Reactance Theory and Persuasion

CHAPTER 11

Brian L. Quick, Lijiang Shen, and James Price Dillard

ir Isaac Newton famously claimed that for by first considering the theory’s original formu- every action there is an equal and opposite lation alongside the empirical research done to Sreaction. It is doubtful that he had persua- test it. The second section focuses on refinement sion in mind when formulating his third law of of the theory, before turning, in the third move- motion, but, the principle alerts us to the parallel ment, to contemporary research. Following from possibility that suasory efforts may too instigate that, directions are offered for future inquiry. reactions that run counter to the intended action of the message. Quite apart from the laws of motion, the literature suggests that opposing Overview of Reactance reactions may vary greatly in degree. Some Theory and Classic Research appeals are so effective that the counter-reaction is scarcely measurable. Others may produce no There are four components to reactance theory: discernable persuasive change, perhaps because Freedom, threat to freedom, reactance, and resto- the action and reaction are closely balanced. In ration of freedom. Freedoms are beliefs about the still other cases, the predominant response is ways in which one can behave. Brehm and Brehm counterpersuasion, a condition that has come to (1981) argue that freedoms are “not ‘abstract be known as a boomerang effect. considerations,’ but concrete behavioral realities” All of these outcomes are of interest to per- (p. 12). Nonetheless, freedom is defined broadly suasion researchers, but boomerangs and failure to include actions as well as emotions and atti- to persuade are both common and pragmatically tudes (Brehm, 1966). In other words, freedom to frustrating. The perspective that is most fre- do, freedom to feel, or freedom to hold a particu- quently called on to give account of them is psy- lar evaluation, or not to. Individuals possess chological reactance theory (PRT; Brehm, 1966). freedoms only to the extent that they have Consequently, this chapter reviews and synthe- knowledge of them and perceive that they are sizes work on this landmark theory. It proceeds capable of enacting the behavior.

167

Given that an individual perceives a specific Corollaries of Reactance Theory freedom, anything that makes it more difficult for him or her to exercise that freedom consti- PRT is built around four principles (Brehm, tutes a threat (Brehm, 1966; Brehm & Brehm, 1966). The first specifies that reactance can only 1981). Even an impersonal event, such as the be aroused if individuals believe they have free- weather can be viewed as a threat, if it renders the dom over a particular outcome. An early study exercise of a freedom more difficult. However, illustrating PRT’s first principle was conducted by social influence as a threat is most pertinent to Wicklund and Brehm (1968), in which partici- questions of persuasive communication. Impor- pants were asked to evaluate a job applicant. Their tant to understanding the theory is recognition results revealed differences among high competent that the term threat is used narrowly to mean individuals following exposure to the freedom- threat-to-freedom, and not threats as used in threatening condition. That is, competent indi- everyday language, nor threats to well-being used viduals exposed to the freedom-threatening in research on fear/threat appeals, or threats as a communication resulted in greater reactance than forewarning of social influence in research on exposure to the non-freedom-threatening condi- inoculation. tion. No significant difference emerged for non- Psychological reactance is “the motivational competent individuals (Wicklund & Brehm, state that is hypothesized to occur when a free- 1968). In other words, for individuals aware of dom is eliminated or threatened with elimina- their abilities, exposure to freedom-threatening tion” (Brehm & Brehm, 1981, p. 37). This is the messages resulted in heightened reactance. key mediator and central explanatory mecha- Principle two pertains to the relationship nism of the theory. Its proposed existence follows between the importance of a freedom and the from the assumption that humans place a high magnitude of reactance. Specifically, PRT asserts value on choice, autonomy, and control. As that as the threatened behavior increases in reflected in many worn phrases, people seek to be attractiveness, so does the amount of reactance the captains of their own destinies, the masters of aroused. Brehm and Weinraub (1977) recruited their own fates. toddlers to participate in an experiment in which The fourth component of the theory is resto- 2-year-olds were presented with two toys. How- ration. PRT contends that when a perceived free- ever, one of these toys was placed behind a trans- dom is eliminated or threatened with elimination, parent barrier, whereas the other was next to the the individual will be motivated to reestablish barrier. As expected, when the toys were different that freedom. Direct restoration of the freedom in shape, as opposed to similar in shape, consis- involves doing the forbidden act. In addition, tent with the second principle, preference for the freedoms may be restored indirectly by increas- obstructed toy was heightened among boys. ing liking for the threatened choice (Brehm, However, this hypothesized finding was not true Stires, Sensenig, & Shaban, 1966; Hammock & for the girls. As will be shown in the subsequent Brehm, 1966), derogating the source of threat pages, more recent work has shown a relation- (Kohn & Barnes, 1977; Schwarz, Frey, & Kumpf, ship between the importance of the threatened 1980; Smith, 1977; Worchel, 1974), denying the freedom and reactance arousal. existence of the threat (Worchel & Andreoli, The third principle states that reactance 1974; Worchel, Andreoli, & Archer, 1976), or by increases as the number of freedom threats exercising a different freedom to gain feeling of increases. Although conceptually cogent, extant control and choice (Wicklund, 1974). (See Quick PRT research is inconsistent in supporting this & Stephenson, 2007b for a measure of various principle (for a review, see Brehm & Brehm, means for restoring reactance.) 1981). One such study finding support for threat

(c) 2013 Sage Publications, Inc. All Rights Reserved. Chapter 11. Reactance Theory and Persuasion——169 intensity on reactance was conducted by Heller, collision between a freedom and a threat to Pallak, and Picek (1973). In their project, in line freedom. Nearer the end of the process, reactance with PRT, as persuasive intent and freedom is the state that is responsible for efforts to restore threats increased, participants’ attitude change in freedom. In either case, it has been defined in the opposite direction spiked. These researchers terms of either what causes or what effect it has, reasoned that perceiving a message as a freedom but not what it is. Apart from a brief mention of threat results in maladaptive responses such as the possibility that individuals “may be aware of unfavorable attitudes toward the advocacy. hostile and aggressive feelings” (Brehm, 1966, The fourth principle maintains that as implied p. 9), if the level of reactance arousal is high, the threats are present, the magnitude of reactance nature of reactance itself is rarely addressed. will increase (Brehm, 1966; Brehm & Brehm, Brehm’s apparent reluctance to explicate the 1981). Implied threats are often communicated in principle mechanism of the theory may have fol- the form of a persuasive message. An earlier study lowed his estimate of its potential for measure- conducted by Reich and Robertson (1979) exam- ment. According to Brehm and Brehm (1981), ined littering behaviors following exposure to “reactance has the status of an intervening, hypo- variations in freedom-threatening language. Spe- thetical variable. . . . We cannot [emphasis added] cifically, in their three-study experiment, they measure reactance directly, but hypothesizing its found individuals presented with messages fea- existence allows us to predict a variety of behav- turing explicit commands such as “Don’t you dare ioral effects” (p. 37). litter” resulted in greater littering than exposure The position that reactance cannot be mea- to messages featuring appeals to social norms. As sured is logically consequential. In effect, it shown earlier, their study provides support for assigns reactance to a black box. Message features the fourth principle of PRT in that as implied go in the box and outcomes emerge, but the threats increase, reactance arousal increases. operations within the box itself remain mysteri- However, it should be noted that the previously ous and unknowable. When results match with mentioned research in support of these four prin- theoretical predictions, then it is assumed that ciples was conducted without measuring reactance. the process-in-a-box must have occurred. But, That is, these researchers merely hypothesized the when results do not work out as expected, then presence of reactance given the outcomes believed the failure must be with the messages, the cir- to be linked to this aversive state. As will be shown cumstances, the sample, or something other than later in this chapter, following a validated opera- the box. Indeed, it has been common to invoke tionalization of psychological reactance (Dillard reactance as the explanation for a failed persua- & Shen, 2005), more recent work has begun and sion attempt. In this regard, it became the default continues to test the four theoretical principles assumption for any persuasive effort that pro- advanced by Brehm (1966). duced a boomerang or a no-effect finding. It is better to have a black-box theory than no theory at all. But, it is vastly preferable to understand Progress in Refining and measure the workings inside the box. Only Reactance Theory then can the entire theoretical process be open to empirical test. Hence, it was important to under- The Nature of Reactance stand the nature of reactance in order that it might be assessed directly. Conceptually speaking, reactance was defined Since the time of Brehm’s writings, scholars primarily in terms of its antecedents and out- have applied ideas drawn from reactance theory comes. For example, reactance is the result of the to a number of different domains. In the course

