Download as pdf or txt
Download as pdf or txt
You are on page 1of 17

IJRM-01243; No of Pages 17

International Journal of Research in Marketing xxx (2018) xxx–xxx

Contents lists available at ScienceDirect

IJRM
International Journal of Research in Marketing
journal homepage: www.elsevier.com/locate/ijresmar

Full Length Article

Sharing product harm information: The effects of self-construal and self-


relevance
Ezgi Akpinar a,⁎, Peeter W J Verlegh b, Ale Smidts c
a
Koç University, Turkey
b
Vrije Universiteit Amsterdam, The Netherlands
c
Rotterdam School of Management, Erasmus University, The Netherlands

a r t i c l e i n f o a b s t r a c t

Article history: Product harm information spreading in the marketplace may have profound consequences for
First received on July 26, 2016 and was under companies, public policy makers and consumer well-being. However, limited research is avail-
review for 6 months able on what makes consumers share such information with others. This paper examines how
Available online xxxx
self-relevance and self-construal affect the sharing of product harm information and the under-
lying processes that shape sharing. Five experiments demonstrate that under independent self-
Guest Editor: Donald R. Lehmann
construal, highly self-relevant product harm information is less likely to be shared. The reluc-
tance to share highly self-relevant information is attenuated under interdependent self-con-
Keywords:
strual. The differential effects of self-construal are related to self-serving processes and
Health information
motivations for sharing product harm information with others. The results apply to the sharing
Product harm
Self-relevance of negative information, but not positive information. Companies may use these findings to
Self-construal manage the diffusion of product harm information.
Word-of-mouth © 2018 Elsevier B.V. All rights reserved.

1. Introduction

Consumers frequently encounter health related information about the products and brands they use. Examples may include
food contaminations (Borah & Tellis, 2016; Lei, Dawar, & Gürhan-Canli, 2012; Liu, Liu, & Luo, 2016), but also scientific discoveries
about harmful effects of certain products (Milkman & Berger, 2014). Such negative information often spreads like a wildfire
(Rosnow, 2001) and disseminates to more people over a longer period than positive information (Hornik, Satchi, Cesareo, &
Pastore, 2015). This process can have a strong impact on both consumers and firms (Van Heerde, Helsen, & Dekimpe, 2007).
For instance, the sales of Ferrero (producer of Nutella) dropped by 3% in less than three months, when it was linked to an in-
creased risk of cancer (Chapman, 2017). Negative information spreading about firms may lower stock prices (Tirunillai & Tellis,
2012), damage brand evaluations and reputation (Ahluwalia, Burnkrant, & Unnava, 2000) and decrease equity and sales (Babić
Rosario, Sotgiu, De Valck, & Bijmolt, 2016; Borah & Tellis, 2016; Hsu & Lawrence, 2016). On the other hand, product harm infor-
mation spreading among consumers can be also useful, as it may help consumers to make better decisions – or at least be better
informed about undesired consequences of their consumption (Keller, 2015). It may also be beneficial for companies, which could
improve their production facilities and communication strategies after receiving consumers' backlash about health-related issues.
Our research contributes to the word-of-mouth literature in two distinct ways. First, although there has been research on un-
derstanding triggers and motives for word of mouth (e.g., Berger, Sorensen, & Rasmussen, 2010; Borah & Tellis, 2016; Eelen,
Özturan, & Verlegh, 2017; Stephen & Lehmann, 2016), there is limited work on understanding why consumers share negative in-

⁎ Corresponding author.
E-mail addresses: eakpinaruysal@ku.edu.tr (E. Akpinar), p.verlegh@vu.nl (P.W.J. Verlegh), asmidts@rsm.nl (A. Smidts).

https://doi.org/10.1016/j.ijresmar.2018.01.001
0167-8116/© 2018 Elsevier B.V. All rights reserved.

Please cite this article as: Akpinar, E., et al., Sharing product harm information: The effects of self-construal and self-relevance, In-
ternational Journal of Research in Marketing (2018), https://doi.org/10.1016/j.ijresmar.2018.01.001
2 E. Akpinar et al. / International Journal of Research in Marketing xxx (2018) xxx–xxx

formation (cf., Berger, 2014). There is a general assumption that individuals are reluctant to spread bad news, because they don't
want to be perceived as a “Debbie Downer” (Berger, 2014), but more research is needed to develop an understanding of what
makes consumers choose to share negative information or not. The present research aims to contribute by exploring some of
the key drivers of product harm information sharing.
Second, we focus on product harm information which has not been explored in negative word of mouth. Previous research on
negative word of mouth has focused on consumers' sharing their experiences with products or services (Berger et al., 2010; De
Angelis, Bonezzi, Peluso, Rucker, & Costabile, 2012). The dissemination of negative information has been shown to have an impor-
tant influence on the success and failure of companies (Ahluwalia et al., 2000; Tirunillai & Tellis, 2012; Van Heerde et al., 2007).
Yet, there has been limited research on consumer dissemination of information from third parties (journalists, scientists) (Berger &
Milkman, 2012; Dubois, Rucker, & Tormala, 2011; Heath, Bell, & Sternberg, 2001). The present research focuses on product harm
information, i.e., information about potential health risks associated with a brand or product, which might be “generated” by an-
other consumer, newspaper, public policy maker (e.g., Greenpeace, 2017). We are interested in understanding when consumers
are more likely to “pass along” this information to other consumers. For example, David might read news about the controversy
around side effects of Nutella on NY Times, and might “share” this product harm information through Twitter, or pass it along dur-
ing a conversation with friends.
The current investigation ties together divergent literatures on word of mouth, health psychology and self-construal, in order to
understand what makes consumers share (or not share) product harm information. A considerable body of literature has investi-
gated consumer responses to health risk information in terms of changing their attitudes (Keller & Block, 1999; Menon, Block, &
Ramanathan, 2002) or consumption (Menon, Raghubir, & Agrawal, 2008; Yan & Sengupta, 2013). Yet, there is not much work that
helps us understand what makes consumers talk about product harm information. Understanding this is important, because word
of mouth might help consumers seek preventive actions for themselves, protect other consumers from health risks or encourage
companies and brands to take action against health hazards.
We also contribute to a growing body of literature on the role of self-construal in consumer behavior. Self-construal has been
linked to the processing of advertising or product related information (Agrawal & Maheswaran, 2005) and more specifically prod-
uct harm information that poses a threat (Block, 2005; White, Argo, & Sengupta, 2012), but has received limited attention as a
moderator of word-of-mouth behaviors. Finally, we hope that our innovative manipulations of self-relevance and behavior-
based measures of sharing are of interest to researchers and practitioners in the fields of marketing, word of mouth and health
psychology.

2. Theoretical background

Although research on negative information sharing has been limited, there is some evidence that individuals are reluctant to
spread bad news (cf., Berger, 2014). Research in psychology indicates that individuals who share negative (versus positive)
news are evaluated less favorably (Forest & Wood, 2012), and are less preferred as interaction partners (Bell, 1978). This might
be one of the reasons why individuals are more likely to share positive (versus negative) news (Berger & Milkman, 2012). Re-
search on word-of-mouth suggests that self-affirmation motives and self-serving biases may play a role in this (Alexandrov,
Lilly, & Babakus, 2013). Specifically, consumers are biased to talk less about their negative consumption experiences. This bias dis-
appears when consumers talk about the experiences of other consumers, which confirms the self-serving nature of this bias (De
Angelis et al., 2012). Sharing content not only helps individuals to self-affirm. It also helps them show that they are knowledgeable
about the issue at hand (Packard & Wooten, 2013). In fact, some consumers give more negative evaluations in reviews to show
that they are experts and have novel and interesting opinions (Moe & Schweidel, 2012). Finally, individuals might share negative
content in order to protect others (Hennig-Thurau, Gwinner, Walsh, & Gremler, 2004; Cheung & Thadani, 2012) or generate help
and social support for themselves (Buechel & Berger, 2015; Hepper, Gramzow, & Sedikides, 2010). In the next sections, we will
explain a) how self-serving biases affect the sharing product harm information and b) how self-construal can change the processes
behind it.

2.1. Self-serving biases for product harm information

There is extensive evidence that individuals display a self-serving bias in their judgments and behaviors (Sherman, Nelson, &
Steele, 2000). Self-serving biases can manifest in several different ways, including perceiving oneself better than others
(Dunning, Meyerowitz, & Holzberg, 1989), focusing more on favorable information about the self (Perloff & Fetzer, 1986) or per-
ceiving negative events as less likely to occur for the self than for others (Weinstein, 1980). Self-serving biases also have important
cognitive and social consequences. Although individuals should behave in their best self-interest, individuals might have difficulties
in avoiding their self-serving biases, which might make them make less ideal decisions in domains such as finance or drug pre-
scription (Dana & Loewenstein, 2003). Further, individuals tend to interpret negative feedback as invalid and inaccurate, and dis-
credit the ones providing it (Ditto & Lopez, 1992; Hepper et al., 2010; Puntoni, Sweldens, & Tavassoli, 2011).
Research in health psychology shows that self-serving biases are especially prevalent in response to information about health
risks (Menon et al., 2002; Raghubir & Menon, 1998). Individuals are often (overly) optimistic in their risk perceptions, and this
helps them to sustain mental well-being and self-affirmation (Taylor & Brown, 1988). Such biases might not always be in one's
best interest. They may lower willingness to take medical screenings, lead consumers to tune out of preventive advertising, or

Please cite this article as: Akpinar, E., et al., Sharing product harm information: The effects of self-construal and self-relevance, In-
ternational Journal of Research in Marketing (2018), https://doi.org/10.1016/j.ijresmar.2018.01.001
E. Akpinar et al. / International Journal of Research in Marketing xxx (2018) xxx–xxx 3

reduce their willingness to take action (Chandran & Menon, 2004; Keller, Lipkus, & Rimer, 2003; Luce & Kahn, 1999; Menon et al.,
2002; Raghubir & Menon, 1998).
Self-serving biases are particularly strong when issues are self-relevant (Block & Williams, 2002; Hepper et al., 2010). As an
ironic consequence of this, consumers are less likely to act on more personally relevant risks (Carvalho, Block,
Sivaramakrishnan, Manchanda, & Mitakakis, 2008). For instance, when women feel that breast cancer is a self-relevant risk for
them, they take fewer actions to deal with the threat (Puntoni et al., 2011).
A growing literature on health psychology suggests that individuals activate a variety of self-serving mechanisms to cope with
self-threatening information. Examples include underweighting self-relevant risk – i.e., simply being “less concerned” about the in-
formation (Ülkümen & Thomas, 2013; Yan & Sengupta, 2013); questioning the credibility of the information (Block, 2005; Puntoni
et al., 2011; Taylor & Brown 1988), downgrading the perceived usefulness of information (Hepper et al., 2010; Menon et al., 2002),
or simply tuning out and ignoring it (Block & Williams, 2002; Liberman & Chaiken, 1992).
Building on this literature, we expect that consumers who are more at risk of products' negative health effects (i.e., those for
whom the information has high self-relevance), will be less likely to share the information with others. This is consistent with the
notion that sharing is a social process that is shaped by several motivations (Berger, 2014), including self-serving biases (De
Angelis et al., 2012; Packard, Gershoff, & Wooten, 2016). From a consumer well-being perspective, self-serving bias is unfortunate:
Although not sharing threatening health information might keep the threat “far from home”, it also prevents consumers from
protecting themselves or letting others know about the potential hazards.
But self-relevance does not affect all consumers in the same way. In particular, we suggest the effect of self-relevance on shar-
ing depends on consumers' self-construal.

