Professional Documents
Culture Documents
Crisis PRinsocialmedia ICA2011
Crisis PRinsocialmedia ICA2011
net/publication/281815504
CITATIONS READS
14 2,651
4 authors, including:
Camiel J. Beukeboom
Vrije Universiteit Amsterdam
31 PUBLICATIONS 692 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
ERC project ReDefTie - Redefining tie strength: How social media (can) help us to get non-redundant useful information and emotional support View project
Mechanisms of linguistic bias: How words reflect and maintain stereotypic expectancies. View project
All content following this page was uploaded by Peter Kerkhof on 16 September 2015.
Please cite as Kerkhof, P. Beugels, D., Utz, S., & Beukeboom, C. J. (2011). Crisis PR in social media. An
experimental study of the effects of organizational crisis responses on Facebook. Paper presented at the 61st
Peter Kerkhof
Dionne Beugels
Sonja Utz
Camiel Beukeboom
VU University Amsterdam
The Netherlands
Paper presented at the 61st Annual ICA Conference, Boston (USA), 26-30 May 2011. Please
University Amsterdam, De Boelelaan 1081, 1081HV Amsterdam, the Netherlands. Tel. 0031
Abstract
In this paper we compare the effects of different organizational crisis responses in social
media after negative publicity regarding the organization. Building on earlier work in crisis
communication (e.g., Bradford & Garrett, 1995; Kim, Avery & Lariscy, 2009), we distinguish
apology and denial as organizational responses. Building on recent work on corporate blogs
by Kelleher (2009), we distinguish a personal and a corporate tone of voice and predict that a
personal tone of voice is more effective in social media because it makes the organization
look more human and more committed to the relationship with its customers.
The hypotheses were tested using a 2 (apologies vs. denial) x 2 (corporate vs.
personal) + 1 control group (no response) between subjects experimental design, building on a
recent case in which H&M got involved in a discussion with its Facebook community after
reports about cutting up and throwing away unsold garments (instead of donating the
garments to the homeless). The 125 participants read an online news report about the incident,
and were then shown the H&M Facebook page on which H&M either apologized or denied
the incident, and did so in a personal or a corporate manner. A control group only read the
The results showed that apologizing led to higher credibility and a more positive
attitude towards the response, but increased the perceived responsibility of the organization
for the crisis event. A personal tone of voice increased conversational human voice,
in the case of denial. The results are discussed in the light of the difficulties that organizations
Key words: crisis communication, PR, social media, apologizing vs. denial, human
conversational voice.
Crisis PR in social media 3
On January 6 2010, a news article appeared in the New York Times stating that the H&M
store in Manhattan discarded bags of unsold and unworn garments. All garments were
destroyed by making holes in them in order to make sure they would not be sold elsewhere
(New York Times, 2009). A few hours later, the Huffington Post – a popular American weblog
– followed the New York Times in reporting about the H&M scandal (Huffingtonpost.com,
2010). Soon after, numerous people made a comment on the Huffington Post website, and the
H&M Facebook page was overflowed with reactions from H&M fans. As a resonse to these
messages H&M came with the following statement: “H&M is committed to take responsibility
for how our operations affect both people and the environment. Our policy is to donate any
that is not representative of our policy. We will follow with more information as soon as we
are able. H&M's US sales operation donates thousands of garments each year through Gifts
In Kind Int.” This statement was posted twice that day and over thousand people 'liked' it but
266 people posted mainly negative reactions to the statement (Facebook, 2010). H&M also
reacted on Twitter in which they referred to their statement on Facebook: “This incident is not
investigating. Please RT.” A lot of retweets followed and the day after, the H&M case was
worldwide trending topic number 2 on Twitter. One day later, newspapers, news websites and
blogs reported about the news. H&M examined the incident and came with the following
statement a few days later: “We have examined the situation surrounding the garments found
outside of our 34th Street store. It is important to note that these garments were already
damaged, and did not meet our safety standards or had been used for in-store display. Going
annually donates hundreds of thousands of garments to charity and aid organizations, and
will continue to do so. For a list of organizations and more information regarding our
corporate social responsibility, please visit www.hm.com/csr.” Again, almost 800 people
'liked' this statement and another 380 H&M fans posted a reaction, the majority being
The H&M incident is not the only example of an organization that uses social media
during times of crisis. Companies like General Motors (during its near bankruptcy in 2009)
and Toyota (the 2010 car recalls) use social media like Twitter and Facebook during times of
crisis to influence public perceptions of the crisis, to answer questions from consumers
regarding the crisis, and to avert the consequences of negative publicity. Negative publicity
has the potential to damage the reputation of an organization, corporate credibility and trust in
an organization (Dean, 2004; Kerkhof, 2009; Perry, Taylor & Doerfel, 2003).