(c) 2013 Sage Publications, Inc. All Rights Reserved. 170——PART II. Theories, Perspectives, and Traditions of analyzing and extending the theory, reactance service announcements (Dillard & Peck, 2000, has, implicitly and explicitly, been defined in 2001; Dillard, Plotnick, Godbold, Freimuth, & several different ways. Dillard and Shen (2005) Edgar, 1996; Stephenson, 2003; Witte, 1994). describe four views. In the first, reactance is A final possibility also suggests that reactance purely cognitive. Research in the cognitive has both cognitive and affective components. response tradition adopted this perspective (e.g., However, unlike the previous position, which Petty & Cacioppo, 1986), as did work on the specified distinct effects, in this fourth perspec- clinical manifestations of reactance (Kelly & tive, cognition and affect are intertwined. In fact, Nauta, 1997). One advantage to conceiving of they are intertwined to such a degree that their reactance in this way is that it immediately effects on persuasion cannot be disentangled. becomes measurable through a variety of self- This view conceives of reactance as an alloy of its report techniques. The means that is most obvi- components, rather than a simple sum of distinct ously relevant to questions of persuasion is the elements (as is implied by the previous position). widely used thought-listing technique (Petty & Dillard and Shen (2005) report on a research Cacioppo, 1986). This purely cognitive view sug- project designed to test each of the four alterna- gests that reactance can be conceived of and tives. The study itself utilized two messages: One operationalized as counter-arguing. that promoted the use of dental floss and the Second, citing similarities between anteced- other that encouraged university students to ents of reactance and cognitive appraisals that curb their alcohol intake. For both messages, the lead to anger, some writers suggest that reactance data clearly favored the intertwined model over might be considered, in whole or in part, as an the alternatives. Hence, the authors concluded emotion (Dillard & Meijnders, 2002; Nabi, 2002). that reactance was in fact measureable and that it This claim aligned well with Brehm’s description could be modeled as a combination of anger and of reactance as the experience of hostile and negative cognitions. aggressive feelings (Seltzer, 1983; White & Several studies provided additional support Zimbardo, 1980; Wicklund, 1974). In this view for this conclusion. For example, Rains and then, reactance is more or less synonymous with Turner (2007) conducted an investigation that the family of concepts that index varying degrees pitted the intertwined model against the dual- of anger (e.g., irritation, annoyance, and rage). process model, as well as a fifth alternative in From this perspective, reactance might be opera- which anger preceded negative cognition. Across tionalized in various ways, including asking indi- their two health-related studies, the intertwined viduals to make a judgment on a close-ended model showed the best fit to the data. In a scale regarding the degree to which they are research project that examined exercise behav- experiencing anger. iors, Quick and Considine (2008) tested and A third logical option holds that reactance found support for the intertwined model in a might be considered as both affect and cogni- sample comprised of members of a health and tion. Though unrelated to reactance per se, one fitness center. Additional, complementary find- example of this type of thinking can be seen in ings can be seen in Quick and Stephenson Leventhal’s (1970) parallel processing model. He (2007a) and Shen (2010a). Given the many varia- posits that individuals have both cognitive and tions across studies in terms of messages and emotional reactions to persuasive health mes- samples, and the apparent consistency of find- sages and that those reactions have unique ings, the intertwined model seems to provide a effects on message acceptance. Evidence consis- workable representation of reactance itself and, tent with both points can be found in studies of thus, an operational means of moving away from cognitive and emotional responses to public black-box theorizing.

Measuring Reactance that the association between the two approaches is low: The four-item measures correlate only The method (discussed earlier) for measuring .23 and .22 with anger and negative cognitions the components of reactance involves two proce- respectively . Most importantly, he recommends dures (Dillard & Shen, 2005). One uses Likert the Dillard and Shen (2005) measure as it distin- scales to assess anger. For example, respondents guishes reactance and freedom threat as distinct were asked: To what extent did the message that entities, whereas Lindsay’s (2005) measure com- you just read make you feel . . . angry/irritated/ bined these concepts. In order to be consistent annoyed/aggravated: 0 = none of this feeling, 4 = a with PRT (Brehm, 1966), a freedom threat must great deal of this feeling. The complementary pro- precede reactance. For this reason, combining a cedure asks respondents to list any thoughts that freedom threat with reactance does not permit they had while reading the message. The results researchers the opportunity to examine the rela- were coded by research assistants in a four-step tionship between them, which is discussed in process that unitized the data, then screened out greater detail later in this chapter. His recom- self-reports of emotion as well as cognitions that mendation, while reasoned, leaves open the ques- were unrelated to the message or topic. tion of how to more efficiently assess reactance. Although not the primary focus of her study, One option explored by Rains and Turner Lindsey (2005) developed a measure of reactance (2007) and by Quick and Stephenson (2008), to bone marrow donation messages using a four- involves participant coding of their own cognitive item scale informed by Hong’s (Hong & Faedda, responses. This generally precludes fine-grained 1996) Psychological Reactance Scale. The items analyses that unitize, then screen out emotions were: I am uncomfortable that I am being told how and irrelevant cognitions, because of the training to feel about bone marrow donation; I do not like time required to carry out the layers of classifica- that I am being told how to feel about bone marrow tion. But, due to the fact that thought listing often donation; It irritates me that the message told me contains ambiguous language, research partici- how to feel about bone marrow donation; I dislike pants may be in a uniquely valid position to judge that I am being told how to feel about bone marrow the valence of their own reports. To wit, when donation. In terms of the economics of research, confronted with a cognitive response such as “Is the potential advantages of the four-item mea- that claim supported by science?” researchers are sure are apparent: It provides a brief index that forced to guess as to whether the query is critical could be easily incorporated into surveys without or supportive. In principle, research participants the expensive and laborious tasks of gathering have more direct access to their own evaluations, and coding free-response data. and can, conceivably, provide more valid judg- Quick (2012) undertook a comparative analy- ments. Of course, focused research is needed to sis of the two methods in terms of reliability and evaluate the veracity of this speculation. validity (by evaluating each in a nomothetic net- work with a freedom threat, attitude, motivation, and source appraisal). Both approaches exhibited Reactance as satisfactory reliability. Both were also sensitive to an Individual Difference variations in forceful language and both pre- dicted attitude, motivation, and source appraisal Psychological reactance was first conceived as in theoretically meaningful ways. Quick recom- situation specific (Brehm, 1966; Wicklund, 1974). mends continued use of the Dillard and Shen Accordingly, most of the classic reactance research (2005) measure on the basis of its relatively was done on situational reactance, such as alterna- greater reliability and validity as well as the fact tive restriction and forced choice and in social

