• No results found

The interaction of negative emotions and power : from interpersonal level to public service advertising

N/A
N/A
Protected

Academic year: 2021

Share "The interaction of negative emotions and power : from interpersonal level to public service advertising"

Copied!
42
0
0

Bezig met laden.... (Bekijk nu de volledige tekst)

Hele tekst

(1)

The interaction of negative emotions and power:

From interpersonal level to Public Service Advertising

Carmen A. Mihu

Student ID-card number: 10494626 Master’s Thesis

Graduate School of Communication Master’s programme Communication Science

Supervisor: dr. S.C.M (Stephanie) Welten

(2)

Abstract

Studies of emotion have been consistent in finding the potential persuasive power of emotions in face-to-face interpersonal interactions. These findings could constitute an appreciable addition to the advertising setting, more specifically to the public service advertising (PSA). In this particular context, the advertiser aims to arouse emotions, an

endeavor which can misfire or even fail altogether. By means of an experiment involving four advertisements, the current study investigated the potential persuasive power of negative emotions (sadness vs. anger) and power (low vs. high) in the PSA context. The results suggest that emotions and power affect the individuals’ attitude towards the advertisement and their perceived control over the behavior promoted in the advertisement. In some respects, these results go against findings in studies of interpersonal interaction in that the advertisement displaying a powerful spokesperson expressing sadness, and a powerless spokesperson expressing anger, are more effective in promoting behavior. This is explained by referring to the moderation role of epistemic motivation and the mediating effect of perceived

appropriateness of expressed emotion. The effect of these variables is therefore hypothesized and tested. The results are finally discussed in reference to practical and theoretical

implications.

(3)

Introduction

It has been shown that individuals tend to make use of their interlocutor’s emotions as a way of achieving various aims in concrete social situations (Sheppes, Scheibe, Suri, & Gross, 2011, Cole, Martin, & Dennis, 2004, Stegge & Terwogt, 2007). We use emotions, either consciously or not, in order to get audience behave in a certain way in response to our behavior, e.g. to forgive us.

To give a simple example, one might express anger towards a colleague who is late and this may lead to the colleague’s apologies and promise of non-repetition. However, expressing anger may in fact ‘activate’ the colleague’s anger in turn. The same expression of anger, however, may not have the same consequence if directed at an individual superior in status or power (say one’s superior in an organization). In such cases, sadness might be ‘safer’ in that it may more likely elicit the colleague’s or the superior’s apology.

With respect to such situations, two aspects might be highlighted. First, discrete negative emotions such as anger and sadness can lead to different forms of behaviors when observed (Van Kleef, 2009). Second, the perceived authority of the person expressing the emotion could influence the responses of the observer to the expressed emotion (Van Kleef, 2009, Lelieveld, Van Dijk, Van Beest, Steinel, & Van Kleef, 2011). These different effects of emotions have been thoroughly studied in the field of interpersonal communication, the most noticeable findings being discovered in the context of negotiation. According to these

findings, people respond to emotional expressions by taking one of two routes: the cognitive route and the affective. The choice is influenced by elements such as motivation to process information (or epistemic motivation), the relationship between the speaker and the listener (perceived power and status) and the perceived appropriateness of the expressed emotion (Van Kleef, 2014).

While these interpersonal findings signal the potential that emotions have as powerful persuasive devices, it remains unclear whether they can be translated to other contexts. More specifically, the question can be asked whether and how the effects of negative emotions displays can be extrapolated from their interpersonal context in other settings focused on persuading recipients, such as public service advertising (PSA).

This study seeks to fill this gap by testing if and how two negative emotions, sadness and anger, can influence the recipient’s intentions, attitudes and beliefs when expressed in a public service advertisement. Further, I investigate how and if perceived social power of the spokesperson plays a moderating role in the effects of these emotions. In doing so, the

(4)

Emotions As Social Information model (EASI model, henceforth) was used, a model applied and confirmed in interpersonal settings (Van Kleef, 2009). This model suggests that the observer’s epistemic motivation moderates, while the perceived appropriateness of the expressed emotion mediates, the direct effect of emotion on people’s cooperative or conflicting reactions.

Anger and sadness are two negative and basic emotions, which are effortlessly recognized and frequently experienced (Ekman, 1971). As showed by Tiedens (2001), these emotions are associated with certain levels of perceived social status by means of stereotypes. This is important in the advertising context in which “advertisers seek to communicate to a target audience with existing beliefs” (Kim & Lowry, 2005, p. 901) in order to avoid

challenging people’s pre-existing beliefs, which may trigger their avoidance of the message or negative attitudes (Knobloch-Westerwick, 2012). Thus, using the stereotype of the angry high-status person or the sad low-status person in an advertisement, the people’s subjective reality is reinforced (Pollay, 1986) and might increase the likeability of advertisement processing and positive evaluation.

Moreover, the PSA context was found suitable for this study mainly because of its focus on negative emotions as persuasive strategies (Dillard & Peck, 2000). A second reason would be their unconfirmed efficiency in eliciting the targeted behavior in their audience (Dillard, Plotnick, Godbold, Freimuth, & Edgar, 1996, Dillard & Peck, 2000) as well as their relatively weak persuasive effects (O'Keefe & Reid, 1990).

In the present study, relevant literature is reviewed and an overview is provided on research in the area of anger and sadness as persuasive tools in the context of interpersonal communication and the potential role played by the perceived power of the emotion expresser (i.e., spokesperson). Further, epistemic motivation and the perceived appropriateness are introduced as potential factors. Based on this theoretical research a set of hypotheses were formulated. Next, the methodology for testing the hypotheses is described, followed by the results and associated discussion. The paper ends with conclusions and suggestions for future research.

(5)

Theoretical background

The effect of anger and sadness expressions

Public service announcements or advertisements (PSAs) differ from regular

advertising in their promotion of social issues (i.e., quitting smoking, changing a sedentary lifestyle, or quitting irresponsible drinking) and in their attempt to influence people’ beliefs, attitudes, and behaviours concerning these issues (O'Keefe & Reid, 1990). Significant

research has been directed at the elements that can help to create persuasive PSAs (Strasser et al., 2009, Sutfin, Szykman, & Moore, 2008, Kang, Cappella, Strasser, & Lerman, 2009), many focusing on the evoking of negative emotional responses such as fear or guit (Dillard & Nabi, 2006). The problem with these emotions and the use of negative emotional arousals in general, as underlined by Dillard and Nabi (2006), is that messages designed to evoke a certain emotion in the target person may arouse not only the desired emotion but others as well. Some examples are ads designed to evoke guilt which also evoked anger (Pinto & Priest, 1991), or ads intending to activate people’s guilt which also produced feelings of shame (Bennett , 1998). In general, these studies employed messages containing a variety of cues (music, color, story, etc.) that were meant to activate a desired emotional state. Although interpersonal research confirmed numerous times the power of negative emotional

expressions as persuasive devices, no research tested the behavioral responses of the audience when exposed to an advertising message displaying a person feeling a particular negative emotion towards an unwanted behavior.

At the interpersonal level, anger and sadness are negative emotions, which can be directed at another individual’s behavior, signaling that the behavior is unwanted and it must be corrected (Fischer & Roseman, 2007, Lelieveld et al., 2011). Sadness is a “supplication emotion” being described by weak levels of arousal and perceived as an indicator of low levels of coping potential need and dependence (Van Doorn, Van Kleef & Van der Pligt, 2015, Clark & Taraban, 1991). On the other hand, anger has a strong level of arousal and the person experiencing it is perceived as having a strong motivation to approach (Carver & Harmon-Jones, 2009) and a tendency to correct an undesired situation (Fischer & Roseman, 2007).