Little is known about the effectiveness of social media PR in restoring trust and in
averting the negative consequences of negative publicity, and little systematic research has
been devoted to this relatively new phenomenon. In this paper we report about an experiment
in which the H&M response was varied among two dimensions: the content of the response
(apologizing vs. denial) and the tone of voice of the response (personal vs. corporate).
communication strategy (Bradford & Garrett, 1995; Kim et al., 2009), whereas denial is
typically considered less effective, but still the most frequently used communication strategy
(Kim et al., 2009). A personal rather than a corporate tone of voice is deemed effective by
several authors. Solis and Breakenridge (2009) argue that in PR 2.0 it is important to
humanize the story and become part of the conversation instead of trying to sell your way into
it. This suggests a personal rather than a corporate response. Similarly, articles in the field of
service recovery and online customer care suggest a personal tone of voice over a corporate
PR in Web 2.0
In just a few years social media have become very popular, both for consumers and
increaslingly for organizations. People like to talk about brands and products in social media
environments (like Twitter, e.g. Jansen, Zang, Sobel, & Chowdury, 2009) and to share
information, insights and opinions, which influences consumer brand perceptions and
purchase intentions (Jansen et al., 2009). Positive online word-of-mouth has become a
Today, a lot of companies have acknowledged the social media's presence by having a
blog, Facebook page or a Twitter account. Organizations can use social media to virtually
meet consumers and create dialogue, for example by receiving feedback or creating an
extension of the customer service. Another important aspect of social media is the opportunity
for organizations to respond swiftly (Kirat, 2007). Wright and Hinson (2009) showed that
88% of the respondents in their study (all PR practitioners) believe that social media make
to criticism. Still, many large organizations struggle with the challenge to communicate
So far there has been very little research about PR 2.0 and effective organizational
responses in social media to negative publicity such as the above mentioned H&M case.
Besides, previous experimental research regarding the effects of crisis responses has almost
exclusively used print media as response stimuli (Coombs & Holladay, 2009). How the public
perceives and experiences these online corporate responses in online media is an important
step in executing more effective online PR. The main research question for this study is how
different types of organizational crisis responses in social media and the tone of voice of those
or from few to many (Hurme, 2001). Wright and Hinson (2009) conclude that social media
have had a huge impact moving PR into the direction of facilitating more two-way
their public without going through news media gate keeping (Kirat, 2007). The interactive
communicative element in PR 2.0 does not exclude traditional PR, but this should according
Previous studies (Perry et al., 2003; Taylor & Perry, 2005) show that traditional public
relation tactics, such as the news release, are still the most prevalent tactics used in crisis and
new communication tactics such as a two-way communication are less frequently used.
However, more recent research about crisis responses suggests that the Internet is already
being used in approximately half of the crisis studied (Taylor & Kent, 2007). The Internet
however has changed rapidly during the last years, with a social network site like Facebook
growing from 100 million in 2007 to over 500 million members in 2011, and with the number
of daily tweets growing from almost none in the beginning of 2007 to over 110 million in
2011.
language, with a focus on human interaction rather than an objective description (Oegema,
Kleinnijenhuis, Anderson & van Hoof, 2008). Solis and Breakenridge (2009) state that PR in
and human manner than organizations are used to operate. Since social media content is
perceived as more personal and less formal, it can be reasoned that a personal response to a
crisis event will be more effective than the kind of corporate response that is typically done in
Kelleher and Miller (2006) and Kelleher (2009) investigated interactive online
human voice as “an engaging and natural style of organizational communication as perceived
building and maintaining a relationship”. In an experimental study Kelleher and Miller (2006)
showed participants either a corporate blog condition or a corporate web page. Participants
assigned to the blog condition perceived the organization's conversational human voice to be
higher than participants assigned to the condition in which they read traditional web pages.
