Master Thesis
This Is Not Funny: the Use of Self-deprecating Humor
in Crisis Communication
Name:
Emilie Holtbach
Student nr: 5822491
Date:
September 8
th2014
Supervisor: Dr. ir. P.W.J. Verlegh
Index
Index ... 2
Abstract ... 4
Acknowledgements ... 6
Introduction ... 7
Organizational reputation and crisis responses ... 9
Self-deprecating humor ... 10
Perceived human character ... 11
Method ... 13 Design ... 13 Participants ... 13 Procedure/Stimuli ... 13 Measures ... 15 Attitude ... 15 Organizational reputation ... 15 Expectancy violation ... 15 Perceived humor ... 16
Organizational Human character ... 16
Perceived bravery ... 16 Results ... 16 Manipulation checks ... 16 Information ... 16 Expectancy violation ... 17 Perceived Humor ... 18
Presence of self-deprecating humor ... 19
Control questions ... 19
Previous knowledge about negative critique on previous Smart models ... 19
Familiarity with the advertisement ... 20
Hypotheses testing ... 20
Effect of advertising strategy on organizational reputation ... 20
Corporate Competence ... 20
Corporate Agreeableness ... 22
Mediation of Perceived Human Character Moderated by Information ... 24
Corporate Competence ... 26 Corporate Agreeableness ... 29 Additional measures ... 30 Perceived bravery ... 30 Discussion ... 31 Summary of finding ... 31 Study 2 ... 34 Method ... 35
Design & procedure ... 35
Design ... 35 Participants ... 35 Procedure/Stimuli ... 35 Measures ... 37 Organizational reputation ... 37 Expectancy violation ... 37 Perceived humor ... 38
Perceived bravery ... 38
Results ... 38
Manipulation Checks ... 38
Presence of self-deprecating humor ... 38
Crisis information (only two experimental groups) ... 39
Control questions ... 39
Joke in the advertisement ... 39
Previous knowledge of the Apple bar incidents ... 40
Familiarity with the advertisement ... 40
Research questions ... 41 Expectancy violation ... 41 Perceived Humor ... 41 Human character ... 42 Corporate Competence ... 42 Corporate agreeableness ... 43 Discussion ... 45 Summary of findings ... 45 General Discussion ... 47 Summary of findings ... 47
Contribution to the literature ... 48
Contribution to practice ... 49
Limitations study 1 and directions for further research ... 49
Limitations study 2 and directions for further research ... 50
General limitations and directions for future research ... 51
References ... 53
Appendices ... 59
A. Stimulus material - Smart commercials ... 59
B. Stimulus material - Apple commercial ... 59
C. Questionnaire study 1: Smart ... 59
Abstract
As crisis-response strategies are activities aimed at maintaining a favorable reputation (Coombs, 2006), it would seem paradoxical that an organization engages in a rhetorical practice directed at deprecating itself with humor (Drewberry & Fox, 2012). Self-deprecating humor emphasizes salient aspects that are criticized by others, by using these aspects as the target of a joke. Self-deprecating humor can be face saving through which an organization can be seen in a more favorable light. But how does it work, and under what
conditions do the positive effects of self-deprecating humor appear?
This study proposes, that since people are aware that organizations are usually keen on maintaining a favorable reputation, self-deprecation by an organization violates expectations of the audience, which will be perceived as humorous by the audience (McGraw & Warren, 2006), and consequently has a positive impact on evaluations of the organization. Additionally, it is proposed that due to the characteristics of this type of humor directed at a true weakness of the organization, the public will perceive the organization as having a human character, which is beneficial for the organizational perceptions (Kelleher, 2009). Next, a possible requirement of information about the organizational crisis prior to exposure to the self-deprecating humor advertisement for the positive effect to occur is studied.
The paper describes two experimental studies. The first study has a 2 x 3 design in which crisis information (negative vs. neutral) is examined as a potential moderator of the effect of self-deprecating humor, while comparing the effect of self-deprecating humor to two control conditions: other-deprecating humor and no humor (N=218). Four dependent variables were measured (expectancy violation, perceived humor, perceived organizational human character, and organizational reputation). Analyses of variance and mediation analyses were employed to gain information about the underlying mechanisms of the effect of self-deprecating humor on organizational reputation. The second study has a 1 factorial between-subjects design with crisis information (not provided vs. provided beforehand vs. provided afterwards) as independent variable (N=102). The same dependent variables as in study 2 were measured. Analyses of variance were executed to investigate the potential requirement of knowledge about the crisis for the positive effect of self-deprecating humor to occur.
Results from study 1 confirm the assumption that self-deprecating humor is a powerful strategy after a crisis situation. Self-deprecating humor has a stronger positive effect than other-deprecating humor and no humor on the organizational reputation, and only after the negative information.Additionally, a formal mediation analysis indicated that perception of organizational human character mediates the relationship between self-deprecating humor and organizational reputation, but only after the negative information. This effect is not found for other-deprecating humor; other-deprecating does not lead to a higher perception of organizational human character. Results of study 2 show that the prior information that people have about the organizational crisis positively influences their evaluation of the self-deprecating humor advertisement, their perception of organizational human character and the organizational reputation. People who first received crisis information and then see the advertisement have more violated expectancies, indicate more perceived
humor as well as more organizational human character than participants who received no crisis information or received the crisis information afterwards. Furthermore, their evaluation of the organizational reputation is more positive than that of people who have no crisis information or received crisis information afterwards.
The key contributions of the paper include expanding self-deprecating humor theories in political communication research and social sciences studies to the field of organizational crisis communication and identifying the underlying mechanism of the effect of self-deprecating humor on organizational reputation. Self-deprecating humor was found to mitigate the negative effects of an organizational crisis on the
organizational reputation. More specifically, when people understand the reason for the use of self-deprecating humor prior to exposure to the humorous self-self-deprecating advertisement, they consider the organization as having a stronger human character and as having stronger positive organizational
characteristics. To conclude, an organization can save its face by making a joke about a genuine weakness of itself as an attempt to change a crisis situation from damage to sympathy.
Keywords: organizatonal reputation, crisis response strategies, self-deprecating humor, expectancy
Acknowledgements
A group of inspiring and supportive people surrounds me, and for that I feel fortunate. I would like to use this opportunity to acknowledge these special people here.
My supervisor Peeter Verlegh has welcomed my enthusiasm about the topic of self-deprecating humor in advertising with open arms. He gave me lots of freedom in my ideas and approach on how to study the positive effect of self-deprecating humor and intervened just at the right times to push me in the right direction. I have grown as a scholar under his tutelage.
Also, a very special thanks to the thesis supervisor of my Bachelors, who also embodies the type of scholar I strive to be. I would like to thank my roommates, who have been so patient during this process and always wished me good luck. I would like to thank my brother, sister, and my close friends, who were all very encouraging. Furthermore, I would like to thank a couple of special people that told me I had twinkles in my eyes when talking about my experiments.