Several theorists contend that the function of emotions is to direct behavior (Frijda, Kuipers, & ter Schure, 1989, Roseman, Wiest, & Swartz, 1994). According to this, each emotion has an associated motivation known as an ‘action tendency’ that can trigger observer’s engagement and/or withdrawal. In this sense, anger is generally viewed as an

(6)

engagement emotion (Harmon-Jones & Allen, 1998), yet, it can lead to attack (i.e., a negative form of engagement) or dismissal (i.e., withdrawal) (Dillard & Nabi, 2006). The same can apply to sadness: one’s sadness can trigger sympathy and the feeling to comfort (i.e.,

engagement), while when sadness is seen as unappropriated in a context this can lead to anger and frustration – a form of withdrawal (Van Doorn, Van Kleef, & Van der Pligt, 2015). Both emotions, although different, can be used as persuasive devices and trigger another person’s compliance in interpersonal interactions. This is of great importance in the advertising context, where people’s engagement is sought and people’s withdrawal is detrimental to the effectiveness of the ad. By knowing under what conditions sadness and anger expressions lead to observer’s engagement one may make use of this knowledge in designating persuasive advertising messages.

Researchers in the interpersonal communication showed a great interest in finding how emotions can determine a person’s likelihood to comply with a request (O'Keefe & Figge, 1997). Compliance can be accurately assessed in interpersonal settings by simply observing the behavior of the target person (Van Doorn, Van Kleef, & Van der Pligt, 2015). However, this is harder to be done in the public service advertising context were many resources must be employed. In consequence, the current study the interpersonal compliance will be adapted to the advertising field by assessing the intention to comply with the behavior requested “verbally” by the spokesperson displayed in the advertisement (“promoted

behavior” henceforth). According to the theory of planned behavior (TPB; Ajzen, 1991), the individual’s intention to perform the behavior is the immediate determinant of behavior.

However, in order to test different levels that may be affected by a spokesperson’s expressed emotions, along with the intention to comply with the promoted behavior, I will also assess people’s attitudes towards the promoted behavior and towards the advertisement, as well as their perceived behavioral control. All these measurements are usually used in the assessment of the persuasiveness of an advertisement targeting behavioral change. Attitudes towards the ad are conceptualized as people’s positive or negative evaluations of the

advertisement, while attitudes towards the behavior refers to the negative or positive way people assess the promoted behavior. Finally, perceived behavioral control (PBC) refers to the degree of control a person believes he or she has over performing the promoted behavior (Ajzen, 1991). These three assessments are highly desirable in the context of PSAs, in which the aim is to promote a positive attitude towards the promoted behavior and enhance people’s perceived ability to perform it.

(7)

Some studies on advertising (e.g., Burke & Edell, 1989) showed that negative feelings decrease one's favorability toward the ad execution. Moreover, McKee et al. (2003) observed that participants in a negative mood were more likely to assess smoking behavior in a

negative way compared to participants in a positive mood (Petty, Schumann, Richman, & Strathman, 1993). In this sense, from an emotion contagion point of view, both the negative emotional displays (i.e. sadness and anger) should lead to more positive attitudes towards the promoted behavior, but more unfavorable attitudes towards the advertisement.

While some research indicates that expressions of negative emotions such as anger elicit more cooperative responses than do expressions of positive emotions such as happiness (e.g., Sinaceur & Tiedens, 2006), more research is consistent in finding that positive rather than negative emotions elicit cooperation (e.g., Kopelman, Rosette, & Thompson, 2006). In line with this, one may suggest that anger (having attack as action tendency) is a more intense emotion than sadness – a passive emotion associated with neediness. Based on the different effects of anger and happiness described above, it can be expected that both anger and sadness would lead to a cooperative response such as intention to comply. However, judging by the perceived intensity of each emotion, it can be predicted that sadness, being a milder emotion, can be more likely perceived as less negative than anger, leading to a greater effect on

people’s compliance.

However, an important element emphasized by the interpersonal finding is that the effects of emotions depend on the target of the emotion, (Steinel, Van Kleef, & Harinck, 2008), such that oriented anger can elicit larger concessions than behaviour-oriented happiness or person-directed anger. Because the current study will focus on

behaviour-oriented emotion, from this perspective, it can be expected that anger would lead to a greater intention to comply than sadness.

Lastly, while no previous literature approached the effect of expressed anger or sadness on a person’s perceived control over his or her behaviour, it is suggested that, based on the low level of coping potential associated with sadness, when people will be exposed to an advertisement showing a spokesperson expressing sadness, this will decrease their

perceived control over the promoted behavior. Conversely, anger suggests a stronger motivation and tendency to correct an undesired situation (Carver & Harmon-Jones, 2009, Fischer & Roseman, 2007). In other words, the spokesperson’s sadness will suggest a weak potential to control the performance of the behavior, thus observers will not feel empowered to take control over the behavior, as they will be when the spokesperson will express anger.

(8)

Anger on the other hand, will imply frustration resulted from an unwanted behaviour which can be avoided but it is not.

Based on these considerations, the following hypothesis emerged:

H1: If the spokesperson in the PSA will express anger, this will lead to a greater intention to comply with the requested behavior (a), a more favorable attitude towards the promoted behavior (b), a higher perceived behavioral control (c), but an unfavorable attitude towards the ad (d) than when the spokesperson will express sadness.

The moderating effect of the spokespersons’ perceived social power

Previous findings have suggested that expressing anger in response to a negative event can be more effective than expressing sadness in creating perceptions of competency about a leader (Tiedens, 2001, Madera, & Smith, 2009). This positive effect on a leader’s competency might have favorable implications for the speaker’s ability to persuade. Although expressions of anger can elicit perceptions of leadership and strength, expressions of sadness are more effective in evoking sympathy from audiences (Labott, Martin, Eason, & Berkey, 1991). Sad people are perceived to be warm, nice, and likable, whereas angry people are perceived to be dominant and unfriendly (Tiedens, 2001). Considering that source likability plays an

important role in persuasion, a spokesperson expressing sadness might be more advantaged in persuading others than when expressing anger (Roskos-Ewoldsen, Bichsel, & Hoffman, 2002).

A study by Lelieveld and colleagues (2012) found that anger and disappointment affect the observers differently when the perceived power of the expresser moderates the effect. More specifically, in their study, anger evoked a complementary emotion (fear) in targets when reported by a high-power bargainer but evoked a reciprocal emotion (anger) when reported by a low-power bargainer. This reciprocal anger led participants to offer less to low-power counterparts who reported anger. Disappointed bargainers, however, evoked a complementary emotion (guilt) in participants and increased offers, regardless of the

bargainer’s power position. In this study, due to the great similarities between disappointment and sadness (Van Doorn, 2014, Van Kleef, De Dreu, & Manstead. 2010), it can be presumed that a spokesperson, regardless of his status, will trigger more intention to comply when expressing sadness than when expressing anger. However, when a high power spokesperson will express anger, this will lead to a greater intention to comply than when a low-power spokesperson will express the same emotion. Many studies testing these effects in settings characterized by clear power differences, such as leader–follower interactions (Van Kleef et

(9)

al. 2009; Lindebaum & Fielden 2011; Van Kleef et al. 2010) support this superior effect of anger and high power over sadness and low power.

Power represents the relative influence of an individual over the outcomes of others (Fast, Halevy, & Galinsky, 2011). High-power negotiators are less affected by the emotions that their counterparts display, presumably because their considerable resources allow them to act ad libitum without costs (Van Kleef & Cote, 2007). For example, Sinaceur and Tiedens (2006) found that displays of anger led to yielding when expressed by low power negotiators, but not from high power negotiators.