2.2. Self-construal and sharing product harm information

Self-construal is defined as the way in which individuals make meaning of the self and see the self in relation to others (Cross,
Hardin, & Gercek-Swing, 2011). It can be conceptualized as a dispositional trait that individuals may tend toward interdependent
or independent self-construal, but also as a state-like characteristic that can be activated by contextual factors (Markus &
Kitayama, 1991). Self-construal is typically characterized by two types of selves: The independent self represents a psychological
tendency to separate the self from the social context and to disengage from others (Kitayama, Markus, Matsumoto, &
Norasakkunkit, 1997). On the other hand, the interdependent self represents a tendency to blur the line between self and other.
Rather than putting oneself at the center, the interdependent self focuses on social requirements and relationships with others
(Sedikides, Gaertner, & Toguchi, 2003; Torelli, 2006). Western cultures are often associated with independent self-construal and
Eastern cultures are often associated with interdependent self-construal. While self-construal can be a chronic disposition, influ-
enced by cultural differences, it can be also temporarily primed by external cues. Several studies have linked the effects of self-con-
strual primes to consumer behavior (Aaker & Lee, 2001; Gardner, Gabriel, & Lee, 1999; Shavitt, Torelli, & Wong, 2009).
Self-construal has been shown to affect cognitive inferences as well as social interactions (Cross et al., 2011; Kitayama, Park,
Sevincer, Karasawa, & Uskul, 2009; Markus & Kitayama, 1991).

2.2.1. Independent self-construal


Research has shown that self-construal is related to self-serving biases in the context of health risks (Fontaine & Smith, 1995).
Those who are under more independent self-construal often behave in a self-serving way with a desire to protect the self from
self-relevant risks (Block, 2005; Heine, Lehman, Markus, & Kitayama, 1999; Sedikides et al., 2003; White et al., 2012) such as be-
lieving that negative events are less likely to happen to oneself.
We expect that under independent self-construal these self-serving biases would make consumers distance themselves from
self-relevant health risks, which may lead them to be less concerned by product harm information, or question the credibility
of the information, which should both make them less likely to share information with others.
Further, while sharing information, consumers are driven by self-serving motivations. For instance, studies with Western par-
ticipants have shown that consumers are less likely to share their own negative experiences but more likely to share others' ex-
periences in order to boost self-enhancement (De Angelis et al., 2012). Further, participants (western) are more likely to show the
knowledge that supports their ideal-selves (Packard & Wooten, 2013). Taken together, we expect that higher self-serving biases
and reduced motivation to show self-relevant and negative knowledge would result in a reduced likelihood of sharing self-relevant
product harm information under independent self-construal.

2.2.2. Interdependent self-construal


Those who are under interdependent self-construal are to a lesser extent driven by biases related to the self, and more by a focus on
others and their opinions (Aaker & Lee, 2001; Heine et al., 1999; White et al., 2012). For these consumers, we expect that higher self-rel-
evance of the product harm information would not lead to higher self-serving biases. Instead, because the “self” is defined by a focus on
other people's opinions, preferences, and responsibilities (Aaker & Lee, 2001), we expect that higher self-relevance should trigger a de-
sire to learn more from others (Price, Feick, & Guskey-Federouch, 1995). So, individuals under interdependent self-construal will display
less self-serving biases and a higher desire to learn about self-relevant health threats. As a result, we expect that the negative effect of
self-relevance on sharing will be attenuated. Formally stated, we expect the following:

H1a. Under independent self-construal, higher self-relevance will decrease sharing of product harm information.

Please cite this article as: Akpinar, E., et al., Sharing product harm information: The effects of self-construal and self-relevance, In-
ternational Journal of Research in Marketing (2018), https://doi.org/10.1016/j.ijresmar.2018.01.001
4 E. Akpinar et al. / International Journal of Research in Marketing xxx (2018) xxx–xxx

H1b. The negative effect of self-relevance on sharing under independent self-construal will be mediated by higher self-serving
biases and reluctance to show knowledge about product harm information.

H2a. Under interdependent self-construal, the negative effect of self-relevance on sharing product harm information will be
attenuated.

H2b. The negative effect of self-relevance on sharing under interdependent self-construal will be attenuated due to weaker self-
serving biases and higher motivation to learn about product harm information.

To illustrate our hypotheses, let's return to the Nutella example. Here, we expect that consumers with independent self-con-
strual, will be less likely to share information about the product's potential harmful effects when they consume more (versus
less) Nutella. This behavior should be driven by self-serving biases and motivations to avoid and deny such knowledge, which re-
duce willingness to share product harm information. On the other hand, consumers with interdependent self-construal will have
weaker self-serving biases, and will be motivated to learn more about it. As a result, they would not share product harm informa-
tion about Nutella less when they consume more.
In this research, we have focused on the sharing of product harm information. We expect our findings to be unique to product
harm information (negative valence) which pose threat and lead to self-serving biases, but not information with positive valence.
Previous research suggests that sharing of positive and negative information are driven by different motivations and processes
(Berger, 2014). Consumers are more likely to share positive word of mouth about products and brands that are highly self-relevant
(Dubois, Bonezzi, & De Angelis, 2016; Sundaram, Mitra, & Webster, 1998), often with the motivations to look good or to provide
useful content to others (Berger & Milkman, 2012; Heath et al., 2001). In order to demonstrate that our predictions are unique to
the sharing of negatively valenced information, we have also explored the effect of self-relevance and self-construal on sharing
positive information about products.

3. The current research

To test our predictions, and to explore a number of boundary conditions, we conduct a set of five experiments. We start by
testing the moderation of self-construal on the effect of self-relevance on sharing intentions (Study 1). Study 2 replicates this
study with actual sharing behavior, and manipulates self-relevance of product-harm information based on gender. In Study 3,
we rely on a lab-based manipulation of self-relevance, which rules out some explanations that may be confounded with actual
consumption or gender. Study 4 is based on an actual “buzz” campaign, and uses real-world behaviors to study consumer re-
sponses to our manipulations. Study 5 again takes place in a lab context, collects more extensive measures on the underlying pro-
cesses, and studies how the information is transmitted by means of linguistic analysis. In Studies 4 and 5, we not only examine the
sharing of negative (product-harm) information, but also test how the effects of self-relevance and self-construal on sharing pos-
itive information. Across our experiments, we use a range of different manipulations and measures that help establish the robust-
ness of our findings (Lynch, Bradlow, Huber, & Lehmann, 2015). An overview of the studies is given in Table 1 below.

4. Study 1: caffeinated drinks (the effects of self-relevance and self-construal on sharing)

Study 1 examines the effect of self-relevance on sharing product harm information under two levels of self-construal (indepen-
dent versus interdependent). We examine whether a) self-relevance decreases sharing information under independent self-con-
strual and b) this negative effect is attenuated under interdependent self-construal.

4.1. Method

Two hundred and two undergraduates at a Western-European university (Mage = 22.4; 132 males) completed this study in
exchange for partial course credits. First, participants were asked to indicate their consumption levels for several product catego-
ries on 5-point scale (1 = not at all, 5 = high amounts). One of the categories was caffeinated drinks (such as coffee, tea and soft
drinks), which was our focal category. Product category use serves as a proxy for measuring the self-relevance of the information
(Keller & Block, 1999; Liberman & Chaiken, 1992).
Second, participants were randomly assigned to one of the two conditions of self-construal (independent vs. interdependent)
priming. The participants read a short story about “a trip to the city” and were asked to circle all the pronouns used within the
text. The same text was presented with either self-oriented (I, me, myself) or other-oriented (we, us, our) pronouns (Ng, 2010;
Yang, Mao, & Peracchio, 2012).
Third, after self-construal priming, participants read the following scenario: “Imagine that a couple of weeks ago, during a con-
versation, you heard that caffeinated drinks (i.e., coffee, tea, and soft drinks such as cola and energy drinks) increase the risk of
losing hair”. Next, they were asked to indicate their likelihood of sharing this information with others during a related conversa-
tion on a 7-point scale (1 = unlikely to pass this along, 7 = likely to pass this along; cf., Akpinar & Berger, 2017; Heath, 1996).
We conducted two manipulation checks. First, a six-item self-construal scale was used as a manipulation check for our self-con-
strual manipulation (Aaker & Lee, 2001), and confirmed that participants in the independent condition thought more about

Please cite this article as: Akpinar, E., et al., Sharing product harm information: The effects of self-construal and self-relevance, In-
ternational Journal of Research in Marketing (2018), https://doi.org/10.1016/j.ijresmar.2018.01.001
ternational Journal of Research in Marketing (2018), https://doi.org/10.1016/j.ijresmar.2018.01.001
Please cite this article as: Akpinar, E., et al., Sharing product harm information: The effects of self-construal and self-relevance, In-

Table 1
Overview of studies.
Self-construal Self-relevance Stimuli Dependent Hypothesis 1: Hypothesis 2: Other results
(Ind SC vs. Inter (SR) (Valence) variable Effect of SR on Sharing under Effect of SR on Sharing under
SC) Ind SC Inter SC