Finally, I would like to acknowledge my parents. My mother supports me every single day to go the extra mile to achieve my goals. My father always supported me to use my full potential and inspired me to think out of the box. I know he would have been extremely proud. I would never be where I am today without my parents’ love and support.
Introduction
Today, it is hard to imagine our streets without Smart cars - the compact cars with seating for two. A great idea led to a small and stylish city car that continues to be a success on the road. However, the development of the car did not go without a struggle. The car was launched successfully in 1998, but the final design had been through many teething troubles. The so-called ‘first generation’ cars had some severe shortcomings. For example, the engine of the car was of inferior quality and Smart started suffering a bad reputation, as many cars needed a new engine after they had driven only 40.000 kilometers. Now, if a company has faced severe troubles and has had a hard time competing in the automobile industry, there are some advertising strategies that can be considered in order to strengthen the company’s reputation in the eyes of the public. One might be to make a commercial which shows in a funny way that a compact car is better than bigger cars which convinces people of the advantages of buying that compact city car. Another might be to make a commercial which jokes about the development problems and charms the public by showing the human character existing behind the organizational façade. Smart did both, and this study examines the power of self-deprecating humor versus other-deprecating humor (relative to a control condition without humor) on the reputation of a company after an organizational crisis.
Organizational crisis-response strategies are goal-directed activities that seek to maintain a favorable reputation (Coombs, 2006). Therefore, it seems paradoxical that an organization would want to use a
rhetorical practice aimed at deprecating itself with humor even more (Drewberry & Fox, 2012). Nevertheless, self-deprecating humor is a strategy that is increasingly used in advertising after an
organizational crisis (“Self-deprecating humor: a trend in advertising”, 2006). This involves emphasizing those aspects that have been criticized by others, by using these aspects as the target of a joke (Greengross & Miller, 2008). Before, little research has been carried out into the impact of self-deprecating humor by an organization after a crisis. However, several organizations (e.g. Hema, Microsoft, Smart Automobile, Apple) have successfully engaged in comic self-deprecation in response to their organizational crises. Therefore, it is essential to recognize the rhetorical force of self-deprecation as a crisis response strategy (Drewberry & Fox, 2012).
After an organization has encountered a crisis situation in which the face of the organization was threatened and vulnerability was high (Sommer & Pearson, 2007), self-deprecating humor can be face saving (Drewberry & Fox, 2012). The rhetorical strategy bears a circular message; the organization admits a
weakness by means of a joke, admitting a weakness means to be strong, so the organization is strong (Zadjman, 1995). This study proposes, that because people know that an organization will want to repair its reputation after a crisis, a self-deprecation strategy will violate their expectations and they will perceive the self-deprecation as humorous (McGraw & Warren, 2006). Consequently, this has a positive impact on how they evaluate the organization. Additionally, it is proposed that due to the characteristics of this type of humor directed at a true weakness of the organization, the public will perceive the organization as having a human character, which in turn benefits their perception of the organization (Kelleher, 2009). Alternately,
other-deprecating humor, directed at something other than the organization or the crisis can be used to distract attention away from the problem or to evoke a positive response. It does not, however, have a further strategic function in repairing the organization’s reputation (Drewberry & Fox, 2012).
The purpose of this research is to investigate the persuasive effects of self-deprecating humor in advertising on organizational evaluations after a crisis and to identify the underlying mechanism of this rhetorical strategy. Up to now there has been no scientific research on the effects of self-deprecating humor in advertising, however researchers did study this communication tactic that is frequently used by U.S. politicians after a personal crisis. Such studies have shown for instance that self-deprecating humor can be a means of image restoration (Drewberry & Fox, 2012), that it is perceived as a sign of goodwill (Bippus, 2007) and that it generates a positive mood among message recipients (Stewart, 2011). In this study, theories concerning self-deprecating humor that have been developed in political communication and social sciences studies are applied to organizational communication to examine the effect of a self-deprecating humor advertisement after an organizational crisis.
While the use of different primary crisis-response strategies aimed at diminishing the negative effects of the crisis are discussed comprehensively (e.g. Fuchs-Burnett, 2002; Coombs, 2006), research about secondary crisis-response strategies aimed at counterbalancing the negative perceptions from the crisis is less extensive. This paper contributes to the existing literature by experimentally examining whether a self-deprecating humor ad following an organizational crisis can counterbalance the negative effects of the crisis on the reputation of the organization. As ‘reputation’ is a fragile multi-dimensional construct based on beliefs about organizational characteristics (Hall, 1993), it is important to identify a strategy that has the power to mitigate the negative effects of the crisis on the organizational reputation. By comparing the effects of an advertisement including self-deprecating humor, an advertisement with other-deprecating humor, and one with no humor at all, the hypothesized persuasive effects of self-deprecating humor can be unraveled. The research questions that will be studied are as follows:
What is the influence of the use of self-deprecating humor in advertising after an organizational crisis on the reputation of the organization, compared with the use of other-deprecating humor and no humor?
To what extent is this influence mediated by the perception of the human characteristics of an organization?
Organizational reputation and crisis responses
To make investment decisions, career decisions and product choices, stakeholders routinely rely on the reputation of organizations (Dowling, 1986). Reputation is often regarded as a valuable, intangible asset relevant for the financial success of the organization (Fombrun & Van Riel, 2004) and is a multidimensional evaluation of the performance and behavior of the company and the ways in which it affects the community (Boulstridge & Carrigan, 2000). Communication plays a crucial role in defining the organizational
reputation, since the public constructs reputations from information originating from interactions with the company or other sources like the media (Coombs & Holladay, 2006).
An organizational crisis can be defined as a specific, unexpected and non-routine event or series of events that create high levels of uncertainty and threaten, an organization’s high priority goals (Seeger, Sellnow, & Ulmer, 2003). The actions taken by organizations affect how the public perceives the
organization and the crisis. In turn, those perceptions shape evaluations of the organizational reputation as well as stakeholders’ emotional response towards and future interactions with the organization (Nerb & Spada, 1997).
As the prospect of losing any value of organizational reputation is extremely costly to an organization (Fombrun & Van Riel, 2004), organizations engage in primary activities to repair their reputation, to reduce negative affect, and to prevent negative behavioral intentions (Coombs, 2007). Most crisis literature has merely focused on primary crisis-response strategies (Coombs, 2000) and image-repair strategies (Benoit, 1997) to handle organizational crisis situations and related negative publicity. After the organization has taken care of those duped by the crisis, organizations can use secondary crisis-response strategies to improve relationships with stakeholders, generate goodwill, and counter-balance the negative perceptions arising from the crisis (Coombs, 2007). However, there is little theoretical research about how organizations can develop secondary crisis-response strategies to combat the negative effects of the crisis (Ahluwalia, Burnkrant & Unnava, 2000).
