THE EFFECT OF POSITIVE EMOTIONS IN CRISIS RESPONSE MESSAGES ON CORPORATE
REPUTATION AND BRAND LOYALTY
A mixed-method research into crisis communication
Nikoletta Mayer
Student number: 12832251 | University of Amsterdam Master’s Thesis
Graduate School of Communication
Master’s Programme Communication Science Supervisor’s name: Rutger de Graaf
Word count: 7460
Date of completion: 04/02/2022
Abstract
In this paper two research methods were combined to gain a deeper insight into the strategic use of (positive) emotions in crisis communication messages. Vignettes aimed to identify whether communicating solidarity and hope in a crisis response message makes stakeholders empathize with the organization, and whether they will have a better perception of the company. Results support previous research regarding the positive influence emotions have on audiences. For example, corporate reputation and brand loyalty were both positively affected by messages with hope or solidarity. By conducting semi-structured interviews, this paper also provides a broader understanding of communication specialists’ attitude towards emotions in crisis communication. Contrary to common belief, most communication specialists do use emotions in crisis response messages, but they mostly follow their intuition which suggests a lack of strategy. Thus, this mixed method research provides support for existing literature about the positive effects of emotions in crisis communication messages, but it also reveals new insights by focusing on only positive emotions, looking at brand loyalty in the context of crisis communication, and by comparing present practices to scientific findings.
Introduction
Communicating emotions in crisis communication is not part of the norm, but it sparked an interest in academia
In October 2021, Facebook came under fire due to a former employee accusing the tech giant of softening its security too soon after the 2020 U.S. elections. The employee further claimed this contributed to the storming of the U.S. Capitol on January 6, 2021 which left many with minor or serious injuries (New York Times, 2021). The accusation also highlights Facebook’s responsibility in the political polarization of U.S. citizens. This can be considered a major crisis, threatening the corporate reputation of Facebook, as well as having a negative
effect on its users’ loyalty to the brand (Cowden & Sellnow, 2002). Nick Clegg, the vice president of policy and global affairs of Facebook responded to the accusations: „Social media has had a big impact on society in recent years, and Facebook is often a place where much of this debate plays out,” he claimed. “But what evidence there is simply does not support the idea that Facebook, or social media more generally, is the primary cause of polarization.” (New York Times, 2021). This is an extraction of a long memo, and an example of many responses to the crisis. However, it is also a clear demonstration of the lack of emotions in the message.
We can observe this pattern in the practice of crisis communication, as most messages released by an organization are formal, informative, and void of emotions. Communication professionals believe that in a negatively charged crisis context, showing emotions might be inappropriate (van der Meer & Verhoeven, 2014, p. 533). However, organizational crises are undeniably emotional for stakeholders and the organization itself (van der Meer & Verhoeven, 2014, p. 526). This research aims to show that contrary to what the majority of communicational professionals claims, embracing the emotional nature of an organizational crisis could benefit the organization in multiple ways, including mitigating the threat to corporate reputation and brand loyalty.
In academia, there has been an increasing interest in the study of emotions in crisis communication, but most studies focused on the public’s emotions, or negative emotions such as anger (Coombs & Holladay, 2005; Jin, Pang, & Cameron, 2010, 2012). Few studies have specifically focused on the emotional nature of organizational messages and corporate reputation. However these focused on negative emotions, such as shame and regret, and involved crises of low intensity (Van der Meer and Verhoeven, 2014). The effect positive emotions, such as expressing hope and trust, can have on more serious crises is an unexplored gap of crisis communication, which this research aims to fill. Moreover, crisis communication professionals could benefit from insights regarding the use of positive emotions in the future.
This research will combine qualitative interviews and an online experiment to distinguish whether or not an effect of positive emotions exists in crisis responses, and also to gain understanding of crisis response strategies. This mixed method is ideal for a deeper, multi- perspective grip on the field.
The emotional level of a crisis, could depend on the crisis type the organization is dealing with (van der Meer & Verhoeven, 2014, p. 526). More serious crises, which involve a financial loss or physical/emotional harm to stakeholders, will likely evoke negative emotions in stakeholders. Therefore, it is more difficult for communicaton professionals to mitigate the damage done to corporate reputation and brand loyalty. Crisis reponse strategies, which guide the organization to express regret or shame, have been developed for these crisis types.
However, it remains unclear how positive emotions could be used to mitigate the threats. Thus, this paper will focus on one crisis type, the accidental cluster:
RQ1: To what extent do positive emotions communicated by an organization in times of crisis have a positive effect on the resulting corporate reputation and brand loyalty when the crisis was unintentionally caused by the organization?
Mitigating the threat to corporate reputation and brand loyalty after a crisis is a difficult task because the crisis response is highly dependent on the specifics and cicumstances of the case. For the same reason, creating a guide for communication professionals is also difficult.
Given the numerous factors affecting the decision-making process of the response strategy, applying the same strategy for multiple cases based on one guide might not be effective.
Therefore, to explore what factors contribute to communicating positive emotions after a crisis, this paper will also investigate:
RQ2: To what extent do external (cultural, demographic) and internal (organization specific) factors affect what emotions are communicated in a crisis reponse strategy?
Theoretical Framework
Communicating positive emotions in crisis response messages can help the audience to emphatise with the organization, and to frame the crisis in a more positive way
Emotions can be categorized into negative and positive emotions. Positive emotions are defined as a “pleasant or desirable situational responses… distinct from pleasurable sensation and undifferentiated positive affect” (Cohn & Fredrickson, 2009). Since crises are not fundamentally positive events, they could potentially evoke negative emotions in stakeholders.
If the organization does not give a proper response to the crises at hand, those negative emotions will likely affect their perception of the company. Thereby, this paper will examine whether these negative perceptions can be mitigated if crisis response messages communicate positive emotions. In the paper’s context, positive emotions will refer to two emotions: hope and solidarity.