Furthermore, it was found that conversational human voice and communicated relational
commitment and control mutuality (Kelleher & Miller, 2006; Kelleher, 2009).
websites). Recently Kerkhof et al. (2011) extended this line of thinking to tone of voice in
online customer care. As is the case for online PR, online customer care, where employees
respond publicly to customer questions and complaints, takes place in consumer forums or
social network sites where consumer rather than organizations set the rules. This makes it
commitment. Indeed, Kerkhof et al. (2011) show that a personal (vs. a corporate) tone of
A recent study by Yang, Kang and Johnson (2010) established results that are in line
with the findings by Kerkhof et al. (2001) and Kelleher (2009). The use of invitational
rhetoric in a blogpost helped to create the impression that the company is open to dialogue.
Crisis PR in social media 8
communication and elicited positive post crisis perceptions. Following this reasoning, we
expect that a personal (vs. a corporate) tone of voice enhances the perception of the
organization as speaking with a human voice and as committed to having a good relationship
A crisis is a sudden and unexpected event that threatens to disrupt an organization's operation
and poses a financial and reputational threat (Coombs & Holladay, 2005). Crises result in
negative publicity for corporations that may threaten corporate reputation (Coombs, 2007;
Dean, 2004) and can deteriorate the relationship between an organization and its publics
(Coombs & Halladay, 2005; Hearit, 1994). Communication following the crisis plays an
integral role in successful crisis management (Ulmer, 2001), which is a process designed to
prevent or mitigate the damage a crisis can have on an organization and its stakeholders.
communicative responses in order to restore the corporate image. Some offer apologies,
thereby admitting their responsibility, others choose to not react, hoping that the attention to
the issue will diminish. However, silence in the eyes of the public is similar to an admission
of guilt. Others deny their responsibility, which may have the unintended consequence of
relations for the past eighteen years is Benoit's (1995) Image Restoration Theory. This theory
offers five broad categories of image repair strategies: denial, evasion of responsibility,
reducing the offensiveness of the wrongful act, corrective action, and mortification, including
an organization to confess and ask for forgiveness (Benoit, 1997). Bradford and Garret (1995)
responses on the consumer perception of corporate image and found that concession was the
This is in line with findings from Conlon and Murray (1996) who conclude that a corporate
response including an apology and accepting responsibility for a negative event positively
Kim et al. (2009) reviewed the effectiveness of response strategies in crisis situations
in 51 articles over the past eighteen years. According to the authors, the most effective crisis
strategy was full apology, 71.4% of the full apology strategies employed were evaluated as
effective. In their analysis, authors evaluated denial as the least effective strategy: only 13.8%
of the denial strategies employed were evaluated as effective. This result was surprising,
because denial was also found to be the second most often used strategy for organizations
(Kim et al., 2009). According to Coombs (2007a) and Benoit (1997) denial may only be
useful when the crisis challenge is unwarranted or when the organization is not held
responsible.
in the case of an apparent wrongdoing. Attribution theory is based on the premise that people
need to assign responsibility for events; people look for the causes of events, especially
unexpected and negative events (Weiner, 1986; Coombs, 2007). Consequently, these
responsibility allocations from stakeholders will have affective and behavioral consequences
for an organization (Coombs, 2007). We will assess the effects of tone of voice and denial vs.
The type of response by an organization may interact with the tone of voice and
together determine consumer evaluations. According to Bell and Zemke (1987) a personal
apologizing response is more effective because it feels sincere and authentic in comparison
with a typical corporate response, which may come across as a standard and impersonal
response. Also, a narrative apology is more effective than a narrative denial after an integrity-
Crisis PR in social media 10
violation in blog posts, because the violator is seen as more ‘human’ (van Laer & de Ruyter,
2010). Because a personal response is also perceived as more ‘human’, the same could be the
case for a personal response and a corporate response in social media. Therefore, it is
expected that a personal apology, in which someone in the organization is acknowledging that
the organization was at fault, is more effective than a corporate, formal apology.