What is clear is that organizational crises and related negative publicity can trigger negative
emotions such as anger, sadness, anxiety or fright, among those directly involved and also among the general public (Jin, Pang, & Cameron, 2007). Emotions regarding a product recall are probably less extreme than in cases when the crisis concerns fatalities, but they will still influence how individuals cognitively and affectively evaluate the organization. Therefore, a primary crisis response should not only include informational repair efforts (e.g. updated information about the crisis) and functional repair efforts (e.g. financial compensation), but also affective repair efforts (e.g. apology) through which compassion and remorse for the negative consequences are expressed (Xie & Peng, 2009).
However, despite the primary crisis-response activities and a decrease in negative publicity, negative emotions can continue over time and carry over from past situations to color future judgments. According to the Appraisal-Tendency Framework, emotions cause an implicit ‘cognitive predisposition’ to appraise future events in line with the central appraisal patterns that characterize the emotions (ATF; Lerner & Keltner,
2000). The ATF provides a useful insight for examining how the public’s affective response to an
organizational crisis influences its subsequent judgment of the organization. Through this appraisal tendency, the negative emotions might continue damaging the reputation of the organization. Therefore, a secondary crisis-response strategy in the form of a post-crisis campaign that evokes positive evaluations about the organization could be effective in breaking through the appraisal tendency of the public and could help to counterbalance the negative emotions arising from the crisis.
The current study examines the appropriateness and power of self-deprecating humor in eliciting positive emotions and evaluations to boost the reputation-healing process of an organization, as the characteristics of this type of humor are expected to turn the appraisal tendency from negative to positive. Contrastingly, it is expected that other-deprecating humor is not strong enough to break through the negative appraisal pattern.
Self-deprecating humor
Almost all forms of humor involve ridicule of something – a person, behavior, belief, group, or possession – at some level (Stewart, 2011). Self- and other-deprecating humor is usually directed towards a specific person (oneself or another) and concerns, for example, their behaviors, utterances, personality traits, or competences (Janes & Olson, 2000).
Self-deprecating humor is a kind of humor that can arouse positive emotions, but it bears more cognitive, strategic advantages compared to humor that does not have a major purpose other than inducing in people a playful state of being (Zajdman, 1995; Drewberry & Fox, 2012). According to the Expectancy Violations Theory (EVT; Cohen 2010), people bring expectancies to communicative acts, anticipating the way a particular group or person will behave in communication. When expectancies are positively violated, the unexpected change of perspective evokes humor (Zijderveld, 1996) and results in an improved evaluation of the communicator. Therefore, acting contrary to what is generally expected of organizations in crisis situations (e.g. managing a positive image) can produce better evaluations of an organization. However, self-deprecating humor can establish more positive effects than merely heightening the perceived humor just as other-deprecating humor can, as it can be used rhetorically for some serious strategic purposes.
First, when ‘the self’ is the target of the joke, the organization attempts to ‘laugh with’ the public instead of being ‘laughed at’ by the public (Martin, 2007). Consequently, by ‘attacking’ the self in the case of a mistake by the organization itself, the laugh of the public may discharge the publics’ hostility towards the organization (Zajdman, 1995) so that they don’t use their ‘weapon’. Second, as self-deprecating humor, in contrast to other-deprecating humor, does acknowledge the weakened position of the communicator, it expresses character (Gruner, 1970) and signals goodwill (Bippus, 2007). A communicator directing humor at himself is usually taken for a courageous person, not afraid to publicly uncover his own weaknesses
(Zajdman, 1995). The organization dares to ridicule its own behavior as a way of coping with the crisis by admitting, rather than denying, the existence of the crisis, which indicates strength, bravery and recovery (Zajdman, 1995). Third, by admitting the obvious instead of covering it up, the organization is admitting
nothing, but it gains back credibility (Buchanan, 2004). As the admission is against self-interest, this results in enhanced trustworthiness (Combs & Keller, 2010); it creates a perception of honesty that makes all subsequent information seem credible to observers (Cialdini, 2001).
These three distinct functions of self-deprecating humor compared to other-deprecating humor, emphasize that self-deprecating humor has the ability to positively affect the publics’ evaluations of the organization. Consequently, it is proposed that self-deprecating humor will lead to a more positive reputation than other-deprecating humor. To be able to examine the positive effect and underlying mechanism of the use of self-deprecating humor by an organization after a crisis, an experiment is executed in which a crisis situation of the car company Smart is experimentally manipulated by providing participants with very negative information (vs. neutral information) about a specific car model. After the participants have read negative information about the car, they are exposed to an advertisement made by the company to
counterbalance the previous negative critique on the car, containing self-deprecating humor (vs. other-deprecating humor vs. no humor). Accordingly, it is proposed that:
H1a: The self-deprecating humor ad will lead to a more positive organizational reputation than the other-deprecating humor ad.
H1b: The other-deprecating humor ad will lead to a more positive organizational reputation than the no-humor ad.
H2: The positive effect of the self-deprecating humor on the organizational reputation will be stronger after the negative information rather than after the neutral information.
Perceived human character
When self-deprecating humor is used by an organization, it can play an important role in communicating the human characteristics existing behind an organizational façade. Firstly, the organization will be perceived as self-aware as it has insight into itself (Zajdman, 1995) and is willing to critically assess the shortcomings of the organization (Drewberry & Fox, 2012). Secondly, the self-deprecating humor conveys the
acknowledgment of a weakness, modesty, and self-disclosure, which can be important for rapprochement and repair of trust (Niwa & Maruno, 2009). Presenting the deprecated self communicates that the
organization is not perfect and manages the organization’s impressions as easy to approach (Niwa & Maruno, 2009). Self-deprecating humor makes the audience more likely to lower their guard and increases the intimacy towards the organization, allowing the organization to come closer to the public. Thirdly, self-deprecating humor will give the organizational image a more human appearance (Kelleher & Miller, 2006) in the fact that it is resistant to jokes just like humans. The organization will appear to be willing to poke fun at itself through which it will look more lifelike. Fourthly, the use of self-deprecating humor can be seen as an indicator of conversational human voice (Kelleher, 2009) which includes characteristics of
communication that are not associated with traditional corporate communication such as a sense of humor or admitting mistakes (Searls & Weinberger, 2001).
The use of self-deprecating humor can be powerful during a crisis because it allows organizations to display the above described human characteristics that have value in efforts aimed at building and
maintaining relationships with stakeholders (Kelleher, 2009). Therefore, the overarching concept of ‘organizational human character’ is used to study the underlying dynamics of self-deprecating humor. This research proposes that the more self-deprecating humor is successful because of the violation of expectancies that break through the negative appraisal tendency, the more humor will message recipients perceive and as the humor is directed at the organization itself, the more they will evaluate the organization as having a human character. Positive perceptions generated by recognizing the organization as having a human
character should transfer to subsequent evaluations of the organization. As a result, the message recipients of self-deprecating humor should have a more positive perception of the organizational reputation. In line with this reasoning, the following hypotheses are formulated:
H3: The positive effect of the self-deprecating humor ad on the organizational reputation is mediated by the higher perception of organizational human character, whereas this effect and mediation is less strong for the other ads.
