‘IT MAKES OR BREAKS YOU!’ THE IMPACT OF EMPLOYEE ENGAGEMENT VIA SOCIAL MEDIA ON YOUR REPUTATION AS EMPLOYER.
Master’s theses
MSc. Human Resource Management MSc. Marketing Management
University of Groningen Faculty of Economics and Business
FREDERIK FOKKE-GEERT HOEKSTRA Student number: 1905309 Voorpiek 22 8322 AC Urk Tel.: +31 (0)6 42 67 77 67 E-mail: f.f.g.hoekstra@gmail.com June 21, 2015 Supervisor MSc. T. Vriend Co-assessor Dr. J. A. Voerman Acknowledgement:
2
‘It makes or breaks you!’ The impact of employee
engagement via social media on your reputation as
employer.
ABSTRACT
Since we live in an ‘internet era’ there is a shift to online social dialogues that has caused a
rapidly increased impact of social media in our lives. Organizations often do not have the power
to control the online communication between individuals who share information about
companies’ brands and products. Employees are essential within this context as ambassadors of the employer, since they can make or break the reputation of employer and company. Although
there are risks associated with social media usage by employees, there are also great opportunities
from a recruitment perspective of the employer to attract potential talents outside the company. In
this study, I investigate whether an employee’s job satisfaction and personality impacts their
social media usage, and whether this social media usage influences the employer’s reputation. Both from marketing and organizational behavior, theories are applied in this study to investigate
the antecedents and effects of employees’ eWOM. Online questionnaires are used to collect data among employees from the field and an experiment is conducted in the lab.
3
INTRODUCTION
Word-Of-Mouth (WOM), which is the interpersonal communication about a brand, product,
service, specific job or an employer that is not under the direct control of the organization (Dichter,
1966; Van Hoye, 2013), has a significant influence on the company’s image and attractiveness as an
employer (Van Hoye & Lievens, 2007; Collins & Stevens, 2002). A lot of research has been conducted
to investigate the effect of WOM on reputation from a consumers’ perspective. However this holds as well for the reputation of employers from the view by (future) employees. With regards to consumers
and WOM, approximately 74% of the consumers base their purchase decisions on WOM as a key
influencer (Ogilvy, 2014) and 92% of people trust recommendations from friends and family over all
other forms of advertising (Rise Smart, 2014). From a recruitment perspective, Van Hoye states:
“Employee referrals (as a example of WOM) have proven to be an effective way to attract and retain the most talented workforce, which is crucial for organizational success and survival” (2013, p. 451). As such, WOM has traditionally been an effective method for recruiters to recruit employees (Breaugh,
2008; Van Hoye & Lievens, 2009; Van Hoye, 2013;). WOM has even a stronger effect on attraction
outcomes than any other recruitment source like recruitment advertising, sponsorships, and positive
publicity (Van Hoye, 2013; Collins & Stevens, 2002).
Although individual communication used to go by channels that favor WOM, the rise of the
internet has changed communication. Instead of relying on traditional channels such as face-to-face
communication and telephone calls, individuals increasingly make use of mobile technologies (e.g.
smartphones, tablets, and laptops) to communicate with others, which is called electronic Word of
Mouth or Word of Mouse (eWOM; Reyt & Wiesenfeld, 2014). eWOM is “any positive or negative
statement made by potential, actual, or former customers about a product or company, which is made
available to a multitude of people and institutions via the Internet” (Hennig-Thurau et al, 2004: 39). The
4 neutral advice or warnings) and not only customers but also employees (as internal customers) utilize
eWOM. Electronic Word-of-Mouth can be perceived as such an exchange between employees and
others since it is any statement about a brand or company that persons share via the Internet (e.g., social
networks etc.) (Kietzmann, & Canhoto, 2013). Electronic Word-of-Mouth (eWOM) allows consumers
or employees to socially interact with one another, exchange information related to companies, and to
decide about informed purchases or to apply for a job via computer-mediated conversations (Blazevic,
Hammedi, Garnefeld, Rust, Keiningham, Andreassen, Donthu, & Carl, 2013; Hoffman & Novak, 1996).
Most common examples of eWOM used by employees are online blogs, posts or updates on social
networking sites about their job and organization. Comparing to traditional ways of advertising, eWOM
is believed to be rather influential because of its naturalistic source (Keller, 2007), the ease of retrieving
(Bakos, 1997; Hoffman & Novak, 1997), and the speed with which eWOM can be transmitted
(Jurvetson, 2000; Mohr, 2001). While traditional WOM can reach a limited audience (e.g. face-to-face,
telephone calls etc.) and is passed on by family and friends in the inner circle, eWOM can reach many
more known or unknown people via different online channels and has the potential to create an
enormous exposure (Cheung & Thadani, 2012; Kozinets, De Valck, Wojnicki, & Wilner, 2010; Zhu &
Zhang, 2010).
Although its prevalence, there is not much known about employees’ motivations that drives
them to engage in eWOM about their employer and its impact on the company’s reputation as well as on employer’s attractiveness perceived by others outside the firm (Van Hoye, 2013). There is a distinction between positively and negatively framed eWOM categorized as the valence of eWOM. Especially, the
investigated impact of negative eWOM has inconsistent results in the literature (Jaidi, Van Hooft, &
Arends, 2011; Kanar, Collins, & Bell, 2010; Van Hoye & Lievens, 2007; Collins & Stevens, 2002) and
therefore it is still unclear what the effect of negatively framed eWOM is on the reputation of employers
5 2008; Gotsi & Wilson, 2001; Lau & Ng, 2001; Shamma & Hassan, 2009) and most resources are
deployed to cope with negative eWOM by organizations (Williams & Buttle, 2014).
Using theories on equity, dispositional approach, and affectivity, I argue that an employee’s
job satisfaction and personality determine whether and when the employee uses eWOM, and that
eWOM influences the perceived employer brand. Specifically, I argue within this research paper that job
satisfaction is positively related to the valence (positive versus negative) of eWOM and that neuroticism
and extroversion moderate this relationship.