Embracing the emotional nature of a crisis can be beneficial for the organization that has to deal with the crisis. Expressing (positive) emotions in the organizational message – in addition to information about the situation – reveals the feelings of the sender (organization) (Ekman, 1993), the motives of the sender (Fridlund, 1991), and how concerned the sender is about the well-being of stakeholders (Knutson, 1996). Research has also shown that those who are subjected to the emotions of the sender take it into consideration when assessing the behavioral intention, so they will act accordingly towards the sender (Lewis, 2000; Van Kleef, De Dreu, & Manstead, 2004). This research will build upon this notion and attempt to identify whether communicating emotions affect stakeholders. According to Mcguire et al. (2020) the
language and tone of messages can help the reader’s sense-making process regarding their assessment of emergency. This means that the organization can help stakeholders to frame the situation in a positive way, and expressing hope could show the stakeholders that the organization can solve the issue at hand and might make them feel safer or calmer. Moreover, Bersade (2020) claims that communicating positive emotions has a contagion effect on building positive attitude. Thus, if the organization communicates hope, and encourages solidarity, people will be more likely to think of the situation similarly.
Mcquire et al. (2020) studied the communication style of Jacinda Ardern during the Covid-19 threat, and showed how an empathetic approach signaling solidarity positioned the Prime Minister on the level of the public. This created a feeling of common ground and acceptance among the public, moreover, communication on social media reinforced authenticity and “informal humanity”. This example supports the idea that positive emotions such as empathy in organizational messages can affect the audience positively and increase acceptance. Although, an empathetic approach can also include negative emotions, which is out of scope for this research, empathy can be evoked by expressing solidarity, as well. Hence, this paper’s findings will not only strengthen this perspective, but also broaden it by looking at specificly crisis response messages.
A crisis communication message also influences corporate reputation and brand loyalty Corporate reputation develops over time based on continuous interactions between stakeholders and the organization (Argenti and Druckenmiller, 2004). A stakeholder’s perception of the organization is affected by many organization-related stimuli, including mass communication, and interaction with people linked to the organization (Abratt and Kleyn, 2010, p. 1050). Thus, an organization’s crisis communication messages also form corporate reputation.
Similarly, brand loyalty develops over time. Here it is defined as a costumer’s repetitive buying and/or (in case of a service) behaviour, which means that as a result of positive bias the customer will choose a branded/labelled product or service over and over again (Jagdish and Sheth, 1974). Over time, customers get attached and committed to the brand, forming a relationship which can normally be hardly affected (Khan and Mahmood, 2012, p. 33).
However, if there is a crisis, the way an organization communicates about the situation might shake the customers’ loyalty to the brand. Corporate reputation and crisis communication have been studied before, however, this research will go further and look at brand loyalty as well to develop a more complex view of stakeholders’ perceptions.
Empathy might also protect brand loyalty and corporate reputation, but effectivess also depends on the crisis type
Coomb (2005 and 2007) provides a thorough overview of crisis types by the SCCT, which is an important factor in this research. A detailed description is in App. 1. The studies elaborate on how stakeholders might react in different crisis situations, including those not directly affected by the crisis. According to Coombs, the way an organization communicates towards the affected parties, affect other stakeholders as well, and thus, has an effect on corporate reputation. SCCT will be incorporated into the research, however, this study will examine emotions in the organizational message and not in stakeholders. This research will also attempt to show that following SCCT is not enough to create an effective crisis response strategy because stakeholders will show greater acceptance when emotions are communicated.
Literature regarding brand loyalty and crisis communication is scarce, however, Ariffin et al. (2019) describe how a crisis and the associated consequences, such as blame attribution, responsible recall and perceived crisis responsibility have an (insignificant) effect on costumers’ purchase intention. However, the study also found that crisis management has a
significant positive affect on purchase intention post-crisis. Thus we can stipulate that crisis type might be a less influential factor than the actual response strategy. Since purchase intention can be a factor of brand loyalty, this insight can be valuable to our research, and emotional factors can be examined as an addition to existing findings.
Shaking it up – a focused approach with the accidental crisis cluster
The study by van der Meer and Verhoeven (2014) caved the way for exploring emotional messages and their effects on stakeholders. Their vignettes based research provides the basis for this paper, as it describes how different crisis response strategies (rebuild vs diminish) and emotions (anger vs regret) can affect corporate reputation. They explain the results by stakeholders’ affective and cognitive responses. They also found that expressing shame or regret with a specific crisis type can have a positive impact on corporate reputation, and suggested examining a greater variety of emotions and crisis types that may be of a more emotional nature. This paper will follow in their footsteps, but focus on crisis clusters rather than response strategies, thus looking at the relationship from a slightly different perspective.
This is also supported by an important study concerning corporate reputation by Schoofs et al. (2019). They researched how empathy affected corporate reputation by conducting three experiments, and found that stakeholders are much more likely to empathize with an organization if they were apologetic in their post-crisis communication, which led to a smaller reputational damage. They also stipulated that a crisis belonging to the accidental cluster does not only flick less reputational damage because stakeholders find the organization less responsible, but also because they empathize with the organization. This is valueable because it shows that emotional reactions can in fact have a positive effect on corporate reputation.
However, the study is limited as it only describes empathy and ignores a range of different emotions. Based on the advice of van der Meer and Verhoeven (2014) this paper will focus on
the accidental cluster, so this will remain a constant in hypothesis. Thus, the hypotheses read as:
H1a: The resulting reputation of an organization communicating hope about the organizational crisis is better than the resulting reputation of an organization communicating no emotions.
H1b: The resulting reputation of an organization communicating solidarity about the organizational crisis is better than the resulting reputation of an organization communicating no emotions.
H2a: The resulting brand loyalty to an organization communicating hope about the organizational crisis is better than the resulting brand loyalty to an organization communicating no emotions.
H2b: The resulting brand loyalty to an organization communicating solidarity about the organizational crisis is better than the resulting brand loyalty to an organization communicating no emotions.
Stakeholder differences – Does gender matter?
Looking at the differentiating qualities of stakeholders is also important to see if a crisis response message will have the same effect on everyone. Research has shown that women are more emphatetic than men considering both emotional empathy and cognitive empathy (Mestre el al., 2009, p. 81). Moreover, their recall of emotional information from a narrative script which contains emotional and neutral information is also better than of men’s (Bloise and Johnson, 2007, pp. 200-201). Based on this, it is assumed that women would be more
susceptible to postitive emotions communicated in the message released by an organization, and thus feel more empathy towards it. Thus, this research looks at whether gender has an influence of the effect of communicated positive emotion on corporate reputation and brand loyalty. The following sub-hypotheses read as:
H3a: The resulting reputation of an organization communicating hope about the organizational crisis is better when perceived by women than when perceived by men.