H4: This mediation effect of perceived human character on the organizational reputation will be stronger after the negative information than the neutral information.
In considering the theory and hypotheses together, this study puts forth the following conceptual model:
Figure 1. Conceptual model
Repairing
Advertising
Strategy
Reputation
Perceived
Human
Character
Information
H1a
H1b
H2
H3
H4
H3
Method
DesignThe experiment has a 2 x 3 factorial between-subjects design with information (negative/neutral) and repairing advertising strategy (self-deprecating humor strategy/humor strategy/no-humor strategy) as independent variables.
Participants
The study has a convenience sample of 218 participants, so the sample cannot be treated as representative of the Dutch population. The data collection for study 1 and 2 started simultaneously, as Qualtrics Survey tool assigned the participants randomly to the conditions of study 1 or study 2 (N=102). Of every three
participants who clicked on the link to the questionnaire, two were directed to study 1 and one was directed to study 2. In total 527 participants started the questionnaire but only 320 participants completed the whole questionnaire (total of study 1 and 2). Friends and relatives were asked to participate by a private request via e-mail or a Facebook message in which the link to the questionnaire was included. Also four requests in the form of persuasive texts and photos were posted on the Facebook wall of the researcher to stimulate
Facebook friends to complete the questionnaire. A total of 758 Facebook friends were reached, and some close friends shared the posts with their friends. Moreover, around 120 people were reached by an e-mail that was sent to friends of the researcher’s family. After nine days, 103 men (47.2%) and 115 (52.8%) women completed the questionnaire of study 1. The age of the participants varied between 17 and 66 years (M= 28.95, SD = 12.24 years).
Procedure/Stimuli
The study took approximately five minutes to complete. Participants could click on a link to the online questionnaire and were randomly assigned to one of the six conditions. First, participants were asked to watch a short video that shows a toucan bird and are asked which sound they hear, to make sure their audio is working well (they should hear the sound of bells ringing). After a short instruction about the
questionnaire, all participants received a short introduction to the Smart car. They were told that the
manufacturer had made improvements to the first Smart after its launch in 1998. To give the participants an impression of the car, they read a review of the second-generation Smart on the next page. The subsequent review of the model Smart ForTwo was experimentally manipulated by either providing negative
information (crisis condition) or neutral information (control condition).
The participants were shown a realistic webpage of Auto Magazine and read a (fictitious) review of the 2007 Smart ForTwo model. Participants in the negative information condition read the critical results of a car evaluation in which was stated that the car was performing dramatically outside the city and that it had several limitations even in the city. Participants in the neutral information condition read the same text but the negative criticism was replaced by mildly negative and positive information. The participants completed
a questionnaire to check the manipulation of the information, including questions concerning their attitude towards Smart, whether they thought the reputation of Smart has fallen into disrepute after the shortcomings of the first model, and whether they thought that it was appropriate to drive the car in the city and outside the city.
Next, participants read a short introduction about the 2011 launch and marketing strategy of the new Smart ForTwo Electric Drive in Germany and were instructed to watch a television commercial. Participants were randomly assigned to one of the three experimental conditions containing a different commercial (all the same length), which was part of Smart’s marketing campaign Open Your Mind. Participants in the self-deprecating humor condition received an advertisement in which the weak features highlighted in the previous review of the Smart ForTwo were the subject of the humor in the advertisement (N=72). The commercial shows the Smart ForTwo driving off-road and it looks as if Smart is trying to convince the public that the car is able to do the same as bigger cars. After a few seconds it becomes clear that the Smart is performing very poorly off-road. Several funny shots are shown in which the Smart fails to drive up to a steep hill, gets stuck on a big rock and is not able to cross a river. At the end of the commercial there is written: „As good offroad, as an offroader in the city”. Next, was shown what a Smart can and an offroader can’t do, namely parking very quickly and easily at a small parking lot in the city. The self-deprecating humor in the advertisement is directed at genuine drawbacks in using the Smart.
The participants in the humor condition received an advertisement in which the indirectly positive features of Smart were emphasized by making fun of the parking problem that people have with bigger cars in cities (N=74). The commercial shows multiple persons trying to park their normal-sized cars in reverse at the same small parking lot. A famous classical song of Andrea Bocelli is compiled of the annoying beeps of all the different parking sensors. On the faces of the persons that are parking you see frustration, people being angry at each other, or cursing and shouting because they fail to park. At the end of the commercial there is written „Bye Bye Beep Beep” and in the next shot one sees how easily the Smart ForTwo can park without any sound of a parking sensor. The humor is directed to a flaw of bigger cars and is triggered by the exaggeration and the recognizing of the frustration that comes with parking in reverse when the parking lot is very small.
The participants in the no-humor condition received an advertisement in which technology and innovation in general were emphasized and brought into relation with Smart, containing no humor at all (N=72). The commercial focuses on future developments in transport and future methods of transport. Different fictive transport inventions are shown, such a „zero-gravity board”, „hyperdrive”, „an aero broom”, „a teleporter" and a „sunspeed superfast”. Along with every fictive mode of transport, the year in which it will be available is given. At the end of the commercial the future development of the Smart ForTwo is presented; that it will be electrically driven, and will be available in 2012 (see appendix A for the links to the commercials). The accompanying music gives the commercial a serious appearance.
After exposure to the experimental stimuli, all participants completed the same follow-up
questionnaire, including questions concerning the reputation of Smart, the perceived human character of the organization and perceived bravery with regard to the commercial. The questions also included manipulation
checks on their attitude towards Smart, expectancy violation, and their perceived humor. Moreover, control questions about familiarity with the advertisements and the previous critique on the first Smart models, and demographic-variable questions were asked. The last page thanked the respondents.
Measures
AttitudeParticipants were asked to indicate their attitude towards Smart on 9-point bipolar scales along the following dimensions: unappealing/appealing, bad/good, unpleasant/pleasant, unfavorable/favorable,
unlikable/likable (Spears & Singh, 2012). All scale items were computed into one index (Cronbach’s alpha=
.915; M=5.05, SD=1.678).
Organizational reputation
As in most reputation research, the organizational reputation is measured on two main dimensions, corporate competence (cognitive component) and agreeableness (affective component) (Schwaiger, 2004). The scale used is based on the Corporate Character Scale (Davies, Chun, da Silva & Roper, 2004), which is
constructed from the idea that people imagine the corporation to be a person and therefore rate the corporate reputation by personal characteristics. The corporate competence dimension consists of eight items: reliable,
secure, hardworking, ambitious, achievement-orientated, leading, technical, and corporate. The corporate
agreeableness dimension, consists of 11 items: friendly, pleasant, open, straightforward, concerned,
reassuring, supportive, agreeable, honest, sincere, trustworthy, and socially responsible. Participants were
asked to indicate on seven-point bipolar scales (1 = highly inapplicable to 7 = highly applicable) to what extent they thought these characteristics are applicable to Smart. Higher values on both dimensions indicate a more favorable corporate reputation of Smart. The scale items of the corporate competence dimension were computed into one index (Cronbach’s alpha= .903; M=4.23 SD=1.12), as well as the scale items of the corporate agreeableness dimension (Cronbach’s alpha .920; M=4.55, SD= .99).