I will test these relationships within a field study among employees and their eWOM usage
(Study 1). The purpose of the first study is to investigate the of job (dis)satisfaction on employees’
eWOM behavior (positively versus negatively framed), and the moderator effects of extroversion and
neuroticism. Subsequently, eWOM communicated by employees influences receivers. Research has
shown that eWOM communication impacts the attitudes and perceptions that receivers have towards a
company. The reputation of an employer among job seekers’ is based on how others affectively view the
company as an (potential) employer and what (positive or negative) others subsequently communicate
about an employer (Cable & Turban, 2001; Chang, 2005; Cheung, Lee, & Thadani, 2009; Cheun &
Thadani, 2012; Rokkaa, Karlsson, & Tienari, 2014). Moreover, receivers’ prior knowledge of the brand
or previous experiences with the organization known as brand awareness, determine whether the valence
of eWOM communication may be less or more effective (Doh & Hwang, 2009; Gupta & Harris, 2005;
Park & Kim, 2008). Therefore, I advocate that the valence of eWOM affects the employer brand of an
organization, and that the personality and prior knowledge of the receiver outside the company
influences the perception process which is based on the theories of regulatory focus, affectivity and
psychology like halo effect and confirmation bias. I will test these relationships in an experimental lab
6 This thesis contributes to the academic literature by investigating the effect of eWOM on
employers’ reputation. Especially, the effect of negative eWOM on reputation is investigated since inconsistent outcomes has been demonstrated from previous research. Moreover, to my knowledge there
is no single study that combines eWOM with employees. Employees can be seen as internal
stakeholders (De Bussy, Ewing, & Pitt, 2003). There is a lot of literature from marketing about WOM
and eWOM between consumers (e.g. online product reviews), however the impact of eWOM between
companies’ employees and external stakeholders (e.g. job seekers) is still ambiguous. Especially, it is unknown how employees influence the perceived reputation by outsiders of the organization determined
by the employee’s (dis)satisfaction combined with their personality. From a practical perspective this
thesis shines a light on how organizations can stimulate eWOM among employees. An increasing
number of organizations are applying employee referral programs to current employees that had been
rewarded (often monetary bonuses) when they have recommended their employer to others (Shinnar,
Young, & Meana, 2004). However, job satisfaction is still a better predictor of employee referral than
extrinsic rewards (Van Hoye, 2011). Moreover, since eWOM has the potential to redistribute the power
to control the social dialogue from organizations to consumers (Kozinets, De Valck, Wojnicki, &
Wilner, 2010) it is important to know as employer to what extend employees’ negative eWOM can
damage your reputation while the bright side means that positive eWOM provides opportunities to build
your brand. Particularly, since eWOM is seen as highly influential for peoples’ decision-making (Keller,
2007, Bakos, 1997; Hoffman & Novak, 1997) where positive employer branding via eWOM by
employees might be utilized optimally as effective recruitment source in the future.
7
ANTECEDENTS OF EWOM
Job satisfaction
Job satisfaction is “a pleasurable or positive emotional state resulting from the appraisal of one's job or job experiences” (Locke, 1976: 1308). As explained by Locke’s affective view on job
satisfaction, employees compare what they have expected/desired and what they have perceived within
their job. The relationship between expected and perceived job components concerns not only concrete
job facets like payment level, bonuses, task variety etcetera, but also emotional aspects like fulfillment,
recognition or attention from the supervisor.
To determine what an employee can or should expect a norm or standard is needed. The
equity theory describes that people are able to define a norm by comparing themselves with others. It is
the balance between inputs that employees bring to a job and the outcomes that they receive from it
compared with the perceived inputs-outputs of others (Adams, 1965). Job satisfaction taken from the
perspective of equity theory can be defined as employees’ overall feelings and experiences with their
job and organization (Mueller & Kim, 2008) based on social comparison to others or previous
input-output ratios. The comparable other could be a colleague, friend, relative, the industry norm, a group of
individuals or even oneself in a past job. According to Adams (1965), people are more concerned about
the perceived fairness of distributed outcomes comparing themselves with others or past experiences
rather than the absolute level of outcomes. As a consequence of job satisfaction, positive experiences
with the company motivate people to give ‘something in return’ because people desire equitable and fair exchanges. However, as a consequence of job dissatisfaction employees might want to compensate the
perceived loss between their input and the received output or loss of expected benefits. For instance in
the perceived fairness of the amount of benefits employees receive, Cohen-Charash and Spector (2001)
have stated that "when employees perceive injustice, they might hurt the organization to make the
8 specifically to users of the Internet it could mean: “If an employee or consumer feels he or she has
received a higher output/input ratio than the company, then helping the firm by recommending its
offerings over the Internet is one way the output/input ratio can be equalized” (Parthasarathy & Forlani, 2010).
Subsequently, they way people deal with perceived (im)balance of input/output ratio can be
also be explained by positively either negatively disconfirmation in the light of equity theory. The equity
theory explains the distributed fairness of efforts invested and benefits received from someone’s job
compared to others, while disconfirmation is rather about individual’s expectations. Positively disconfirmation means that a person’s expectations are exceeded however when a person’s experiences are worse than expected there will be negatively disconfirmation (Oliver, 1985) and this might end up
into frustration or compensation behavior by this person to restore the unequal exchange. For instance, a
new employee has certain expectation of his new job and organization. By investing some extent of
effort he expects to receive certain rewards or at least a fair exchange between his input (effort, hard
work, enthusiasm, skills, flexibility etc.) and output (salary, recognition, job security etc.) given by the
organization (Janssen, 2001). In the case of negatively disconfirmation employees might be prone to
restore this balance by de-motivation, reduced effort, withdrawal, becoming disgruntled, or even
disruptive by spreading negative feelings about their employer or work to others (Adams, 1965). With
positively disconfirmation there is likely for a ‘fair exchange’ reason, the desire to help the company to find appropriate employees (Van Hoye, 2011; Van Hoye & Lievens, 2009). An individual’s satisfaction
is strongly related to disconfirmation (Niedrich, Kiryanova, & Black, 2005). Employees who are highly
satisfied with their job are more prone to experience positive emotions compared with those who are
less satisfied (Agho, Mueller, & Price, 1991). Since job (dis)satisfaction leads to (negative) positive
9 goes without saying that these referrals not only holds for offline communication (WOM) but also for
online referrals (eWOM).
A motive to use eWOM as employee is to find appropriate employees or to praise your
employer in order to restore a ‘fair exchange’ balance because of a higher perceived output-input ratio
(Van Hoye, 2011; Van Hoye & Lievens, 2009). From marketing research we know more about motives
for making referrals or spreading information to others (De Matos & Rossi, 2008). For instance,
individuation and self-identification are motives in order to signal the communicator his distinctiveness
to others and to establish his identity (Ho & Dempsey, 2010; Mangold & Miles, 2007). This desire for
being included and to stand out of the crowd is especially strong when the individual wants to be
identified with an organization as a specific social group of employees (Griskevicius, Tybur, Sundie,
Cialdini, Miller, & Kenrick, 2007). Extroversion as personality trait is another important influencer that
determines how people engage in eWOM (Ryan & Deci, 2000; Shinnar, Young, & Meana, 2004;
Picazo-Vela, Chou, Melcher, & Perason, 2010). One of the strongest motives for engaging into positive
referral was the pro-social desire to help other people with finding good fitting jobs together with the
motive of employees’ own job satisfaction (Grant, 2008; Hennig-Thurau et al., 2004; Van Hoye & Lievens, 2009). Satisfaction influences individuals to engage in eWOM in order to balance the
psychological inside tension coming from the input-output ratio (Hennig-Thurau et al., 2004; Adams,
1965). Satisfied individuals with positive experiences can restore balance through communicating
positive emotions to reduce his or her psychological tension caused by the joy coming from experienced
satisfaction (Sundaram, Mitra, & Webster, 1998; Dichter, 1966). However, there is still a debate
concerning eWOM valence as a consequence of satisfaction as research has found that engaging into
negative eWOM is even possible when the individual is satisfied (Parthasarathy & Forlani, 2010). While
East, Hammond, and Wright (2007) have found that positive referrals are three time as common as
10 satisfaction on the valence of eWOM. Based on the literature the following statement can be
hypothesized:
Hypothesis 1a. Job satisfaction is positively related to positive eWOM.