H3b: The resulting reputation of an organization communicating solidarity about the organizational crisis is better when perceived by women than when perceived by men.
H4a: The resulting brand loyalty of women to an organization communicating hope about the organizational crisis is better than the resulting brand loyalty of men.
H4b: The resulting brand loyalty of women to an organization communicating solidarity about the organizational crisis is better than the resulting brand loyalty of men.
Diving deeper into other factors
The interview conducted for this research revealed multiple factors that affect what emotions should be communicated in the crisis response message. A detailed outline will be described in the Results section, which were used to formulate the following hypotheses:
H5: Communicating both negative and positive emotions in a crisis response message results in higher corporate reputation compared to communication only positive emotions.
H6: Communicating the emotions of the CEO in a crisis response message damages corporate reputation more than not communicating the emotions of the CEO.
H7: Communicating emotions in a crisis response message in addition to operative actions generates more trust in stakeholders compared to communicating emotions without operative actions.
H8: The effect of emotions in a crisis response message on corporate reputation is moderated by organization type, crisis type, and nationality.
Conceptual model
Method
Double-trouble – semi-structures interviews mixed with vignette study
In this paper, two research designs were combined to understand both communication specialists and real audiences. Semi-structured interviews were conducted with 6
communication professionals who specialize in crisis communication who have years of experience behind them, to ensure that their insights are valuable. Participants were contacted by the researcher and recruited by email through snowballing technique, but since they differed demographically and since this research has an expolaritve nature, their insights were still valid.
The interviews took place via zoom since at the time of the fieldwork all kinds of Covid prevention measures were in place. Five questions were prepared beforehand e.g. ‘Do you strategically use emotions in crisis communication when the crisis was unintentionally caused by the organization?’ (see. App. 2), but in order to gain a deeper understanding of how practitioners use emotions in crisis communication other questions arose based on the conversation.
Paralell to the interviews, a between-subjects experimental design was conducted using a vignette study. An experimental vignette/survey embedded experiment was chosen to adhere to the current health measures, and because it can be distributed to more participants. In this study, each participant was randomly assigned to one of the three conditions created by a 3 (hope vs trust vs no emotion) X 1 (crisis type: accidental) factorial design.
Participants and Sampling – 315 finished surveys
Participants were recruited online with a recruitment message (App. 3) by sending out an URL link to the experiment and through snowballing technique. Convenience sampling may be limited but since the link was distributed on multiple platforms, including in multiple subreddits which consist of people from all over the world, the experiment reached a magnitude of different people. There were no filters required for the experiment since crisis communication messages affect a strongly heterogenous audience, thus the sampling technice provided a varity of people and opinions. At the end of the data collection 390 responses were recorded. After deleting uncomplete cases, 315 responses remained. There was no attention
check in the experiment to filter out participants in order to strengthen the ecological validity of the study because even in reality people tend to lose their attention when reading or browsing. Since the online experiment was designed in a way that questions could not be skipped or left unanswered, those who arrived to the last page had answered all questions and no missing values were left in the dataset. All personal data was deleted from the datafile.
Demographics – On average 29 year old, an equal number of men and women, of 45 nationalities, with mainly Bachelor’s Degrees
Source: charts were made by Nikoletta Mayer using Visme
Procedure
Each participant was randomly assigned to either of the 3 conditions by Qualtrics. First, participants had to read a consent form and agree to the terms to continue the study, after which they answered demographic questions of gender, age, highest educational level, and nationality. Participants were notified about the completion of the first part of the study and then instructed to read the information carefully in the second part of the study. All participants were shown a description of the fictitious organization, Blyss, which was created for the study (see App. 4). Thus, this text was kept constant across all conditions. Then, participants were required to answer a set of questions about their perceived corporate reputation of Blyss and their brand loyalty as a part of the pre-test.
Next, they read about a crisis, belonging to the accidental crisis type, which Blyss was facing. This was kept constant across all conditions to further appreciate how the manipulation affected repondents in the same circumstances. This was followed by the manipulation of the independent variable. Participants saw a Tweet from the company reacting to the crisis either containing no emotions/hope/solidarity. Following the manipulations, a manipulation check question was asked. Finally, participants were instructed to fill out a questionnaire measuring their levels of perceived corporate reputation about the company, and their brand loyalty. At the end of the session, participants reconsented to the uses of their data, and were debriefed about the purpose of the study.
Independent variable
Emotions: Crisis communication message with hope/solidarity/no emotions was manipulated with different photoshopped texts. Participants assigned to the hope, solidarity or no emotions conditions, respectively, read different messages. The messages were photoshopped to look like Tweets posted by the ficticious organization (see App. 5). The study by van der Meer et al.
(2014) used similar messages as their manipulation; moreover photoshopping the messages to look like Tweets strengthens the ecological validity of the research because they accurately represent real messages.
Dependent measures
Corporate reputation: To assess the (resulting) corporation reputation of the organization, participants were asked to indicate their level of agreement with five items taken from existing measures (The Organizational Reputation Scale), eg.. ‘The organization is concerned with the well-being of its public. ’(Coombs & Holladay, 2002), which originally consists of ten items but was shortened to prevent survey fatigue. This specific scale was used because the study of van der Meer et al. (2014) used it in their study, as well. Detailed description and reliability results in App. 6.
Brand loyalty: To asses the (resulting) brand loyalty of respondents to the organization, participants were asked to indiciate their level of agreement with 12 items taken from a measure e.g. ‘I like switching from one brand of shampoo to another.’ (Odin, Odin and Valette-Florance, 2001), originally consisting of 17 items. The content of the items were also rephrased to address the organization present in the experiment. This scale was used as it measures multiple dimensions of brand loyalty and thus gives a broader, more reliable understanding of respondents’ attitude. Detailed description and reliability results in App. 7.