Expectancy violation
In order to assess expectancy violation participants rated the degree to which they were surprised by the advertisement and the extent to which the source’s behavior was unexpected. Both items are based on the Expectancy Violation Theory (Cohen, 2010) and measured on 9-point antonym scales along the following dimensions: surprising-unsurprising and expected-unexpected (Cronbach’s alpha= .845; M=5.66 SD=2.11).
Perceived humor
In order to assess perceived humor, participants rated the degree to which they found the advertisement humorous, funny and amusing on 9-point antonym scales. All three items were computed into one index (Cronbach’s alpha= .965; M=6.37, SD=1.90).
Organizational Human character
The item measuring organizational human character is ‘To what extent do you perceive the organization as
human?’. Participants were asked to rate the organizational human character on a 9-point bipolar scale (1 =
in a very small extent to 9 = in a very great extent), (M=5.71, SD=1.85).
Perceived bravery
In order to assess perceived bravery to launch an advertisement with self-deprecating humor, participants rated the degree to which they find the advertisement gutsy, strong and brave on 9-point antonym scales. All three items were computed into one index (Cronbach’s alpha= .877; M=5.65, SD=1.92).
Results
Manipulation checks Information
The manipulation of the negative evaluation was checked in an ANOVA with information as independent variable (negative vs. neutral) and attitude towards Smart as dependent variable. As expected, the attitude towards Smart of the participants who had read the negative information was less positive (M=4.28,
SD=1.50) than that of the participants who had read the neutral information (M=5.82, SD=1.50). This
difference was statistically significant, F(1, 216) = 57.412, p < 0.001, η2 = 0.21.
Moreover, three more questions to check the manipulation of the negative information were asked and assessed with independent samples t-tests with information as independent variable (negative vs. neutral). Firstly, participants indicated to what extent they thought the reputation of Smart had fallen into disrepute after the shortcomings of the first model were acknowledged. As expected, participants who had read the negative evaluation indicated more reputation damage (M= 5.52, SD=2.05) compared to participants who had read the neutral evaluation (M=4.08, SD=1.91). This difference was statistically significant,
F(1,216) =28.889, p < 0.001, η2 = 0.12). Secondly, the participants were asked whether they thought that the second generation of Smart was an appropriate car to drive in the city, and thirdly whether they thought that the car was appropriate to drive outside the city. The expectation was that after the negative information, participants would indicate that they found the car appropriate for use in the city (M=3.01, SD=1.62), but that they would find it less appropriate than the participants in the neutral condition (M= 3.85, SD=2.07). This expectation was confirmed by the results and the difference was statistically significant, F(1,216) =38.826, p
< 0.001, η2 = 0.05). Lastly, in line with expectations, participants in the negative condition indicated that they found the Smart less appropriate for driving out of the city (M=6.55, SD=1.79), and they found it significantly less appropriate than the participants in the neutral condition (M=7.83, SD=1.25), F(1,216) =88.628, p < 0.001, η2 = 0.007).
Expectancy violation
The manipulation of the self-deprecating humor and humor was checked in two ANOVAs with repairing advertising strategy (self-deprecating humor, humor and no humor) and information (negative vs. neutral) as independent variables and expectancy violation and perceived humor as dependent variables. The
expectations of the participants who saw the advertisement with humor were the least violated (M=4.28,
SD=1.93), while the expectations of the participants who saw the self-deprecating humor advertisement were
the most violated (M=6.99, SD=1.53). The participants who saw the no-humor advertisement, scored in between the others on the expectancy-violation variable (M=5.74, SD=1.90). These outcomes are not completely in line with the expectations, as it was expected that the no-humor advertisement would lead to the least expectancy violation. A possible explanation is that the participants in the no-humor condition interpreted the items of the expectancy violation scale differently. The no-humor commercial is about future ways of transport and contains several examples of creative means of transport, which could be perceived as surprising and unexpected. Expectancy violation was expected to be the least in the no-humor condition, as the commercial did not contain an unexpected change of perspective with regards to the previous information that the participants had read about the organization. Nevertheless, the fact that the self-deprecating humor has lead to the most expectancy violation is important for the manipulation (F (2, 212) = 41.76, p < 0.001, η2 = 0.28).
The post-hoc multiple comparisons test showed significant differences between the three different advertisements on the expectancy violation variable; between the participants who saw the no-humor advertisement and the humor advertisement (Mdiff= 1.46, p < 0.001), between the participants who saw the no-humor advertisement and the self-deprecating humor advertisement (Mdiff= -1.26, p < 0.001), and between the participants who saw the humor advertisement and the self-deprecating humor advertisement (Mdiff=-2.72, p < 0.001).
The expectations of the participants who read the negative information (M=5.65, SD=2.18) was not violated in a greater extent than the expectations of the participants who read the neutral information (M=5.67, SD=2.04), F (1, 212) = .0779, p=0.778. However, an interaction effect between the advertisement and the information on the expectancy violation was expected and was marginally significant, F (2, 212) = 2.45, p =0.089, η2 = 0.02). Participants who read the negative information were more surprised by the self-deprecating humor advertisement (M=7.27, SD=0.30) than the participants who read the neutral information (M=6.70, SD=0.30), however this difference was not statistically significant (Mdiff= .570, p=0.179). Also the means of the participants who saw the humor advertisement did not differ significantly between the negative (M= 4.26, SD=2.13) and neutral (M=4.29, SD=1.76) information groups (Mdiff= -.026, p=0.951). The expectations of the participants who saw the no-humor advertisement after the negative information were
less violated (M=5.36, SD=0.30) than the participants who read the neutral information (M=6.11, SD=0.30), and this difference was marginally significant (Mdiff= -0.75, p = 0.077). The manipulation of the three advertisements was successful, as the means of the different advertising strategy groups differed on the expectancy violation in the expected direction although there was no statistically significant difference in expectancy violation between the three advertisements in the negative and the neutral condition.