E. Akpinar et al. / International Journal of Research in Marketing xxx (2018) xxx–xxx


Manipulated Measured Side effects of Sharing H1a supported: H2a supported: High SR information was
(pronoun circling) (caffeinated caffeinated Likelihood The effect of caffeinated The effect of caffeinated shared more under Inter
Study 1
drinks drinks drinks consumption on drinks consumption on sharing SC than Ind SC.
(Caffeinated
consumption) (Negative) sharing information about the information about the side Results cannot be
Drinks)
side effects of caffeinated effects of caffeinated drinks explained by pre-existing
drinks was negative. was not significant. beliefs about the stimuli.
Manipulated Manipulated Side effects of Actual Shares H1a supported: H2a supported: High SR information was
Study 2 (story writing) (gender BPA The effect of self-relevance The effect of self-relevance shared more under Inter
(BPA match) (Negative) (based on gender) on sharing (based on gender) on sharing SC than Ind SC.
Plastic) the side effects of BPA Plastic the side effects of BPA Plastic
was negative. was not significant.
Manipulated Manipulated Side effects of Sharing H1a supported: H2a supported: Results persist
(pronoun circling) (saliva fictitious food Likelihood The effect of self-relevance The effect of self-relevance controlling for article
Study 3 reports) component (based on saliva reports) on (based on saliva reports) on evaluation.
(Saliva (Negative) sharing the side effects of a sharing the side effects of a
Reports) food component was negative. food component was positive
but not significant.
Measured Measured Handling toxic Actual Shares H1a supported: H2a supported: For positive information,
(brand textile The effect of self-relevance The effect of self-relevance the effect of shopping
Study 4
purchase) garments (based on brand buying (based on brand buying frequency on sharing
(Detox
(Negative and frequency) on sharing tweets frequency) on sharing tweets tweets was positive and
Catwalk)
Positive) about the brand was negative. about the brand was not significant.
significant.
Measured and Measured Side effects Sharing H1a & H1b supported: H2a & H2b supported: Less authentic language
Manipulation (chocolate and benefits of Likelihood The effect of chocolate The effect of chocolate is used under Ind SC
(exposed to consumption) chocolate + Mediators consumption led to higher consumption led to lower self- when sharing high SR
Study 5
pronouns in (Negative and self-serving biases and serving biases and higher negative information.
(Chocolate
content) Positive) reluctance to show knowledge desire to seek information
Study)
about the side effects of about the side effects of For positive information,
chocolate, which together chocolate, which explained the the effect of consumption
explained less shares. insignificant effect on shares. did not explain sharing.

Note: Ind SC = independent self-construal, Inter SC = interdependent self-construal, SR = self-relevance.

5
6 E. Akpinar et al. / International Journal of Research in Marketing xxx (2018) xxx–xxx

Likelihood of Sharing
6 Independent
Interdepedent
5
4
3
2
1
Low Self-relevance High Self-relevance

Fig. 1. The moderating role of self-construal on sharing.

themselves than those in the interdependent condition (M = 4.69 vs. M = 3.75, F (1,200) = 41.48, p b .001). Conversely, partic-
ipants in the interdependent condition thought more about others (M = 5.01 vs. 4.49, F (1,200) = 12.55, p b .001).
A second manipulation check confirmed that consumption level was related to the self-relevance of the product-harm informa-
tion. At the end of the experiment, participants rated personal relevance of the information regarding caffeinated drinks' effect on
hair loss (1 = not at all, 7 = extremely). Participants who consumed caffeinated drinks more frequently found the information to
be more relevant (r = 0.29, p b .001).

4.2. Results

We examined the effect of self-relevance and self-construal on sharing likelihood using an OLS regression, mean centering self-
relevance. Confirming our theory and earlier results, self-construal moderated the effect of self-relevance on sharing. The interac-
tion effect between self-relevance and self-construal on sharing likelihood was significant (β = 0.12, t (198) = 2.01, p = .045). As
predicted, the follow-up slope tests showed that under independent self-construal, higher self-relevance decreased sharing
(β = −0.32, t (98) = −3.37, p = .001), and this negative effect disappeared under interdependent self-construal (β = −
0.018, t (100) = −0.15, p = .87), see Fig. 1.
We tested a number of alternative explanations. First, one could argue that because interdependent self-construal focuses on in-
terconnectedness (Markus & Kitayama, 1991), we might expect that participants would share all types of information more. As can be
observed in our results (Fig. 1), there is a main effect of self-construal, so that information is shared less under independent self-con-
strual than under interdependent (p b .001). This trend is driven by the high self-relevance condition. Spotlight analyses showed a
significant difference between independent and interdependent self-construal when information is highly self-relevant (M = 2.57
vs. 3.79, F (1,45) = 8.21, p = .006), but not for low self-relevant information (M = 3.66 vs. 3.91, F (1,60) = 0.46, p = .49).
Second, one could argue that participants who consume caffeinated drinks more often might have pre-existing beliefs about
their effects on health, which might affect sharing likelihood. To examine this notion, we conducted a post-study with another
sample of participants (n = 40) from the same population. Participants were asked to indicate the extent to which they thought
there was a strong association between hair loss and caffeinated drinks (1 = strongly disagree, 7 = strongly agree). Participants
also indicated the extent to which they consumed caffeinated drinks on a 7-point scale (1 = not at all, 7 = high amounts). There
was no correlation between consuming caffeinated drinks and the perceived strength of association between hair loss and caffein-
ated drinks (r = −0.087, p = .59), which suggests that caffeine consumption is not confounded with general beliefs (outside the
context of the present study) about the effect of caffeine consumption on hair loss.

4.3. Discussion

Study 1 shows that higher self-relevance of information makes consumers less likely to share negative information when inde-
pendent self-construal was activated. This effect was attenuated when an interdependent self-construal was activated.

5. Study 2: BPA plastic (the effects of self-relevance and self-construal on actual sharing)

In Study 2, we conceptually replicate Study 1. We (a) manipulate both self-construal and self-relevance, and (b) examine actual
information sharing behavior rather than relying on self-reported intentions to share.

5.1. Method

Ninety-nine undergraduates (Mage = 23; 52 males) at a Western-European university completed this study as part of a set of
experiments in exchange for partial course credits. Participants were randomly assigned to one of four conditions in a 2 (self-rel-
evance: low vs. high) vs. 2 (self-construal: independent vs. interdependent) between-subjects design.

Please cite this article as: Akpinar, E., et al., Sharing product harm information: The effects of self-construal and self-relevance, In-
ternational Journal of Research in Marketing (2018), https://doi.org/10.1016/j.ijresmar.2018.01.001
E. Akpinar et al. / International Journal of Research in Marketing xxx (2018) xxx–xxx 7

Fig. 2. Stimuli used in Study 2.

First, self-construal of the participants was manipulated through a story-writing task. Participants wrote a short story in which
they were instructed to either use the pronouns “I, me, myself, mine” to prime independent self-construal or “we, our, ourselves,
ours” to prime interdependent self-construal (Zhang & Shrum, 2009).
Second, participants were informed that they would be participating in an online conversation task where they would read
some newspaper articles and chat with a randomly chosen participant through instant messaging. Participants received a New
York Times article that was allegedly shared by another person who was simultaneously participating in the same study.
The article discussed the harmful health effects of BPA plastic, and was based on actual news (cf., Grady, 2010). A control ques-
tion showed that only 11% of the participants indicated that they had heard about the information beforehand. This did not have
any effect on sharing (p = .28).
Self-relevance of the article was manipulated using a procedure adapted from Puntoni et al. (2011). There were two versions of the
article. In the first version, the effects of BPA plastics are discussed to be harmful for female consumers (i.e., to breast cancer). In the sec-
ond version, the effects of BPA plastics are discussed to be harmful for male consumers (i.e., to prostate cancer). The article discussing
harmful effects of BPA plastic on breast cancer, is highly self-relevant for females and lowly self-relevant for males, whereas the article
discussing harmful effect of BPA plastic on prostate cancer, is highly self-relevant for males and lowly self-relevant for females (Fig. 2).
Thus, we were able to keep everything else about the articles the same, and just manipulate self-relevance.
Next, participants were provided with the option to share the article with people outside the lab, by clicking on the share but-
tons via email, Twitter, or a social network site. This decision to share the article was our key dependent variable (for the flow of
the behavioral sharing set-up and stimuli, see Web Appendix, Part 1). If participants clicked on any of the “sharing” buttons, they
saw a screen informing them that the Internet connection was not available. This setup allowed us to record actual shares while
preventing participants from sharing fictitious information outside the lab. Participants received an apology for the technical incon-
venience. Before leaving the lab, they were debriefed and probed for suspicion. None suspected the actual purpose of the study.

5.2. Results

We examined sharing as a binary dependent variable and self-construal and self-relevance as the independent variables, using
a logistic regression. Consistent with our theorizing, self-construal moderated the effect of self-relevance on sharing. The interac-
tion effect between self-construal and self-relevance on sharing was significant, (β = 1.97, p = .04). More important and consis-
tent with findings of Study 1, under independent self-construal, high self-relevant information (M = 10.3%) was less likely to be
shared than low self-relevant information (M = 33%; β = −2.15, p b .001). Further, as expected, the reluctance to share high self-
relevant information was attenuated under interdependent self-construal (Mhigh = 66.70% vs. Mlow = 54.50%; β = 0.69, p =
.109). Overall, highly self-relevant information was more likely to be shared under interdependent self-construal (M = 66.7%)

Please cite this article as: Akpinar, E., et al., Sharing product harm information: The effects of self-construal and self-relevance, In-
ternational Journal of Research in Marketing (2018), https://doi.org/10.1016/j.ijresmar.2018.01.001
8 E. Akpinar et al. / International Journal of Research in Marketing xxx (2018) xxx–xxx

than under independent self-construal (M = 10.3%), (β = 2.85, p b .05). An additional analysis for gender differences showed no
main or interaction effect of gender (χ2 (1) b 1, n.s.).

5.3. Discussion

Study 2 replicates and extends Study 1, by using manipulations rather than measures of self-relevance and using actual sharing
behaviors rather than self-reported sharing intentions as the dependent variable. Study 3 therefore tests whether the results can
be replicated with different manipulations of self-relevance and self-construal.

6. Study 3: saliva sample report (ruling out alternative explanations)

Study 3 tests the central hypothesized effects. To rule out a potential alternative explanation under which the effect would be
driven by knowledge or a priori assumptions about food that are consumed at higher levels, we used fictitious stimuli created spe-
cifically for this study, so that the participants could not have any prior knowledge about the harmful effects of the product. Fur-
ther, self-relevance is manipulated independent of participant characteristics, such as consumption behavior (Study 1) or gender
(Study 2).