To summarize, we propose that the organizational response to the crisis event consists of
three response types: apologizing, denial and no response. It is expected that these three
corporate responses will directly influence corporate credibility and responsibility for the
event. Furthermore, two different types of tone of voice are distinguished, a personal and a
corporate tone of voice. It is expected that the tone of voice will affect communicated
relational commitment and conversational human voice. Moreover, tone of voice may
moderate the relationship between the type of response and corporate credibility. Apart from
assessing attitudes we will also assess the cognitive responses to the organizational messages
we use in our study. Petty and Cacioppo (1979) state that direct cognitive responses mediate
Method
Design
We tested our hypotheses using an experimental design. Participants were randomly assigned
to a 2 (type of response: apologizing vs. denial) x 2 (tone of voice: personal vs. corporate)
experimental between subjects research design plus a control group (the no response
condition, see Table 1 for an overview of the 5 conditions). The five conditions can be seen in
Table 1. Participants first saw an online news message in a nu.nl (a Dutch news site) layout, in
which the H&M incident was shortly described. Then, participants in all the experimental
conditions (but not the “no response” condition), saw one of the four manipulated responses
Participants
The online questionnaire was made available online on May 2, 2010 and closed on May 14,
sampling and snowball sampling. First, participants (family and friends) were approached by
e-mail, social network sites and MSN messenger. Furthermore, the potential participants were
asked to send the link to the questionnaire to 5-10 other people. Besides, the questionnaire
A total of 125 participants filled out the questionnaire completely. The age of
respondents ranged from 18 until 63 years old (M=26.27 years). More women than men
participated in this study, 76% (N=95) of all respondents was female and only 14% (N=30)
was male. Most respondents in this study were highly educated: the majority (44%, N=55) of
the respondents had a bachelor’s degree and 27.2% (N=34) had a master’s degree.
Respondents in this study used the Internet between 1-600 minutes a day, with a mean of
124.53 minutes a day. More than 90% of the respondents use social media, spending on
average 27.70 minutes per day on social network sites. Hyves (the largest social network site
in the Netherlands) is most frequently used as a social medium; 76.2% (N=96) of the
respondents use Hyves. Facebook and You Tube are also popular social media in this study,
59.9% (N=75) of the respondents use Facebook and 53.2% (N=67) use You Tube.
Stimulus materials
The participants were introduced to the H&M case with a short Dutch language news article
based on the article that appeared in the New York Times on January 6. The news article used
in this study was formatted to a “nu.nl” layout which is the largest Dutch online news website.
Participants in the four experimental conditions (formal apologizing, formal denial, personal
apologizing and personal denial) then saw the Facebook fan page of H&M including one of
the four manipulated responses of H&M. The Facebook layout was kept in order to keep the
Crisis PR in social media 12
look and feel of a real corporate response in social media. Each participant saw only one of
the four responses. In manipulating the corporate responses, a first distinction was made
between a response with a corporate and a personal tone of voice. In the personal conditions a
photograph of a fictitious person named Anna de Vries was shown accompanying the
response. The two corporate responses used the H&M logo as profile picture. The content of
the formal and personal responses were kept as similar as possible. The main difference
between the corporate and the personal conditions was the perspective from which the
response was written. In the personal condition, both the apologizing and the denial response
were written in the name of Anna de Vries, who speaks using the first person for H&M. In the
corporate condition, the response was written in the name of H&M, using the third person.
A second manipulation included the distinction between the apologizing and the denial
response. The denial response was based on the actual response of H&M. The main message
in the denial conditions was that the clothes were already damaged and did not meet the safety
requirements anymore. In the apologizing condition, the main message was that H&M takes
Pretest
Before the questionnaire was made available online, a pretest questionnaire was developed in
order to check whether the manipulations were successful. A total of 34 participants filled out
the pretest. Participants were randomly assigned to one of the four conditions: corporate
apologizing (N=10), corporate denial (N=8), personal apologizing (N=9) and personal denial
(N=7). Since respondents in the “no response” condition only saw the news article and not
one of the four manipulated responses, only four pretest questionnaires were developed.