Neuroticism as a personality trait has been demonstrated as a predictor of negative referrals
(De Matteo, Bonezzi, Peluso, Rucker, & Costabile, 2011). Moreover, dissatisfaction has been found as
the best predictor of negative referral together with the desire to help job seekers avoid bad fitting jobs
as a motive (Van Hoye, 2011). People aim to restore their originally balanced state of being after it has
been unbalanced (Hennig-Thurau et al., 2004). By venting a negative experience caused by job
dissatisfaction the worker may restore this balance that helps to reduce the discontent associated with
the employee’s negative feelings (Parthasarathy & Forlani, 2010). Therefore, following hypothesis can be drafted:
Hypothesis 1b. Job satisfaction is negatively related to negative eWOM.
Personality
Not only the perceived fairness of distributed outcomes compared to others determines the
consequences of job (dis)satisfaction but also someone’s personality makes people respond differently
to job satisfaction. Highly extroverted persons care about their social long-term relationships with others
and therefore prefer a positive balance of ‘fair’ exchanges (Raja, Johns, & Ntalianis, 2007). However, neurotics will not engage into relationships that require long-term commitments, trust in others, and
11 that contains exchanges for specific aspects that do not demand confidence and high initiative according
to Raja, Johns, and Ntalianis (2007).
The dispositional approach explains this difference between extroversion and neuroticism
since it suggests that job satisfaction is an individual trait, because persons differ in their tendency to be
satisfied with their jobs (Staw, Bell, & Clausen, 1986; Judge & Larsen, 2001) and differences in
personality have a significant influence on how individuals respond to their job satisfaction (Agho,
Mueller, & Price, 1993; Barsade & Gibson, 2007; Connolly & Viswesvaran, 2000; Jex, 2002; Judge &
Kammeyer-Mueller, 2008; Judge & Larsen, 2001; Kim & Chung, 2014; Staw & Ross, 1985).
Extroversion can be described as outgoing, talkative, energetic behavior, whereas low levels
of extroversion (introversion) is manifested in more reserved and solitary behavior (Thompson, 2008;
Costa & McCrae, 1992). It has been demonstrated that highly extroverted individuals show especially
pronounced reactions to positive experiences (Bolger & Schilling, 1991; Larsen & Ketelaar, 1991;
Luhmann & Eid, 2009; Verduyn & Brans, 2012). Extroverts have a strong tendency to approach because
they are primarily motivated by pleasure or reward, even in novel situations (Gray, 1994).
From previous research it has been demonstrated that extroversion is often referred to as
positive affectivity and neuroticism is referred to as negative affectivity since both have found to be
related to affectivity (Judge, Heller, & Mount, 2002; Landy & Conte, 2010; Clark & Watson, 1992). A
person with positive affectivity is prone to perceive things through ‘pink lens’ (Barsade & Gibson, 2007). Landy and Conte (2010) describe this proverbial difference as someone “who sees the glass half
full instead of as half empty” (p.429). Even more, difference in affectivity may indirectly influence job
satisfaction since affectivity likely impacts the way (positive or negative) and how strong individuals
perceive job circumstances like working conditions and pay (Brief & Weiss, 2002). Therefore, this
dispositional affect theory as an extension of the dispositional approach explains the opposite effect of
12 overall tendency to respond to situations in stable, predictable ways (Barsade & Gibson, 2007; Judge &
Larsen, 2001; Staw & Ross, 1985; Watson & Clark, 1984).
The rise of Internet and popularity of social networking sites as modern way of
communication supports extroverted people to connect with more colleagues, friends, and even
strangers independent of distance to the other person (Wilson, Fornasier, & White, 2010). Extroverted
persons are more likely to communicate with others compared to introverts (low level of extroversion)
and extroverts have the tendency to be more involved in social activities. Furthermore, extroverted
people desire to have strong interpersonal relations, which promotes the positive aspects of an
individual’s psychological feelings while introverts do not have this desire or at least to a lesser extent (Wilson, Fornasier, & White, 2010). Wang, Jackson, Zhang, and Su (2012) have found that extroverts
have significantly more (online) social network friends relatively to introverts and use social networks
more to communicate with others. As extroverted people desire to extend their social network, social
media can facilitate this need and therefore extroverted people actively seek social relationships (Oh,
Ozkaya, & LaRose, 2014). Social media do not only promote interpersonal interactions but also enhance
positive self-evaluations of the sender. Since extroverts are sociable, they prefer to have significant
interactions with others by expressing their own thoughts and feelings (Chiaburu, Stoverink, Li, &
Zhang, 2013; Panaccio & Vandenberghe, 2012). Moreover, extroverts preferably interact and exchange
their thoughts and feelings to enhance positive self-evaluations either self-promotion, social media will
be used particularly for positive eWOM. Therefore, the following hypothesis can be formulated:
13 Extroversion and neuroticism have been argued and shown to influence the strength
(frequency) and direction (valence) of engaging into eWOM (Chiaburu, Stoverink, Li, & Zhang, 2013;
Correa, Hinsley, & De Zúñiga, 2010 ). Contrary to extroversion, neuroticism (low emotional stability) is
an individual’s tendency to exhibit poor emotional adjustment and experience negative affects (e.g. hostility, fear, and depression) (Goldberg, 1990). Neurotic individuals are inclined to be fearful of novel
situations and vulnerable to feelings of helplessness and dependence (Kim & Chung, 2014; Wiggings,
1996). According to Gray (1994) neuroticism is responsible for individual differences in response to
cues of punishment and frustration. “They are vigilant for signs of impending punishment or frustration
in the environment and as such are sensitive to aversive stimuli” (Gray, 1994; 330) like low job
satisfaction. Individuals with high neuroticism are prone to perceive situations as more unpleasant and
process information in a negative way as persons with high levels of negative affectivity are more prone
to experience less job satisfaction and have the tendency to perceive things through the ‘black lens’
(Barsade & Gibson, 2007). Moreover, neurotics have more negative experiences with their work and
therefore rather sensitive to and more focused on processing these negative emotions (Zhai, Willis,
O'Shea, Zhai, & Yang, 2013; Judge, Locke, & Durham, 1997) as highly neurotic individuals
demonstrate pronounced psychological reactions in particular to negative experiences (Bolger &
Schilling, 1991; Larsen & Ketelaar, 1991; Luhmann & Eid, 2009; Verduyn & Brans, 2012).