The dominance signaled by expressions of anger may therefore be seen as incongruent when expressed by a low power spokesperson as opposed to a high power spokesperson. On the other hand, sadness may be seen as more stereotype-congruent when expressed by a low power status (Tiedens, Ellsworth, & Mesquita, 2000). However, while expressions of sadness coming from a high power person (i.e. leader) might signal inappropriateness, this

inappropriateness in turn could “humanize” the spokesperson and emphasize the

“sensitiveness” of the topic. This idea has been supported in the crisis literature, revealing that a leader expressing sadness was evaluated more favorably than a leader expressing

anger (Madera & Smith, 2009). Considering the literature reviewed in the previous

paragraphs, which support the existence of an interaction effect between emotion and power, the following hypothesis was derived:

H2a-b. The effect of spokespersons’ emotional expressions on perceiver’s intention to comply with the behavior and attitude towards the behavior and ad is moderated by the spokesperson’s perceived power, such that (a) a spokesperson’s angry expression engenders less intention to comply and less positive attitudes toward the ad and behavior, as well as a less perceived behavioral control when he has a perceived high power than when he is perceived as having less power, whereas (b) a spokesperson’s sadness expression engenders more intention to comply and positive attitudes toward the ad and behavior, as well as a less perceived behavioral control when he is perceived as having less power than when he is perceived to have low power.

Leading the response to emotion expressions: the affective and inferential route

If we take a look at the effects of emotional expressions in advertising we can see that these can be explained by interpersonal mechanisms. One of them is “emotion contagion”, regarded as an interpersonal influential mechanism in which emotional expressions transfer to other individuals (Howard, & Gengler, 2001). According to this view, the observers

(10)

spontaneously and unconsciously mimic the nonverbal signals (e.g., facial expressions) of the person expressing the emotion, this being followed by the experience of the mimicked

emotion. These findings hold also in the advertising context, people tending to mimic the emotions expressed by the spokesperson displayed in the advertisement (Howard & Gengler, 2001).

By testing emotion contagion in charity advertisements, Small and Verrochi (2009) showed that people “catch” the emotions displayed on a victim's face and that they are particularly sympathetic and likely to donate when they see sad expressions versus happy or neutral expressions. In line with this, it may also be expected that if a person will see an advertisement displaying an angry spokesperson he will also experience anger. Thus emotion contagion could advantage the advertising displaying the sad spokesperson and disadvantage the angry spokesperson.

However, these emotion contagion effects are diminished when people start to process the advertisement and begin to make inferences about the reasons of the emotional display. What these results pinpoint is the existence of two processes that might occur when an individual is exposed to an advertisement: affect contagion (that is the emotion contagion described above) and inferential processing.

These two processes are extensively tested in the interpersonal literature concerning emotions as social information (EASI model). Van Kleef (2009) suggests that the likelihood that one of these processes will occur depends on two factors: epistemic motivation and the perceived appropriateness of the expressed emotion.

Epistemic motivation and perceived appropriateness

Epistemic motivation is rooted in individual differences and refers to the observer’s innate motivation to process the information around him, such as emotional displays (Van Kleef, De Dreu, & Manstead, 2004b). When an individual has a high epistemic motivation he is more predisposed to process the information about the situation, and is more likely

influenced by others emotions, as opposed to an individual with low epistemic motivation (Côté & Hideg, 2011, Van Kleef, De Dreu, & Manstead, 2004b).

Evaluation of emotion appropriateness is a second factor proposed by Van Kleef (2009) as influencing the response to expressed emotions. This is defined as the degree to which an observer judges an emotion expression to be in line with his expectations (Shields, 2005). The importance of emotion expression appropriateness is also underlined by Bucy’s study (2000), who showed that participants’ evaluations of a President (i.e., a powerful

(11)

person) displaying negative emotions evaluated the emotions as more appropriate than the positive displays, and as significantly more honest and credible. By this, one may expect that a negative emotion displayed by a powerfully spokesperson may be seen as more appropriate, and may trigger more favorable attitudes towards the spokesperson and the advertisement.

Appropriateness was found to also play a role in people’s willingness to offer help and donate to charity. When a request was accompanied by disappointment, rather than anger, this led to more cooperation as the emotion was perceived to be more appropriate for the context (Van Doorn, Van Kleef, & Van der Pligt, 2015). Anger, on the other hand, was perceived as inappropriate and unjust, triggering aggressive responses (Barclay et al. 2005; Van Kleef & Cote, 2007).

Taking into account epistemic motivation and appropriateness, through a series of studies in the negotiation field, Van Kleef (2009) discovered that inferential processes occur to the degree that the observer engages in more thorough information processing (i.e. high level of epistemic motivation) and perceives the emotional expression as appropriate. On the other hand, affective processes become more predictive when information processing is low (i.e. low level of epistemic motivation) and the emotional expression is evaluated as

inappropriate.

In light of this, it is expected that epistemic motivation will play a moderating role in the relationship between the emotion expressed by the spokesperson and the power, while the perceived appropriateness will mediate this emotion-power interaction:

H3a-b. The effect of a spokesperson’s perceived power on the negative emotion expression is moderated by observer’s epistemic motivation, such that: (a) a high power spokesperson expressing anger triggers more intention to comply and more positive attitudes toward the ad and behavior, as well as a more perceived behavioral control than does a high power spokesperson expressing sadness when the observer has a high epistemic motivation than when he has a low epistemic motivation, whereas; (b) a low power spokesperson expressing anger triggers less intention to comply and attitudes toward the ad and behavior, as well as a more perceived behavioral control than does a high power spokesperson

expressing sadness when the observer has a high epistemic motivation than when he has a low epistemic motivation.

H4a-b. H5. Perceived appropriateness will mediate the moderation effect of power on the effect of emotion on DVs.

(12)

Method

Participants and Design

To test the hypotheses summarized in Fig. 1, an experiment was conducted with a 2 (Negative emotion: sadness vs. anger) x 2 (Power: low vs. high) between-subjects design, while measuring for subjects’ epistemic motivation and perceived appropriateness. In total, 205 English speaking participants (67.3% female) having a mean age of 22.39 years

(SD=.19), were randomly exposed to one of the four conditions. Moreover, 71.7% indicated The Netherlands as their country of origin, followed by Germany (5.4%), China (4.9%) and Italy (2.4%). Also, 53.7% had a Bachelor’s degree, 23.9% had a high school degree, 43% a Master or equivalent, and 3% had a doctorate, law or medical degree.

Data collection & procedure

Participants were recruited face-to-face from two canteens belonging to the University of Amsterdam. They were asked by the researcher to fill in a 5-minute pen-and-paper survey about “bike safety advertising” and were offered chocolate mini-bars as a thank-you act for their participation.

The survey consisted of 14 questions (See Appendix) and a final question asking participants to provide their email address in order to take part in a potential online follow-up study1. After presenting the topic and purpose of the survey and asking participant’s consent for participation, the survey began with socio-demographic questions, followed by a question assessing the innate epistemic motivation of the participants and their involvement with the “bike safety” topic. After these, participants were exposed to one of the four advertisements representing the four conditions tested in this experiment. Once they had seen the ad, they were asked to fill out a set of questions regarding the advertisement (i.e., dependent variables, perceived appropriateness and covariates), followed by a set of questions designed to check the successfulness of the manipulation.

1

I decided not to conduct the follow-up study, as I do not believe that one exposure to an advertisement is sufficient in persuading its viewers over time. Thus, it would have been non-realistic to assume that the advertisement presented in this study influenced the significance of any result that might have appeared in the follow-up study.

(13)

Stimulus Materials

The experimental material was a fictitious public service advertisement part of a European bike safety campaign. The bike safety topic is highly knowledgeable and relevant to people living in Amsterdam. In addition, it was mentioned that the ad is part of a European campaign in order to create a “neutral ground” and diminish the influence of potential pre-existent cultural-specific beliefs which may get undesirable differences within an international sample.