Participants first saw the news article which was directly followed by the one of the four
voice manipulation (corporate versus personal) and concerning the type of response
The first question with regard to the denial response was 'In this response, H&M denies
the news message'. Respondents had to indicate on a 7 point Likert scale (1= 'totally agree',
7= 'totally disagree') to what extent they agreed or disagreed with this statement. There was a
significant difference between respondents who saw an apologizing response and respondents
who saw a denial response (t(33)= 2.980, p = .050). Respondents who saw the denial response
(M=3.00, SD=1.710) agreed significantly more with this statement that respondents who saw
the apologizing response (M=4.95, SD=2.012). The second pretest question with regard to the
apologizing response was: 'In this response apologies are offered'. There was again a
significant difference between respondents who saw the apologizing response and
respondents who saw the denial response (t(33)= -6.499, p = .000). Respondents who saw the
apologizing response (M=2.29, SD=1.554) agreed significantly more with this statement than
respondents who saw the denial response (M=5.71, SD=1.490). Therefore, the type of
response manipulations were successful as indicated by the pretest results and no adaptations
statement 'The sender of this message reacts in a formal manner’ on a 7 point Likert scale (1=
'totally agree', 7= 'totally disagree') as all the other questions described below. There was a
marginally significant difference between the personal and corporate condition (t(33) = -1.77,
p = .087), with respondents in the corporate condition (M=3.56, SD=1.92) rating this
statement as more formal than respondents in the personal condition (M=4.59, SD=1.502).
The second question was 'The sender of the response is a person /an employee of H&M.'
SD=0.64) significantly more often indicated that the sender of the response was an employee
than respondents in the corporate condition (M=4.00, SD=2.22). Finally, there was a
Crisis PR in social media 14
marginally significant difference between the two groups on the question: 'I think the sender
of the message is a nice person' (t(33) = 1.785, p = .083), with respondents in the personal
condition (M=3.12, SD=0.86) agreeing more with this statement than respondents in the
Dependent variables
The manipulation check consisted of one item measuring the type of response (apologizing
versus denial): 'How do you experience the response of H&M on Facebook'. The answer
options for this question were 'H&M is apologizing' (1) and 'H&M denies the situation' (2).
Another item was used to check whether the corporate versus personal manipulation was
successful. Respondents were asked to indicate whether the sender of the Facebook response
Conversational human voice was measured with 10 items adapted from Kelleher
(2009). Examples of items are 'H&M invites people to conversation' and 'H&M attempts to
make communication enjoyable'. Respondents had to answer on a 5 point Likert scale (1=
'totally disagree', 5= 'totally agree'). Items were compiled into a mean index (Cronbach's α =
.84).
Also adapted from Kelleher & Miller (2006) and Kelleher (2009) was the scale
measuring communicated relational commitment. The scale consisted of the following six
items: 'H&M attempts to demonstrate they are committed to maintaining the relationship',
relationship', 'H&M stresses commitment to me and others', 'H&M implies that our
relationship has a future/is a long-term commitment', 'H&M directly discusses the nature of
the organization' and 'H&M emphasizes the quality of our relationship'. Respondents had to
indicate to what degree they 'strongly disagreed' (1) or 'strongly agreed' (5) with each of these
statements on a 5 point Likert scale. Items were again compiled into a mean index
(Cronbach's α = .82).
Crisis PR in social media 15
Responsibility was measured with three items adapted from Coombs and Holladay
(2002). In contrast to the two above-mentioned questions, this question was not only available
for the four experimental conditions, but also for the control or 'no response' condition. The
three questions about responsibility were: 'Circumstances, not the organization, are
responsible for the crisis', 'The blame for the crisis lies with the organization' and 'The blame
for the crisis lies in the circumstances, not the organization'. The three items were measured
on a 5 point Likert scale (1='strongly disagree', 5= 'strongly agree') and combined into a scale
(Cronbach's α = .86).