(c) 2013 Sage Publications, Inc. All Rights Reserved. 172——PART II. Theories, Perspectives, and Traditions influence settings (see Burgoon, Alvaro, Grandpre, revised Merz’s (1983) questionnaire, thereby & Voulodakis, 2002, for a review). However, Brehm creating the 14-item Hong Psychological and Brehm (1981) recognized that reactance could Reactance Scale (HPRS). The first investigation be conceptualized as a trait too, a position consis- of the scale structure yielded a four-factor solution tent with the theory’s assumption that people vary whose factors were labeled Freedom of Choice, in the strength of their needs for autonomy and Conformity Reactance, Behavioral Freedom, and self-determination (Wicklund, 1974). Reactance to Advice and Recommendations. Several attempts have been made to develop Subsequent efforts also produced structurally scales to measure reactance proneness (Dowd, similar results after discarding three items (Hong, Milne, & Wise, 1991; Hong, 1992; Hong & Faedda, 1992; Hong & Faedda, 1996). Although it is not 1996; Hong & Page, 1989; Merz, 1983). Merz clear that the individual factors are sufficiently (1983) developed the first self-report measure, reliable to justify scientific application, there is which showed promising results, but later work evidence that the overall sum of the items is a yielded inconsistent factor structures (see Donnell, useful, single indicator of proclivity to experience Thomas, & Buboltz, 2001; Hong & Ostini, 1989; reactance (Brown, Finney, & France, 2011; Shen Tucker & Byers, 1987). Subsequently, Dowd et al. & Dillard, 2005). The set of items formed well- (1991) developed the Therapeutic Reactance fitting, second-order factor model that exhibited Scale for clinical purposes. However, the two- parallelism across both theoretically relevant and factor solution lacked a simple structure, and tangential reference indicators (e.g., attitude vs. several of the items exhibited poor correspon- biological sex; Shen & Dillard), or in a bifactor dence with the theoretical construct of reactance framework, they formed a single general factor (cf. Buboltz, Thomas, & Donnell, 2002). that captured the variances beyond the common Dissatisfied with existing measures, Hong and residual variances explained by specific factors Page (1989) translated (German to English) and (Brown et al.; Table 11.1. provides the items).

Table 11.1 Items for the Hong Psychological Reactance Scale (Hong & Faedda, 1996)

1. I become frustrated when I am unable to make free and independent decisions.

2. It irritates me when someone points out things which are obvious to me.

3. I become angry when my freedom of choice is restricted.

4. Regulations trigger a sense of resistance in me.

5. I find contradicting others stimulating.

6. When something is prohibited, I usually think, “That’s exactly what I am going to do”.

7. I resist the attempts of others to influence me.

8. It makes me angry when another person is held up as a role model for me to follow.

9. When someone forces me to do something, I feel like doing the opposite.

10. I consider advice from others to be an intrusion.

11. Advice and recommendations usually induce me to do just the opposite.

Source: Abridged and modified from “Refinement of the Hong Psychological Reactance Scale,” by S. M. Hong and S. Faedda, 1996, Educational & Psychological Measurement, 56, p. 177, Copyright 1996 by Sage Publications.

Research indicates that trait reactant individ- Table 11.2 Items to Measure Perceived uals are autonomous, independent, nonconform- Threat to Freedom ist, self-determined, and somewhat rebellious (Dowd, Wallbrown, Sanders, & Yesenosky, 1994; Hong & Faedda, 1996; Miller, Burgoon, Grandpre, 1. The message threatened my freedom to & Alvaro, 2006; Miller & Quick, 2010). In line choose. with these characteristics, research supports a 2. The message tried to make a decision for me. positive association between reactance proneness 3. The message tried to manipulate me. and anger, autonomy, denial, dominance, inde- 4. The message tried to pressure me. pendence, interpersonal mistrust, intolerance, negative cognitions, nonconformity, and self- sufficiency (e.g., Dowd et al., 1994; Seibel & The reasoning behind including the freedom Dowd, 2001). threat induction check is that individuals could For the sake of clarity, it is worth emphasizing reasonably express anger and negative cogni- one final point regarding reactance as an indi- tions for any number of reasons having nothing vidual difference. Specifically, the HPRS does not to do with a threatened or eliminated freedom capture a steady state form of reactance. Rather, (e.g., fallacious reasoning or noncredible the HPRS assesses sensitivity and responsiveness sources). To fully test the process proposed by to circumstances and messages that may be seen PRT, researchers should ensure that the ante- as curtailing autonomy. Individual differences in cedents under investigation are directly related reactance are differences in propensity to experi- to a freedom threat, which in turn predicts reac- ence the process specified by reactance theory. tance. Inclusion of the threat induction check Accordingly, measurements via the HPRS might creates a more demanding and theoretically be expected to contribute to predicting responses precise test of the theoretical process under to persuasive messages either (1) above and scrutiny. Modeling this portion of the process is beyond messages effects alone or (2) in conjunc- akin to opening another viewing window on the tion with them. Research pertaining to trait reac- contents of the black box. It has considerable tance as both a main effect and as a moderator is practical application too in that it permits per- considered in more detail later in this chapter. suasion practitioners to trace the effects of their appeals or campaigns through the process from beginning to end. Recent demonstrations of Modeling the Reactance Process the utility of this two-step logic can be found in Quick and Kim (2009); Quick, Scott, and Although research has modeled the freedom Ledbetter (2011), and Quick (2012). threat induction check as both an endogenous and exogenous variable, Quick and his colleagues encouraged PRT researchers to treat freedom Summary threat and reactance measures both as mediators (Quick & Considine, 2008; Quick & Stephenson, Whereas classic reactance research generally 2008). That is, they advocate modeling reactance adopted a black box approach to process, con- as a two-step process featuring a freedom threat temporary inquiry has made significant strides as induction check followed by reactance. toward understanding the process more directly. Although many alternatives are possible, Quick Recent advances directly measuring threat to and colleagues have used four items (drawn from freedom and reactance allow researchers to trace Dillard & Shen, 2005) and provided consistent the effects of individual differences and message evidence of their reliability and validity. The four variations through the multiple steps proposed items are given in Table 11.2. by PRT.