Perceived Humor
In line with the expectations, the participants who saw the advertisement with no humor perceived the least humor (M=5.58, SD=2.1), while the participants who saw the self-deprecating humor advertisement
perceived the most humor (M=7.15, SD=1.39). The participants who saw the humor advertisement, scored in between the others on the perceived humor variable (M=6.37, SD=1.83). These differences were statistically significant, F (2, 212) = 87.91, p < 0.001, η2 = 0.11. The post-hoc multiple comparisons test showed a significant difference between the participants who saw the no-humor advertisement and the humor advertisement (Mdiff= -0.79, p < 0.05), between the participants who saw the no-humor advertisement and the self-deprecating humor advertisement (Mdiff= -1.57, p < 0.001), and between the participants who saw the humor advertisement and the self-deprecating humor advertisement (Mdiff= -0.78, p < 0.05). Actually, it would have been better if there had been no difference between self-deprecating humor and the humor advertisement on the perceived humor variable, to be able to exclude the suggestion that the positive effect of self-deprecating humor is just a result of better humor. However, the post hoc analysis of the interaction between advertising strategy and information (see next paragraph) provides an explanation that these
differences are the result of the moderation effect of information. A significant difference was found between self-deprecating humor and humor within the negative information condition, whereas this difference was not present in the neutral condition. Thus, the advertising strategy manipulation was successful, because the self-deprecating humor group indicated they perceived more humor and their score differed significantly from the humor and no-humor group. Apparently this result is driven by the differences within the negative condition.
The perceived humor of the participants who read the negative information (M=6.28, SD=1.90) was slightly less than the perceived humor of the participants who read the neutral information (M=6.46,
SD=1.84), but this difference was not statistically significant (F (1, 212) = 1.926, p=0.440). Moreover, there
was no significant interaction effect between the advertising strategy and the information, F (2, 212) = 11.73,
p = 0.164. However, the post-hoc pairwise comparisons test showed that the differences between the three
advertisements were statistically significant within the negative condition; between the participants who saw the no-humor advertisement and the self-deprecating humor advertisement (Mdiff= -2.02, p < 0.001) and between the participants who saw the humor advertisement and the self-deprecating humor advertisement (Mdiff= -1.30, p < 0.01). However, the difference between the participants who saw the no-humor
advertisement and the humor advertisement was not significant (Mdiff= -.71, p=0.280). In the neutral condition, the difference between the participants who saw the no-humor advertisement and the
difference between the participants who saw the no-humor advertisement and the humor advertisement is significant (Mdiff= -.85, p = .127), nor the difference between the participants who saw the humor
advertisement and the self-deprecating humor advertisement was significant (Mdiff= -.26, p = 1.000). The manipulation of the three advertisements was successful, as the means of the different advertising strategy groups differed on the perceived humor variable in the direction as expected, even though not all
advertisements differed significantly from each other in terms of perceived humor.
Presence of self-deprecating humor
A question about the degree of self-deprecating humor in the advertisement was asked to the participants to assess whether the participants perceived self-deprecating humor in the advertisement of Smart. The presence of self-deprecating humor was checked in an ANOVA with advertising strategy (self-deprecating humor, humor and no humor) as independent variable and the self-deprecating humor question as dependent variable. In line with the expectations, as the advertisement exposed to the no humor and humor contained no self-deprecating humor, the difference between the mean of the self-deprecating humor group (M=7.89,
SD=1.04) and the mean of the no-humor group (M=3.63, SD=1.83) was large and highly significant (Mdiff =
4.26; p <0.001). Also, the difference between the mean of participants who saw the self-deprecating humor advertisement and the participants who saw the humor advertisement (M=3.86, SD=2.12) was large and highly significant (Mdiff = 4.02; p <0.001). The mean difference between the no humor and humor group was not significant (Mdiff = -0.24; p=1.000). The self-deprecating humor manipulation was successful, since the outcomes show the self-deprecating humor group had a strong opinion that the advertisement contained self-deprecating humor (F (2, 212) = 139.192, p < 0.001, η2 = 0.09). The participants who saw the no-humor and humor advertisements indicated less self-deprecating humor in the advertisement.
Control questions
Previous knowledge about negative critique on previous Smart models
Although the negative information about Smart was experimentally manipulated in this study, Smart indeed had problems with the first models of the car in real life. Therefore, participants were asked whether they already had knowledge about the negative critique on the previous Smart models. 72.2% of the participants in the self-deprecating humor condition indicated they had not heard of the critique (N=52), whereas 25.0% had heard of the criticism (N=18) (2 participants answered ‘don’t know’). 77.0% of the participants in the humor condition had not heard about the criticism (N=57), whereas 20.3% already had knowledge about the criticism (N=15) (2 participants answered ‘don’t know’). 69.4% of the participants in the no-humor
condition indicated they had not heard of the criticism (N=50), whereas 29.2% of the participants in the no- humor condition did know about the criticism (N=21) (1 participant answered ‘don’t know’). The
percentages of participants who did not know about the critique did not differ to a great extent, which was beneficial for the results of the experiment.
Familiarity with the advertisement
A control question about the advertisement was asked to assess whether the participants had already seen the advertisement. 94.4 % of the participants in the self-deprecating humor condition (N=68), 86.5% in the humor condition (N=64), and 93.1% in the no-humor condition indicated that they had not already seen the advertisement (N=67). Thus, there were no big differences between the experimental groups in terms of familiarity with the advertisements.
Hypotheses testing
Effect of advertising strategy on organizational reputation
In two two-way ANOVAs, hypothesis 1 and 2 were tested. Repairing advertising strategy (self-deprecating humor, humor and no-humor) and information (negative vs. neutral) were the independent variables.
Organizational reputation was the dependent variable. Organizational reputation is divided into the corporate competence dimension and the corporate agreeableness dimension. Therefore, these two variables were the dependent variable of the first and the second two-way ANOVA. Multiple comparison tests were used to test the differences between the levels of the independent variables.
Corporate Competence
According to hypothesis 1a the self-deprecating humor ad will lead to a more positive organizational reputation than the humor ad, and according to hypothesis 1b the humor ad will lead to a more positive organizational reputation than the no-humor ad. The participants who saw the advertisement with no humor scored the lowest on the competence dimension of the reputation of Smart (M=4.03, SD=1.21), while the participants who saw the self-deprecating humor advertisement scored the highest (M=4.55, SD=1.04). The participants who saw the humor advertisement, scored in between the others on the competence variable (M=4.10, SD=1.04). These differences were statistically significant and confirm the expectations, F (2, 212) = 4.799, p < 0.05, η2 = 0.04. The post-hoc multiple comparisons test showed significant differences between the different advertisements on the competence variable; between the participants who saw the no-humor advertisement and the self-deprecating humor advertisement (Mdiff= -0.52, p < 0.05), and between the participants who saw the humor advertisement and the self-deprecating humor advertisement (Mdiff= -0.44,
p < 0.05). These results support hypothesis 1a. Only the difference between the participants who saw the
no-humor advertisement and no-humor advertisement was not significant (Mdiff= -0.08, p=1.00), therefore hypothesis 1b has to be rejected.