Manipulation check item
In order to see whether the changes in corporate reputation and brand loyalty were indeed caused by the different type of emotions, the following manipulation check question was asked:
‘What emotion did the message express?’. Respondents had to choose from three answer options: (1) Hope, (2) Solidarity, and (3) No emotions.
Control Variable
The resulting corporate reputation and the resulting brand loyalty were also controlled before the manipulation itself in order to account for any differences among participants. They were measured separately by the same scales and items that were used for the dependent variables.
This was crucial because if respondents had different attitudes to the company before the crisis reponse message, it would skew the results. Reliability results in App. 8.
Results
Positive and negative emotions should be communicated in addition to operative actions in crisis response messages
The interviews revealed that communication practitioners do not exclude communicating emotions when an organization faces a crisis, however, a lack of strategy can be observed. All specialists interviewed claimed that including emotions in crisis communication messages can be beneficial for corporate reputation and brand loyalty, but most could not describe exactly when or in which crisis situation it is advised. It seems that in practice we include positive and negative emotions in crisis communication based on human intuition rather than a guide. However, we should use mixed emotions, rather than only positive or negative ones. Also, „emotions in crisis communication should occur only as an addition to operative actions” because when there is no actual help or plan of action, only emotions, it is only virtue-signaling, which damages corporate reputation (Practitioner 6). Practitioner 3 and 6 argued that our first priority is assessing the circumstances and the responsibility of the
organization, for which Coomb’s SCT provides a reliable guide. Only after establishing how much responsibility the organization has over the crisis and consider the initial reputation or crisis history of the organization, should we think about what kind of emotion to include in the message.
Organization and crisis type should be weighed in before choosing what emotions we want to communicate in a crisis response
There are several factors which affect what kind emotion we should communicate in crisis communication. First, the type of organization influences the decision-making process because governmental organizations, such as fire departments or the police, mostly respond to crises belonging to the victim cluster (e.g. natural disasters, terror attacks). The organization then should communicate negative emotions – expressing empathy with saying ‘sorry’ – but end the message with positive emotions, expressing hope or solidarity (Practitioner 2).
However, it is essential to combine positive emotions with communicating what needs to be done by the organization or the people. Here, including operative actions is essential. Some counterargue that “governmental organizations forget that they are not only responsible for helping after a disaster but also for preventing a disaster (e.g. terrorist attack), which they leave out of their communication” (Practitioner 6). Since they incorrectly assume to respond to a “victim cluster crisis” instead of an “accidental” crisis type, they choose the incorrect response strategy. Assessing whether positive or negative emotions help mitigate corporate reputation or brand loyalty in this case is difficult because the initial strategy is flawed.
Nevertheless, it is evident that crisis type should matter the most when deciding on whether to include emotions, and if yes, what kind. Audiences sympathise and emphasize more with organizations when the crisis involves human mistakes of the accidental kind (Practitioner 1, 3, 4, 5, 6). Thus communicating emotions when the organization faces a crisis belonging to
the ‘accidental cluster’ is encouraged. Again, depending on the context, negative and positive emotions should be used together, paying attention to ending the message with a positive emotion – so as not to leave the audience with negative connotations. A factor that should be considered when including positive emotions in crisis communication messages is not to sound ingenuine or dishonest. If the context requires communicating negative emotions such as regret or shame and the organization does not yet have a plan of action how to move forward, then communicating hope or solidarity can sound artificial (Practitioner 2, 4, 5, 6).
CEO equals the organization, but not when it comes to communicating emotions
Additionally, one mistake in regards to communicating anger, sorrow, regret, hope, or other kind of emotions often occurs when the CEO speaks in the name of the organization.
Frequently, it is difficult to differentiate between the two because the CEO represents the organization and the brand of CEO and the brand of the organization cannot be separated.
Problems occur when the CEO feels personally attacked when faced with an organizational crisis. The CEO should not communicate their own emotion, and it should not be mistaken for strategic use of emotions in crisis communication (Practitioner 3, 6).
Contrary to real-life practices, the online experiment in this study looked only at positive emotions to see if they have a positive effect when the crisis belongs to the accidental cluster.
Successful manipulation check
χ2 analysis was employed to check the effectiveness of the experimental treatments.
Results showed that participants who viewed the post containing no emotions were more likely to indicate the post contained no emotions (43%) compared to those who viewed the post containing hope (27.1%) or solidarity (29.3%). Participants who viewed the post containing
hope were more likely to indicate the post contained hope (70.8%) compared to those who viewed the post containing solidarity (12.5%) or no emotions (16.7%). Finally, participants who viewed the post containing solidarity were more likely to indicate the post contained solidarity (39.1%) compared to those who viewed the post containing hope (31.8%) or no emotions (29.1%). The effect of the manipulations on the respondents perceptions in statistically significant, χ2(4)= 24.19, p < .001, with a weak effect, as Goodman and Kruskal’s Tau = 0.28. The percentages across all groups are somewhat similar, and since the manipulation contained subtle differences and the image disappeared before the manipulation question, it is possible that participants did not remember the exact wording. Nevertheless, the reader should be mindful of this when interpreting the results.
Data Analysis
Crisis response messages containing hope or solidarity increase perceived corporate reputation after a crisis
In order to answer the RQ, and to test the main effect of hope (H1a), and the main effect of solidarity in a crisis communication message (H1b) on perceived corporate reputation, a one-way analysis of covariance was carried out, controlling for pre-crisis corporate reputation to account for differences among participants whose attitude differed before the manipulation.
Results are presented in Table 2. All assumptions were met.
Table 1
Levels of Perceived Corporate Reputation – highest CR after solidarity message
n M SD
Crisis message
With No Emotions Total: 109 2.60 0.71
With Hope Total: 103 2.71 0.62
With Solidarity Total: 103 2.91 0.58
*Controlling for pre-crisis corporate reputation (M = 2.75)
Table 2
Results of a one-way analysis of covariance
Sum of Squares df Mean Square F p η2
Pre-Crisis Corporate reputation 21.63 1 21.63 63.68 0.000 0.16
Crisis message 6.50 2 3.25 9.56 0.000 0.048
Error 105.64 311 0.34
(Corr.) Total 132.52 314
Note: N = 315
The main effect of the control variable pre-crisis corporate reputation is statistically significant, F(1, 314) = 63.68, p = .000, η2= 0.16. Thus, 16% of the variance in participants’
level of perceived corporate reputation can be explained by their initial predispositions about pre-crisis corporate reputation.