Contemporary Research on student (Quick & Stephenson, 2008), and adult Reactance Effects samples (Quick & Considine, 2008) across multiple topics. Additionally, Henriksen, Dauphinee, A good deal of recent research examines message Wang, and Fortmann (2006) discovered that and audience features associated with psycho- adolescents exposed to ads containing strong logical reactance. This section reviews that litera- language, such as “Buy our product. It will kill ture for the purpose of highlighting factors that you,” “Think. Don’t smoke!” and “Tobacco is are likely to increase or decrease reactance. whacko if you’re a teen,” received poor evaluations, thus resulting in increased curiosity toward cigarettes. Overall, while the labels identifying Inducing Reactance this language feature are inconsistent (e.g., controlling, dogmatic, forceful, threat-to-choice lan- Domineering Language guage), each demonstrates the ineffectiveness of intense language in persuasion. One of the most obvious predictions of PRT is that high pressure communicators will induce Intent to Persuade reactance. But, how exactly, is pressure exerted? One common means of answering this question Several classic studies demonstrate that manip- has been to experimentally manipulate the lan- ulations of perceived intent to persuade produces guage of the persuasive appeal. For example, in results that are consistent with the theoretical high pressure conditions, individuals are told to predictions of reactance theory (e.g., Heller et al., “Stop the denial!” or “Any reasonable person 1973; Kohn & Barnes, 1977). More contemporary would agree that not exercising is a problem!” evidence of the negative impact of intent can be (Dillard & Shen, 2005; Grandpre, Alvaro, Burgoon, found in Benoit’s (1998) meta-analysis, which Miller, & Hall, 2003; Quick & Kim, 2009; Rains & reported a homogeneous effect across 12 investi- Turner, 2007). The ineffectiveness of using this gations. Wood and Quinn (2003) describe a sub- type of language has been reported within the sequent meta-analysis that shows diminished context of alcohol consumption (Bensley & Wu, postmessage attitude change in forewarned condi- 1991; Rains & Turner, 2007), drug use (Burgoon, tions. Their results suggest that counter-persua- Alvaro, Broneck, Miller, Grandpre, Hall, et al., sion effects may be limited to high-involvement 2002), exercise (Miller, Lane, Deatrick, Young, & topics, a side condition that seems to have been Potts, 2007; Quick & Considine, 2008; Quick & anticipated by PRT’s claim that importance of the Stephenson, 2008), meningitis (Rains & Turner, freedom enhances reactance. Evidence can also be 2007), sunscreen (Buller, Borland, & Burgoon, drawn from the impact of threat (of influence 1998; Quick & Stephenson, 2008), strep throat attempt) on inducing resistance in inoculation (Rains & Turner, 2007), and tobacco use (Grand- research (chapter 14 in this volume). pre et al., 2003). There are, however, two important qualifica- The research assessing the ineffectiveness of tions to the research on intent. One is that the data controlling language has benefited from several base is relatively small. Thus, it precludes strong studies sampling from various demographics. tests of distinctions among types of forewarning. For example, Rains and Turner (2007) found That is, warning of a pending message, warning of exposure to freedom-threatening language the topic of a pending message, and warning of resulted in greater reactance in her two-study the topic of the message as well as its position (i.e., experiment utilizing college students. Similarly, counter- vs. pro-attitudinal). Second, none of the Quick and colleagues discovered that domineer- available studies utilized a methodology that ing language was perceived as a freedom threat directly assessed reactance or located it in a pro- among adolescent (Quick & Kim, 2009), college cess model. While this is true of much reactance

research, as discussed earlier in this chapter, it losses of performing an act (see Quick & Bates, limits the confidence with which one can claim 2010; Reinhart, Marshall, Feeley, & Tutzauer, that reactance is implicated in persuasion or boo- 2007), influences the degree of reactance arousal. merang effects. For instance, Rains and Turner’s (2007) results suggest that as the magnitude of the request increased [ranged from small requests (e.g., Narrative reporting unclean conditions on campus) to large requests (e.g., donating $250 each semester Many of the messages that have been used to to fund a sanitization company)], psychological study reactance consist of arguments and evi- reactance also increased. dence. But, research on narrative tells us that sto- ries may also be an effective means of inducing change in beliefs and attitudes (chapter 13 in this Reactance Proneness as an volume). Moyer-Gusé (2008) theorized that one Audience Segmentation Variable of the reasons for narrative effectiveness is the As noted by Atkin and Salmon (chapter 17 relative inhibition of reactance. This is thought to this volume), the purpose behind segmenting occur to the extent that the narrative form obfus- an audience is to create groups of similar indicates persuasive intent. Indeed, Moyer-Gusé and viduals. One reason to seek similarity is that Nabi (2010) found that perceived persuasive group membership may correspond with the intent was positively associated with reactance for target of change. Another is that group similar- individuals exposed to a dramatic narrative. ity may indicate a propensity for members of Hence, individual differences in perceived persua- the group to make similar responses to any sive intent were in evidence even though a single given message. narrative message was presented. Those differ- On the first point, recent work suggests reac- ences corresponded directly with reactance. tance proneness should be considered when One complicating factor is that their index of segmenting at-risk audiences in public health reactance was composed of items such as The campaigns (Miller et al., 2006; Miller & Quick, show tried to pressure me to think a certain way 2010; Quick, Bates, & Quinlan, 2009). Specifically, and The show tried to force its opinions on me. The Miller et al. (2006) made a strong case for seg- item content suggests a closer correspondence menting audiences by reactance proneness as they with threat to freedom (see Table 11.2.) than to found it to be a “prominent predictor” of smok- reactance as it has been discussed in this chapter. ing in adolescents (p. 246). More recently, Miller But, either way, the measure taps an important and Quick (2010) report that reactant-prone piece of the reactance process. And, the findings individuals were more likely to use tobacco prod- suggest that when individuals perceived that an ucts and engage in risky sex than low-reactant entertainment program was attempting to influ- prone individuals. Importantly, these researchers ence their health attitudes, then threat-to-freedom found reactance proneness to be a better predic- or reactance was likely to follow. In sum, even if, tor of certain risky behaviors than sensation seek- on average, narratives are less likely to induce ing, an individual, characterized as individuals reactance than argument, individuals exposed to seeking out novel, arousing, emotionally com- narrative may still experience something like plex, or intense situations (Stephenson et al., reactance as a result of trait-type differences. 1999). For years, sensation seeking has been a primary segmentation variable for illicit drug use Magnitude of the Request as well as other hedonic behaviors, including risky sex and tobacco use. Thus, despite a positive cor- Recent PRT research suggests the magnitude relation between the traits (r = .34–.38 in Miller & of the request, as well as emphasizing the gains or Quick, 2010, and in Quick & Stephenson, 2008),