6.1. Method

6.1.1. Materials
We used a 2 (self-relevance: high vs. low) × 2 (self-construal: independent vs. interdependent) between-subjects design. As in
Study 2, self-construal was primed using a pronoun-circling task.
We created two versions of a newspaper article discussing the health risks of a (fictional) food additive that we called “Stevia,”
and that was used in breakfast cereals, soft drinks and snack foods. The article reported that the consumption of Stevia could result
in damage to organs such as liver, stomach and kidney. One version of the article stated that Stevia posed risks for people with
high levels of duclin (a fictitious hormone), and the other version stated that Stevia posed risks for people with low levels of duclin
(for stimuli, see Web Appendix, Part 2). Using a fictional additive and linking it to fictional hormone levels allowed us to create a
health risk manipulation about which consumers could not have prior knowledge.
To manipulate self-relevance of the information (which was based on levels of duclin), a quick saliva test was allegedly con-
ducted (cf., Gal, 2012). When signing up for the study, participants who agreed to take a quick saliva test were allowed to partic-
ipate. To collect saliva samples, participants were provided with a swab (dental cotton roll) and asked to rub it between their
cheeks and lower gums; the swabs were then sealed in plastic bags. Participants were provided with a (fictitious) report of
their basic hormone levels, and levels of duclin. The layout and design of the report was based on actual test reports that can
be obtained from online testing companies.
While participants were waiting for their reports, they completed the pronoun-circling task (self-construal priming). Partici-
pants were randomly exposed to reports that indicated either high or low levels of duclin. They were instructed to examine the
report after reading the newspaper article. The negative information was highly self-relevant for participants who received reports
with a) high levels of duclin and newspaper article discussing side effects of Stevia for people with high levels of duclin and b) low
levels of duclin and newspaper article discussing side effects of Stevia for low levels of duclin.
To confirm that the manipulations were effective, participants were asked to indicate the extent to which they believed that
the risks associated with Stevia were relevant for them (1 = strongly disagree, 7 = strongly agree). Participants in the high
self-relevance condition evaluated the negative information as more self-relevant (M = 5.02) compared to those under the low
self-relevance condition (M = 3.14), F (1, 82) = 48.46, p b .001.

6.1.2. Participants and design


Eighty-four undergraduates completed a short survey as part of a larger group of questionnaires. The number of participants
was relatively low because the experiment was laborious: individuals' saliva samples had to be collected, and subsequently “proc-
essed” (which involved a waiting time to make the procedure believable), after which the participants received and read their per-
sonalized test report.
First, participants were randomly exposed to one of the self-construal priming conditions. Second, participants were randomly
exposed to one of the newspaper articles that manipulated self-relevance, in combination with their saliva reports. Next, partici-
pants indicated their likelihood of sharing the information in the article with others (e.g., friends, acquaintances), imagining that a
related conversation comes up on a 7-point Likert scale (1 = not likely to share, 7 = very likely to share; cf., Akpinar & Berger,
2017; Heath, 1996). To test whether self-relevance influenced the evaluation of the article, participants were asked to evaluate the
articles (1 = bad, 7 = good). In addition, we collected two measures about the evaluation of the content (credibility, difficulty to
process), and one item to measure the desire to help others. The outcomes of this ancillary analysis were consistent with the more
elaborate analysis in Study 5 (See Web Appendix, Part 2).

Please cite this article as: Akpinar, E., et al., Sharing product harm information: The effects of self-construal and self-relevance, In-
ternational Journal of Research in Marketing (2018), https://doi.org/10.1016/j.ijresmar.2018.01.001
E. Akpinar et al. / International Journal of Research in Marketing xxx (2018) xxx–xxx 9

6.2. Results

A 2 (self-relevance) × 2 (self-construal) ANOVA was conducted on sharing. Similar to Studies 1 and 2, self-construal moderated
the effect of self-relevance on sharing (F (1, 80) = 7.83, p = .006). Consistent with the predictions and earlier results, under in-
dependent self-construal, high self-relevant product harm information (M = 4.30) was less likely to be shared than low self-rel-
evant information (M = 5.39, F (1,80) = 5.78, p = .02). This effect disappeared under interdependent self-construal. In fact,
under interdependent self-construal, there was a positive but not significant difference between sharing of high self-relevant
(M = 5.24) and low self-relevant (M = 4.55) information, F (1,80) = 2.38, p = .12.
Further, we tested the effect of article evaluation on sharing. While self-relevance did not have a significant effect on article eval-
uation (M = 4.94 vs. M = 4.63; F (1, 82) = 1.54, p = .22), there was a main effect of article evaluation on sharing (β = 0.37, F (1,
82) = 13.77, p b .001). The moderation effect of self-construal on sharing was still significant (F (1, 79) = 9.59, p = .003), when
we included article evaluation as a control variable. These results confirm that the effects due to self-relevance and self-construal per-
sist controlling for the effect of article evaluation.

6.3. Discussion

Study 3 confirms and extends the findings of the first two studies. Taken together, self-relevance reduces sharing under inde-
pendent self-construal, but not under interdependent self-construal. The focus of our research is on sharing product harm infor-
mation. Yet, one could wonder whether it is the negativity of the content or self-relevance that drives our results.
In Studies 4 and 5, we therefore we not only examine the sharing of negative (product-harm) information, but also test how
the effects of self-relevance and self-construal play a role in sharing positive information.

7. Study 4: Detox the Catwalk (actual sharing of positive and negative information)

Study 4 is an online field experiment in which participants made actual online sharing decisions. Further, rather than manip-
ulating self-construal, we measure chronic self-construal and examine its effects on sharing. We measure self-relevance of the in-
formation, similar to Study 1. Further, we test whether our results are specific to sharing negative information but not positive
information about products.

7.1. Method

7.1.1. Participants and design


Two hundred and four participants from a large Western European university participated in this study in exchange for
course credit. First, self-relevance was operationalized via shopping frequency of fashion brands. Participants were asked
to indicate their shopping frequency for several fashion brands (Nike, Esprit, Zara and HM) on a six-point scale (1 = never,
6 = every week). This measure served as an indicator of self-relevance, with higher frequency translating into greater self-
relevance. To measure chronic self-construal, the Singelis (1994) scale was administered. Chronic self-construal was calculat-
ed as an index of independence relative to interdependence, following the procedure described by Escalas and Bettman
(2005).

7.1.2. Procedure
Participants were told that they would be taking part in two unrelated studies. In the first study, they answered a list of ques-
tions that included questions about shopping frequency for several brands.
Next, participants were informed about an actual social media campaign that is run by Greenpeace: The Detox Catwalk. In
this campaign, brands are evaluated on their efforts to eliminate hazardous chemicals. Valence was manipulated in a within-
subject design, manipulating the evaluation of the brands (Zara, HM, Nike and Esprit), as provided in the Greenpeace campaign.
For Zara and HM, positive information was provided. For Nike and Esprit, negative information was provided. Participants were
asked whether they would like to support the campaign by sending related tweets to the information they received (e.g., pos-
itive tweet: “Keep up the good work @Zara and lead the textile sector towards a toxic-free future.”; negative tweet: “Toxic
#fashion is so last season. #ESPRIT it is time to #Detox!”). The ready-to-send tweets were presented in a randomized order.
To facilitate sending the tweets, we created a Twitter account and participants clicked on send buttons, which recorded their
sharing behavior in the online experiment set-up and match this data with self-relevance and self-construal measures (See
Web Appendix, Part 3).

7.1.3. Manipulation checks


We conducted a series of tests to check whether our manipulations worked as intended. First, participants rated the tweets
about H&M and Zara more positive than the tweets about Nike and Esprit (Mpositive = 4.63 vs. Mnegative = 3.99; t = 7.93,
p b .001). Second, participants who buy the brands more frequently find the information more relevant for the self (rHM =
0.49; rZara = 0.67, rNike = 0.35, rEsprit = 0.42, ps b .001). Relevance for others was less related to shopping frequency (rHM =
0.14, p = .03; rZara = 0.16, p = .01; rNike = 0.04, p = .49; rEsprit = 0.08, p = .24). Third, we ran a post-test to check whether
our stimuli were perceived as related to health effects. Results confirmed that participants (n = 70) perceived the campaign to

Please cite this article as: Akpinar, E., et al., Sharing product harm information: The effects of self-construal and self-relevance, In-
ternational Journal of Research in Marketing (2018), https://doi.org/10.1016/j.ijresmar.2018.01.001
10 E. Akpinar et al. / International Journal of Research in Marketing xxx (2018) xxx–xxx

be related to health effects (M = 4.97, SD = 1.49; which was higher than the scale midpoint (4), t (l, 69) = 5.44, p b .001). Fur-
ther, participants were asked to indicate what extent they felt threatened by the information. Perceived self-threat was significant-
ly higher among participants who read about negative (Nike and Esprit), versus positive information (H&M and Zara), Mnegative =
4.68 vs. Mpositive = 3.44; F (1,69) = 30,58, p b .001. Further, perceived self-threat was positively related to shopping frequency
(r = 0.32, p b .001). Taken together, these findings provide evidence that the stimuli – although of a different nature than the
stimuli in Studies 1, 2 and 3 - were perceived to be self-threatening and related to health risks, especially for participants to
whom the brands had greater self-relevance (i.e., who had a higher purchase frequency).

7.2. Results

We examined how self-relevance influenced sharing negative and positive information under varying levels of chronic self-con-
strual. A binary logistic regression was conducted to predict whether participants tweeted the information or not (0 = did not
send a tweet, 1 = sent a tweet). Our independent variables were level of chronic self-construal, self-relevance (i.e., the extent
to which participants shop for the brands) and valence. We categorized self-relevance as low and high using a median-split for
each brand.1
Our results for sharing information showed that the three-way interaction between self-construal, self-relevance and valence
(β = 4.47, p = .19) was not significant. There was a significant interaction effect of valence and self-relevance (β = 0.77, p =
.01). Further, planned contrasts showed that for negative information, participants with higher self-relevance (i.e., shopped the tar-
get brands more frequently), although not significantly, were less likely to share tweets (β = −0.31, p = .15), whereas for pos-
itive information, participants with higher self-relevance were more likely to share (β = 0.48, p = .015).
Most importantly, we explored the predicted effects on sharing product harm information. We conducted a binary logistic re-
gression that tested the effect of self-relevance and self-construal on sharing negative information, controlling for the effect of
brands. In line with our hypotheses, and the results of Studies 1–3, there was a significant interaction effect between the self-rel-
evance and self-construal on shares (β = −5.21, p = .04). We also found a marginal negative effect of self-relevance on sharing
(β = −0.40, p = .07), and a not-significant effect of self-construal (β = 5.26, p = .17). Follow-up contrasts confirmed our hy-
potheses: for participants with independent self-construal (higher than the mean of centered self-construal index),2 negative in-
formation with higher self-relevance was less likely to be shared (β = −0.84, p = .018). Under interdependent self-construal,
this effect was not significant (β = −0.043, p = .88).
For the negative stimuli, although there was no significant three-way interaction between brand, self-relevance and self-con-
strual (β = −1.34, p = .60), and all the two-way interactions with the brand were also nonsignificant (p N .15), we did conduct
follow-up analyses at the brand level, using a split sample analyses. For Nike, the interaction effect between self-relevance and
self-construal on shares was negative and significant (β = −6.56, p = .05), and for Esprit, it was also in the expected direction,
yet not significant (β = −3.86, p = .33).
For the positive stimuli, a binary logistic regression testing the effect of self-relevance and self-construal on shares, controlling
for the effect of the brands, revealed significant positive main effects of self-relevance (β = 0.40, p = .05), yet the interaction was
not significant (β = −0.16, p = .94). This is in line with the general notion that consumers are more likely to share positive in-
formation about the product benefits if the products or brands are more relevant to them and this effect was more prevalent
under interdependent self-construal (p = .05). The two-way interactions and the three-way interaction between the brand,
self-relevance and self-construal were not significant (p N .23).