Figure 2. Organizational reputation (competence dimension) as a function of the information and of the type of advertising strategy
According to hypothesis 2 the positive effect of the self-deprecating humor on the organizational reputation will be stronger after the negative information than after the neutral information. It is expected that the positive effect of the self-deprecating humor advertisement on the competence variable will be different for the two information groups. Compared to the other two advertisements, the self-deprecating humor will have a larger positive effect in the negative condition than in the neutral condition. As can be seen in figure 2, the participants in the negative condition who saw the self-deprecating humor advertisement scored substantially higher on the competence dimension (M=4.73, SD= .95) than the participants who saw the humor
advertisement (M=3.81, SD= 1.14), and the no-humor advertisement (M=3.80, SD= 1.17). In the neutral condition however, this great difference between the self-deprecating humor advertisements and the other advertisements on the competence variable was not present, nor did the self-deprecating humor ad have a stronger effect than the other ads. The participants in the humor condition scored the highest on the competence variable (M=4.40, SD= .87), whereas the no humor condition scored the lowest on the
competence variable (M=4.26, SD= 1.22). The participants in the self-deprecating humor scored in between the others (M=4.36, SD= 1.10). Thus, the positive effect of self-deprecating humor on the competence dimension had a different effect in the negative condition compared to the neutral condition, which confirmed the moderation effect of the variable ‘information’, F (2, 212) = 4.260, p < 0.05, η2 = 0.02.
The post-hoc pairwise comparisons test showed that the differences between the three
advertisements were only statistically significant within the negative condition; between the participants who
4,55 4,1 4,03 0 0,5 1 1,5 2 2,5 3 3,5 4 4,5 5 C or p or ate C omp ete n ce Advertising strategy
Results H1a & H1b
Competence
4,73 4,36 3,81 4,4 3,8 4,26 0 0,5 1 1,5 2 2,5 3 3,5 4 4,5 5Negative information Neutral information
C or p or ate C omp ete n ce Advertising strategy
Results H2
Competence
Self-deprecating humor Humor No humorsaw the no-humor advertisement and the self-deprecating humor advertisement (Mdiff= -.93, p < 0.001) and between the participants who saw the humor advertisement and the self-deprecating humor advertisement (Mdiff= -.94, p < 0.001). However, the difference between the participants who saw the no-humor
advertisement and the humor advertisement was not significant (Mdiff= -.01, p=0.255). Thus, this result demonstrates that the information moderated the effect of the advertising strategy on the competence
dimension of the organizational reputation. In the negative information condition, the self-deprecating humor strategy had a stronger positive effect compared to the other two advertising strategies than in the neutral condition, whereas this stronger effect of self-deprecating humor was not present in the neutral condition (see figure 2). This result supports H2.
Corporate Agreeableness
A similar analysis was executed with the corporate agreeableness dimension as dependent variable. According to hypothesis 1a the self-deprecating humor ad will lead to a more positive organizational reputation than the humor ad, and according to hypothesis 1b the humor ad will lead to a more positive organizational reputation than the no-humor ad. The participants who saw the advertisement with no humor scored the lowest on the agreeableness variable (M=4.14, SD=1.15), while the participants who saw the self-deprecating humor advertisement scored the highest (M=4.85, SD= .85). The participants who saw the humor advertisement, scored in between the others on the agreeableness dimension (M=4.66, SD= .81). These differences were statistically significant, F (2, 212) = 10.903, p < 0.001, η2 = 0.05. The post-hoc multiple comparisons test showed no significant difference between the participants who saw the humor
advertisement and the self-deprecating humor advertisement (Mdiff= -.19, p=0.673), so hypothesis 1a has to be rejected. However, there was a significant difference between the participants who saw the no-humor advertisement and the humor advertisement (Mdiff= -.52, p < 0.01), which supports hypothesis 1b. The difference between the participants who saw the no-humor advertisement and the self-deprecating humor advertisement was also statistically significant (Mdiff= -.71, p < 0.001).
Figure 3. Organizational reputation (agreeableness dimension) as a function of the information and of the type of advertising strategy
According to hypothesis 2 the positive effect of the self-deprecating humor on the organizational reputation will be stronger after the negative information than the neutral information. It was expected that the positive effect of the self-deprecating humor advertisement on the agreeableness variable would be different for the two information groups. Compared to the other two advertisements, the self-deprecating humor would have a larger positive effect in the negative condition than in the neutral condition. As can be seen in figure 3 (left graph), the participants in the negative condition who saw the self-deprecating humor advertisement scored substantially higher on the agreeableness dimension (M=5.03, SD= .80) than the participants who saw the humor advertisement (M=4.50, SD= .91), and the no-humor advertisement (M=3.90, SD=1.14). In the neutral condition however, the differences between the three advertisements were different. The participants in the humor condition scored the highest on the agreeableness variable (M=4.81, SD= .67), whereas the no-humor condition scored the lowest on the agreeableness variable (M=4.38, SD= 1.13). The participants in the deprecating humor scored in between the others (M=4.66, SD= .89). Thus, the positive effect of
self-deprecating humor on the agreeableness dimension had a different effect in the negative condition compared to the neutral condition, which confirmed the moderation effect of the variable ‘information’, F (2, 212) = 4.171, p < 0.05, η2 = 0.02.
The post-hoc pairwise comparisons test showed that the differences between the three
advertisements were only statistically significant within the negative condition; between the participants who saw the no-humor advertisement and the humor advertisement (Mdiff= -.60, p < 0.05) and between the
4,85 4,66 4,14 0 0,5 1 1,5 2 2,5 3 3,5 4 4,5 5 5,5 C or p or ate A gr ee ab le n es s Advertising strategy
Result H1a & H1b
Agreeableness
5,03 4,66 4,5 4,81 3,9 4,38 0 0,5 1 1,5 2 2,5 3 3,5 4 4,5 5 5,5Negative information Neutral information
C or p or ate A gr ee ab le n es s Advertising strategy
Results H2
Agreeableness
Self-deprecating humor Humor No humorparticipants who saw the nohumor advertisement and the selfdeprecating humor advertisement (Mdiff= -1.13, p < 0.001). Moreover, the difference between the participants who saw the humor advertisement and the self-deprecating humor advertisement was significant (Mdiff= -.53, p < 0.05). Thus, the information moderated the effect of the advertising strategy on the agreeableness dimension of the organizational reputation. Consistent with hypothesis 2, the results show that in the negative information condition, the self-deprecating humor strategy had a stronger positive effect compared to the other two advertising strategies than in the neutral condition, whereas this stronger effect of self-deprecating humor was not present in the neutral condition (see figure 3).
Mediation of Perceived Human Character Moderated by Information
According to hypothesis 3, the positive effect of self-deprecating humor on the organizational reputation is mediated by perceived organizational human character. Moreover, hypothesis 4 assumes that this mediation effect of perceived human character is stronger after the negative information than after the neutral
information (H4). When the a path is moderated by W (see Figure 4), there is an effect traditionally termed ‘moderated mediation’ (Preacher, Rucker & Hayes, 2007). When there exists an indirect effect of X on Y through M, the question is whether mediation exists for different conditional values of W. Using the PROCESS macro as discussed in Preacher and Hayes (2004) and Hayes and Preachers (in press), the coefficients of a moderated mediation model in which this indirect effect of perceived human character was presumed to be moderated by the information (negative versus neutral), is estimated.