The analysis shows a statistically significant main effect of crisis message, F(2, 314) = 9.56, p = .000, η2= 0.048. Thus, 4.8% of the variance in respondents’ levels of perceived corporate reputation can be explained by the crisis message. This amounts to a small to medium effect. Providing support for H1a (controlling for pre-crisis
corporate reputation), the results indicate that among those exposed to the message containing hope, the average level of corporate reputation is slightly higher (M = 2.71, SD = 0.62) than among those exposed to the message containing no emotions (M = 2.6, SD = 0.71). Support was also found for H1b (controlling for pre-crisis corporate reputation) as results show that among those exposed to the message containing solidarity, the average level of perceived corporate reputation is also higher (M = 2.91, SD = 0.58) than among those exposed to the message containing no emotions.
Crisis response messages containing hope or solidarity increase brand loyalty after a crisis In order to test the main effect of hope (H2a) and solidarity (H2b) in a crisis communication message on brand loyalty, another one-way analysis of covariance was carried out, controlling for pre-crisis brand loyalty to account for differing initial attitudes among participants before the manipulation. Results are presented in Table 4. All assumptions were met here as well.
Table 3
Levels of Brand Loyalty – higher BL after seeing Hope & Solidarity message
n M SD
Crisis message
With No Emotions Total: 109 2.47 0.65
With Hope Total: 103 2.58 0.63
With Solidarity Total: 103 2.59 0.65
*Controlling for pre-crisis brand loyalty (M = 2.51)
Table 4
Results of a one-way analysis of covariance
Sum of Squares df Mean Square F p η2
Pre-Crisis Brand Loyalty 97.27 1 97.27 941.48 0.000 0.75
Crisis message 1.34 2 0.67 6.48 0.002 0.01
Error 32.12 311 0.1
(Corr.) Total 130.32 314
Note: n = 315
The main effect of the control variable pre-crisis brand loyalty is statistically significant, F(1, 314) = 941.48 p = .000, η2= 0.75. Thus, 75% of the variance in participants’
level of perceived corporate reputation can be explained by their initial predispositions about pre-crisis brand loyalty. This should be taken into account when interpreting the results.
The analysis shows a statistically significant main effect of crisis message, F(2, 314) = 6.48, p = .002, η2= 0.01. Thus, 1% of the variance in respondents’ levels of perceived brand loyalty can be explained by the crisis message. This amounts to an almost non-existent effect.
Providing (weak) support for H2a (controlling for pre- crisis brand loyalty), the results indicate that among those exposed to the message containing hope, the average level of brand loyalty is slightly higher (M = 2.58, SD = 0.63) than among those exposed to the message containing no emotions (M = 2.47, SD = 0.65). Support was also found for H2b (controlling for pre-crsis brand loyalty) as results show that among those exposed to the message containing solidarity, the average level of brand loyalty is also higher (M = 2.59, SD = 0.65) than among those exposed to the message containing no emotions.
It should not be forgotten that the manipulation was not evident in many cases and repondents believed to see different emotions from what they were orginially shown. Thus, further or stronger support could be found if more observable conditions were shown to participants.
Gender has no effect on corporate reputation or brand loyalty
In order to test the main effect of hope (H3a), or solidarity (H3b) on the resulting corporate reputation of the organization when perceived by men or women, two Independent- Samples t-tests were designed to show the differences of levels of resulting corporate reputation between men and women. All assumptions were met. Results are shown in Table 5 and 6.
Table 5
Results of Independent-Samples t-test – No significant difference Perceived Corporate Reputation Men
(N=52)
Women (N=49)
t(99) p Cohen’s d
M SD M SD
When communication hope 2.65 0.65 2.8 0.58 -1.250 .214 0.251 Note: n = 101
Table 6
Results of Independent-Samples t-test – No significant difference Perceived Corporate Reputation Men
(N = 53)
Women (N = 47)
t(98) p Cohen’s d
M SD M SD
When communication solidarity 2.84 0.58 3.0 0.62 -1.357 .178 0.274 Note: n = 100
Supporting H3a, the resulting corporate reputation of an organization when communication hope in a crisis communication message is slightly higher when perceived by females (M = 2.8, SD = 0.58) than when perceived by males (M = 2.65, SD = 0.65). However, the mean difference (of –0.15) is statistically not significant, t (99) = -1.250, p =.214, 95% CI [–0.40, 0.09], and represents a small effect, d = 0.25.
The same can be said about H3b, as the resulting corporate reputation of an organization when communicating solidarity in a crisis communication message is also slightly higher when perceived by females (M = 3.0, SD = 0.62) than when perceived by males (M = 2.84, SD = 0.58). However, the mean difference (of –0.16) is statistically not significant, t (98)
= -1.357, p =.178, 95% CI [–0.39, 0.07], and represents a small effect, d = 0.27.
Another set of Independent-Samples t-tests were conducted to test the main effect of hope (H4a) or solidarity (H4b) in crisis communication messages between men and women.
Results are shown in Table 7 and 8. All assumptions were met.
Table 7
Results of Independent-Samples t-test – No significant difference
Brand Loyalty Men
(N=52)
Women (N=49)
t(99) p Cohen’s d
M SD M SD
When communication hope 2.53 0.66 2.63 0.61 -0.780 .437 0.157 Note: n = 101
Table 8
Results of Independent-Samples t-test – No significant difference
Brand Loyalty Men
(N = 53)
Women (N = 47)
t(98) p Cohen’s d
M SD M SD
When communication solidarity 2.60 0.57 2.61 0.74 -0.02 .984 0.004 Note: n = 100
Again, the results do not show statistically significant differences in perceived brand loyalty between men and women. There is very weak support for H4a, as the brand loyalty of the organization when communicating hope is slightly higher when this message is perceived by women (M = 2.63, SD = 0.61) than when perceived by males (M = 2.53, SD = 0.66). However, the mean difference (of –0.1) is statistically not significant, t (99) = -0.780, p =.437, 95% CI [–
0.35, 0.15], and represents a very small effect, d = 0.16. Considering H4b, no support was found, as the brand loyalty of the organization when communicating solidarity is no higher when this message is perceived by women (M = 2.61, SD = 0.74) than when perceived by men (M = 2.60, SD = 0.57). The mean difference (of –0.002) is statistically not significant, t (98) = -0.02, p =.984, 95% CI [–0.26, 0.26], and represents a non existent effect, d = 0.004.