Each advertisement (see Fig. 2a-d in Appendix) displayed a man aged 40-50 (i.e., spokesperson), a quote suggesting his words, the name of the spokesperson and his function, and a headline saying: “Two people die in traffic accidents every day. Stop rushing! Don’t pass the red light!”. Another element constant across the conditions was the European Transport Safety Council logo, meant to enhance the credibility of the advertisement. Apart from these elements, which were kept constant across all four conditions, the advertisements differed in power (hospital janitor as low power vs. police officer as high power) and the negative emotion displayed by the spokesperson (anger vs. sadness).

Negative emotion manipulation. In order to create two conditions differing in the negative emotion displayed, with the use of Adobe Photoshop, the facial expression of the spokesperson was altered based on Facial Action Coding System (FACS) (Ekman, Friesen, & Hager, 2002) so that the spokesperson in the angry condition expressed anger, while the spokesperson in the sadness condition expressed sadness. To enhance the recognition of the expressed emotions by the observers, the ad also included a quote of the spokesperson who was explicitly expressing his emotion towards passing a red light i.e. “I get angry when I see people like you biking through the red light as if it’s not there” (for the anger condition), “I get sad when I see people like you biking through the red light as if it’s not there” (for the sadness condition). The choice of using both non-verbal and verbal elements in the

manipulation of emotion is justified by previous research, which illustrated that people engage both non-verbal and verbal cues when detecting an emotion (Buchanan, Lutz, Mirzazade, Specht, Shah, Zilles, & Jäncke, 2000).

Power manipulation. In order to manipulate the power of the spokesperson, the low power spokesperson was depicted as a janitor, wearing a blue overall, and the quote was accompanied by a short description of the spokesperson – “John, hospital janitor”. On the other hand, the high power spokesperson was wearing a policeman uniform and the quote was completed by a short textual description - “John, police officer”. Relying on the distinction made by Fast, Halevy and Galinsky (2011) it can be contended that a policeman has a greater

(14)

perceived power than a janitor (i.e., asymmetric outcome control), and both jobs are seen as lacking status (i.e., respect and admiration).

Measurements

The four dependent variables assessed in the study are described in the following paragraphs.

Intention to not cross the red light. After seeing the advertisements, participants were asked to give their answers regarding their intention to not cross the red light in the forthcoming month on a series of three 7-point Likert-type items (1=Strongly disagree, 7=Strongly agree) which started with: “I intend to stop at the red light”, “I will try to stop at the red light”, “I plan to stop at the red light”, and ended with “each day in the forthcoming month” (Ajzen, 2011). The items proved to constitute a reliable scale and were averaged to form a composite score (M=3.83, SD=1.93, α=.94).

Attitude toward the behaviour. In order to assess participants’ attitudes towards crossing a red light, they had to indicate how they assess the behaviour on three 7-point Likert-scale items anchored Strongly Disagree (1) and Strongly Agree (7). The items were adopted from Ajzen’s recommendations to construct a Theory of Planned Behaviour Questionnaire (2011). The scale included items negatively formulated which were recoded accordingly. The question was formulated as follows: “For me, stopping at the red light each day in the forthcoming month is… beneficial/a waste of time/a bad idea”. After recoding the items so that for all items a score of “7” indicate a very favourable attitude and a score of “1” indicate a very unfavourable attitude, as in previous research, the scale proved reliable, and, hence, the items were averaged to form an index measure (M=4.46, SD=1.46, α=.77).

Attitude towards the advertisement. Respondents’ attitudes towards the

advertisement were assessed by a question containing three items evaluated through 7-point Likert scales. Thus, when asked to answer “To what extent do you think the company is...”, participants had to choose from a 7-point scales having as minimum and maximum ranges the following adjectives: Bad/Good, Unpleasant/Pleasant, Disagreeable/Agreeable, and

Unsatisfactory/Satisfactory. All four items loaded on one factor and showed sufficient reliability to be combined into one item (M=3.69, SD=1.28, α=.83).

Perceive behavioural control. Perceived behavioural control was measured by asking respondents to indicate their agreement with the following sentences: “For me to stop at the red light each day in the forthcoming month would be impossible” and “If I wanted to I could

(15)

stop at the red light each day in the forthcoming month”. Both items were significantly correlated (r=.29, p<.001) and averaged to form a single measure.

In this study two factors were hypothesized to influence the interaction effect between sadness and anger and the perceived power of the spokesperson:

Epistemic motivation. Respondent’s innate epistemic motivation was measured as did by Van Kleef, Van Doorn, Heerdink and Koning (2011), that is in terms of need for structure by using an 11 items scale developed by Neuberg and Newsom (1993). After recoding two negative questions, a PCA analysis showed one item having a low belonging to the scale, thus when removed, the reliability of the scale increased from .81 to .83. In

consequence, the remaining ten items were averaged to form an index measure (M=3.98, SD=0.94, α=.83). After this, a median split was performed resulting in one item consisting in high levels of epistemic motivation (equal or more than the median score 4), and low level of epistemic motivation (less than the median score 4). This median split resulted in 106

respondents with high epistemic motivation and 99 participants with low epistemic motivation.

Perceived appropriateness. The perceived appropriateness of the spokesperson’s expressed emotion has been measured with the question “The reaction of the person depicted in the advertisement is...” to which participants had to choose on a 7 scale how

Appropriate/Inappropriate, Justified/Unjustified, Correct/Incorrect they assessed the spokesperson’s reaction. The items were recoded so that all positive adjective to equal “7”, and all negative adjective to equal “1”. The three items resulted in a reliable component (M=4.62, SD=1.32, α=.86).

Covariates and potential confounds

Perceived emotion intensity, involvement with the topic and gender were measured for exploratory reasons to assess their impact on the dependent variables. To measure

involvement with the topic, respondents were asked to assess on a 7-point scale ranging from Not at all (1) to To a great extent (7) whether they find the bike safety campaign as relevant, meaningful to them or if it’s worth remembering. All three items were combined into one reliable item (M=3.92, SD=1.42, α=.85).

The perceived intensity of the emotion expressed by the spokesperson is another element which may account for variance in the dependent variables measured in this study. This has been measured with one question asking respondents to indicate the level of the intensity on a 7-point scale ranging from not at all intense to extremely intense. Gender was

(16)

assessed within the demographic set of question, through a single standard question. All these variables were tested as potential covariates and their results can be seen in the Results

section.

Manipulation checks

Perceived Negative emotions. To test whether sadness and anger displayed by the spokesperson were successfully observed by the respondents exposed to the ad, they were asked to assess the extent to which the spokesperson expressed anger, disappointment,

sadness or happiness. For each word, participants had to indicate on a 7-point scale the extent to which they felt that each of the specific emotion was expressed: Not at all (1) and To a great extent (7). Disappointment was measured as this is usually used interchangeably with sadness (Van Doorn, Van Kleef, & Van der Pligt, 2015), while happiness was measured for control purposes.

Perceived power. To measure the extent to which respondents perceived the

spokesperson’s power (i.e., asymmetric outcome control), a complex scale used by Bachman, Smith, and Slesinger (1966) was adopted. The scale contained five bases of social power: referent, expert, reward, coercive and legitimate. However, from this scale, the “referent” item was used as indication of status. Relying on Fast, Halevy, Galinsky’s hypotheses and findings (2011) “admiration” doesn’t determine the level of power, but of status (i.e. “The person in the ad is a person I admire”). As suggested by Fast, Halevy, Galinsky (2011), power and status are two distinct constructs, although used interchangeably for many years, which can actually interact, leading to different effects in both the people having them and in the people observing them2. Because of the complexity of the scale, a Principal Component Analysis was conducted showing that another item (i.e. “influence” – “The person in the ad is a person with a lot of influence over people”) was not related to the scale. After removing this item, Cronbach’s Alpha increased from .77 to .82. The three remaining items were averaged and resulted in a new item (M=4.03, SD=1.73, α=.82).