When dissatisfied people are highly neurotic they do often engage into eWOM via social
media since they may feel misinterpreted in face-to-face social situations and, therefore, might prefer
online interactions since in this online context they feel less restrained (Correa, Hinsley, & De Zúñiga,
2010; Amichai-Hamburger, Wainapel, & Fox, 2002). People are more likely to write about negative
disconfirmed expectations if the experience was an electronic or personal-but-distant interaction
(Kietzmann & Canhoto, 2013). Furthermore, persons who feel less comfortable with themselves and
14 Internet. Moreover, highly neurotic persons have greater instant messaging use because it permits
additional time to contemplate responses compared to face-to-face interaction that makes it easier to
communicate with others for highly neurotics (Ehrenberg, Juckes, White, & Walsh, 2008; Ross, Orr,
Sisic, Arsenault, Simmering, & Orr, 2009). Another explanation when (dis)satisfied employees engage
into eWOM is that neuroticism is often related to loneliness and therefore highly neurotics with anxiety
and nervousness may use social media to seek support and company (Ehrenberg, et al., 2008). Expected
is a neurotic person with low job satisfaction (dissatisfied) will have the tendency to engage into
negative eWOM because they are more prone to experience negative emotions and rather process
information in a negative way (‘the glass is half empty’).Therefore, the next statement can be hypothesized:
Hypothesis 3. The negative relationship between job satisfaction and negative eWOM
(Hypothesis 1b) is moderated by neuroticism, such that the relationship becomes stronger when neuroticism is high rather than low.
15
FIGURE 1: CONCEPTUAL MODEL STUDY 1: INTERNAL STAKEHOLDER (THE EMPLOYEE)
STUDY 1
Method
The purpose of Study 1 was to investigate whether job satisfaction would have an effect on
employees’ eWOM behavior, and whether this would be moderated by the extraversion and neuroticism of the employee. As explained by Cheung and Giu (2006) many job-related factors that could determine
overall job satisfaction only arise from concrete experience with the real-life job, rather than from an
experiment with manipulated tasks. Hence, a field study was chosen to sustain a high external validity
(Pelham & Blanton, 2007: 63). Data were collected through an online survey amongst a sample of from
a medium sized organization in The Netherlands that operates in the food industry (Sample A) and a
16 Sample A. Data for Sample A were gathered from a medium sized organization in the
Netherlands that operates in the food industry as a fish-processing factory. The organization employs
approximately 250 staff. An online survey was administrated to all employees. A total of 98
completed responses were returned (a response rate of 39.2%). Out of all respondents there were 54
males (55.1%) and the mean respondent’s age was 25.82 years (SD = 11.18). Their average organizational tenure was 7.73 years (SD = 8.52), which ranged between 2 weeks and 40 years.
Furthermore, among all respondents the majority had lower (34.7%, VMBO or secondary school) or
intermediate vocational education (38.8%, MBO) where the minority has college (12.1%, HBO) or
university level of education (5.1%, WO) and among 9.2% of all respondents their highest level of
education was unclear.
Sample B. Data for Sample B were gathered among friends and family who all work in
different organizations. Sample B has functioned as a pilot study in order to test the general effects of
job satisfaction on eWOM. Respondents have filled in the online survey with their smartphone,
laptop, tablet and/or personal computer. The companies differ from profit to non-profit and from
private to public organizations. An online survey was administered to 204 persons. A total of 96
completed responses were returned (a response rate of 47.1%). Out of all respondents there were 64
males (66.7%) and the mean respondent’s age was 36.86 years (SD = 13.93). Their average
organizational tenure was 17 years (SD = 13.39).
Measure. The survey was administrated in Dutch (see Appendix 1A). All measurement
instruments were in English originally and were translated to Dutch using a back translation procedure.
All items could be answered on a Likert scale from 1 (strongly disagree) to 5 (strongly agree).
17 with my present job,” and “I consider my job to be rather unpleasant.” Internal consistency was
acceptable (Sample A: α = 0.85; Sample B: α = 0.80).
Extroversion. To measure extroversion all six items from the extroversion subscale of the shortened version of the Big Five Personality Inventory of Hendrik, Hofstee, and De Raad (1999) were
used, with for each subscale the six highest loading items from it original Big Five Personality Inventory
factor. On the positive pole the subscale for extroversion included three items such as “loves to chat,”
and on the negative pole three items as for example “keeps apart from others.” Internal consistency was
acceptable (Sample A: α = 0.73; Sample B: α = 0.70).
Neuroticism. To measure neuroticism all six items of emotional stability subscale of the shortened version of the Big Five Personality Inventory (Hendrik, Hofstee, & De Raad, 1999) were
used. On the positive pole there were three items such as “is always in the same mood,” and on the
negative pole there were three items like “gets overwhelmed by emotions”. Internal consistency was low
(Sample A: α = 0.50, Sample B: α = 0.63). Usually, the cutoff score for Cronbach’s alpha is 0.70
(Cronbach, 1951), however assessing personality often provides lower reliability scores since people
answer rather humble or moderate instead of extremely. In previous literature an alpha of 0.65 was
found (Bakker, Van Oudenhoven, & Van Der Zee, 2004). Within exploratory studies for instance, an
alpha of 0.60 is recommened (Nunnally, 1978). Even after performing a factor analysis there were still
two components with a positive and negative pole each consisting of the three items suggested based on
the highest loadings. The low internal consistency estimate of reliability of Sample A (α = 0.50) was caused presumably by the back translation procedure from English to Dutch. As a consequence, the
average covariance between the item-pairs and the explained variance of neuroticism measured by the
six items together was undermined.
Valence (eWOM). To measure the valence of employees’ eWOM the six-item measurement
18 and the other three items assess the valence of negative eWOM. The items has been applied to the
online context and social media platforms within this study. An example of positively framed eWOM is
“I only have good things to tweet/post about my organization/job.” An example of questions about negatively framed eWOM is “I mostly tweet/post negatively about my job/employer at my current
organization.” Internal consistency for positive valence (Sample A: α = 0.78; Sample B: α = 0.79) and
negative valence (Sample A: α = 0.62; Sample B: α = 0.88) was acceptable.
Data analysis
Various hierarchical lineair regressions were conducted to measure the potential main effects
and interaction effects, and to thereby test the hypotheses. For both Sample A and Sample B, similar
hierarchical lineair regressions were performed by assessing whether there was a significant effect on
the valence of eWOM predicted by control variables, job satisfaction as the independent variable
(Model 1), the moderators: extroversion and neuroticism (Model 2), and the interaction between job
satisfaction and moderator (Model 3). Since there were two dependent variables (positive and negative
eWOM) with each one moderator (extroversion; Hypothesis 2 vs. neuroticism; Hypothesis 3) the
hierarchical linear regression was conducted twice: once for each dependent variable. The interaction
effect of extroversion (Hypothesis 2) and neuroticism (Hypothesis 3) was computed by multiplying the
19
Results
Sample A. Table 1 represents the means, standard deviations, Cronbach’s alphas and the
zero-order Pearson correlations among the variables. A hierarchical lineair regression analysis was
conducted twice to test the Hypotheses. With regards to Hypothesis 1a and 2, positive eWOM was
regressed on the control variables, job satisfaction (see Table 2, Model 1), extroversion (see Table 2,
Model 2), and the interaction between job satisfaction and extroversion (see Table 2, Model 3). Results
indicate that job satisfaction is significantly related to positive eWOM, B = 0.42, p = .00 (see Table 2,
Model 1), thereby accepting Hypothesis 1a. Furthermore, results indicate that extroversion does not
moderate this effect, B = 0.09, p =.60 (see Table 2, Model 3), thereby, Hypothesis 2 could not be
confirmed. A significant negative effect has been found between age and positive eWOM (B = -0.04, p
=.01). The overall models were significant (Model 1, p = .00; Model 2, p = .00; Model 3, p = .00).