7.3. Discussion

Study 4 extends the findings of the first three studies using an actual social media campaign run by a global NGO. For negative
information, the results show a negative effect of self-relevance on sharing under independent self-construal, but not under inter-
dependent self-construal. Further, we obtain different effects for positive stimuli than negative stimuli, which show us that it is not
just self-relevance that explains sharing but rather negativity of the content, which is moderated by self-construal.

8. Study 5: chocolate (understanding the drivers of sharing)

Study 5 examines the extent to which the effects observed in Studies 1–4 are related to self-serving biases and reduced mo-
tivations to share the information, as proposed in our theory section. Our theoretical framework proposes that self-serving biases
may prevent consumers from sharing product harm information. To capture self-serving bias, we include measures from prior
(marketing) research on consumers' defensive processing of perceived (health) risks. One common measure is to ask participants
about the extent to which they feel concerned about the risks. Individuals that engage in defensive processing report lower levels of
concern when a threat is personally relevant to them (Ülkümen & Thomas, 2013; Yan & Sengupta, 2013). In addition, we mea-
sured the perceived believability of the information. Individuals who are at risk tend to defend themselves by questioning the be-
lievability of negative (but not positive) health-related information (Block, 2005; Puntoni et al., 2011; Taylor and Brown 1998).

1
We used median split of shopping frequency as it varied across different brands (MNike = 3.08, SD Nike = 0.99; MEsprit = 1.54, SD Esprit = 0.808;
M H&M = 3.42, SD H&M = 1.1; M Zara = 3.20, SD Zara = 1.46).
2
The sample sizes under 1SD above and below the mean are relatively lower (around 20), therefore we could not conduct spotlight.

Please cite this article as: Akpinar, E., et al., Sharing product harm information: The effects of self-construal and self-relevance, In-
ternational Journal of Research in Marketing (2018), https://doi.org/10.1016/j.ijresmar.2018.01.001
E. Akpinar et al. / International Journal of Research in Marketing xxx (2018) xxx–xxx 11

In addition to these measures, we captured the extent to which higher levels of relevance impact consumers' motivations to
share health-related information. Specifically, we measured whether consumers were motivated (1) to show off their knowledge
and expertise to others, which boosts self-enhancement (Berger, 2014; Berger & Milkman, 2012; Packard & Wooten, 2013), and
(2) to receive information from others about the effects of chocolate. Research has shown that consumers who are in doubt
about risky and uncertain situations, use word of mouth to seek advice from others (Hennig-Thurau et al., 2004). Thus, someone
who consumes large amounts of chocolate and learns about potential side effects, might want to talk with others to reflect upon
the situation and receive suggestions about what to do. In Study 5, we assess how these measures mediated the effects of self-rel-
evance on the intention to share, under interdependent and independent self-construal. In addition to these self-report measures,
we asked those participants who indicated that they were likely to share the information with others, to write down what they
would tell others about the health effects of chocolate. This allowed us to examine the effects of our variables on (a proxy for)
the content of word of mouth.
In Study 5, we both manipulate and measure self-construal. Self-relevance was measured using levels of consumption, similar
to Studies 1 and 4. Similar to Study 4, we also manipulated the valence of the information, to confirm that the predicted effects of
self-relevance and self-construal are limited to the context of negative (product-harm) information.

8.1. Method

8.1.1. Participants and design


One hundred and ninety-five participants from a large Western European university participated in this online study in ex-
change for course credit. This study utilized a 2 (self-construal: independent vs. interdependent) × 2 (valence: positive vs. nega-
tive) between-subjects design. Self-relevance was operationalized as a continuous measure based on reported levels of chocolate
consumption. Further, the chronic self-construal Singelis (1994) scale was administered, which allowed us to control for differ-
ences in chronic self-construal.

Fig. 3. Stimuli used in Study 5.

Please cite this article as: Akpinar, E., et al., Sharing product harm information: The effects of self-construal and self-relevance, In-
ternational Journal of Research in Marketing (2018), https://doi.org/10.1016/j.ijresmar.2018.01.001
12 E. Akpinar et al. / International Journal of Research in Marketing xxx (2018) xxx–xxx

8.1.2. Procedure
Participants were told that they would be taking part in two unrelated studies. In the first study, they answered a set of ques-
tions that first included the measures of self-construal, and then their consumption of chocolate, as well as a number of filler prod-
ucts on 7-point scale (not at all/quite a lot).
Next, participants were presented with a newspaper article containing information about the effects of chocolate which sup-
posedly appeared in the New York Times. To manipulate self-construal, the articles featured either independent appeals or inter-
dependent appeals, adapted from Agrawal and Maheswaran (2005). Specifically, the article text contained 27 pronouns. The same
text was presented with almost all of the pronouns being either self-oriented (your, you, I) or collective-oriented (our, we, us) to
prime self-construal. The headlines read: “Chocolate Benefits Your (vs. Our) Health”. The text featured the effects of chocolate as
either focused on self or collective benefits (e.g., “I might avoid some diseases” or “We might avoid some diseases”).
Besides self-construal, we manipulated the valence of the information in the article. In the positive condition, health benefits of
chocolate were presented. In the negative condition, harmful effects of chocolate were presented. The stimuli are presented in
Fig. 3.

8.1.3. Manipulation checks


After reading the article, participants answered three items on 7-point scales anchored by “not at all/a lot”, that measured the
extent to which participants focused on and thought about the self, others, and their social circle, adapted from Agrawal and
Maheswaran (2005). A composite scale was calculated to measure focus on others, including a reversed item for focus on the
self (Cronbach alpha = 0.68). Confirming the efficacy of the self-construal prime, participants in the interdependent (vs. indepen-
dent) prime condition focused more on others (Minter = 3.49 vs. Mind = 3.12; F (l, 190) = 5.54, p = .02). This effect also persisted
when we control for the effect of chronic self-construal.
Second, participants rated the valence of the information in the article on two items (negative/positive; bad/good; r = 0.91).
Participants in the positive (vs. negative) condition rated the health effects of the chocolate more favorably (Mpositive = 5.39 vs.
Mnegative = 2.29; F (l, 193) = 536.01, p b .001). We also checked whether consumption of chocolate was related to the self-rele-
vance of the information in the articles. Participants rated the extent to which the article was personally relevant on three items
(not personally relevant/personally relevant; not personally involving/personally involving; not likely to happen to you/very likely
to happen) on 7-point scales (Cronbach alpha = 0.76; Block, 2005). Results confirmed that participants who consume more choc-
olate found the information more personally self-relevant (r = 0.27, p b .001).
Finally, evaluation of the articles (ratings of ‘informative’, ‘exciting’, ‘useful’, ‘scary’ on a 7-point scale) was not influenced by the
self-relevance and self-construal manipulations. For negative information, the evaluation of articles did not differ on being
informative (Minter = 4.53 vs. Mind = 4.58; F (l, 97) = 0.043, p = .83), exciting (Minter = 3.61 vs. Mind = 3.61; F (l, 97) = 0,
p = 1.0), useful (Minter = 4.12 vs. Mind = 3.96; F (l, 97) = 0.30, p = .58), or scary (Minter = 2.61 vs. Mind = 2.69; F (l, 97) = 0.062,
p = .80) across self-construal conditions. The results persisted for positive information on being informative (Minter = 4.58 vs.
Mind = 4.75; F (l, 94) = 0.49, p = .48), exciting (Minter = 4.56 vs. Mind = 4.13; F (l, 94) = 2.61, p = .10), useful (Minter = 4.00 vs.
Mind = 3.98; F (l, 94) = 0.005, p = .94), scary (Minter = 1.25 vs. Mind = 1.19; F (l, 94) = 0.25, p = .61). Further, the interactions be-
tween self-relevance and self-construal on these measures were not significant (p N .21). As expected, negative articles were evaluated
as scarier than positive articles (p b .001) and positive articles were evaluated as more exciting than negative articles (p b .001).

8.1.4. Dependent variables


Participants were asked for their likelihood of sharing the article with others (“Imagine your friends/family are talking about
the effects of chocolate. How likely are you to pass along (share) what you have read in the article?”) on a 7-point scale (not
very likely/very likely) which served as our main dependent variable (cf., Akpinar & Berger, 2017; Heath, 1996).
Second, we measured the potential processes that explain sharing. To assess self-serving biases, we included measures of
the extent to which participants (1) felt concerned about the health risks (Ülkümen & Thomas, 2013; Yan & Sengupta,
2013), and (2) found the information believable (Block, 2005; Puntoni et al., 2011; Taylor and Brown 1998). To capture
motivations for sharing, we measured the extent to which consumers were motivated to share this information (1) to
show their knowledge to others (i.e., self-enhancement; Berger & Milkman, 2012; Packard & Wooten, 2013), and (2) to
receive information from others (Berger, 2014; Hennig-Thurau et al., 2004). We provide details of the measures in the
Web Appendix (Part 4).
In addition to these self-report measures, we used an open-ended method to further examine the processes behind sharing.
Participants who indicated that they were likely to share the information with others (i.e., score higher than 4 out 7), were
asked to write down what they would tell others about the health effects of chocolate. We used text analysis (i.e., Linquistic In-
quiry and Word Count [LIWC], Pennebaker, Boyd, Jordan, & Blackburn, 2015) to examine these comments for the occurrence of
self-serving biases. Research suggest that when individuals discuss personal topics in a defensive way (i.e., self-deceptively),
they use a less authentic tone (Barrett, Williams, & Fong, 2002; Newman, Pennebaker, Berry, & Richards, 2003). Standard linguistic
analysis software (LIWC) allows researchers to examine the “authenticity” of a text. Higher authenticity scores are associated with
a more honest, personal, and disclosing text; lower numbers suggest a more guarded, distanced form of discourse (Pennebaker et
al., 2015). Because we propose that individuals under independent self-construal should display more self-serving biases regarding
the product harm information when this information is more relevant to them, we predict that they will use less authentic lan-
guage when they communicate such information to others.