This insignificance of gender is reflected in real life practices, as the interviewed communication specialists revelead that they do not segment audiences in such a way. From the interviews, we can deduct that one should also consider the target audience in regards to communicating positive or negative emotions, but since most organizations have a highly
mixed audience (gender, age and ethnic-wise) we should not assume to speak to only one section of the stakeholders (Practitioner 1, 2, 3, 5, 6, 7).
Culture matters when communicating emotions in crisis response messages
The culture in which the organization operates also affects what kind of emotions should be communicated in a crisis communication message. Although the interview sample consisted of only Dutch and Hungarian practitioners, it was evident that the Dutch consider using positive emotions at times to be more beneficial than Hungarians. Southern European cultures are also more emotional in nature and might embrace emotional language in crisis communication more than others (Practitioner 1, 4). Furthermore, since the rise of social media people can freely interact with each other and organizations/brands as well, allowing them to personalize the relationships between the two. This gave way to expressing emotions on both sides, most audiences even find it unappealing and ‘soulless’ if an organization does not express emotions in their messages. Practitioners thus are encouraged to use emotions in their crisis communication, now more than ever, in order to humanize themselves (Practitioner 4, 5).
Admission of guilt does not equal to communicating emotions
Some practitioners did not recommend using emotions in crisis communication messages because it conveys an admission of guilt, which could leave the organization exposed to liability and lawsuits (Practitioner 2). However, one should not confuse admission of guilt with communicating emotions in an organizational message. Phrasing is important in these cases, but using emotions only humanizes the organization in times when (great) loss occurred.
When the organization is responsible for the crisis somehow, communicating emotions is
recommended. This can not only generate understanding or empathy, but also mitigate the damaged corporate reputation and/or brand loyalty (Practitioner 3, 5, 6).
An alternative solution: “holding-statement”
One practitioner emphasised that this process, which could allow us to calmly assess every factor of the crisis, could lead us to the best communication strategy, but works only in theory. In practice, time is of the upmost importance and organizations want to react as fast as possible to make a stance, inform stakeholders and mitigate reputational damage. This usually causes more damage than good because incorrect response strategies are employed as there is no time to plan. If negative or positive emotions are used when they should not have been, they cause damage to corporate reputation and a loss of brand loyalty. A solution for this is a control strategy, called “holding-statment”, in which the organization tells its stakeholders that it is aware of the crisis and investigates the details, and provides an exact date on which an extensive response will be given. This reassures stakeholders that the crisis is not swept under a rug, and also gives time to the organization to examine and consider every factor of the crisis (strategy).
This way, the crisis communication response will not be based on intuition but on facts (Practitioner 6).
Discussion and Limitations
To answer RQ1 ‘To what extent do positive emotions communicated by an organization in times of crisis have a positive effect on the resulting corporate reputation and brand loyalty when the crisis was unintentionally caused by the organization?’ results of the two studies show that communicating positive emotions, more specifically hope or solidarity, in a crisis response message can help mitigate the damages done to an organization’s corporate reputation
or stakeholders’ brand loyalty. Practitioners claim that communicating both negative and positive emotions are usually the most effective; and since support was found for H1a,b and H2a,b, this is now also supported by literature. First of all, findings suggest that Bersade (2020) was right and communicating positive emotions help building positive attitude. Moreover, these findings also compliment the study of van der Meer’s (2014) which focused on negative emotions, and together they show that if a positive or negative emotion is communicated in a crisis response message, stakeholders will perceive the organization more positively. What would we find if we conducted an experiment using mixed emotions? An interesting parallel can be found to the research of Mcquire et al. (2020) assessing Ardern’s crisis leadership. Her crisis management started with a firm, evidence-based tone and shifted to ‘a more empathetic approach encouraging solidarity among the community and a common understanding on how people should make sense of the situation, and also move forward with it ’. This paper found supporting evidence that a crisis response like this can be successful, however, one wonders to what extent Ardern’s success was related to her personality. An extensive qualitative study would discover the assessment process of the audience, and also help reinforce the notion made by Lewis (2000) and Van Kleef, De Dreu, & Manstead (2004) that audiences will more likely act accordingly to the sender if they communicate emotions. Also, results of this paper elaborate on the connection made by Ariffin et al. (2019) that claim a crises and its conseqencies can negatively affect purchase intention, but crisis management will counteract these damages. This research showed that crisis management with positive emotions do indeed help keep stakeholders’ loyalty and purchase intention.
Findings of both the experiment and the interviews suggest that the gender of the audience should not play a role in choosing what emotions to include in the crisis response message. Although practitioners pointed out that most organizations have a very heterogenous audience, thus they need to address a variety of people, other demographic factors that were
not tested here such as age, cultural differences, level of education should also be further investigated to provide data-driven support. The answer to RQ2 ‘To what extent do external (cultural, demographic) and internal (organization specific) factors affect what emotions are communicated in a crisis reponse message?’ can be derived from the interview results.
People’s nationality can affect how they react to emotions in crisis response messages as their attitude towards emotions might be different. Also, governmental organizations often respond to crises which they have not caused, but since they are responsible for helping the population they need to be firm but empathetic, hence communicating positive emotions with a set of actions is highly encouraged. However, the most important factor according to practitioners is the crisis type. SCCT offers a solid framework, but strategies should be assessed case-by-case.
People are more empathetic towards crises which are caused by human error, so even if the crisis was (un)intentially caused by the organization, corporate reputation might take less damage if the crisis response message included emotions in these cases.
