-on-older-adults#Old-age- and-preexisting-health-conditions

Shen, L. (2015). Antecedents to psychological reactance: The impact of threat, message frame, and choice. Health Communication, 30(10), 975–985. doi:10.1080/10410236.2014.910882

Silvia, P. J. (2006). Reactance and the dynamics of disagreement: Multiple paths from threatened freedom to resistance to persuasion. European Journal of Social Psychology, 36(5), 673–685. doi:10.1002/ejsp.309

Stein, R. A., Ometa, O., Pachtman Shetty, S., Katz, A., Popitiu, M. I., & Brotherton, R. (2021). Conspiracy theories in the era of COVID-19: A tale of two pandemics. International Journal of Clinic Practice, 75 (2), e13778. doi:10.1111/ijcp.13778

Susskind, J. M., Lee, D. H., Cusi, A., Feiman, R., Grabski, W., & Anderson, A. K. (2008). Expressing fear enhances sensory acquisition. Nature Neuroscience, 11(7), 843–850. doi:10.1038/nn.2138

Tannenbaum, M. B., Hepler, J., Zimmerman, R. S., Saul, L., Jacobs, S., Wilson, K., & Albarracín, D. (2015). Appealing to fear: A meta-analysis of fear appeal effectiveness and theories. Psychological Bulletin, 141(6), 1178–1204. doi:10.1037/a0039729

Van Hooff, J. C., Devue, C., Vieweg, P. E., & Theeuwes, J. (2013). Disgust-and not fear-evoking images hold our attention. Acta Psychologica, 143(1), 1–6. doi:10.1016/j.actpsy.2013.02.001

Walster, E., & Festinger, L. (1962). The effectiveness of “overheard” persuasive communications. The Journal of Abnormal and Social Psychology, 65(6), 395–402. doi:10.1037/h0041172

Wechsler, H., Nelson, T. E., Lee, J. E., Seibring, M., Lewis, C., & Keeling, R. P. (2003). Perception and reality: A national evaluation of social norms mar- keting interventions to reduce college students’ heavy alcohol use. Journal of Studies on Alcohol, 64(4), 484–494. doi:10.15288/jsa.2003.64.484

Witte, K. (1992). Putting the fear back into fear appeals: The extended parallel process model. Communications Monographs, 59(4), 329–349. doi:10.1080/03637759209376276

Witte, K., & Allen, M. (2000). A meta-analysis of fear appeals: Implications for effective public health campaigns. Health Education & Behavior, 27(5), 591–615. doi:10.1177/109019810002700506

Woody, S. R., & Teachman, B. A. (2000). Intersection of disgust and fear: Normative and pathological views. Clinical Psychology: Science and Practice, 7(3), 291–311. doi:10.1093/clipsy.7.3.291

Yang, C. (2017). The impact of disgust on threat appeals: Enhancement or attenuation of persuasion? [Doctoral dissertation]. Pennsylvania State University.

“I Felt Completely Turned off by the Message” 435

Appendix Sample Messages High controlling language, high fear, high disgust

436 H. Ma and C. Miller

Low controlling language, low fear, low disgust

“I Felt Completely Turned off by the Message” 437

high fear/low disgust message low fear/high disgust message

● Many people infected by COVID-19 (which is caused by SARS-CoV-2, i.e., the coronavirus) have been examined by physicians who have found their lungs to look abnormal and unhealthy. For the most part, their lungs are relatively inflamed, irregular, nonuniform, and atypical in form, with an overall discolored appearance throughout.

● COVID-19 has brought more than 28,000,000 confirmed cases and 500,000 deaths across the U.S.; the death count continues to rise. When infected, people may experience severe illness and complications from the virus, such as heart disease requiring hospitalization, including admission to the ICU, intubation, and mechanical ventilation to prevent suffocation and death.

● Upon closer examination, there are many small clots in the lungs, and in the surrounding veins and arteries, as well as in the smaller blood vessels of COVID-19 patients. Also, in some cases, these clots have appeared in the brain scans of patients with advanced disease.

● With no immunity and without vaccination, there is a high chance of exposure, and it is highly likely that you will get infected! There have been reports of a greater-than-expected number of young patients being hos- pitalized for, and sometimes dying from, serious strokes. So, even if you are young, when you are infected with the virus, this could happen to you!

● Some physicians have observed that the lungs of people who’ve recovered from COVID-19 look very similar to those of heavy tobacco smokers.

● If you smoke, or if you have certain medical issues, such as being overweight or having high blood pressure, you are at extremely high risk for severe illness and death from the virus!

● Forensic autopsies performed on deceased victims of COVID-19 (which is caused by SARS-CoV-2, i.e., the coronavirus), show their lungs cut open revealing what appears to be clotted, scarred tissues that are enlarged, and completely firm, with a viscid, gummy substance plastered throughout. Much of the lung tissue shows festering inflammation with clotted blood, and sticky, oozing pale yellow glutinous fluid, white pus, and red-brown matter.

● COVID-19 has reached many confirmed cases, with hundreds of thousands of people getting sick and many hospitalized across the U.S. If infected, people may experience mild to moderate symptoms, such as fever, fati- gue, body aches, loss of appetite, dry cough, and in many cases, shortness of breath.

● Closer forensic examination reveals hundreds of thousands of microclots in the lungs of victims who have been mortally contaminated with the virus. The same vile microcloting also appears to spread throughout the surrounding veins and arteries, and even in the brains of some of those who have died from COVID-19.

● With no immunity and without vaccination, there is a real chance of exposure, and it is likely that you may get infected. There have been reports of young people being at risk for hospitalization, and sometimes experi- encing other medical complications. So, even if you are young, if you were to become infected with the virus, this could conceivably happen to you.

● Post-COVID-19 lungs in particular look even more revolting and disgusting than cancerous smokers’ lungs.

● If you smoke, or if you have certain medical issues, such as obesity or high blood pressure, you may be at elevated risk for illness and hospitalization from the virus.

438 H. Ma and C. Miller