TABLE 1: CORRELATION TABLE
Descriptive statistics and Pearson zero-order correlations among the study variables
M SD 1 2 3 4 5 6 7 1. Gender1 1.45 0.50 2. 3. 4. Age IV (Job satisfaction) Moderator 1 (Extroversion) 25.82 3.54 4.03 11.93 0.74 0.57 0.18 0.13 0.29*** 0.36*** 0.05 (0.80) 0.13 (0.70) 5. Moderator 2 (Neuroticism) 2.44 0.50 -0.11 -0.25** -0.20 0.28*** (0.50) 6. 7. DV1 (Positive eWOM) DV2 (Negative eWOM) 2.94 1.40 1.06 0.65 0.03 -0.15 0.10 -0.12 0.34*** 0.12 -0.10 (0.79) -0.37*** -0.15 0.18 -0.11 (0.88) 1
Dummy coded, 1 = male, 2 = female. Note: Cronbach’s alpha is in parentheses along the diagonal. * p < .10, ** p < .05, *** p < .01.
20 TABLE 2: HIERARCHICAL REGRESSION TABLE (Moderation; Sample A; N=98)
With regards to Hypothesis 1b and 3, negative eWOM was regressed on the control variables,
job satisfaction (see Table 3, Model 1), neuroticism (see Table 3, Model 2), and the interaction between
job satisfaction and neuroticism (see Table 3, Model 3). Results indicate that job satisfaction is
significant negative related to negative eWOM, B = -0.24, p = .00 (see Table 3, Model 1), thereby
confirming Hypothesis 1b. However, results indicate that neuroticism does not moderate this effect, B =
0.07, p = .61 (see Table 3, Model 3), thereby Hypothesis 3 is rejected. The overall models were
significant (Model 1, p = .00; Model 2, p = .00; Model 3, p = .01).
Positive eWOM
Predictors Model 1 Model 2 Model 3
1 Control variables
Gender1 H1a: 0.11 (0.21) 0.07 (0.22) H2: 0.06 (0.22) Age H1a: -0.04 (0.02)** -0.04 (0.02)** H2: -0.04 (0.02)**
2 Main effects
Job satisfaction (Z-scored) H1a: 0.42 (0.11)*** 0.41 (0.11)*** H2: 0.42 (0.11)*** Extroversion (Z-scored) 0.08 (0.11) H2: 0.09 (0.11)
3 Interaction
Job satisfaction (Z-scored) *
Extroversion (Z-scored) H2: 0.04 (0.10) R2 0.18*** 0.19*** 0.19**
Adjusted R2 0.14*** 0.14*** 0.13**
1
Dummy coded, 1 = male, 2 = female. Note: N = 98. Unstandardized B coefficients are reported with Standard Errors.
21 TABLE 3: HIERARCHICAL REGRESSION TABLE (Moderation; Sample A; N=98)
Sample B. Various hierarchical linear regression analyses were estimated to test the
hypotheses after the descriptive statistics (i.e., means, standard deviations, Cronbach’s alphas and the
zero-order Pearson correlations among the variables) were consulted (see Table 4). First, with regards
to Hypothesis 1a and 2, positive eWOM was regressed on the control variables, job satisfaction (see
Table 5, Model 1), extroversion, and the interaction between job satisfaction and extroversion (see
Table 5, Model 2 and 3). Results indicate that job satisfaction is significantly related to positive
eWOM, B = 0.35, p = .00 (see Table 5, Model 1), thereby providing support for Hypothesis 1b.
Furthermore, results indicate that extroversion does not moderate this effect, B = 0.05, p = .48 (see
Table 5, Model 3), thereby not providing support for Hypothesis 2. The overall models were
significant (Model 1, p = .01; Model 2, p = .02; Model 3, p = .03).
Negative eWOM
Predictors Model 1 Model 2 Model 3
1 Control variables
Gender1 H1b: -0.14 (0.13) -0.16 (0.14) H3: -0.18 (0.14) Age H1b: -0.01 (0.01) 0.01 (0.01) H3: -0.01 (0.01)
2 Main effects
Job satisfaction (Z-scored) H1b: -0.24 (0.07)*** -0.22 (0.07)*** H3: -0.24 (0.08)*** Neuroticism (Z-scored) 0.06 (0.07) H3: 0.07 (0.07)
3 Interaction
Job satisfaction (Z-scored) *
Neuroticism (Z-scored) H3: 0.06 (0.09) R2 0.15*** 0.16*** 0.16***
Adjusted R2 0.11*** 0.11*** 0.10*** 1
Dummy coded, 1 = male, 2 = female. Note: N = 98. Unstandardized B coefficients are reported with Standard Errors.
22 TABLE 4: CORRELATION TABLE
Descriptive statistics and Pearson zero-order correlations among the study variables
M SD 1 2 3 4 5 6 7 1. Gender1 1.33 0.47 2. Age 36.86 13.93 -0.11 3. IV (Job satisfaction) 4.22 0.63 0.14 0.08 (0.85) 4. 5. Moderator 1 (Extroversion) Moderator 2 (Neuroticism) 4.07 3.52 0.50 0.59 0.17 -0.16 -0.10 0.09 0.43** 0.28** (0.73) 0.28** (0.63) 6. 7. DV1 (Positive eWOM) DV2 (Negative eWOM) 1.80 1.02 1.04 0.12 0.02 -0.25* 0.11 0.11 0.29** -0.16 0.03 0.09 (0.78) -0.15 0.02 0.12 (0.62)
Second, with regards to Hypothesis 1b and 3, negative eWOM was regressed on the control variables,
job satisfaction (see Table 6, Model 1), neuroticism, and the interaction between job satisfaction and
neuroticism (see Table 6, Model 2 and 3). Results indicate that job satisfaction is not significantly
related to negative eWOM, B = -0.07, p = .20 (see Table 6, Model 1), thereby rejecting Hypothesis 1b.
Furthermore, results indicate that neuroticism does not moderate this effect, B = -0.05, p =.21 (see
Table 6, Model 3), thereby not providing support for Hypothesis 3. The overall models were significant
(Model 1, p = .04; Model 2, p = .08; Model 3, p = .08).
1
Dummy coded, 1 = male, 2 = female. Note: Cronbach’s alpha is in parentheses along the diagonal. * p < .10, ** p < .05, *** p < .01.