2

According to Fast, Halevy, Galinsky (2011), a person may have both status and power, have status but lack power and vice-versa.

(17)

Results

Manipulation check

Perceived Negative emotions. The manipulation of the emotions proved successful: people in the anger condition scored significantly higher on anger (M=5.83, SD=1.21) than did the people in the sadness condition (M=3.16, SD=1.68) (F(1,204) =92.80; p<.001). People in the sadness condition perceived the advertisement as more sad (M= 5.24, SD=1.41) than people in the angry condition (M=3.16, M=1,68) (F(1,204)=91.80, p<.001).

Disappointment and happiness also showed significant differences between the two groups, more specifically, people in the sadness condition perceived the spokesperson as more disappointed (M=4.19, SD=1.77) than people in the angry condition (M=4.19, SD=1.77) (F(1,204)=24.95, p<.001). Taking into account the significant mean difference between anger and sadness on both anger and sadness measurements, it may be concluded that the

manipulation of the emotion proved successful.

Perceived power and status. A simple analysis of variance comparison with sadness and anger as independent variable and perceived power as dependent variable showed that people in the low power condition perceived the spokesperson as less powerful (M=2.67, SD=1.20) as opposed to people in the high power condition who perceived the high power condition accordingly (M=4.63, SD=1.29) (F(1,204)=118,841, p<.001). Regarding the perceived status of the two spokesperson, no significant difference between the conditions occurred, and both conditions scored quite low: low power (M=2.46, SD=1.51), high power (M=2.17, SD=1.38) F(1,203)=2,012, p=.158). Thus, the ads produced the hypothesized differences in perceived power of the spokesperson.

Confound checks

Confound checks revealed that topic involvement was significantly related to the attitude towards the ad (r=.29, p<.01), attitude towards behaviour (r=.27, p<.01) and intention (r=.27, p<.001). As expected and confirmed by previous studies, gender also plays a

significant role in how people react to different emotional expressions (Wild, Erb, & Bartels, 2001). In line with this, gender showed a significant correlations with both intention to comply (r=.25, p<.01) and attitude towards the behaviour (r=.28, p<.01). Both variables were therefore included as covariates in subsequent analyses. In addition to this, the perceived intensity of the emotion was assessed. An analysis of variance showed that there were

(18)

significant differences between the two conditions. More specifically, people in the angry condition perceived the emotion as more intense (M=5.04, SD=1.67) than people exposed to the sadness condition (M=4.43, SD=1.66) (F(1,204) =6.96, p=.009). This was in line with previous research indicating anger as an “active” emotion and “sadness” as a passive one (Truong & Raaijmakers, 2008). However, the perceived intensity of emotion had no significant correlation with the dependent variables, thus this was not added as covariate in further analyses.

Hypotheses Testing

In order to test the first three hypotheses (see Fig. 1) a 2 (emotion: sadness vs. anger) x 2 (power: low vs. high) x 2 (epistemic motivation: low vs. high) Multivariate Analysis of Covariance (MANCOVA) was conducted, with intention to comply, attitude towards the ad, attitudes towards behaviour and perceived behavioural control as dependent variables, while controlling for gender and involvement with the topic of bike safety. The choice of

performing all the analysis at once is justified by the concern of avoiding an inflated experiment-wise error rate.

However, before performing the MANCOVA, a series of Pearson correlations were performed between all of the dependent variables with the purpose of testing the Multivariate Analysis of Variance assumption described by Meyers, Gampst, and Guarino (2006), that all the dependent variables analysed should be correlated with each other. A significant pattern of correlations was observed amongst all the dependent variables analysed (see Table 3),

suggesting the appropriateness of a MANOVA. Moreover, it should be noted that Levene’s test produced non-significant p-values, suggesting that the null hypothesis regarding equal variances for the dependent variables cannot be rejected. The results for all the free

hypotheses are further reported.

Hypothesis 1. With the first hypothesis (H1) it was proposed that expressions of anger, will lead to a greater intention to comply with the requested behavior, a more favorable attitude towards the promoted behavior, a higher perceived behavioral control, but an

unfavorable attitude towards the ad than when the spokesperson will express sadness. The MANCOVA analysis described above showed that no significant differences for the respondents in the angry (n=108) versus sad condition (n=97), Pillai’s Trace=.016,

F(4,192)=0.77, p=.555, η=0.02, regarding their perceived behavioural control in their attitude towards the ad, attitude towards the behaviour and intention to comply with the promoted behaviour (see Table 1). This means that the emotion alone doesn’t have an impact over a

(19)

person’s attitudes, intentions and perceived control over behaviour as suggested by the first hypothesis. Based on these results, it can be concluded that the first hypothesis was not confirmed.

Hypothesis 2. The second hypothesis predicted that a spokesperson’s angry

expression will lead to less intention to comply and unfavorable attitudes toward the ad and behavior, as well as a less perceived behavioral control when he has a high power than less power. Moreover, a spokesperson’s sadness expression is supposed to lead to more intention to comply and attitudes toward the ad and behavior, as well as a less perceived behavioral control when he is perceived as having less power than when he has high levels of power. The Multivariate Analysis of Covariance showed a non-significant interaction effect between emotion and power, Pillai’s Trace=.037, F(4, 192)=1.86, p=.118, η=0.04. However, tests of Between-Subjects effect showed that the interaction effect between emotion and power approached significance for two of the four dependent variables: attitude towards the ad and perceived behavioural control. In order to check whether the power of the interaction effect might have been affected by the third independent variable, the same MANCOVA was performed, but this time epistemic motivation was excluded. As shown by the univariate results, this time the interaction effect between emotion and power proved to approach significance, F(1,199)=2.23, p=0.67, η=0.04, for attitude towards the ad (F(1,199)=5.01, p=.026, η=0.03) and perceived behavioural control (F(1,199)=3.90, p=.050, η=0.02).

Follow-up planned comparisons showed that people feel more empowered to think they have control over their behaviour (i.e., PBC) when exposed to a sad spokesperson with high perceived power (M=5.91, SD=1,27) than low perceived power (M=5.33, SD=1.32) (p= .031)(see Fig. 3), and express a more favourable attitude towards the ad when exposed to an angry spokesperson with low power (M=3.85, SD=1,31) rather than high power (M=3.33, SD=1.26) (p=.033) (see Fig. 4).

Despite the two significant differences described above, there were no significant results for intention (F(1,199)=.037 , p=.847, η=0.00) or attitude towards the behaviour (F(1,199)=.172 , p=.679, η=0.01). Hence, hypothesis two was partially confirmed, the results showing the existence of an interaction effect between emotion and power, however, this interaction existed for two out of four dependent variables, and for the significant ones the effects were contradicting the hypothesized expectations. Possible explanations are discussed in the discussion section of this paper.

Hypothesis 3. Hypothesis 3 stated that epistemic motivation would moderate the interaction effect between emotions and perceived power of the spokesperson on observer’s

(20)

perceived behavioural control, intention to comply, attitude towards the behaviour and attitude towards the ad. The same MANCOVA performed for the first two hypotheses, showed that the three-way interaction didn’t have a significant effect on intention to comply, perceived behavioural control and attitude towards the ad and behaviour (F<1, ns). Thus, hypothesis three was disconfirmed.