(c) 2013 Sage Publications, Inc. All Rights Reserved. 176——PART II. Theories, Perspectives, and Traditions it is intriguing that reactance seems to predict with low issue involvement toward the issue, certain risky behaviors with greater power than exposure to the message increased perceptions of sensation seeking. a freedom threat compared to individuals with The second reason to utilize reactance prone- high issue involvement. A criticism of this study ness as an audience segmentation variable is their broad conceptualization of involvement. emphasizes the potential for high or low reac- Specific types, such as impression-, outcome-, tance prone individuals to respond similarly and value-relevant involvement (Johnson & within group but differently between groups. In Eagly, 1989, 1990), might render different effects other words, trait reactance may interact with on reactance arousal. message features. A few studies report just such an effect, but sometimes in a form more com- Diminishing Reactance plex than anticipated. For instance, Dillard and Shen (2005) found that trait reactance moder- Choice-Enhancing Postscripts ated the effect of domineering language on reactance, such that high reactant individuals were Just as domineering and controlling language significantly more sensitive to linguistic varia- induces reactance, the presence of choices tends tions than were low reactant individuals. How- to diminish it. Brehm and Brehm (1981) reviewed ever, the interaction was observed for a message the impact of choice on reactance. Choice can on the use of dental floss, but not one on reduc- impact reactance from the dispositional as well as ing binge drinking. situational perspectives. The dispositional Quick and Stephenson (2008) applied the approach considers individual differences in multigroup method to test for differences in the external versus internal locus of control. The association between message features, freedom basic argument is there is an interaction between threat, reactance, and a motivation to perform locus of control and language used in counseling: the advocated behavior. In doing so, they discov- Controlling and directive language is more effec- ered that the association between a freedom tive for externals; and nondirective and unstruc- threat and reactance was stronger for high trait tured language works better for internals reactant individuals than low trait reactant indi- (Abramowitz, Abramowitz, Roback, & Jackson, viduals within the context of sunscreen promo- 1974; Kilman, Albert, & Sotile, 1975). The situation. However, this finding was not replicated tional approach examines the outcomes of the within the context of exercise. From this research, presence of choice. Gordon (1976) found that it appears that reactance proneness influences choice led to significantly more perceived value how individuals process persuasive messages, of a clinical treatment, as well as more relaxation. although these differences are not consistent Kanfer and Grimm (1978) found that free choice across contexts/topics. participants (choices offered and preference hon- With partial empirical support for reactance ored) improved significantly more than lost proneness as a moderator of PRT, questions choice participants (choice offered, but preference remain about why this individual difference vari- not honored); however, no choice participants able impacts the message processing of freedom- (choices not offered) did not differ significantly threatening messages within some contexts but from either group. not others. One explanation for this inconsis- However, without a measure of reactance, tency could be the importance of an issue. these studies cannot rule out the possibility that Recently, Quick et al. (2011) discovered that trait these results could have been produced due to reactance was positively associated with a free- cognitive dissonance, which unfortunately did dom threat within the context of organ donation. not have a measure either. Miller et al. (2007) Moreover, among high trait reactant individuals examined the impact of postscripts that restore

(c) 2013 Sage Publications, Inc. All Rights Reserved. Chapter 11. Reactance Theory and Persuasion——177 one’s freedom. Basically, a restoration postscript and mine, Lazarus, 1991). Identification with the tells the recipient, “the choice is yours. You are persuasive message means that the recipients free to decide for yourself”; while a filler post- tend to consider the persuasive attempt as less script does not contain such an argument. Their external, which reduces perceived threat to free- results showed that the restoration postscript dom, hence mitigating psychological reactance. significantly reduced perceived threat to free- There has been empirical evidence that in addi- dom. However, the authors did not report the tion to a direct impact on persuasion, state impact of the postscript on negative cognition empathy indeed mitigates psychological reac- or anger, or on the persuasion outcomes. tance (Shen, 2010b, 2011). Shen also identified Although the evidence could have been stronger, three message features that are believed to induce this research has important implications for empathy: the degree to which a message portrays research and practice at the same time. Theo- characters’ pain and suffering, the degree to retically speaking, such research is consistent which the message is realistic, and the degree to with the theorization of Brehm and Brehm which the message is affect-laden. These studies (1981). Practically speaking, this offers a cost- provided initial evidence that empathy can be an effective means of message design to reduce effective means to reduce reactance. psychological reactance. Sensation Value Empathy Message sensation value refers to “the degree Given that reactance can be considered as an to which formal and content audio-visual fea- alloy of anger and negative cognition (Dillard & tures of a message elicit sensory, affective, and Shen, 2005), messages that inhibit unintended arousal responses” (Palmgreen, Donohew, Lorch, anger and reduce counter-argument should Rogus, Helme, & Grant, 1991, p. 219), with sensa- diminish reactance as well. It has been proposed tion value being a critical determinant in message that empathy-arousing messages fall into that effectiveness. Messages high in sensation value strategy (Shen, 2010b, 2011). Empathy during elevate arousal or stimulation due to being dra- message processing is defined as a perception- matic, exciting, and novel (Morgan, Palmgreen, action process that occurs when the perception Stephenson, Lorch, & Hoyle, 2003). Similarly, of the characters’ state automatically activates the perceived message sensation value refers to a recipients’ vicarious experience of the characters’ message recipient’s judgments of message sensa- state, situation, and objectives. Such empathic tion value along three dimensions (Palmgreen, responses are conceptualized to have three com- Stephenson, Everett, Baseheart, & Francies, 2002): ponents: perspective taking, emotional conta- dramatic impact, emotional arousal, and novelty. gion, and identification (Shen, 2010a). Perspective Kang, Cappella, & Fishbein (2006) advance the taking means that the message recipients adopt intriguing argument that high sensation value the viewpoint of the message and its source, messages can distract audiences from refuting the which means counter-argument is less likely. argument, that is, reduce counter-arguing, thereby Emotional contagion means that the recipient’s enhancing ad persuasiveness. The results are emotional experiences would be similar to those complex. They show an interaction between argu- portrayed in the message, thus less unintended ment quality, message sensation value, risk for emotions (including anger). marijuana use such that high risk adolescents The function of empathy lies in social bonding judged the high quality–high sensation value and relationship development, which indicates messages as particularly ineffective. As the authors that it contradicts or inhibits the core relational note, one challenge to interpretation lies with theme for anger (a demeaning offense against me the fact that the high quality argument messages

(c) 2013 Sage Publications, Inc. All Rights Reserved. 178——PART II. Theories, Perspectives, and Traditions contained five times more commands than the here to tell people not to smoke, because, well, it low quality argument messages. A reactance per- doesn’t work” (truth®, 2012). However, the anti- spective might suggest that this aspect of the mes- smoking ads themselves leverage reactance sage reduced persuasive impact especially among effects by portraying the tobacco industry as the pro-attitudinal (i.e., high risk group). controlling and deceptive (Farrelly, Davis, Working in the same area, Quick (in press) Haviland, Messeri, & Healton, 2005). One mes- explicitly examined the relationship between the sage underscores industry efforts to manipulate three dimensions of perceived message sensation adolescents in this way: value and reactance. His study of anti-marijuana messages showed no observable effect for either In response to the fact that 5 million people dramatic impact and emotional arousal. How- around the world die annually due to ever, perceived novelty did significantly reduce tobacco, individuals outside of a major reactance levels. Hence, the available data, though tobacco company sing and dance to a catchy few in number, indicate that message novelty song suggesting the ingredients in cigarettes, may be an effective means of preventing reac- the health costs associated with smoking, tance effects. This knowledge provides direction and the number of deaths annually must be for research and application, but cannot alone a typo. At the end of the ad, one of the per- solve the context-bound problem of developing formers say, “Wait, there is no way there is genuinely new reasons for behavior change. that many typos.”