23 TABLE 5: HIERARCHICAL REGRESSION TABLE (Moderation; Sample B; N=96)
Positive eWOM
Predictors Model 1 Model 2 Model 3
1 Control variables
Gender1 H1a: -0.05 (0.23) -0.04 (0.23) H2: -0.03 (0.23) Age H1a: 0.00 (0.00) 0.01 (0.01) H2: 0.01 (0.01)
2 Main effects
Job satisfaction (Z-scored) H1a: 0.35 (0.11)*** 0.34 (0.12)*** H2: 0.43 (0.13)*** Extroversion (Z-scored) -0.12 (0.12) H2: -0.11 (0.12)
3 Interaction
Job satisfaction (Z-scored) *
Extroversion (Z-scored) H2: 0.05 (0.07)
R2 0.12** 0.13** 0.13**
Adjusted R2 0.09** 0.09** 0.09**
1
24 TABLE 6: HIERARCHICAL REGRESSION TABLE (Moderation; Sample B; N=96)
Negative eWOM
Predictors Model 1 Model 2 Model 3
1 Control variables
Gender1 H1b: -0.24 (0.11)** -0.24 (0.11)** H3: -0.25 (0.11)** Age H1b: 0.00 (0.00) 0.00 (0.00) H3: 0.00 (0.00)
2 Main effects
Job satisfaction (Z-scored) H1b: -0.07 (0.05) -0.07 (0.05) H3: -0.10 (0.06)* Neuroticism (Z-scored) 0.00 (0.06) H3: -0.02 (0.06)
3 Interaction
Job satisfaction (Z-scored) *
Neuroticism (Z-scored) H3: -0.05 (0.04)
R2 -0.09** 0.09* 0.11*
Adjusted R2 0.06** 0.05* 0.06*
1
25
Discussion
Both, Sample A and B had consistent results. Namely, a positive significant relationship
between job satisfaction and positive eWOM (Hypothesis 1a). With regards to the relationship
between job satisfaction and negative eWOM (Hypothesis 1b) there were different findings between
Sample A and B. No significant effect was found in Sample B, however a significant negative effect
was found in Sample A. In contrast to expectations, no support was found for the moderating roles
of extraversion (Hypothesis 2) and neuroticism (Hypothesis 3). Therefore, personality could not be
demonstrated as determining the valence or frequency of eWOM predicted by job satisfaction.
Next to communicating positive or negative eWOM about the employer, receivers’
attitude towards the company are influenced by employees’ (online) messages. As has been demonstrated in Study 1, job (dis)satisfaction predicts positive (negative) eWOM. We now turn to
26
CONSEQUENCE(S) OF EWOM: EMPLOYER BRAND
Employer brand
Employer brand contains “the perceptions of a company's employment offering among
potential employees” (Rosengren & Bondesson, 2014: 255) as it has been defined as an organization’s
reputation as an employer (Barrow & Mosley, 2005). Employer branding is “the entire process of
promoting a company or an organization as the employer of choice to a desired target group, one which
a company needs and wants to recruit and retain” (Mosley, 2007: 124).
There are many antecedents of employer brand known from literature. Besides financial
rewards (salaries, bonuses etc.) and employees’ relationships, companies with highly engaged employees have a better reputation compared to employers without engaged employees. Engaged
employees are more intellectually and emotional connected to the organization and therefore rather
promoters of the employer brand (Foster, Punjaisri, & Cheng, 2010) who may affect the way potential
employees evaluate the company as employer. Moreover, familiarity with the organization together with
external ratings of the corporate reputation are drivers of the employer brand (Cable & Turban, 2003).
Another key driver of employer brand is the positive value from employee referral or company’s
recruitment messages because when expectations of potential employees based on expressed values fits
with actual organizational values there are high levels of trust and loyalty created among applicants.
Consequences of eWOM
Employees can write about their employer via eWOM through a positive or negative message.
With regards to the receiver of the message, there are many motives why people seek and evaluate
eWOM. For instance, to balance their informational disadvantage that leads to increased reflexivity in
27 information by companies (Askegaard, Gertsen, & Langer, 2002). Moreover, to reduce efforts to search
and evaluate efforts (Goldsmith & Horovitz, 2006), to reduce risk (Kim, Mattila, & Baloglu, 2011;
Sweeney, Soutar, & Mazzarol, 2008), and to find social assurance (Bailey, 2005).
Especially, the online availability to a multitude of people and institutions is an essential aspect
of eWOM because negative and positive statements made by employees have certain consequences. As
has been demonstrated by Van Hoye and Lievens (2009) receiving positive employment information
through employees’ referrals leads to higher perceived organizational attractiveness and actual
application decision. Another outcome of eWOM is the influence on receivers’ willingness to invest or
pay for a brand and the perceived quality of it (Bickart & Schindler, 2001; Pavlou & Dimoka, 2006).
Moreover, someone’s levels of trust and loyalty towards a brand are influenced by eWOM (Awad &
Ragowsky, 2008; Ba & Pavlou, 2002; Gauri, Bhatnagar, & Rao, 2008). Another consequence of eWOM
is increased engagement with the company or brand by the receiver of the message (Algesheimer, Borle,
Dholakia, & Singh, 2010; Nambisan & Baron, 2007; Schau & Muniz, 2002). It has been found that job
seekers are more attracted to firms with strong positive reputations compared to firms with either no or
negative reputations (Cable & Turban, 2003).
Employer’s brand may be influenced by eWOM since employees reflect their own experience
with the company to others outside the organization (Gotsi & Wilson, 2001; Hatch & Schultz, 2003;
King & Grace, 2008). The communicated eWOM by employees can result into positive publicity for
employers’ reputation but employees can behave as brand ‘saboteurs’ as well in the worst-case scenario by acting against the organization in their eWOM (Richards, 2008; Wallace & De Chernatony, 2007).
As has been demonstrated by Gotsi and Wilson (2001) reputation is influenced by the way how
employees create (or destroy) their organization’s reputation in their everyday exchanges and social
28 (Collins & Han, 2004; Saks, 2005; Zottoli & Wanous, 2000) because talented and many job seekers are
attracted by the positive and prestigious image of the organization. In general, positive eWOM has a
positive effect on company’s reputation and image (Chen, Wang, & Xie, 2011). Therefore, the
following statement could be hypothesized:
Hypothesis 4a. There is a positive relationship between positive eWOM and employer branding.
While positive eWOM builds company’s reputation and image, it has been argued that stakeholders who spread negative messages about their company might seriously damage organization’s
reputation (Gensler, Völckner, Liu-Thompkins, & Wiertz, 2013). It has been found that negative eWOM
can be detrimental for companies’ image (Van Hoye & Lievens, 2007; Kanar, Collins, & Bell, 2010). Yet, negative reviews are perceived as less informational because many receivers may find negative
emotions irrational (Kim & Gupta, 2012). On the contrary, negative eWOM may have a stronger impact
than positive eWOM on employer brand since these negative referrals are evaluated as rather
trustworthy in the presence of a large number of positive referrals (Ba & Pavlou, 2002; Park & Lee,
2009). Receivers of the negative eWOM do not only pay more attention but even tend to weigh negative
information more than positive information in their evaluation and decision-making process (Herr,
Kardes, & Kim, 1991).