Hypothesis 4. The fourth hypothesis suggested that the effect of a spokesperson’s expressed emotion on perceived appropriateness is moderated by the perceived power of the spokesperson, such that an expression of anger is perceived as more appropriate when the spokesperson has a low power, whereas an expression of sadness is perceived as less appropriate when the spokesperson has a high power. In order to test this hypothesis, an Univariate Analysis of Variance performed with appropriateness as dependent variable, emotion and power as independent variables and gender and involvement as covariates. For this, no hypothesized effect approached significance (F(1,205)=0.35, p=.555). However, although not hypothesized, the perceived appropriateness differed significantly on emotion (F(1,199)=4,76, p=.030, η=0.02) such that sadness was perceived as more appropriate

(M=4.84,SD=1.24) than anger (M=4.43, SD=1.37). Implications of this result are discussed in the Discussion section.

Hypothesis 5. To test the last hypothesis (H5), a formal test of mediated moderation was performed through a series of hierarchical regression analyses following the procedure recommended by Muller, Judd, and Yzerbyt (2005). According to this procedure, to

demonstrate mediated moderation, three conditions must be met: (1) there is a significant interaction effect between emotion and power on the dependent variables; (2) there is a significant interaction effect between emotion and power of the spokesperson on the mediator appropriateness; (3) there is a significant effect of the mediator appropriateness on the

dependent variables controlling for emotion and perceived power; and (4) the interaction effect between emotion and power on the dependent variables should be smaller than the effect resulted when the mediator is included in the model.

Although the first condition has been confirmed by hypothesis two, which showed a significant interaction effect between emotion and status on perceived behavioural control and attitude towards the ad, no significant effect has been found for the interaction effect on the mediator perceived appropriateness (b*=.04, t=.51, p=.608, 95% CI [-0,54, 0,92]). Because the second requirement described by Muller, Judd, and Yzerbyt (2005) was not met, the analysis was stopped being concluded that appropriateness does not have a mediating role in the emotion-power relationship. In consequence, the last hypothesis was not confirmed.

(21)

Discussion

This study examined the potential effects of negative emotions and power in public service advertisements by drawing upon research concerning the use of emotions in

interpersonal interaction. For this purpose, the interpersonal context was ‘re-created’ in an advertisements varying in expressed negative emotion (sadness vs. anger) and spokesperson power (low vs. high).

The results of the current experimental study showed that people will not intend to comply with the promoted behaviour, have favourable attitudes towards the advertisement and the promoted behaviour, or believe they can achieve the promoted behaviour to not pass the red light when they observe the spokesperson’s expressions of anger or sadness,

regardless of his power. Disconfirming H1, these results may suggest that, in line with Burke and Edell (1989) and Aaker, Stayman, and Vezina (1988), negative emotions alone do not lead to significant effects in the context of advertising.

Several explanations have been suggested for these results. First, negative emotions may trigger negative moods which in turn can influence people’s evaluations in an

unfavourable way (see also McKee et al., 2003). Second, the contrast brought by negative emotions in a world full with positive advertisements my work against the advertiser’s aims. A viewer is presumably exposed daily to many advertisements meant to induce positive moods and create favourable associations with the advertised product. An advertisement presenting negative emotions, as usually happens with public service advertisement, may contrast in a negative way leading people to avoidance and resistance to persuasion. Third, the medium in which people perceive these advertisements may influence their reaction. If an individual does not have the time or the motivation to fully engage, her or she may more likely rely on heuristics and emotion contagion (Howard & Gengler, 2001). Nevertheless, while emotion contagion can prove to be a favourable way to persuasion for positive emotions (Petty, Schumann, Richman, & Strathman, 1993), this may not hold also for advertisements presenting expressions of negative emotions. Future research should further investigate the relationship between the medium and the expression of specific negative emotions.

Another finding of this study was that while emotion alone may not affect people’s evaluations and intentions, this occurs when the spokesperson differs also in social power. Although the second hypothesis was confirmed in that there is indeed an interaction effect, these effects did not have the expected direction. More specifically, a sad powerful

(22)

spokesperson performed better in enhancing people’s perceived ability to control their behaviour than a sad spokesperson with a low power. Although against the interpersonal findings, which suggest that a powerful person will be more likely associated with anger than sadness, the results could suggest that in the advertising setting the sadness expressed by a powerful person may be evaluated against people’s existing stereotypes (Tiedens, 2001). In light of this, a policeman is assessed as a tough person more likely to become angry. Hence, his sadness expression may signal the gravity of the issue promoted. However, this would imply that people took the processing route and started to make inferences about why a spokesperson with that level of power would express that certain emotion (Van Kleef, 2009). According to the interpersonal findings (EASI model) this can occur if the observer of the emotional expression has a high epistemic motivation (i.e., the innate tendency to make use of all the available information), while in the advertising context this is more likely to occur when the person has a high involvement with the topic at hand (i.e., Elaboration Likelihood Model). If indeed people drew inferences they may have thought that the sadness expression of the police officer implies that the deviant behaviour of passing a red light can be controlled for, however, many people don’t do it thus the policeman feels powerless towards this

situation (Van Doorn, Van Kleef, & Van der Pligt, 2015).

The interaction between power and emotion also had an effect over attitude towards the advertisement in that an angry spokesperson with low power leads to more favourable attitudes towards the advertisement than an angry spokesperson with high power. This result comes also against the interpersonal findings and can be explained by inferential processing. By taking the path of inferences, it can be supposed that people inferred that the hospital janitor has reasons to be angry because he has no expertise on the matter, but he is a common citizen annoyed by people’s irresponsibility. In light of this, it can be believed that the anger expressed by the janitor is more sincere being expressed by a person in symmetrical power position. Thus his genuine emotional expressions may have triggered people’s sympathy and favourable attitudes towards the advertisements (Bucy, 2000).

As described above, these results could stress that the underlined effects are caused by a high epistemic motivation of the observer (i.e., high motivation to engage in extensive processing). Therefore, the role of epistemic motivation as a moderator of emotion and power was further investigated, as suggested by interpersonal literature (Van Kleef, 2009). However, this did not play a significant role in this interaction. One explanation could be that

advertising differs from interpersonal settings in that people may not engage fully when assessing an advertisement for the first time, regardless of their level of epistemic motivation.

(23)

In the advertising context, an advertisement must be displayed several times in order to trigger people’s interest and motivate them to engage in cognitive elaboration, as suggested by the Elaboration Likelihood Model. Thus in this case, the inferences described above could be caused by other factors such as involvement with topic, which showed a significant correlation with the dependent variables and consequently was controled for.

Furthermore, although appropriateness proved to be a reliable mediator in

interpersonal settings (Van Doorn, Van Kleef, & Van der Pligt, 2015), the findings of this study suggest that this is not the case for the interaction of emotion and power in the public service advertising setting. However, further analysis showed that perceived emotion expressed by the spokesperson, regardless of his power, does affect how appropriate people perceive the expression. This could be explained by the appropriateness of the expressed emotion to the context, such that anger may not be perceived as an appropriate emotion to be expressed in a bike safety advertising campaign. As suggested by Warner and Shields (2009), appropriateness is a multidimensional construct and future studies should take this into consideration.

Despite the interesting findings, this study was not without limitations. One such limitation was that participants were international students, mostly Dutch. While this can be a good sample because of its randomness, a more specific and homogenous sample could lead to different effects as a results of cultural norms and conventions. As showed by Adam, Shirako, and Maddux (2010), cultural backgrounds influence the way people react to emotion expressions in negotiation context, stressing the urgency of investigating these effects in other types of social interactions.

A second limitation was the printed ads used as stimulus material. The choice of using printed advertisements was deliberate in order to isolate emotion and power from the sum of elements (e.g., music, action, colours, setting, etc.) which may compose a video or audio advertisement, and which may interact with the manipulated factors. However, while printed ads were used for convenience purposes, I acknowledge the importance of testing these in video and audio advertisements, mediums which may enhance the recognition and power of the emotion-power interaction.