Please cite this article as: Akpinar, E., et al., Sharing product harm information: The effects of self-construal and self-relevance, In-
ternational Journal of Research in Marketing (2018), https://doi.org/10.1016/j.ijresmar.2018.01.001
E. Akpinar et al. / International Journal of Research in Marketing xxx (2018) xxx–xxx 13

8.2. Results

8.2.1. Main effects


First, we tested the effect of valence, self-relevance and self-construal (independent versus interdependent) on sharing. Consis-
tent with the results in Study 4, the two-way and the three-way interactions between self-construal, self-relevance and valence
were not significant (ps N .52) except for the interaction between self-relevance and valence (β = 0.31, p = .15) and self-rele-
vance and self-construal for negative information (β = −0.55, p = .15).
Next, we tested the effect of self-relevance and self-construal on sharing negative information about chocolate. Follow-up
contrasts confirmed our hypotheses: for participants under independent self-construal, higher self-relevant information
was less likely to be shared (β = −0.32, p = .015). Under interdependent self-construal, this effect was not significant
(β = −0.18, p = .19), controlling for the effect of chronic self-construal. Further, we tested the effect of self-relevance
and self-construal on sharing positive information about chocolate. The effect of self-relevance on shares was not significant
under both independent (β = 0.064, p = .68) and interdependent self-construal (β = 0.035, p = .81), controlling for the
effect of chronic self-construal.

8.2.2. Mediation analyses using self-reports


We explored four possible mediators for the effects of self-relevance and self-construal on willingness to share. These
mediators include two indicators for self-serving bias (concern about the topic and believability of the information), and
two known drivers of information sharing (the desire to show knowledge and the desire to learn more about a topic).
Because self-relevance only influenced the sharing of negative information, this discussion focuses only on the results for
this half of the design. For the effects of self-relevance on mediators for positive information, please see Table 1A, Web
Appendix Part 4.
First, in line with our theorizing, we found that in the independent negative condition, participants for whom the information
had higher self-relevance rated the article as less believable (β = −0.33, p = .02) and were less concerned about the related
health effects (β = −0.29, p = .03). These findings are consistent with the self-serving biases discussed in literature. In the inde-
pendent self-construal condition, we found a negative effect of self-relevance on the motivation “to show knowledge” (β = −0.36,
p = .01), while in the interdependent condition, we found a positive effect of self-relevance on the motivation “to learn more”
(β = 0.34, p = .015),
Next, in order to test the mediating role of self-serving biases and motivations on sharing negative information under
different levels of self-construal, we used the moderated mediation method (PROCESS macro; Hayes, Preacher, & Myers,
2011). Our results showed that the effect of self-relevance on sharing was mediated differently under independent versus
interdependent self-construal. First, the significant effect of self-relevance on sharing (β = −0.40) became insignificant
under independent self-construal after including the mediators in the model (β′ = −0.069), total indirect effect =
− 0.33, SE = 0.11, 95% CI [−0.57, −0.12]. This suggests mediation under independent self-construal. The negative effect
of self-relevance on sharing was driven by lower perceived believability and less concern about the health effects and
less reluctance to show knowledge. On the other hand, under interdependent self-construal, the effect of self-relevance
on sharing was insignificant (β = −0.13). Further, confirming our predictions, there was no significant effect of the me-
diators on sharing, except for a partial mediation of motivation to ‘learn more’ under interdependent self-construal (indi-
rect effect = 0.15, SE = 0.09, 95% CI [0.02, 0.41]); which meant higher self-relevance increased the motivation to learn
more, which in turn had a partial positive effect on sharing (for further details of the results, Web Appendix Part 4 and
Fig. A4).
Taken together, our results confirm that under independent self-construal, the negative effect of self-relevance on sharing was
mediated by higher self-serving biases (i.e., lower scores on perceived believability and concern about the topic), and higher re-
luctance to show their knowledge about the health effects of chocolate. For participants primed with interdependent self-constru-
al, there was no significant effect of self-relevance on self-serving biases and motivations, which explained the non-significant
effect of self-relevance on shares.
Further ancillary tests showed that variables such as believability and concern about health effects of chocolate did not mod-
erate the effect of self-relevance, neither the effect of self-construal priming on shares (Web Appendix Part 4).

8.2.3. Text analysis


Participants who were willing to share the content (rated 4 and higher on a 7-point scale) were probed to write in an open-
ended format how they would share this content with others. We analyzed these answers in a text analysis using the LIWC pro-
gram (Pennebaker et al., 2015) to explore whether the effects of self-relevance and self-construal reflected on the authenticity of
the shared content. As expected, consumers with independent self-construal used less authentic language to communicate more
self-relevant product harm information (β = −0.35, p = .04), controlling for the effect of chronic self-construal. This result is con-
sistent with the self-serving biases, where consumers share negative content that is highly self-relevant in a less honest and per-
sonal way. The effect of self-relevance on authenticity was not significant under any of the other conditions (p N .35). We have
also tested the effect of self-relevance and self-construal on other LIWC dimensions (e.g., pronouns, social words, emotional
words; for detailed results, see Table 2A, Web Appendix Part 4).

Please cite this article as: Akpinar, E., et al., Sharing product harm information: The effects of self-construal and self-relevance, In-
ternational Journal of Research in Marketing (2018), https://doi.org/10.1016/j.ijresmar.2018.01.001
14 E. Akpinar et al. / International Journal of Research in Marketing xxx (2018) xxx–xxx

8.3. Discussion

Study 5 extends the findings of the first four studies using both a manipulation of self-construal and a chronic measure of self-
construal. We used a self-construal priming procedure different than circling pronouns (used in Study 1 and Study 3) or writing
stories using pronouns (Study 2). Participants were exposed to pronouns while reading the stimuli, which could be an easier pro-
cedure to be adapted by practitioners. Although manipulation check showed that the self-construal prime had a significant effect,
it was relatively weak, possibly because self-construal priming was less salient while reading compared to circling pronouns or
using them in writing.
For product harm information, the results again show a negative effect of self-relevance on sharing under independent self-
construal, but not under interdependent self-construal. We did not observe any moderation effect of self-construal for positive in-
formation, showing that our results are specific to negative information. Further, we find that less shares for highly self-relevant
information under independent self-construal are explained by higher self-serving biases and lower motives for sharing the
knowledge. This behavior seems counterintuitive, as one could expect that when individuals are concerned about others, they
would share in order to protect others or to show that they care about them. When encountering product harm information, in-
dividuals under independent self-construal, should still have the drive to protect others, but this might well be overshadowed by
their self-serving biases. This explains why under independent self-construal, individuals are more likely to share content that has
low self-relevance. Finally, we observed that individuals shared more self-relevant product harm information in a less authentic
way, which is consistent with the self-serving biases against threats.

9. General discussion

Product harm information spreading in the marketplace has become increasingly important for companies, public policy
makers and consumer welfare. But what makes consumers share such information with others? The present work investigated
this question, contributing to both health psychology and word of mouth literature.
Taken together, the findings demonstrate that consumers who are more at risk of products' negative health effects (i.e., those
for whom the information has high self-relevance), are less likely to share the information with others, but only under indepen-
dent self-construal (see Table 2). The results also suggest that this behavior is the outcome of self-serving biases and reduced mo-
tivations to demonstrate knowledge, that helps consumers cope with self-threatening information. Although not sharing
threatening health information might keep the threat ‘far from home’, it also prevents consumers from protecting themselves
or letting others know about the potential hazards. Under interdependent self-construal, the negative effect of self-relevance on
sharing was attenuated, due to lower levels of self-serving biases and higher motivations to learn more.
The findings are consistent across a variety of manipulations and measurements of the core constructs as well as measures for
sharing, which underscores their generalizability. Further, we have shown that the effect of self-relevance and self-construal on
sharing were specific to product harm information (negative content), but not to product benefits (positive content).

9.1. Managerial implications

This research has important implications for managers and public policy makers as well as consumer well-being. Policy makers
might design campaigns that encourage consumers to share their concerns about products (see Stephen & Lehmann, 2016, for a
similar suggestion). Our research suggests that sharing can be influenced through manipulating self-construal and self-relevance,
which can be accomplished in several ways, some of which are suggested below.
First, as shown in Study 5, in order to influence sharing, policy makers may craft content that primes interdependent self-con-
strual. The simple use of plural pronouns (“We should be careful about our consumption of X”) may, for example, encourage shar-
ing self-relevant negative information among consumers (cf., Agrawal & Maheswaran, 2005).
Second, sharing can be influenced by changing the focus of the target audience, which might trigger interdependence (e.g.,
how consuming food high in cholesterol affects one's health versus showing how people worry about the consumption behaviors
of a relative or a friend). Highlighting how products harm other people might trigger the desire to help others, and reduce self-serv-
ing biases. As a result, dissemination of such information might increase.

Table 2
Summary of studies: effect of self-relevance on sharing under independent and interdependent self-construal.

Study 1 Study 2 Study 3 Study 4 Study 5

Sharing Negative info Negative info Negative info Negative info Positive info Negative info Positive info

β (Sig.) β (Sig.) Means (Sig) β (Sig.) β (Sig.) β (Sig.) β (Sig.)

Ind SC −0.32 (.001) −2.15 (.001) Mlow: 5.39 −0.84 (.018) 0.18 (.55) −0.32 (.015) 0.064 (.68)
Mhigh: 4.30 (.02)
Inter SC −0.018 (.87) 0.69 (.10) Mlow: 4.55 0.043 (.88) 0.54 (.05) −0.18 (.19) 0.035 (.81)
Mhigh: 5.24 (.12)

The table displays the results (β = regression coefficient, p-value) from analyses across the five studies testing the effect of self-relevance on sharing information
under independent self-construal (Ind SC) and interdependent self-construal (Inter SC).