Contrary to many other studies, Berger, Sorensen, & Rasmussen (2010) have found empirical
evidence to say that “any publicity is good publicity”. They have demonstrated that even negative
publicity (e.g. negatively framed eWOM) can increase the likelihood of attractiveness and sales of a
brand. Other research has found that negative eWOM has a strong negative effect on the organizational
29 valence or negativity bias, is loss prevention since people are more alert when there is a threat to lose
something rather than gaining the benefits (Aggarlwal, Gopal, Gupta, & Singh, 2012). Based on this
literature the following hypothesis could be formulated:
Hypothesis 4b. There is a negative relationship between negative eWOM and employer brand.
Personality
As has been argued in Study 1, employees’ experiences with their job or their individual traits
(highly extroverted and/or highly neurotic) determines the manner (valence) they engage into eWOM,
but the receiver’s evaluation process of a message is influenced when the receiver’s personality plays a role. Personality can (next to e.g. extroversion and neuroticism as used in Study 1) be described by the
dispositional affect which is “a personality trait referring to a person’s relatively stable, underlying
tendency to experience positive and/or negative moods and emotions” (Watson & Clark, 1984: 466).
Watson and Clark (1984) categorized positive and negative affectivity. The process of persuasive
messages can profoundly be influenced by the two affective states.
A person with a high levels of positive affectivity has the tendency to be energetic, cheerful,
and experiences positive moods (e.g. pleasure or well-being) in different situations (Barsade & Gibson,
2007). Receivers of messages who have high levels of positive affect often use a less effortful and
schema guided processing style because their affect signals that the situation is harmless (Schwarz,
1990). Moreover, persons with high levels of positive affect views the world in a good light.
Extroversion is closely associated with positive affectivity (Judge, Heller, & Mount, 2002;
Landy & Conte, 2010; Clark & Watson, 1992). As has been argued by Gomez, Gomez, and Cooper
30 and current mood. When the receiver is highly extroverted he or she has the tendency to process
pleasant information (Gomez, Cooper, & Gomez, 2002; Gray, 1981). Therefore, following hypothesis
can be drafted:
Hypothesis 5. The positive relationship between positive eWOM and employer brand (Hypothesis 4a) is moderated by extroversion, such that the effect becomes stronger when the receiver is highly rather than less extroverted.
On the contrary, neuroticism is associated with negative affect characterized by negative
effect on trusting others and more sensitive to negative information compared to extroverted
individuals (Keller, Mayo, Greifeneder, & Pfattheicher, 2015). When a person feels negative affect
during message exposure he or she will process more systematically than individuals compared to
recipients who feels highly positive affect (Bless, Mackie, & Schwarz, 1992; Bohner, Crow, Erb, &
Schwarz, 1992; Sinclair, Mark, & Clore, 1994). Furthermore, an effortful and analytic processing
style is adopted since the negative affect signals the individual that the situation is problematic or
threatening.
When the receiver is highly neurotic he or she has the tendency to ‘see the glass half empty’,
through a black lens, together with higher anxiety and loss aversion from a defensive mindset and
will be triggered by negative eWOM which is perceived as more persuasive than positive news
(Bolger & Schilling, 1991; Larsen & Ketelaar, 1991; Luhmann & Eid, 2009; Verduyn & Brans,
2012; Zhang, Craciun, & Shin, 2010). When the receiver is highly neurotic he or she has the
tendency to process unpleasant information (Gomez, Cooper, & Gomez, 2002). Therefore, the
31
Hypothesis 6. The negative relationship between negative eWOM and employer brand (Hypothesis 4b) is moderated by neuroticism, such that the effect becomes stronger when the receiver is highly rather than less neurotic.
Brand awareness
Brand awareness contains the ability of a consumer to recognize and recall a brand in
different situations. Brands awareness is the ability to identify the brand under different conditions by
the strength of the brand node or trace in memory, where brands with high levels of awareness and
strong, favorable and unique associations are high equity brands (Keller, 1993; 2008). However, low
levels of brand awareness and therefore unknown brands are evaluated as less positive compared to
strong, high equity brands.
Brand awareness affects perceivers’ attitude because individuals have the tendency to favor
and rather positively evaluate familiar and well-known products (Keller, 1993; Jacoby, Szybillo, &
Busato-Schach, 1977; Roselius, 1971). Within online communication contexts the information adoption
is affected by the receivers' experience and knowledge of the brand or company that moderates both the
central (the nature of arguments in the message) and peripheral (the subject matter of the message)
(Sussman & Siegel, 2003). This evaluation’s error of favoring well-known brands can be caused by the
halo effect which is a cognitive bias of someone’s overall impression of a person, company, brand, or
product that influences someone’s feelings and thoughts about that entity's character or properties
(Murphy, Jako, & Anhalt, 1993; Thorndike, 1920). The halo effect results into a tendency to confirm
existing attitudes wherein positive feelings to a brand from a first experience cause ambiguous or neutral
traits to evaluate the same brand in the future as positive as well.
32 Especially, when the transmitted message is positive about a brand or company it is likely that it has a
positive effect on the employer’s brand particularly if the brand is highly familiar to someone. Therefore, when strong brands (high brand awareness) are favorable because of one aspect or attribute,
the person has a positive predisposition toward everything about it. But when a weak or unfamiliar
brand (low brand awareness) has one aspect or attribute that has been disliked there is a negative
predisposition toward everything about it. The following hypothesized states:
Hypothesis 7a. The positive relationship between positive eWOM and employer brand (Hypothesis 4a) is moderated by brand awareness, such that the relationship becomes stronger when brand awareness is high rather than low.
This can be argued by the confirmation of established association(s) with a well-known brand
by the individual in line with the psychological theory about ‘selective recall’ and ‘confirmation bias’. From psychological research it is found that people often have a confirmation bias which is the tendency
to interpret or to recall information in a way that confirms one’s existing beliefs or expectations (Nickerson, 1998). According to Nickerson (1998) individuals give undue weight to evidence that
supports their existing beliefs or positions, while they discount or neglect to gather evidence that would
contradict their established hypotheses. Sometimes people do gather and interpret evidence in an
unbiased manner, but still they remember it selectively in order to confirm their existing beliefs and
expectations. This is called ‘selective recall’ (Costley & Brucks, 1992). Although, this holds for both effects of positive and negative eWOM on employer brand, especially high brand awareness (strong
brands) can act as a buffer against the detrimental effect of negative eWOM since positive associations
and brand nodes are already established in memory. For this reason it is likely that the effect of negative
33 brand awareness, since the person will neglect or discount these negative thought that contradict his
existing beliefs about the brand. This means that negative eWOM has a greater impact on individuals’
evaluations of unknown or unfavorable brands compared to familiar or favorable brands (Laczniak,
DeCarlo, & Ramaswami, 2001). Furthermore, unknown brands are more vulnerable to change in their
evaluations. When receivers of the eWOM are less familiar with a brand, they are more amenable to
change their brand evaluations based on the direction of the information and to process new
brand-related information (Sundaram & Webster, 1999). Therefore, negative eWOM harms the employer
brand when the brand is unknown to the receiver of the online message. The hypothesis sounds:
Hypothesis 7b. The negative relationship between negative eWOM and employer brand (Hypothesis 4b) is moderated by brand awareness, such that the relationship becomes stronger when brand awareness is low rather than high.