The campaign itself was a notable applied suc- Reactance as a Persuasive Strategy cess. More intensive study of the messages using contemporary methods for assessing reactance Quick and colleagues (2009) argued that could provide insight into the basis for that arousing reactance could serve as a motivator for achievement. individuals to support clean indoor air policies. They discovered that as individuals’ anger increased following exposure to secondhand New Directions smoke, they were more likely to support clean for PRT Research indoor air policies. Therefore, as a rhetorical strategy, Quick et al. advocated the use of mes- In keeping with the themes developed through- sages emphasizing the violation of one’s right to out this chapter, we see several promising possi- breathable air could be an effective strategy, bilities for future research. Foremost among which in turn would motivate individuals to them is working toward a better understanding actively support clean indoor air initiatives. Min- of the message features that enhance and dimin- ing a similar conceptual vein, Turner (2007) ish reactance. As currently formulated, many of presents a framework for understanding political the message concepts discussed in this chapter action that is motivated by choice-reduced anger. have their roots in lay perceptions of language A similar application of reactance-based per- rather than crafted scientific theories. A turn suasion can be seen in the “truth” campaign, a toward sociolinguist frameworks, such as polite- successful long-term effort to discourage smok- ness theory, or basic interpersonal communi- ing among Florida youth (Bauer, Johnson, cation theory might provide a perspective on Hopkins, & Brooks, 2000; Siegel & Biener, 2000). message design that would move PRT in a for- Interestingly, the website notes explicitly the ward direction. potential for reactance effects: “Tell someone not Future research might also assess reactance to do something and they will. . . . We’re not proneness as a moderator variable in large-scale

(c) 2013 Sage Publications, Inc. All Rights Reserved. Chapter 11. Reactance Theory and Persuasion——179 campaigns in much the same way sensation Conclusion seeking has been examined in the past. In theory, reactance proneness appears to be an especially As of this writing, reactance theory is almost half important segmentation strategy for adoles- a century old. The fact that it still inspires ques- cents and young adults, in particular because tions and motivates research is testimony to the this is an age when reactance arousal is height- power of its ideas. Looking back, there is ample ened (see Brehm & Brehm, 1981; Burgoon, evidence that PRT has made signal contributions Alvaro, Grandpre, et al., 2002; Woller, Buboltz, to our understanding of persuasion. As efforts are & Loveland, 2007). Answers to at least two ques- made to further refine the theory and to meld it tions are needed before it can be concluded that with ongoing research, we anticipate that its value trait reactance is a useful general strategy: (1) To to future generations of persuasion researchers what extent, if any, is it empirically distinct will equal or exceed its past achievements. from sensation seeking? (2) Can trait reactance predict patterns of message choice and consumption (as has been done so effectively for References sensation seeking)? Abramowitz, C. V., Abramowitz, S. I., Roback, H. B., & Future research should seek to understand Jackson, C. (1974). Differential effectiveness of reactance as a strategy of empowerment. By directive and non-directive group therapies as a broadening depictions of reactance as a poten- function of client internal-external control. Jour- tially useful state, not just an aversive state, social nal of Consulting and Clinical Psychology, 42, influence researchers might be able to capitalize 849–853. on opportunities to infuse change, particularly Bauer, U. E., Johnson, T. M., Hopkins, R. S., & Brooks, among adolescents who value their freedom and R. G. (2000). Changes in youth cigarette use and strongly oppose persuasive attempts aimed at intentions following implementation of a tobacco reducing these freedoms. Understanding reac- control program: Findings from the Florida tance as the basis for change is surely equal in Youth Tobacco Survey, 1998–2000. The Journal of the American Medical Association, 284, 723–728. importance to its role as a force for counter- Benoit, W. L. (1998). Forewarning and persuasion. In persuasion. M. Allen & R. W. Preiss (Eds.), Persuasion: Finally, we reemphasize the value of studying Advances through meta-analysis (pp. 139–154). reactance as a process. Mass communication Cresskill, NJ: Hampton Press. research might also benefit from application of Bensley, L. S., & Wu, R. (1991). The role of psycho- PRT. Extant work has hypothesized that PRT can logical reactance in drinking following alcohol play a pivotal role in media selection, particularly prevention messages. Journal of Applied Social among adolescents (Bushman & Cantor, 2003; Psychology, 21, 1111–1124. Scharrer & Leone, 2008; Sneegas & Plank, 1998). Brehm, J. W. (1966). A theory of psychological reactance. Essentially, these researchers assert that media New York, NY: Academic Press. ratings restricting individuals from viewing par- Brehm, J. W., & Brehm, S. S. (1981). Psychological reac- ticular programs or playing certain videogames tance: A theory of freedom and control. San Diego, CA: Academic Press. enhance the attractiveness of the content (referred Brehm, J. W., Stires, L. K., Sensenig, J., & Shaban, J. to as forbidden-fruit effect). Like other classic (1996). The attractiveness of an eliminated choice studies of PRT, media selection is limited by its alternative. Journal of Experimental Social Psy- reliance on outcomes associated with reactance chology, 2, 301–313. rather than measuring the state directly. Media Brehm, S. S., & Weinraub, M. (1977). Physical barriers effect research would benefit from this approach, and psychological reactance: 2-year-olds’ responses just as classic PRT inquiry has advanced by focus- to threats to freedom. Journal of Personality and ing on mediating processes. Social Psychology, 35, 830–836.