Suggestions for further research
The manipulation of the crisis response message was not completely effective as numerous participants failed to recall what emotions they had seen, thus a more visible manipulation could be developed. This would strengthen the validity of the findings as a stronger causality could be drawn between type of emotions and corporate reputation or brand loyalty. The scale of brand loyalty was lengthy thus participants might have gotten tired during the survey, causing respondents to answer without reading the questions. Consequently, this probable respondent fatigue might have caused incorrect results, which should be further improved in future research to strengthen the validity of the results. The majority of the participants had at least a Bachelor’s Degree, which might influence the results because these people are overrepresented in the sample compared to people from other educational
backgrounds – who might perceive crisis messages and emotions differently. Thus, in order to make the sample and finding more representative to the population, participants with different demographics should also be assessed. Furthermore, the interviewees were either Dutch or Hungarian, which limits the generalizability of the findings, hence communication specialists from different countries should also be asked. Finally, the progress of choosing the right crisis response message is highly case specific and depends on a dozen factors. Formulating a hypothesis – and guide – that can be generally accepted and applied is challenging, and a lot of further research is needed to specify other factors that can mitigate the threats made by organizational crises.
Conclusion
In this paper two research methods were combined to gain a deeper insight into the strategic use of (positive) emotions in crisis communication messages. This paper aimed to find out whether communicating solidarity and hope in a crisis response message makes stakeholders empathize with the organization, and whether they will have a better perception of the company. Literature suggested that communication practitioners rather steer clear of using emotions in crisis response strategies, but in order to see if emotions are indeed not used in practice, the scope of this research was extended to the knowledge of communication professionals. Exploration of present practices was carried out by online semi-structured interviews, while a between-subjects onlince experiment was conducted to examine the effect of hope and solidarity in crisis response messages on corporate reputation and brand loyalty.
Since there are so many factors that play a hand in what constitutes as an effective crisis response message, these two studies also looked into possible cultural, gender, and other differences. Results support previous research about the positive influence emotions have on
audiences, corporate reputation and brand loyalty; but it also provided a more broader understanding of communication specialists’ attitude towards emotions in crisis communication. Contrary to common belief, most communication specialists do use emotions in crisis response messages, but they mostly follow their intuition which suggests a lack of strategy. Those who do think about it strategically, claim that both positive and negative emotions should be used, but choosing the specific emotion depends on the circumstances of the crisis. In conclusion, this mixed method research provided support for existing literature about the positive effects of emotions in crisis communication messages, and it also revealed new insights by focusing on only positive emotions, looking at brand loyalty in the context of crisis communication, and by comparing present practices to scientific findings.
References
Abratt, R., & Kleyn, N. (2012). Corporate Identity, corporate branding and corporate reputations. European Journal of Marketing, 46(7/8), 1048–1063.
https://doi.org/10.1108/03090561211230197
Argenti, P.A. and Druckenmiller, B. (2004). Reputation and the corporate brand. Corporate Reputation Review. 6(4) 368-74.
Ariffin, S. K., Ali, N. N. K., & Kamsan, S. N. A. (2019). The influence of crisis management on Customer Purchase intention. Global Business and Management Research: An International Journal , 10(1), 12–29. https://doi.org/10.15405/epsbs.2019.08.22
Barsade, S. (2020, August 14). The contagion we can control. Harvard Business Review.
Retrieved June 10, 2021, from https://hbr.org/2020/03/the-contagion-we-can-control
Bloise, S. M., & Johnson, M. K. (2007). Memory for emotional and neutral information:
Gender and individual differences in emotional sensitivity. Memory, 15(2), 192–204.
https://doi.org/10.1080/09658210701204456
Cohn, M. A., & Fredrickson, B. L. (2009). Positive emotions. In S. J. Lopez & C. R. Snyder (Eds.), Oxford handbook of positive psychology. 13–24. Oxford University Press.
Coombs, W. T., & Holladay, S. J. (2002). Helping crisis managers protect reputational assets:
Initial tests of the situational crisis communication theory. Management Communication Quarterly, 16, 165–186.
Coombs, W. T. (2007). Protecting organization reputations during a crisis: The development and application of situational crisis communication theory. Corporate Reputation Review, 10(3), 163–176.
Coombs, W. T., & Holladay, S. J. (2005). An exploratory study of stakeholder emotions: Affect and crises. Research on Emotion in Organizations, 1, 263-280.
Cowden, K., & Sellnow, T. L. (2002). Issues advertising as crisis communication: Northwest Airlines’ use of image restoration strategies during the 1998 pilots’ strike. Journal of Business Communication, 39, 193-219. doi:10.1177/002194360203900203
Ekman, P. (1993). Facial expression and emotion. American Psychologist. 48(4), 384–
392. https://doi.org/10.1037/0003-066X.48.4.384
Isaac, M. (2021, October 2). Whistle-blower to accuse facebook of contributing to Jan. 6 riot, memo says. The New York Times. Retrieved October 22, 2021, from
https://www.nytimes.com/2021/10/02/technology/whistle-blower-facebook- memo.html.
Jagdish, N. and Sheth, C.W. (1974). A theory of multidimensional brand loyalty. Advances in Consumer Research. 1(1), 449-459.
Jin, Y., Pang, A., & Cameron, G. T. (2010). The role of emotions in crisis responses: Inaugural test of the integrated crisis mapping (ICM) model. Corporate Communications: An International Journal, 15, 428-452. doi:10.1108/13563281011085529
Jin, Y., Pang, A., & Cameron, G. T. (2012). Toward a publics-driven, emotion-based conceptualization in crisis communication: Unearthing dominant emotions in multi- staged testing of the Integrated Crisis Mapping (ICM) model. Journal of Public Relations Research, 24, 266-298. doi:10.1080/1062726X.2012.676747
Khan, M. A., & Mahmood, Z. (2012). Impact of brand loyalty factors on brand equity. International Journal of Academics Research, 4, 33–37. Retrieved from https://connection.ebscohost.com/c/articles/80229663/impact-brand-loyalty- factors-brand-equity [Google Scholar]
Mcguire, D., Cunningham, J. E., Reynolds, K., & Matthews-Smith, G. (2020). Beating the virus: An examination of the crisis communication approach taken by New Zealand Prime Minister Jacinda Ardern during the Covid-19 pandemic. Human Resource
Development International, 23(4), 361–379.
https://doi.org/10.1080/13678868.2020.1779543
Mestre, M. V., Samper, P., Frías, M. D., & Tur, A. M. (2009). Are women more empathetic than men? A longitudinal study in adolescence. The Spanish Journal of Psychology, 12(1), 76–83. https://doi.org/10.1017/s1138741600001499
Odin, Y., Odin, N., & Valette-Florence, P. (2001). Conceptual and operational aspects of Brand Loyalty: An empirical investigation. Journal of Business Research, 53(2), 75–
84. https://doi.org/10.1016/s0148-2963(99)00076-4
Pam, M. S. N. (2015, June 24). What is negative emotion? Definition of negative emotion.