Please cite this article as: Akpinar, E., et al., Sharing product harm information: The effects of self-construal and self-relevance, In-
ternational Journal of Research in Marketing (2018), https://doi.org/10.1016/j.ijresmar.2018.01.001
E. Akpinar et al. / International Journal of Research in Marketing xxx (2018) xxx–xxx 15

Third, this research suggests that companies can respond to product harm information spreading in the marketing place by
pursuing different response strategies to targeted segments based on their self-relevance (i.e., heavy and light users, age, gender,
geography) and self-construal. This suggestion is in line with research suggesting several efficient ways to tailor communications
about health issues to different market segments (Keller & Lehmann, 2008). Online platforms such as Twitter and Facebook facil-
itate the targeting of specific (segments of) consumers with promoted content. For instance, negative information might be
targeted especially toward those who are more likely to spread (i.e., high self-relevance under interdependent self-construal).
Finally, our linguistic analysis findings provided initial evidence on the way information is communicated based on self-rele-
vance, self-construal and valence. There is some existing work that shows that self-construal and related cultural characteristics
influence usage of language and can be partially captured through LIWC categories (De Andrea, Shaw, & Levine, 2010). Further
research could establish ways to predict consumers' level of self-construal based on their online textual content such as Twitter
posts.

9.2. Directions for future research

Certain limitations in the present research suggest opportunities for future study. First, our research focused on a very specific
area related to product harm information, yet may apply to other areas such as negative consumption experiences due to product
side effects. Furthermore, our research has focused mostly on the effect of self-relevance based on consumption, whereas products
could also be self-relevant based on extent product signal their identity (e.g., signaling to be environmental friendly) in sharing
negative product information. Second, we have tried to simulate real sharing behavior through two controlled experimental de-
signs where individuals were given a chance to share online (Study 2) or participate in an actual sharing campaign (Study 4).
While field evidence on sharing negative information such as product harm crises (cf., Borah & Tellis, 2016) would have been
ideal, manipulating self-relevance of the product harm information and self-construal at an aggregate level with field data remains
as a challenge for future research.
Next, future research could examine how the type of audience affects sharing. There is evidence that the in-group versus out-
group distinction is more important to individuals with interdependent self-construal than to those with independent self-constru-
al (Triandis, 1989). Recent work has shown that depending on whether the audience is a stranger versus a close friend, valence of
the content might change during word of mouth (Dubois et al., 2016). Further research could examine whether the effect of self-
relevance would lead to higher levels of sharing under interdependent self-construal, especially when with close friends (i.e., in-
group) and this effect might disappear for the audience with strangers (i.e., out-group).
Third, future research could look at the extent to which the roles of self-relevance and self-construal differ between online and
offline sharing. Several recent studies have studied differences between word of mouth online and offline (Eelen et al., 2017;
Kannan & Li, 2017; Lamberton & Stephen, 2016). Although Study 2 and Study 4 - which examine the sharing of information in
an online context - replicate the general pattern of effects of self-relevance under independent and interdependent self-construal,
online and offline sharing may differ in terms of the motives that underlie sharing product harm information. Online sharing of
product harm information – which reaches a large audience - may be driven to a greater extent by the desire to show one's
knowledge about a topic (cf., Barasch & Berger, 2014), while offline shares might be driven by reciprocal sharing of information.
In conclusion, this research demonstrates how product harm information is shared based on self-relevance of the information
and self-construal of the consumers. We shed light on the self-serving biases that makes consumers not share information that
poses a threat for them, and how this reluctance in sharing can be mitigated.

Acknowledgments

The authors thank the editor and the anonymous reviewers for their constructive comments and valuable feedback. We
gratefully acknowledge financial support from the Erasmus Research Institute of Management (ERIM) and Vrije University
Amsterdam.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.ijresmar.2018.01.001.

References

Aaker, J. L., & Lee, A. Y. (2001). ‘I’ seek pleasures and ‘we’ avoid pains: The role of self-regulatory goals in information processing and persuasion. Journal of Consumer
Research, 28(1), 33–49.
Agrawal, N., & Maheswaran, D. (2005). The effects of self-construal and commitment on persuasion. Journal of Consumer Research, 31(4), 841–849.
Ahluwalia, R., Burnkrant, R. E., & Unnava, H. R. (2000). Consumer response to negative publicity: The moderating role of commitment. Journal of Marketing Research,
37(2), 203–214.
Akpinar, E., & Berger, J. (2017). Valuable virality. Journal of Marketing Research, 54(2), 318–330.
Alexandrov, A., Lilly, B., & Babakus, E. (2013). The effects of social-and self-motives on the intentions to share positive and negative word of mouth. Journal of the
Academy of Marketing Science, 41(5), 531–546.
Babić Rosario, A., Sotgiu, F., De Valck, K., & Bijmolt, T. H. A. (2016). The effect of electronic word of mouth on sales: A meta-analytic review of platform, product, and
metric factors. Journal of Marketing Research, 53(3), 297–318.
Barasch, A., & Berger, J. (2014). Broadcasting and narrowcasting: How audience size affects what people share. Journal of Marketing Research, 51(3), 286–299.

Please cite this article as: Akpinar, E., et al., Sharing product harm information: The effects of self-construal and self-relevance, In-
ternational Journal of Research in Marketing (2018), https://doi.org/10.1016/j.ijresmar.2018.01.001
16 E. Akpinar et al. / International Journal of Research in Marketing xxx (2018) xxx–xxx

Barrett, L. F., Williams, N. L., & Fong, G. T. (2002). Defensive verbal behavior assessment. Personality and Social Psychology Bulletin, 28(6), 776–788.
Bell, P. A. (1978). Affective state, attraction, and affiliation: Misery loves happy company, too. Personality and Social Psychology Bulletin, 4(4), 616–619.
Berger, J. (2014). Word of mouth and interpersonal communication: A review and directions for future research. Journal of Consumer Psychology, 24(4), 586–607.
Berger, J., & Milkman, K. L. (2012). What makes online content viral? Journal of Marketing Research, 49(2), 192–205.
Berger, J., Sorensen, A. T., & Rasmussen, S. J. (2010). Positive effects of negative publicity: When negative reviews increase sales. Marketing Science, 29(5), 815–827.
Block, L. G. (2005). Self-referenced fear and guilt appeals: The moderating role of self-construal. Journal of Applied Social Psychology, 35(11), 2290–2309.
Block, L. G., & Williams, P. (2002). Undoing the effects of seizing and freezing: Decreasing defensive processing of personally relevant messages. Journal of Applied Social
Psychology, 32(4), 803–830.
Borah, A., & Tellis, G. J. (2016). Halo (spillover) effects in social media: Do product recalls of one brand hurt or help rival brands? Journal of Marketing Research, 53(2),
143–160.
Buechel, E. C., & Berger, J. (2015). Motivations for engaging in social media. In C. Dimofte, C. Haugtvedt, & R. Yalch (Eds.), Consumer psychology in a social media world
(pp. 3–23). New York; London: Routledge, Taylor & Francis.
Carvalho, S. W., Block, L. G., Sivaramakrishnan, S., Manchanda, R. V., & Mitakakis, C. (2008). Risk perception and risk avoidance: The role of cultural identity and per-
sonal relevance. International Journal of Research in Marketing, 25(4), 319–326.
Chandran, S., & Menon, G. (2004). When a day means more than a year: Effects of temporal framing on judgments of health risk. Journal of Consumer Research, 31(2),
375–389.
Chapman, B. (2017). Nutella maker fights back against fears over cancer-causing palm oil. Retrieved from http://www.independent.co.uk/news/business/news/
nutella-causes-cancer-palm-oil-supermarket-ban-italy-cadburys-a7522291.html, Accessed date: February 2017.
Cheung, C. M. K., & Thadani, D. R. (2012). The impact of electronic word-of-mouth communication: A literature analysis and integrative model. Decision Support
Systems, 54(1), 461–470.
Cross, S. E., Hardin, E. E., & Gercek-Swing, B. (2011). The what, how, why, and where of self-construal. Personality and Social Psychology Review, 15(2), 142–179.
Dana, J., & Loewenstein, G. (2003). A social science perspective on gifts to physicians from industry. Journal of the American Medical Association, 290(2), 252–255.
De Andrea, D. C., Shaw, A. S., & Levine, T. R. (2010). Online language: The role of culture in self-expression and self-construal on Facebook. Journal of Language and Social
Psychology, 29, 425–442.
De Angelis, M., Bonezzi, A., Peluso, A. M., Rucker, D. D., & Costabile, M. (2012). On braggarts and gossips: A self-enhancement account of word-of-mouth generation and
transmission. Journal of Marketing Research, 49(4), 551–563.
Ditto, P. H., & Lopez, D. F. (1992). Motivated skepticism: Use of differential decision criteria for preferred and non-preferred conclusions. Journal of Personality and Social
Psychology, 63(4), 568–584.
Dubois, D., Bonezzi, A., & De Angelis, M. (2016). Sharing with friends versus strangers: How interpersonal closeness influences word-of-mouth valence. Journal of
Marketing Research, 53(5), 712–727.
Dubois, D., Rucker, D. D., & Tormala, Z. L. (2011). From rumors to facts, and facts to rumors: The role of certainty decay in consumer communications. Journal of
Marketing Research, 48(6), 1020–1032.
Dunning, D., Meyerowitz, J. A., & Holzberg, A. D. (1989). Ambiguity and self-evaluation: The role of idiosyncratic trait definitions in self-serving assessments of ability.
Journal of Personality and Social Psychology, 57(6), 1082–1090.
Eelen, J., Özturan, P., & Verlegh, P. W. J. (2017). The differential impact of brand loyalty on traditional and online word of mouth: The moderating roles of self-brand
connection and the desire to help the brand. International Journal of Research in Marketing, 34(4), 872–891.
Escalas, J. E., & Bettman, J. R. (2005). Self-construal, reference groups, and brand meaning. Journal of Consumer Research, 32(3), 378–389.
Fontaine, K. R., & Smith, S. (1995). Optimistic bias in cancer risk perception: A cross-national study. Psychological Reports, 77(1), 143–146.
Forest, A. L., & Wood, J. V. (2012). When social networking is not working: Individuals with low self-esteem recognize but do not reap the benefits of self-disclosure on
Facebook. Psychological Science, 23(3), 296–305.
Gal, D. (2012). A mouth-watering prospect: Salivation to material reward. Journal of Consumer Research, 38(6), 1022–1029.
Gardner, W. L., Gabriel, S., & Lee, A. Y. (1999). ‘I’ value freedom, but ‘we’ value relationships: Self-construal priming mirrors cultural differences in judgment.
Psychological Science, 10(4), 321–326.
Grady, D. (2010). In feast of data on BPA plastic, no final answer. Retrieved from http://nytimes.com/2010/09/07/science/07bpa.html, Accessed date: 9 November
2011.
Greenpeace (2017). Detox Catwalk campaign. Retrieved from http://www.greenpeace.org/international/en/campaigns/detox/fashion/detox-catwalk/, Accessed date:
9 January 2017.
Hayes, A. F., Preacher, K. J., & Myers, T. A. (2011). Mediation and the estimation of indirect effects in political communication research. In E. P. Bucy, & R. L. Holbert
(Eds.), Sourcebook for political communication research: Methods, measures, and analytical techniques (pp. 434–465). New York: Routledge.
Heath, C. (1996). Do people prefer to pass along good or bad news? Valence and relevance of news as predictors of transmission propensity. Organizational Behavior
and Human Decision Processes, 68(2), 79–94.
Heath, C., Bell, C., & Sternberg, E. (2001). Emotional selection in memes: The case of urban legends. Journal of Personality and Social Psychology, 81(6), 1028–1041.
Heine, S. J., Lehman, D. R., Markus, H. R., & Kitayama, S. (1999). Is there a universal need for positive self-regard? Psychological Review, 106(4), 766–794.
Hennig-Thurau, T., Gwinner, K. P., Walsh, G., & Gremler, D. D. (2004). Electronic word-of-mouth via consumer-opinion platforms: What motivates consumers to ar-
ticulate themselves on the internet? Journal of Interactive Marketing, 18(1), 38–52.
Hepper, E. G., Gramzow, R. H., & Sedikides, C. (2010). Individual differences in self-enhancement and self-protection strategies: An integrative analysis. Journal of
Personality, 78(2), 781–814.
Hornik, J., Satchi, R. S., Cesareo, L., & Pastore, A. (2015). Information dissemination via electronic word-of-mouth: Good news travels fast, bad news travels faster!
Computers in Human Behavior, 45, 273–280.
Hsu, L., & Lawrence, B. (2016). The role of social media and brand equity during a product recall crisis: A shareholder value perspective. International Journal of Research
in Marketing, 33(1), 59–77.
Kannan, P. K., & Li, H. A. (2017). Digital marketing: A framework, review and research agenda. International Journal of Research in Marketing, 34(1), 22–45.
Keller, P. A. (2015). Social marketing and healthy behavior. In D. W. Stewart (Ed.), The handbook of persuasion and social marketing (pp. 9–38). Routledge: New York.
Keller, P. A., & Block, L. (1999). The effect of affect-based dissonance versus cognition-based dissonance on motivated reasoning and health-related persuasion. Journal
of Experimental Psychology: Applied, 5(3), 302–313.
Keller, P. A., & Lehmann, D. R. (2008). Designing effective health communications: A meta-analysis. Journal of Public Policy and Marketing, 27(2), 117–130.
Keller, P. A., Lipkus, I. M., & Rimer, B. K. (2003). Affect, framing, and persuasion. Journal of Marketing Research, 40(1), 54–64.
Kitayama, S., Markus, H. R., Matsumoto, H., & Norasakkunkit, V. (1997). Individual and collective processes in the construction of the self: Self-enhancement in the
United states and self-criticism in Japan. Journal of Personality and Social Psychology, 72(6), 1245–1267.
Kitayama, S., Park, H., Sevincer, A. T., Karasawa, M., & Uskul, A. K. (2009). A cultural task analysis of implicit independence: Comparing North America, Western Europe,
and East Asia. Journal of Personality and Social Psychology, 97(2), 236–255.
Lamberton, C., & Stephen, A. T. (2016). A thematic exploration of digital, social media, and mobile marketing: Research evolution from 2000 to 2015 and an agenda for
future inquiry. Journal of Marketing, 80(6), 146–172.
Lei, J., Dawar, N., & Gürhan-Canli, Z. (2012). Base-rate information in consumer attributions of product-harm crises. Journal of Marketing Research, 49(3), 336–348.
Liberman, A., & Chaiken, S. (1992). Defensive processing of personally relevant health messages. Personality and Social Psychology Bulletin, 18(6), 669–679.
Liu, A. X., Liu, Y., & Luo, T. (2016). What drives a firm's choice of product recall remedy? The impact of remedy cost, product hazard, and the CEO. Journal of Marketing,
80(3), 79–95.
Luce, M. F., & Kahn, B. E. (1999). Avoidance or vigilance? The psychology of false-positive test results. Journal of Consumer Research, 26(3), 242–259.
Lynch, J. G., Bradlow, E. T., Huber, J. C., & Lehmann, D. R. (2015). Reflections on the replication corner: In praise of conceptual replications. International Journal of
Research in Marketing, 32(4), 333–342.