A third limitation was that being the first study of its kind, the dependent variables assessed consisted of measures that may indicate changes in the behaviour as these are the main effects targeted by the public service advertisements. A full spectrum of potential persuasive effects which can occur in the advertising setting might have also indicated interesting results.

(24)

As showed by the results of the current study, people indeed process both the emotion expressed and the perceived power of the spokesperson and draw inferences. Further research should try to identify the underlying inferences which justify people’s reactions to the

interaction between emotion and power.

Another element which should be further investigated is that of the interaction between negative and positive emotions and factors which describe the social position of the spokesperson such as status, leadership or authority. Gender has also been proved to be

stereotypically associated with specific emotions. According to this, women are more likely to be perceived as expressing milder emotions such as sadness, while men are expected to

express anger (Lewis, 2000). Thus, the gender of the spokesperson could also account for differences in people’s reactions to the observed expressed emotion.

In their day to day life people make use of other people’s expressed emotions to inform their behaviour, while others take advantage of the persuasive power of emotions and use them to direct their peers’ behaviour in a way favourable to them. A great deal of research has been dedicated to find out what emotions and under what conditions certain behaviours are acquired. In line with Fischer and Van Kleef (2010), it can be said that other disciplines using emotion expressions as persuasive device could take advantage of these findings. This study was a first attempt to apply interpersonal findings concerning the persuasive use of emotion in a context such as advertising. I believe that the findings of this study have shown the potential persuasive power of explicitly expressing emotions, a potential that is not fully explored in other academic disciplines interested in persuasion. By identifying people’s reactions to specific emotional expressions, one may determine patterns of stimulus-reaction which can be fruitfully used in predicting and directing recipient’s behaviour.

(25)

References

Aaker, D. A., Stayman, D. M., & Vezina, R. (1988). Identifying feelings elicited by advertising. Psychology & Marketing, 5(1), 1-16.

Adam, H., Shirako, A., & Maddux, W. W. (2010). Cultural variance in the interpersonal effects of anger in negotiations. Psychological Science, 21(6), 882-889.

Ajzen, I. (2011). Constructing a theory of planned behavior questionnaire.Unpublished manuscript. Retrieved, 1.

Ajzen, I. (1991). The theory of planned behavior. Organizational behavior and human decision processes, 50(2), 179-211.

Bachman, J. G., Smith, C. G., & Slesinger, J. A. (1966). Control, performance, and

satisfaction: an analysis of structural and individual effects. Journal of personality and

social psychology, 4(2), 127.

Barclay, L. J., Skarlicki, D. P., & Pugh, S. D. (2005). Exploring the role of emotions in injustice perceptions and retaliation. Journal of Applied Psychology, 90(4), 629. Bennett, R. (1998). Shame, guilt and responses to non-profit and public sector

ads. International Journal of Advertising, 17, 483-500.

Buchanan, T. W., Lutz, K., Mirzazade, S., Specht, K., Shah, N. J., Zilles, K., & Jäncke, L. (2000). Recognition of emotional prosody and verbal components of spoken language: an fMRI study. Cognitive Brain Research, 9(3), 227-238.

Burke, M. C., & Edell, J. A. (1989). The impact of feelings on ad-based affect and cognition. Journal of Marketing Research, 69-83.

Bucy, E. P. (2000). Emotional and evaluative consequences of inappropriate leader displays. Communication Research, 27(2), 194-226.

Carver, C. S., & Harmon-Jones, E. (2009). Anger is an approach-related affect: evidence and implications. Psychological bulletin, 135(2), 183.

Clark, M. S., & Taraban, C. (1991). Reactions to and willingness to express emotion in communal and exchange relationships. Journal of Experimental Social

Psychology, 27(4), 324-336.

Cole, P. M., Martin, S. E., & Dennis, T. A. (2004). Emotion regulation as a scientific construct: Methodological challenges and directions for child development research. Child development, 75(2), 317-333.

(26)

Côté, S., & Hideg, I. (2011). The ability to influence others via emotion displays A new dimension of emotional intelligence. Organizational Psychology Review,1(1), 53-71 Dillard, J. P., & Nabi, R. L. (2006). The persuasive influence of emotion in cancer prevention

and detection messages. Journal of Communication, 56(s1), S123-S139.

Dillard, J. P., & Peck, E. (2000). Affect and Persuasion Emotional Responses to Public Service Announcements. Communication Research, 27(4), 461-495.

Dillard, J. P., Plotnick, C. A., Godbold, L. C., Freimuth, V. S., & Edgar, T. (1996). The Multiple Affective Outcomes of AIDS PSAs Fear Appeals Do More Than Scare People. Communication Research, 23(1), 44-72.

Ekman, P., Friesen, W. V., & Hager, J. C. (2002). FACS investigator’s guide.A human face. Ekman, P., & Friesen, W. V. (1971). Constants across cultures in the face and

emotion. Journal of personality and social psychology, 17(2), 124.

Fast, N. J., Halevy, N., & Galinsky, A. D. (2012). The destructive nature of power without status. Journal of Experimental Social Psychology, 48(1), 391-394.

Fischer, A. H., & Roseman, I. J. (2007). Beat them or ban them: the characteristics and social functions of anger and contempt. Journal of personality and social psychology, 93(1), 103.

Frijda, N. H., Kuipers, P., & Ter Schure, E. (1989). Relations among emotion, appraisal, and emotional action readiness. Journal of personality and social psychology, 57(2), 212. Harmon-Jones, E., & Allen, J. J. (1998). Anger and frontal brain activity: EEG asymmetry

consistent with approach motivation despite negative affective valence. Journal of personality and social psychology, 74(5), 1310.

Howard, D. J., & Gengler, C. (2001). Emotional contagion effects on product attitudes. Journal of Consumer Research, 28(2), 189-201.

Kang, Y., Cappella, J. N., Strasser, A. A., & Lerman, C. (2009). The effect of smoking cues in antismoking advertisements on smoking urge and psychophysiological

reactions. Nicotine & Tobacco Research, ntn033.

Kim, K., & Lowry, D. T. (2005). Television commercials as a lagging social indicator: Gender role stereotypes in Korean television advertising. Sex Roles,53(11-12), 901-910.

Knobloch‐Westerwick, S. (2012). Selective exposure and reinforcement of attitudes and partisanship before a presidential election. Journal of Communication, 62(4), 628-642.

(27)

Kopelman, S., Rosette, A. S., & Thompson, L. (2006). The three faces of Eve: Strategic displays of positive, negative, and neutral emotions in negotiations.Organizational Behavior and Human Decision Processes, 99(1), 81-101.

Labott, S. M., Martin, R. B., Eason, P. S., & Berkey, E. Y. (1991). Social reactions to the expression of emotion. Cognition & Emotion, 5(5-6), 397-417.

Lelieveld, G. J., Van Dijk, E., Van Beest, I., & Van Kleef, G. A. (2012). Why Anger and Disappointment Affect Other’s Bargaining Behavior Differently The Moderating Role of Power and the Mediating Role of Reciprocal and Complementary

Emotions. Personality and Social Psychology Bulletin, 38(9), 1209-1221. Lelieveld, G. J., Van Dijk, E., Van Beest, I., Steinel, W., & Van Kleef, G. A. (2011).

Disappointed in you, angry about your offer: Distinct negative emotions induce concessions via different mechanisms. Journal of Experimental Social

Psychology, 47(3), 635-641.

Lewis, K. M. (2000). When leaders display emotion: How followers respond to negative emotional expression of male and female leaders. Journal of Organizational Behavior, 21(2), 221-234.

Lindebaum, D., & Cartwright, S. (2011). Leadership effectiveness: the costs and benefits of being emotionally intelligent. Leadership & Organization Development

Journal, 32(3), 281-290.