Psychology Dictionary. Retrieved October 3, 2021, from https://psychologydictionary.org/negative-emotion/.
Schoofs, L., Claeys, A. S., De Waele, A., & Cauberghe, V. (2019). The role of empathy in crisis communication: Providing a deeper understanding of how organizational crises and Crisis Communication Affect Reputation. Public Relations Review, 45(5), 101851.
https://doi.org/10.1016/j.pubrev.2019.101851
van der Meer, T. G. L. A., & Verhoeven, J. W. M. (2014). Emotional crisis communication.
Public Relations Review, 40(3), 526–536. https://doi.org/10.1016/j.pubrev.2014.03.004
Appendix
App. 1: SCCT identifies three crisis clusters: (1) the victim cluster, where the organization is seen as the victim of the event, and has low responsibility; (2) the accidental cluster, where the event is seen unintentional or uncontrollable by the organization, and thus the organization has minimal crisis responsibilities; and (3) the intentional cluster, where the event is seen as purposeful and the organization has a high crisis responsibility.
App. 2: Interview questions
1. Do you use emotions strategically in organizational messages in times of crisis when communicating to stakeholders? Why (not)?
2. What kind of emotions do you communicate in these organizational messages?
3. Do you communicate positive emotions in these messages on purpose? Why (not)?
4. Do you communicate (positive) emotions to stakeholders when the crisis is accidentally caused by the organization? Why (not)?
5. Do you think communicating (positive) emotions in crisis communication effects the resulting corporate reputation, and customer’s brand loyalty?
App. 3: Recruitment message for experiment
“Dear Sir or Madame,
I am contacting you with the kind request to participate in my study which aims to contribute to the understanding of crisis communication messages. This study is created by Nikoletta Mayer, a Master student of the University of Amsterdam (supervised by Rutger de Graaf) as a part of a Master’s thesis and should take about 10 minutes. To start the survey, please follow the link included below:
https://uva.fra1.qualtrics.com/jfe/form/SV_eP9Vxz2PJxgqmsC
It is important that you answer the questions honestly. Any information that you provide will remain completely anonymous. I will not collect your name or other personally identifiable information and you are free to stop at any time; your answers and expressed views will be treated in strictest confidence and neither the research team, nor the University of Amsterdam will be able to identify you from your response. If you have any questions regarding the study, now or later, you may contact Nikoletta Mayer via nikoletta.mayer@student.uva.nl.
I thank you in advance for participating in my survey!
Kind regards,
Nikoletta Mayer”
App. 4: Description of fictitious company “Blyss”
Blyss is a multinational consumer good corporation headquartered in Los Angeles, California, and founded in 1971. It has a range of products belonging to personal/consumer health, personal care, and hygiene, which are categorized into segments such as Beauty, Baby &
Family care, Grooming, and Health Care. Blyss has been leading innovation and challenging the status quo for generations in order to improve its consumers lives and have a meaningful impact on society. Consumers trust that Blyss brands are quality brands, the products are safe to use, and environmental friendly. They constantly test new products, or try and improve their
existing ones based on research and consumer feedback. They are market leaders and their brands can be found in any shop all around the world.
App. 5: Manipulation messages
App. 6: Corporate reputation scale
Respondents had to answer on a 4-point Likert scale, answers ranging from ‘Completely Disagree’ (1) to ‘Completely Agree’ (4). To measure the reliability of the scale, item responses were averaged into an index, the coefficient Cronbach’s alpha was 0.86. The average corporate reputation was (M = 2.74, SD = .65), the skewness was -0.25, while the kurtosis was -0.07.
Items:
1. The organization is concerned with the well-being of its public 2. The organization is basically dishonest
3. I do not trust the organization to tell the truth about the incident
4. Under most circumstances, I would be likely to believe what the organization says 5. The organization is not concerned with the well-being of its public
App. 7: Brand loyalty scale
Respondents had to answer on a 4-point Likert scale, answers ranging from ‘Completely Disagree’ (1) to ‘Completely Agree’ (4). To measure the reliability of the scale, item responses were averaged into an index, the coefficient Cronbach’s alpha was 0.91. The average corporate reputation was (M = 2.54 SD = .64), the skewness was 0.15, while the kurtosis was -0.58.
Items:
1. I like switching from one brand of shampoo to another.
2. I am loyal to only one brand of shampoo.
3. If the brand of shampoo I usually buy is not available in a shop, I go to another shop.
4. On several purchase occasions, it is likely that I will buy each time the same brand of shampoo.
5. During my next purchase, I will buy the same brand of shampoo as the last time.
6. Even if the price of the brand of shampoo I am used to buying strongly increases, I'll still buy it.
7. Generally, I am loyal to a small number of brands of shampoo.
8. Even when I hear negative information about the brand of shampoo I usually buy, I still stick to that brand.
9. I've been buying the same brand of shampoo for a long time.
10. I like trying several brands of shampoo.
11. I always buy the same brand of shampoo.
12. If I like a brand of shampoo, I rarely switch from it.
App. 8: Control variable scales: reliability tests
To measure the reliability of the scales, responses were averaged into an index. The coefficient Cronbach’s alpha for CR was 0.82, and the average CR was (M = 2.75, SD = .56), the skewness was -0.03, and the kurtosis was -0.16. The coefficient Cronbach’s alpha for BL was 0.90, and the average BL was (M = 2.51, SD = .64), the skewness was 0.08, while the kurtosis was -0.64.