Please cite this article as: Akpinar, E., et al., Sharing product harm information: The effects of self-construal and self-relevance, In-
ternational Journal of Research in Marketing (2018), https://doi.org/10.1016/j.ijresmar.2018.01.001
E. Akpinar et al. / International Journal of Research in Marketing xxx (2018) xxx–xxx 17

Markus, H. R., & Kitayama, S. (1991). Culture and the self: Implications for cognition, emotion, and motivation. Psychological Review, 98(2), 224–253.
Menon, G., Block, L. G., & Ramanathan, S. (2002). We're at as much risk as we are led to believe: Effects of message cues on judgments of health risk. Journal of
Consumer Research, 28(4), 533–549.
Menon, G., Raghubir, P., & Agrawal, N. (2008). Health risk perceptions and consumer psychology. In C. Haugtvedt, P. Herr, & F. Kardes (Eds.), The handbook of consumer
psychology (pp. 981–1010). Lawrence Erlbaum and Associates: NJ.
Milkman, K. L., & Berger, J. (2014). The science of sharing and the sharing of science. Proceedings of the National Academy of Sciences, 111, 13642–13649.
Moe, W. W., & Schweidel, D. A. (2012). Online product opinions: Incidence, evaluation, and evolution. Marketing Science, 31(3), 372–386.
Newman, M. L., Pennebaker, J. W., Berry, D. S., & Richards, J. M. (2003). Lying words: Predicting deception from linguistic styles. Personality and Social Psychology
Bulletin, 29(5), 665–675.
Ng, S. (2010). Cultural orientation and brand dilution: Impact of motivation level and extension typicality. Journal of Marketing Research, 47(1), 186–198.
Packard, G., Gershoff, A. D., & Wooten, D. B. (2016). When boastful word of mouth helps versus hurts social perceptions and persuasion. Journal of Consumer Research,
43(1), 26–43.
Packard, G., & Wooten, D. B. (2013). Compensatory knowledge signaling in consumer word-of-mouth. Journal of Consumer Psychology, 23(4), 434–450.
Pennebaker, J. W., Boyd, R. L., Jordan, K., & Blackburn, K. (2015). The development and psychometric properties of LIWC2015. Research report. Austin, TX: University of
Texas At Austin.
Perloff, L. S., & Fetzer, B. K. (1986). Self–other judgments and perceived vulnerability to victimization. Journal of Personality and Social Psychology, 50(3), 502–510.
Price, L. L., Feick, L. F., & Guskey-Federouch, A. (1995). Everyday market helping behavior. Journal of Public Policy and Marketing, 255–266.
Puntoni, S., Sweldens, S., & Tavassoli, N. T. (2011). Gender identity salience and perceived vulnerability to breast cancer. Journal of Marketing Research, 48(3), 413–424.
Raghubir, P., & Menon, G. (1998). AIDS and me, never the twain shall meet: The effects of information accessibility on judgments of risk and advertising effectiveness.
Journal of Consumer Research, 25(1), 52–63.
Rosnow, R. L. (2001). Rumor and gossip in interpersonal interaction and beyond: A social exchange perspective. In R. M. Kowalski (Ed.), Behaving badly: Aversive be-
haviors in interpersonal relationships (pp. 203–232). Washington, DC: American Psychological Association.
Sedikides, C., Gaertner, L., & Toguchi, Y. (2003). Pancultural self-enhancement. Journal of Personality and Social Psychology, 84(1), 60–79.
Shavitt, S., Torelli, C. J., & Wong, J. (2009). Identity-based motivation: Constraints and opportunities in consumer research. Journal of Consumer Psychology, 19(3),
261–266.
Sherman, D. A. K., Nelson, L. D., & Steele, C. M. (2000). Do messages about health risks threaten the self? Increasing the acceptance of threatening health messages via
self-affirmation. Personality and Social Psychology Bulletin, 26(9), 1046–1058.
Singelis, T. M. (1994). The measurement of independent and interdependent self-construal. Personality and Social Psychology Bulletin, 20(5), 580–591.
Stephen, A. T., & Lehmann, D. R. (2016). How word-of-mouth transmission encouragement affects consumers' transmission decisions, receiver selection, and diffusion
speed. International Journal of Research in Marketing, 33(4), 755–766.
Sundaram, D. S., Mitra, K., & Webster, C. (1998). Word-of-mouth communications: A motivational analysis. Advances in Consumer Research, 25(1), 527–531.
Taylor, S. E., & Brown, J. D. (1988). Illusion and well-being: A social psychological perspective on mental health. Psychological Bulletin, 103(2), 193–210.
Tirunillai, S., & Tellis, G. J. (2012). Does chatter really matter? Dynamics of user-generated content and stock performance. Marketing Science, 31(2), 198–215.
Torelli, C. J. (2006). Individuality or conformity? The effect of independent and interdependent self-concepts on public judgments. Journal of Consumer Psychology,
16(3), 240–248.
Triandis, H. C. (1989). The self and social behavior in differing cultural contexts. Psychological Review, 96(3), 506–520.
Ülkümen, G., & Thomas, M. (2013). Personal relevance and mental simulation amplify the duration framing effect. Journal of Marketing Research, 50(2), 194–206.
Van Heerde, H., Helsen, K., & Dekimpe, M. G. (2007). The impact of a product-harm crisis on marketing effectiveness. Marketing Science, 26(2), 230–245.
Weinstein, N. D. (1980). Unrealistic optimism about future life events. Journal of Personality and Social Psychology, 39(5), 806–820.
White, K., Argo, J. J., & Sengupta, J. (2012). Dissociative versus associative responses to social identity threat: The role of consumer self-construal. Journal of Consumer
Research, 39(4), 704–719.
Yan, D., & Sengupta, J. (2013). The influence of base rate and case information on health-risk perceptions: A unified model of self-positivity and self-negativity. Journal
of Consumer Research, 39(5), 931–946.
Yang, X., Mao, H., & Peracchio, L. A. (2012). It's not whether you win or lose, it's how you play the game? The role of process and outcome in experience consumption.
Journal of Marketing Research, 49(6), 954–966.
Zhang, Y., & Shrum, L. J. (2009). The influence of self-construal on impulsive consumption. Journal of Consumer Research, 35(5), 838–850.

Please cite this article as: Akpinar, E., et al., Sharing product harm information: The effects of self-construal and self-relevance, In-
ternational Journal of Research in Marketing (2018), https://doi.org/10.1016/j.ijresmar.2018.01.001

You might also like