Figure 4. Conceptual diagram of moderated mediation
Because PROCESS allows only a single independent variable that must be either dichotomous or continuous, “it must be tricked into estimating a model with a multi-categorical independent variable” according to Hayes and Preacher (in press, supplement p.4). With the strategic use of covariates, manual construction of the indicator codes prior to execution, and multiple executions of the macro, PROCESS can estimate a model as shown in Figure 5. As the independent variable X is multi-categorical, mean differences can be estimated with a linear model by representing groups with a set of k – 1 variables, where k is the number of groups. So, two dummy variables were constructed, with D1 set to 1 if a case was in the group of self-deprecating humor,
and 0 otherwise. D was set to 1 if a case is in the group of humor, and 0 otherwise. The group of no humor
X Y M W b c’1 a1
was not explicitly coded, meaning all k – 1 dummy variables were set to 0 for cases in that group. This group functioned as the reference category in the analysis, and parameters in the model pertinent to group
differences were quantifications relative to this reference group. Thus, D1 coded the self-deprecating humor
condition, D2 coded humor condition, and the no-humor group functioned as the reference group and
received a code of 0 on D1 and D2. Moreover, a dummy variable for the moderator variable W (information)
was created, with Wnegative set to 1 if a case is in the group of the negative information, and 0 otherwise. The group of the neutral information was not explicitly coded, meaning the dummy variable was set to 0 for cases in the neutral-information group. This group functioned as the reference category in the analysis, and parameters in the model pertinent to group differences were quantifications relative to this reference group (Hayes & Preacher, in press).
Figure 5. Statistical diagram of moderated mediation model
The statistical diagram as displayed in figure 5 can be estimated by running PROCESS k – 1 times, where k is the number of levels of the independent variable, and using k – 1 dummy variables. At each run of the mediation model (with moderator W), one of the dummies was used as X and the other as covariate, with the dummy serving as X being swapped with the covariate at subsequent PROCESS runs. When the moderator variable was inserted in PROCESS, it created automatically an interaction variable between the independent variable X and the moderator W. As in the first PROCESS run D1 was inserted as independent variable and D2 was inserted as covariate, PROCESS only estimated the interaction effect of D1*W. Therefore, the interaction variable D2*W needed to be manually created and inserted as covariate in the first run. This
D1 Self-‐Depr Y Competence/ Agreeableness M Human Character D2 Humor eM eY 1 1 b= 0.242 c’2 c’1 a2 a1 W Negative D1W Self-‐Depr * Negative Humor * D2W Negative a3 a4 =1.436 ** a5 = .284 c’3 c’4 c’5
needed to be done in the second run for D1*W too, because PROCESS now automatically estimated the interaction effect of D2*W as D2 was inserted as the independent variable.
This procedure generated the estimates of a1, a2, a3, a4, a5, b, c'1, c'2, c'3, c'4, and c'5 corresponding to the values in Table 1. a1 and a2 correspond to the mean differences on Y between the self-deprecating humor and humor conditions, respectively, relative to the control condition. a3 corresponds to the mean difference on Y relative to the neutral information condition. a4 and a5 correspond to the mean differences on Y caused by the interactions between the self-deprecating humor dummy and the negative information dummy and the humor dummy and the negative information dummy, relative to the neutral information no-humor control condition. b is the relative total effect of M on Y, holding the other variables constant. c'1 and c'2 are the relative direct effects of self-deprecating humor and humor on Y, relative to the no humor condition, whereas
c'3 is the relative direct effect of negative information on Y, relative to the neutral condition. c'4, and c'5 are the relative direct effects of the interactions on Y. The conditional indirect effects of X on Y through M were constructed by multiplying a4 and a5 by b, however PROCESS provides these conditional indirect effects automatically (a4b in the first PROCESS run and a5b in the second run) and gives a bias-corrected bootstrap confidence interval (CI) for a test of equality of the conditional indirect effects in the two groups based on 10.000 bootstrap samples. This tests whether the different values of the indirect effects of the two conditions (negative and neutral information) are statistically significant.
To test the predictions regarding to the mediating role of perceived human character (H3) and the moderating effect of information on this mediation effect (H4), the causal links between the independent variable (advertising strategy; indicator coded into two dummy variables), the mediator variable (human character), the moderator variable (information; indicator coded into one dummy variable) and the outcome variable (reputation, divided into two competence and agreeableness) were examined.
Corporate Competence
In the first PROCESS run, the self-deprecating humor dummy was inserted as X, perceived human character as mediator M, the negative information dummy as moderator W and the corporate competence dimension as Y. There was controlled for the humor dummy and the manually created interaction variable of the humor dummy*negative information dummy. In the second PROCESS run, the Xs were swapped and the humor dummy was inserted as X, perceived human character as M, the negative information dummy as W and the corporate competence dimension as Y. Now, there was controlled for the self-deprecating humor dummy and the manually created interaction variable of the self-deprecating humor dummy*negative-information
dummy. Figure 5 depicts the statistical diagram (combining the estimations of a4 and a5 by two PROCESS runs) used to investigate two conditional indirect effects. The statistically significant interaction between the self-deprecating humor dummy and the negative information dummy in the model for human character implies that the indirect effect of self-deprecating humor on competence through perceived human character was moderated by the negative information (see table 1). The sign of the interaction was consistent with the hypothesis that the indirect effect is larger for participants who have read the negative information.
M= Y = Y = Outcome Human character Competence Agreeableness
Coefficient Coefficient Coefficient
(SE) (SE) (SE)
if X=D1/X=D2 if X=D1/X=D2 if X=D1/X=D2 Constant i1 5.361*** i2 2.965*** i2 2.673*** (.296) (.264) (.199) D1self-depr humor a1 .410 / -1.026 c'1 -.006 / -.500 c'1 .148 / - .245 (.422) / (.941) (.236) / (.527) (.178) / (.397) D2humor a2 .461 / .744 c'2 .140 / -.049 c'2 .120 / .196 (.927) / (.414) (.518) / (.233) (.390) / (.175) W= Dnegative a3 -.667 c'3 -.304 c'3 -.265 (.419) (.235) (.177) D1self-depr*W a4 1.436** c'4 .494 c'4 .393 (.593) (.336) (.252) D2humor*W a5 .284 c'5 -.189 c'5 .076 (.589) (.329) (.248) M = Human character b .242*** b .318*** (.038) (.029)
Table 1. Moderated mediation model, with human character as mediator and information as moderator
The conditional indirect effects of X on Y (corporate competence) through M are constructed by multiplying
a4 and a5 by b:
a4 * b = 1.436*.242 = .348 a5 * b = .284*.242 = .069
Relevant components of the summary table of the PROCESS outputs are displayed in Table 2. Moderated mediation is addressed by examining the significance of the product a4b and a5b, which is the case if the 95% bias-corrected bootstrap CIs for the relative indirect effects do not straddle zero. Moreover, PROCESS provided the conditional indirect effects of X on Y at values of the moderator (see table 5).