FIGURE 2: CONCEPTUAL MODEL STUDY 2: EXTERNAL STAKEHOLDER (JOB SEEKER)
34
STUDY 2
Method
The purpose of Study 2 was to investigate the effect of positive either negative eWOM on the
reputation of companies as employers perceived by others outside the firm. To empirically investigate
the hypotheses in Study 2, an experimental design has been set up (see Appendix 1B). Moreover, this
study was conducted to test whether there is a different effect of the valence of eWOM (positive or
negative) on employer brand between familiar and unfamiliar organizations. Twitter was used as social
media platform to model the positive or negative eWOM message to participants.
Participants and Design
One-hundred-and-nineteen undergraduates at the Faculty of Economics and Business of the
University of Groningen participated in this study. Participants were compensated with course credits or
€8 in exchange for their participation. Among all participants 77 were female (65.4%) and the mean respondent’s age was 21.71 years (SD = 2.29). With regards to nationality, 54 participants were Dutch (44%) and 16 students (13%) were non-EU with a majority of 9 Asian students. Participants were
randomly assigned to one of three conditions of a 3 (message valence: positive vs. negative vs. neutral
eWOM message [between-subjects]) x 2 (brand awareness: high vs. low [within-subjects])
mixed-subject design. The factor valence (message valence: positive, negative, or neutral) was manipulated as
a between-subject design since every participant has seen two tweets with one kind of valence (two
positively, neutrally, or negatively framed messages) and the factor brand awareness was a
within-subject design since each participant has seen one tweet of a company with high brand awareness and a
35
Procedure
Every participant was placed in a single room of the research lab with only a desk and
computer. After they had read the instructions on the computer screen they continued the experiment by
reading two fictional tweets. These fictional tweets were either both positive, negative, or neutral, which
constituted the manipulation for message valence. Both tweets had similar content except the mentioned
company name and the employee name with his twitter profile picture were different for both tweets.
For the positive valence the tweet was “Monday morning the first day of a new week of work at (company name). Feeling excited! I love this company”. With regards to the neutral valence, the tweet
was “Monday morning the first day of a new week of work at (company name),” and for negative valence it was “Monday morning the first day of a new week of work at (company name). Looking forward to the weekend! I hate this company!” According to Stieglitz and Dang-Xuan (2013) the
amount of sentiment (so called: valence) can be determined by the tool SentiStrength which classifies
texts for positive sentiment on a scale of 1 (neutral) to 5 (strongly positive) and for a negative sentiment
on a scale of –1 (neutral) to –5 (strongly negative). Words as ‘love’ or ‘happy’ have a high positive
sentiment score while ‘hate’ or ‘sad’ are classified as strong negative sentiment. The two tweets that were presented to the participant were either of a well-known (high brand awareness) or unknown
company (low brand awareness), in random order. After each of the two tweets, participants were asked
to answer the online survey questions about the tweet. Google was selected as company with high brand
awareness in this study based on its high Fortune’s ranking (cf. Fombrun & Shanley, 1990), and Jr. Gerolm Ltd. was selected as company with low brand awareness as fictional company in order to ensure
low brand awareness since a fictional company should logically not be familiar to people.
Extroversion. Similarly to Study 1, all six items from the extroversion subscale of the shortened version of the Big Five Personality Inventory of Hendrik, Hofstee, and De Raad (1999) has
36 excluded (α = 0.59). Therefore, five-item measure was chosen without the item “I slap people on the
back”. Just like all the other variables within this study, the measurement items were assessed on a five-point Likert scale (1 = strongly disagree, 5 = strongly agree).
Neuroticism. Neuroticism was assessed by the six items of emotional stability subscale of the shortened version of the Big Five Personality Inventory (Hendrik, Hofstee, & De Raad, 1999) similar to
Study 1. Internal consistency was acceptable (α = 0.67).
Employer brand. Employer brand was measured by the ten-item measurement instrument developed by Turban and Keon (1993). One of the items is for example “for me, this company would be
a good place to work,” and “employees are probably proud to say they work at this company.” Internal consistency was acceptable (α = 0.92).
Familiarity. To assess whether brand awareness was manipulated, participants were asked whether they were familiar or not with the brand that was posted in the tweet through three items of
37
Results
Descriptive statistics
With regards to high brand awareness, there is a significant difference between the mean
of employer brand for the three conditions of eWOM valence (positive, neutral, and negative)
determined by one-way between-subjects ANOVA (F(2,119) = 8.78, p = .00). Table 7 represents the
means and standard deviations for the different conditions.
TABLE 7: DESCRIPTIVES (N = 119)
Valence
Brand awareness
N (positive) = 41, N (neutral) = 39, N (negative) = 39.
Hypotheses testing
To assess the hypotheses, I conducted a repeated measures ANOVA (see Table 7) with a
lower-bound correction (Girden, 1992) in order to measure the two conditions, since participants
had to evaluate two tweets (tweet with high brand awareness company and a tweet with low brand
awareness company) in the same way. This analysis showed that there was no significant main
effect of valence on employer brand, F = 0.88, p = .92, indicating that the positive message (M =
2.66, SD = 0.07) and negative message (M = 2.65, SD = 0.07) did not differ significantly from the Positive Neutral Negative
38 neutral message (M = 2.70, SD = 0.07). Therefore, Hypotheses 4a and 4b could not be confirmed
based on this study. Moreover, no significant main effects were found by extroversion F = 2.37, p =
.13, or neuroticism F = 0.42, p = .52, on employer brand. Moreover, no significant interaction effect
by valence and extroversion was found, F = 0.66, p = .52, nor of valence and neuroticism, F = 0.13,
p = .20. Based on these findings, Hypotheses 5 and 6 were not confirmed. Although, brand
awareness in itself significantly change employer brand, F = 5.64, p = .02, indicating that high brand
awareness (M = 4.05, SD = 0.07) significantly differs from low brand awareness (M = 1.29, SD =
0.06), no significant interaction effect by brand awareness on the relationship between valence and
employer brand was found, F = 0.22, p = .81. Therefore, Hypothesis 7a and 7b can not be
confirmed.
Discussion
In contrast to expectations, there was no significant relationship found between positive
eWOM (valence) and employer brand (Hypothesis 4a) or between negative eWOM (valence) and
employer brand (Hypothesis 4b). The moderating roles of extroversion (Hypothesis 5) and
neuroticism (Hypothesis 6) were not found in this study. Therefore, personality could not be
demonstrated as determining the employer brand predicted by valence. Although, brand awareness
in general significantly affects employer brand, no significant interaction effect of brand awareness
on the relationship between valence and employer brand has been demonstrated (Hypothesis 7a and
7b). With regards to the effect of personality based on affectivity, there was no evidence found to
conclude that extroversion moderates the effect of positive eWOM on employer brand as