Corporate credibility was assessed in all 5 conditions using eight items on a 7 point
Likert scale adapted from Newell and Goldsmith (2001). According to Newell and Goldsmith
corporate credibility is “the extent to which consumers feel that the firm has the knowledge or
ability to fulfill its claims and whether the firm can be trusted to tell the truth or not”. The
items in this scale measure the organizational credibility by two dimensions, namely the
credibility was assessed in all conditions, including the control condition. Examples of items
are 'H&M is skilled in what they do' and 'H&M makes trustful claims', referring to expertise
and trustworthiness respectively. Answer categories ranged from 'totally disagree' (1) to
'totally agree' (7). This scale was compiled into a mean index (Cronbach's α= .80).
Attitude towards the response is only measured in the four experimental conditions.
Participants indicate to what degree they think the response of H&M is 'totally adequate'/'not
adequate at all', 'very honest'/'not very honest', 'very sincere'/'not very sincere' and 'very
trustworthy'/ 'not very trustworthy'. The scale, which consisted of a mean index, was reliable
(Cronbach's α = .71).
Cognitive responses. Directly after respondents saw the one of the four organizational
responses, their cognitive responses were assessed. The cognitive response question was an
open-ended question in which respondents had to write down all thoughts that came up after
Crisis PR in social media 16
reading the organizational response. This cognitive response variable was recoded into a
variable measuring positive cognitive response (0= 'no positive cognitive response', 1=
'positive cognitive response) and a variable measuring negative cognitive response (0 = 'no
coded as positive when people said something positive about the response. For example, “a
positive reaction of H&M”, “an appropriate response of H&M” or” H&M reacts good, I
believe them”. A cognitive response was coded as negative when people said something
negative about the response or about the company H&M. Examples are: “This is nonsense,
damaged clothes can also be sent to charities”, “a very simple response of H&M which does
not have a content” or “It is very easy to respond like this, probably a way to attract the
Results
Manipulation check
The manipulation of apologies vs. denial was successful (Pearson chi square = 29.43, p =
.000). In the apology condition respondents perceived the response more often as an apology
Also, there was a marginally significant difference between a personal and a corporate
tone of voice on the question concerning the sender of the message (Pearson chi square =
3.25, p = .071. In the personal condition respondents perceived the response more often as
In the first analysis, we compared the effects of responding vs. not responding using an
ANOVA comparing the mean scores in all five groups on credibility and responsibility. There
was a marginally significant difference between the responsibility scores in the 5 conditions,
F(4,120)=2.072, p = .089. However, the Bonferroni post-hoc test showed that none of the
Crisis PR in social media 17
To examine the effects of the four different types of responding, a 2 (type of response:
apologizing versus denial) x 2 (tone of voice: corporate versus personal) between subjects
ANOVA was performed. The respondents in the no response condition (control condition)
were not taken along in this analysis. A summary of the results of this 2 x 2 ANOVA can be
A marginally significant main effect of the type of response on corporate credibility was
found, (F(1,95) = 2.812, p = .097). Respondents in the apologizing condition (M= 4.75, SD =
.78) find H&M more credible than respondents in the denial condition (M=4.48, SD = .75).
Also, type of response (apologizing versus denial) had a significant main effect on the
mediating variable responsibility (F(1,95) = 5.985, p = .016). This indicates that respondents
in the apologizing condition (M= 3.81, SD = .88) find H&M more responsible for the incident
than respondents in the denial condition (M=3.37, SD = .96). Furthermore, it appeared that
type of response had a marginally significant main effect on the attitude towards the response
SD = .99) have a more positive attitude towards the response than respondents in the denial
condition (M=3.88, SD = 1.04). No significant main effects of type of response were found on
Counter to expectations, there was no significant main effect of tone of voice (personal versus
corporate) on corporate credibility. As expected, there was a significant main effect of tone of
voice on conversational human voice (F(1,95) = 8.213, p = .005). Respondents in the personal
conditions rated the H&M response as more human (M=3.21, SD = .54) than respondents in
Crisis PR in social media 18
the corporate conditions (M=2.84, SD = .71). Furthermore, there was a main effect of tone of
the personal conditions also rated H&M’s relational commitment higher (M=3.55, SD = .69)
than respondents in the corporate conditions (M=3.24, SD = .75). Additionally, there was a
Respondents in the corporate conditions reported almost twice as many (M=0.47, SD = .50)
negative cognitive responses as the respondents in the personal condition (M=0.24, SD = .43).