Brown, A. R., Finney, S. J., & France, M. K. (2011). Dillard, J. P., & Shen, L. (2005). On the nature of reac- Using the bifactor model to assess the dimension- tance and its role in persuasive health communi- ality of the Hong Psychological Reactance Scale. cation. Communication Monographs, 72, 144–168. Educational and Psychological Measurement 71(1), Donnell, A. J., Thomas, A., & Buboltz, W. (2001). 170–185. Psychological reactance: Factor structure and Brugoon, M., Alvaro, E., Broneck, K., Miller, C., internal consistency of the questionnaire for the Grandpre, J. R., Hall, J. R., & Frank, C. A. (2002). measurement of psychological reactance. The Using interactive media tools to test substance Journal of Social Psychology, 14, 679–687. abuse prevention messages. In W. D. Crano & Dowd, E. T., Milne, C. R., & Wise, S. L. (1991). The M. Burgoon (Eds.), Mass media and drug preven- therapeutic reactance scale: A measure of psycho- tion: Classic and contemporary theories and logical reactance. Journal of Counseling & Devel- research (pp. 67–87). Mahwah, NJ: Erlbaum. opment, 69, 541–545. Buboltz, W. C., Thomas, A., & Donnell, A. J. Dowd, E. T., Wallbrown, F., Sanders, D., & Yesenosky, J. (2002). Evaluating the factor structure and inter- M. (1994). Psychological reactance and its rela- nal consistency reliability of the therapeutic reac- tionship to normal personality variables. Cogni- tance scale. Journal of Counseling & Development, tive Therapy and Research, 18, 601–612. 80, 120–125. Farrelly, M. C., Davis, K. C., Haviland, M. L., Messeri, Buller, D. B., Borland, R., & Burgoon, M. (1998). P., & Healton, C. G. (2005). Evidence of a dose- Impact of behavioral intention on effectiveness of response relationship between “truth” antismok- message features: Evidence from the family sun ing ads and youth smoking prevalence. American safety project. Human Communication Research, Journal of Public Health, 95, 425–431. 24, 433–453. Gordon, R. M. (1976). Effects of volunteering and Burgoon, M., Alvaro, E., Grandpre, J., & Voulodakis, responsibility on perceived value and effective- M. (2002). Revisiting the theory of psychologi- ness of a clinical treatment. Journal of Consulting cal reactance: Communicating threats to attitu- and Clinical Psychology, 44, 799–801. dinal freedom. In J. P. Dillard & M. Pfau (Eds.), Grandpre, J. R., Alvaro, E. M., Burgoon, M., Miller, C. H., The persuasion handbook: Developments in the- & Hall, J. R. (2003). Adolescent reactance and ory and practice (pp. 213–232). Thousand Oaks, anti-smoking campaigns: A theoretical approach. CA: Sage. Health Communication, 15, 349–366. Bushman, B. J., & Cantor, J. (2003). Media ratings Hammock, T., & Brehm, J. W. (1966). The attractive- for violence and sex: Implications for policy- ness of choice alternatives when freedom to makers and parents. American Psychologist, choose is eliminated by a social agent. Journal of 58, 130–141. Personality, 34, 546–554. Dillard, J. P., & Meijnders, A. (2002). Persuasion and Heller, J. F., Pallak, M. S., & Picek, J. M. (1973). The the structure of affect. In J. P. Dillard and M. Pfau interactive effects of intent and threat on boo- (Eds.), The persuasion handbook: Developments in merang attitude change. Journal of Personality theory and research (pp. 309–327). Thousand and Social Psychology, 26, 273–279. Oaks, CA: Sage. Henriksen, L., Dauphinee, A., Wang, Y., & Fortmann, Dillard, J. P., & Peck. E. (2000). Affect and persuasion: S. (2006). Tobacco companies’ antismoking ads Emotional responses to public service announce- and adolescent reactance: Test of a boomerang ments. Communication Research, 27, 461–495. effect. Tobacco Control, 15, 13–18. Dillard, J. P., & Peck. E. (2001). Persuasion and the Hong, S. M. (1992). Hong’s Psychological Reactance structure of affect: Dual systems and discrete Scale: A further factor analytic validation. Psycho- emotions as complementary models. Human logical Reports, 70, 512–514. Communication Research, 27, 38–68. Hong, S. M., & Faedda, S. (1996). Refinement of the Dillard, J. P., Plotnick, C. A., Godbold, L. C., Freimuth, Hong psychological reactance scale. Educational V. S., & Edgar, T. (1996). The multiple affective and Psychological Measurement, 56, 173–182. consequences of AIDS PSAs: Fear appeals do Hong, S. M., & Ostini, R. (1989). Further evaluation of more than scare people. Communication Research, Merz’s psychological reactance scale. Psychologi- 23, 44–72. cal Reports, 64, 707–710.

Hong, S. M., & Page, S. (1989). A psychological reac- restoration of freedom. Human Communication tance scale: Development, factor structure and Research, 33, 219–240. reliability. Psychological Reports, 64, 1323–1326. Miller, C. H., & Quick, B. L., (2010). Sensation seeking Johnson, B. T., & Eagly, A. H. (1989). Effects of involve- and psychological reactance as health risk predic- ment on persuasion: A meta-analysis. Psychologi- tors for an emerging adult population. Health cal Bulletin, 106, 290–314. Communication, 25, 266–275. Johnson, B. T., & Eagly, A. H. (1990). Involvement and Morgan, S. E., Palmgreen, P., Stephenson, M. T., Lorch, persuasion: Types, traditions, and the evidence. E. P., & Hoyle, R. H. (2003). The relationship Psychological Bulletin, 107, 375–384. between message sensation value and perceived Kanfer, F. H., & Grimm, L. G. (1978). Freedom of message sensation value: The effect of formal choice and behavioral change. Journal of Consult- message features on subjective evaluations of ing and Clinical Psychology, 46, 873–878. anti-drug public service announcements. Journal Kang, Y., Cappella, J. N., & Fishbein, M. (2006). The of Communication, 53, 512–526. attentional mechanism of message sensation Moyer-Gusé, E. (2008). Toward a theory of entertain- value: Interaction between message sensation ment in persuasion: Explaining the persuasive value and argument quality on message effective- effects of EE messages. Communication Theory, ness. Communication Monographs, 73, 351–378. 18, 407–425. Kelly, A. E., & Nauta, M. M. (1997). Reactance and Moyer-Gusé, E., & Nabi, R. L. (2010). Explaining the thought suppression. Personality and Social Psy- effects of narrative in an entertainment television chology Bulletin, 23, 1123–1132. program: Overcoming resistance to persuasion. Kilman, P. R., Albert, B. M., & Sotile, W. M. (1975). Human Communication Research, 36, 26–52. Relationship between locus of control, structure Nabi, R. L. (2002). Discrete emotions and persuasion. of therapy, and outcome. Journal of Consulting In J. P. Dillard & M. Pfau (Eds.), The persuasion and Clinical Psychology, 43, 588. handbook: Developments in theory and research Kohn, P. M., & Barnes, G. E. (1977). Subject variables (pp. 289–308). Thousand Oaks, CA: Sage. and reactance to persuasive communications Palmgreen, P., Donohew, L., Lorch, E, P., Rogus, M., about drugs. European Journal of Social Psychol- Helme, D., & Grant, N. (1991). Sensation seeking, ogy, 7, 97–109. message sensation value, and drug use as media- Lazarus, R. S. (1991). Emotion and adaptation. New tors of ad effectiveness. Health Communication, 3, York, NY: Oxford University Press. 217–227. Leventhal, H. (1970). Findings and theory in the study Palmgreen, P., Stephenson, M. T., Everett, M. W., of fear communications. In L. Berkowitz (Ed.), Baseheart, J. R., & Francies, R. (2002). Perceived Advances in experimental social psychology (Vol. 5, message sensation value (PMSV) and the dimen- pp. 119–186). New York, NY: Academic Press. sions and validation of a PMSV scale. Health Lindsey, L. L. M. (2005). Anticipated guilt as behavioral Communication, 14, 403–428. motivation: An examination of appeals to help Petty, R. E., & Cacioppo, J. T. (1986). Communication unknown others through bone marrow donation. and persuasion: Central and peripheral routes to Human Communication Research, 31, 453–481. attitude change. New York, NY: Springer-Verlag. Merz, J. (1983). A questionnaire for the measurement Quick, B. L. (2012). What is the best measure of psy- of psychological reactance [in German]. Diagnos- chological reactance? An empirical test of two tica, 29, 75–82. measures. Health Communication, 27, 1–9. Miller, C. H, Burgoon, M., Grandpre, J. R., & Alvaro, Quick, B. L. (in press). Perceived message sensation E. M. (2006). Identifying principal risk factors for value and psychological reactance: A test of the the initiation of adolescent smoking behaviors: dominant thought disruption hypothesis. Journal The significance of psychological reactance. of Health Communication. Health Communication, 19, 241–252. Quick, B. L., & Bates, B. R. (2010). The use of gain- or Miller, C. H., Lane, L. T., Deatrick, L. M., Young, A. M., loss-frame messages and efficacy appeals to dis- & Potts, K. A. (2007). Psychological reactance and suade excessive alcohol consumption among college promotional health messages: The effects of con- students: A test of psychological reactance theory. trolling language, lexical concreteness, and the Journal of Health Communication, 15, 603–628.