Madera, J. M., & Smith, D. B. (2009). The effects of leader negative emotions on evaluations of leadership in a crisis situation: The role of anger and sadness. The Leadership Quarterly, 20(2), 103-114.

McKee, S. A., Wall, A. M., Hinson, R. E., Goldstein, A., & Bissonnette, M. (2003). Effects of an implicit mood prime on the accessibility of smoking expectancies in college

women. Psychology of Addictive Behaviors, 17(3), 219.

Meyers, L.S., Gamst, G., & Guarino, A. (2006). Applied multivariate research: Design and interpretation. Thousand Oaks, CA: Sage Publishers.

Muller, D., Judd, C. M., & Yzerbyt, V. Y. (2005). When moderation is mediated and mediation is moderated. Journal of personality and social psychology,89(6), 852. Neuberg, S. L., & Newsom, J. T. (1993). Personal need for structure: Individual differences in

the desire for simpler structure. Journal of Personality and Social Psychology, 65(1), 113.

O'Keefe, G. J., & Reid, K. (1990). The Uses and Efiects of Public Service Advertising. Journal of Public Relations Research, 2(1-4), 67-91.

(28)

Petty, R. E., Schumann, D. W., Richman, S. A., & Strathman, A. J. (1993). Positive mood and persuasion: Different roles for affect under high-and low-elaboration

conditions. Journal of personality and social psychology, 64(1), 5. Pinto, M. B., & Priest, S. (1991). Guilt appeals in advertising: An exploratory

study. Psychological reports, 69(2), 375-385.

Pollay, R. W. (1986). The distorted mirror: Reflections on the unintended consequences of advertising. The Journal of Marketing, 18-36.

Roseman, I. J., Wiest, C., & Swartz, T. S. (1994). Phenomenology, behaviors, and goals differentiate discrete emotions. Journal of Personality and Social Psychology, 67(2), 206.

Roskos-Ewoldsen, D. R., Bichsel, J., & Hoffman, K. (2002). The influence of accessibility of source likability on persuasion. Journal of Experimental Social Psychology, 38(2), 137-143.

Sinaceur, M., & Tiedens, L. Z. (2006). Get mad and get more than even: When and why anger expression is effective in negotiations. Journal of Experimental Social

Psychology, 42(3), 314-322.

Sheppes, G., Scheibe, S., Suri, G., & Gross, J. J. (2011). Emotion-regulation choice. Psychological Science, 0956797611418350.

Shields, S. A. (2005). The Politics of Emotion in Everyday Life:" Appropriate" Emotion and Claims on Identity. Review of General Psychology, 9(1), 3.

Small, D. A., & Verrochi, N. M. (2009). The face of need: Facial emotion expression on charity advertisements. Journal of Marketing Research, 46(6), 777-787.

Stegge, H., & Terwogt, M. M. (2007). Awareness and regulation of emotion in typical and atypical development. Handbook of emotion regulation, 269-286.

Strasser, A. A., Cappella, J. N., Jepson, C., Fishbein, M., Tang, K. Z., Han, E., & Lerman, C. (2009). Experimental evaluation of antitobacco PSAs: Effects of message content and format on physiological and behavioral outcomes.Nicotine & Tobacco Research, ntn026.

Steinel, W., Van Kleef, G. A., & Harinck, F. (2008). Are you talking to me?! Separating the people from the problem when expressing emotions in negotiation. Journal of Experimental Social Psychology, 44(2), 362-369.

Sutfin, E. L., Szykman, L. R., & Moore, M. C. (2008). Adolescents' responses to anti-tobacco advertising: exploring the role of adolescents' smoking status and advertisement theme. Journal of health communication, 13(5), 480-500.

(29)

Tiedens, L. Z. (2001). Anger and advancement versus sadness and subjugation: the effect of negative emotion expressions on social status conferral. Journal of personality and social psychology, 80(1), 86.

Tiedens, L. Z., Ellsworth, P. C., & Mesquita, B. (2000). Sentimental stereotypes: Emotional expectations for high-and low-status group members.Personality and Social

Psychology Bulletin, 26(5), 560-575.

Truong, K. P., & Raaijmakers, S. (2008). Automatic recognition of spontaneous emotions in speech using acoustic and lexical features. In Machine Learning for Multimodal

Interaction (pp. 161-172). Springer Berlin Heidelberg.

Van Doorn, E. A., van Kleef, G. A., & van der Pligt, J. (2015). How emotional expressions shape prosocial behavior: Interpersonal effects of anger and disappointment on compliance with requests. Motivation and Emotion, 39(1), 128-141

Van Kleef, G. A. (2014). Understanding the positive and negative effects of emotional expressions in organizations: EASI does it. Human Relations, 0018726713510329. Van Kleef, G. A. (2009). How emotions regulate social life the emotions as social information

(EASI) model. Current Directions in Psychological Science,18(3), 184-188

Van Kleef, G. A., & Côté, S. (2007). Expressing anger in conflict: when it helps and when it hurts. Journal of Applied Psychology, 92(6), 1557.

Van Kleef, G. A., De Dreu, C. K., & Manstead, A. S. (2010). An interpersonal approach to emotion in social decision making: The emotions as social information

model. Advances in experimental social psychology, 42, 45-96.

Van Kleef G.A., De Dreu C.K.W., Manstead A.S.R. (2004a). The interpersonal effects of anger and happiness in negotiations. Journal of Personality and Social Psychology, 86, 57–76.

Van Kleef G.A., De Dreu C.K.W., Manstead A.S.R. (2004b). The interpersonal effects of emotions in negotiations: A motivated information processing approach. Journal of Personality and Social Psychology, 87, 510–528.

Van Kleef, G. A., Homan, A. C., Beersma, B., Van Knippenberg, D., Van Knippenberg, B., & Damen, F. (2009). Searing sentiment or cold calculation? The effects of leader

emotional displays on team performance depend on follower epistemic motivation. Academy of Management Journal, 52(3), 562-580.

Van Kleef, G. A., Van Doorn, E. A., Heerdink, M. W., & Koning, L. F. (2011). Emotion is for influence. European Review of Social Psychology, 22(1), 114-163.

(30)

Wild, B., Erb, M., & Bartels, M. (2001). Are emotions contagious? Evoked emotions while viewing emotionally expressive faces: quality, quantity, time course and gender differences. Psychiatry research, 102(2), 109-124.

Warner, L. R., & Shields, S. A. (2009). Judgements of others’ emotional appropriateness are multidimensional. Cognition and Emotion, 23(5), 876-888.

(31)

Appendix

Referenties

GERELATEERDE DOCUMENTEN

Figure 4.9: Estimates using importance sampling with a uniform distribution over Euler angles (red), a uniform distribution over quaternions (blue) and an almost uniform

Since schools tend to be the first environment where children with complex learning difficulties and disabilities are identified (Millar, Thompson, Schwab, Hanlon-Dearman,

De factoren die het meeste invloed lijken uit te oefenen op de uitvoering van het rolstoelenbeleid zijn: starheid, flexibiliteit en het beleidsnetwerk van de gemeenten,

In order to safeguard depth and quality in such a ‘firehose society’ in which focused and prolonged attention is the exception rather than the norm, BMS learning research needs

In Almería wordt zowel bij tomaat, paprika als komkommer naar schatting drie tot vier keer meer werkzame stof per m 2 kas verbruikt dan in Nederland.. Bij tomaat en kom- kommer

Andere deelprogramma’s zijn nog niet begonnen met kosten-baten analyses en voor de fase van mogelijke strategieën (gereed: voorjaar 2012) gaat dit ook niet meer lukken.. FASE

To sum up, the present research aims to show that leaders with high levels of SDO tend to abuse their power in order to protect their social status or power positions and maintain

To conclude, by showing that power has a negative relationship with COIs, this study is able to contribute to the literature focusing on the positive social effects that power can