There were no effects of tone of voice on responsibility, credibility, attitude towards the
There was a significant interaction effect of type of respons and tone of vice on
personal denial condition (M=3.70, SD = .70) scored higher on commitment than respondents
in the corporate denial condition (M=3.06, SD = .74). Bonferroni post-hoc tests in a simple
ANCOVA with the four experimental groups as factor and conversational human voice as the
dependent variable, showed that this difference was significant and there were no other
significant differences between the groups. Thus, when H&M denied the incident,
communicated relational commitment was higher in the personal condition than in the
corporate condition. There was no difference between a personal and a corporate apology.
There was another interaction effect on conversational human voice (F(1,95) = 4.263,
a personal or corporate manner when H&M denies the incident. Bonferroni post-hoc tests in a
simple ANOVA with the four experimental groups as factor and conversational human voice
as the dependent variable, show that there is a significant difference between a formal denial
response (M=2.63, SD = .69) and both a personal apologizing response (M=3.12, SD = .55)
Crisis PR in social media 19
and a personal denial response (M=3.25, SD = .55). A personal apology and a personal denial
Some limitations should be noted before drawing conclusions from the results [presented in
this paper. This study only investigates one case, within one particular social network site, and
with two types of corporate responses to crisis situations. A variation in each of these may
alter the results and one should thus be careful not to generalize to easily to other situations
Notwithstanding the limitations, we believe there are some lessons to be learned from
our study. Tone of voice matters in social network sites like Facebook and probably in other
social network sites too. A personal tone of voice led to higher communicated relational
commitment includes the concepts ‘assurance’ and ‘openness’. Assurances include those
efforts an organization takes to assure that the concerns of the public are acknowledged
(Kelleher, 2009). A personal tone of voice makes the public believe that an organization cares
about the ideas, opinions and reactions of the public on the incident. The other concept,
openness, refers to the fact that people in an organization can freely discuss their own
thoughts about the organization (Kelleher, 2009). A personal tone of voice may lead to more
perceived openness by the public, since the response is from a person within the organization
who discusses an important and sensitive event. A corporate response may signal distance
between the organization and its employees and therefore less assurances and openness. This
which it was found that a personal response was perceived as more human and led to higher
tone of voice on conversational human voice were also found in the current study. Kelleher
Crisis PR in social media 20
(2009) already pointed out in his study that blogs have a positive effect on conversational
human voice and this study adds that responding personally on the social medium Facebook
positively affects conversational human voice. Kuhn (2005) stated that the interactive element
Surprisingly, a personal denial was the most effective response, leading to more
apology did not differ from any of the other responses. With regard to conversational human
voice, a corporate denial was seen as significantly less human compared to a personal apology
and a personal denial. Both personal responses led to higher communicated relational
commitment and conversational human voice. A possible explanation for these findings could
be that consumers already feel sympathy for the organization that apologizes. The public may
judge the apology as an open and honest response, which implies taking the concerns of the
public into account, regardless of whether the response is corporate or personal. Denial on the
other hand does not lead to sympathy and in that case a personal style of communicating adds
credibility. Even though H&M is perceived as more responsible for the incident when
apologizing, apologizing led to a more positive attitude towards the response when compared
takes the blame for the incident and admits its own wrongdoings. The opposite is the case for
denial: an organization indirectly pronounces that the incident was not their fault and is not
taking responsibility. Previous research regarding integrity violations in blog posts (van Laer
& de Ruyter, 2010) already indicated that an apology is positively related to guilt and people
regard guilt admissions as a negative sign, especially when the wrongdoing is a moral
transgression. Still, in our study as in many other studies, apologizing is positively evaluated
Crisis PR in social media 21
and leads to higher credibility. Apologizing may lead to the impression that the company is
In sum, the effects of a personal (vs. a corporate) tone of voice are more consistently
positive than the effects of an apology (vs. denial) and may even compensate for the negative
effects of denial. Although crisis PR in social media may in many ways be similar to
traditional crisis PR (e.g., the effects of apologies), the tone of voice effect may be more
characteristic of social media than of traditional media. Future research should aim at
unraveling the precise mechanisms behind the effects of a personal tone of voice.