Quick, B. L., Bates, B. R., & Quinlan, M. R. (2009). The Seibel, C. A., & Dowd, E. T. (2001). Personality charac- utility of anger in promoting clean indoor air teristics associated with psychological reactance. policies. Health Communication, 24, 548–561. Journal of Clinical Psychology, 57, 963–969. Quick, B. L., & Considine, J. R. (2008). Examining the Seltzer, L. F. (1983). Influencing the “shape” of resis- use of forceful language when designing exercise tance: An experimental exploration of paradoxi- advertisements for adults: A test of conceptualiz- cal directives and psychological reactance. Basic ing reactance arousal as a two-step process. Health and Applied Social Psychology, 4, 47–71. Communication, 23, 483–491. Shen, L. (2010a). Mitigating psychological reactance: The Quick, B. L., & Kim, D. K. (2009). Examining reactance role of message-induced empathy in persuasion. and reactance restoration with Korean adoles- Human Communication Research, 36, 397–422. cents: A test of psychological reactance within a Shen, L. (2010b). On a scale of state empathy during collectivist culture. Communication Research, 36, message processing. Western Journal of Commu- 765–782. nication, 74, 504–524. Quick, B. L., Scott, A. M., & Ledbetter, A. (2011). A Shen, L. (2011). The effectiveness of fear- vs. empathy- close examination of trait reactance and issue arousing anti-smoking PSAs. Health Communica- involvement as moderators of psychological reaction, 26, 404–415. tance theory. Journal of Health Communication, Shen, L., & Dillard, J. P. (2005). Psychometric proper- 16, 660–679. ties of the Hong Psychological Reactance Scale. Quick, B. L., & Stephenson, M. T. (2007a). Further Journal of Personality Assessment, 85, 74–81. evidence that psychological reactance can be Siegel, M., & Biener, L. (2000). The impact of an anti- modeled as a combination of anger and nega- smoking media campaign on progression to tive cognitions. Communication Research, 34, established smoking: Results of a longitudinal 255–276. youth study. American Journal of Public Health, Quick, B. L., & Stephenson, M. T. (2007b). The reac- 90, 380–386. tance restoration scale (RSS): A measure of direct Smith, M. J. (1977). The effects of threats to attitudinal and indirect restoration. Communication Research freedom as a function of message quality and Reports, 24, 131–138. initial receiver attitude. Communication Mono- Quick, B. L., & Stephenson, M. T. (2008). Examining graphs, 44, 196–206. the role of trait reactance and sensation seeking Sneegas, J. E., & Plank, T. A. (1998). Gender differences on reactance-inducing messages, reactance, and in pre-adolescent reactance to age-categorized reactance restoration. Human Communication television advisory labels. Journal of Broadcasting Research, 34, 448–476. & Electronic Media, 42, 423–434. Rains, S. A., & Turner, M. (2007). Psychological reac- Stephenson, M. T. (2003). Examining adolescents’ tance and persuasive health communication: A responses to antimarijuana PSAs. Human Com- test and extension of the intertwined model. munication Research, 29, 343–369. Human Communication Research, 33, 241–269. Stephenson, M. T., Palmgreen, P., Hoyle, R. H., Reich, J. W., & Robertson, J. L. (1979). Reactance and Donohew, L., Lorch, E., & Colon, S. E. (1999). norm appeal in anti-littering messages. Journal of Short-term effects of an anti-marijuana media Applied Social Psychology, 9, 91–101. campaign targeting high sensation seeking ado- Reinhart, A. M., Marshall, H. M., Feeley, T. H., & lescents. Journal of Applied Communication Tutzauer, F. (2007). The persuasive effects of mes- Research, 27, 175–195. sage framing in organ donation: The mediating truth®. (2012). About Us. Retrieved March 9, 2012, role of psychological reactance. Communication from http://www.thetruth.com/about/ Monographs, 74, 229–255. Tucker, R. K., & Byers, P. Y. (1987). Factorial validity of Scharrer, E., & Leone, R. (2008). First-person shooters Merz’s psychological reactance scale. Psychologi- and the third-person effect. Human Communica- cal Reports, 61, 811–815. tion Research, 34, 210–233. Turner, M. M. (2007). Using emotion to prevent risky Schwarz, N., Frey, D., & Kumpf, M. (1980). Interactive behavior: The anger activism model. Public Rela- effects of writing and reading a persuasive essay tions Review, 33, 114–119. on attitude change and selective exposure. Journal White, G. L., & Zimbardo, P. G. (1980). The effects of of Experimental Social Psychology, 16, 1–17. threat of surveillance and actual surveillance on

expressed opinion toward marijuana. Journal of Wood, W., & Quinn, J. M. (2003). Forewarned and Social Psychology, 111, 49–61. forearmed? Two meta-analytic syntheses of fore- Wicklund, R. A. (1974). Freedom and reactance. warnings of influence appeals. Psychological Bul- Potomac, MD: Lawrence Erlbaum. letin, 129, 119–138. Wicklund, R. A., & Brehm, J. W. (1968). Attitude Worchel, S. (1974). The effect of three types of arbi- change as a function of felt competence and trary thwarting on the instigation to aggression. threat to attitudinal freedom. Journal of Experi- Journal of Personality, 42, 300–318. mental Social Psychology, 4, 64–75. Worchel, S., & Andreoli, V. (1974). Attribution of cau- Witte, K. (1994). Fear control and danger control: A test sality as a means of restoring behavioral freedom. of the Extended Parallel Process Model (EPPM). Journal of Personality and Social Psychology, 29, Communication Monographs, 61, 113–134. 237–245. Woller, K. M. P., Buboltz, Jr., W. C., & Loveland, J. M. Worchel, S., Andreoli, V. A., & Archer, R. (1976). When (2007). Psychological reactance: Examination is a favor a threat to freedom: The effects of attri- across age, ethnicity, and gender. American Journal bution and importance of freedom on reciprocity. of Psychology, 120, 15–24. Journal of Personality, 44, 294–310.