References
Bell, C. R. & Zemke, R. E. (2004). Service breakdown: the road to recovery. Management
Benoit, W. L. (1997). Image repair discourse and crisis communication. Public Relations
892.
39(4), 1040-1056.
Coombs, W. T. & Holladay, S. J. (2002). Helping crisis managers protect reputational assets.
Coombs, W.T. & Holladay, S. J. (2005). Exploratory study of stakeholder emotions: affect and
Coombs, W.T. & Holladay, S.J. (2009). Further explorations of post-crisis communication:
Galloway, C. (2005). Cyber-PR and 'dynamic touch'. Public Relations Review, 31, 572-577.
Hearit, K. M. (1994). Apologies and public relations crises at Chrysler, Toshibu and Volvo.
Huffington Post (2010). H&M, Wal-Mart destroy unsold clothing. Retrieved from the World
hm-wal-mart-destroy-nsol_n_413234.html.
6(2), 71-75.
Jansen, B. J., Zang, M., Sobel, K. & Chowdury, A. (2009). Twitter power: Tweets as
electronic word of mouth. Journal of the American society for Information Science
188.
Kelleher, T., & Miller, B.M. (2006). Organizational blogs and the human voice: relational
395-414.
Crisis PR in social media 23
Retrieved from the World Wide Web on January 14, 2010 from
http://www.allacademic.com/meta/p300319_index.html
Kerkhof, P., Vonkeman, C., Beukeboom, C. & Utz, S. (2011). Customer service as PR:
Kim, S., Avery, E. J. & Lariscy, R. W. (2009) Are crisis communicators practicing that we
Kirat, M. (2007). Promoting online media relations: Public relations departments' use of
Kuhn, M. (2005). C.O.B.E: A Proposed Code of Blogging Ethics. Paper presented at the
New York Times (2010). A clothing clearance where more than just the prices have been
slashed. Retrieved from the World Wide Web on February 10, 2010 from
http://www.nytimes.com/2010/01/06/nyregion/06about.html?ref=nyregion.
Newell, S. J., & Goldsmith, R. E. (2001). The development of a scale to measure perceived
Oegema, D., Kleinnijenhuis, J., Anderson, K. & van Hoof, A. (2008). Flaming and blaming:
the influence of mass media content interactions in online discussions. In: Konijn, E.
A., Utz, S., Tanis, M. & Barnes, S. B. (Eds.). Mediated interpersonal communication.
Petty, R. E. & Cacioppo, J. T. (1979). Issue involvement can increase or decrease persuasion
Solis, B. & Breakenridge, D. (2009). Putting the public back into public relations. How social
media is reinventing the aging business of PR. Upper Saddle River, New Jersey:
Taylor, M. & Kent, M. L. (2007). Taxonomy of mediated crisis responses. Public Relations
Taylor, M. & Perry, D. C. (2005). Diffusion of traditional and new media tactics in crisis
11, 51-73.
van Laer, T., & de Ruyter, K. (2010). In stories we trust: How narrative apologies provide
Weiner, B. (1986). An attributional theory of motivation and emotion. New York, NY:
Springer-Verlag.
Wright, D. K. & Hinson, M. D. (2009). An updated look at the impact of social media on
Yang, S, Kang, M., & Johnson, P. (2019). Effects of narratives, openness to dialogic
Xie, Y., & Peng, S. (2009). How to repair customer trust after negative publicity: The roles of
572 – 589.
Crisis PR in social media 25
Table 2: Results of the 2 (type of response: apologizing versus denial) x 2 (tone of voice
(personal versus corporate) ANOVA