Content uploaded by ko de ruyter
Author content
All content in this area was uploaded by ko de ruyter on Jan 19, 2015
Content may be subject to copyright.
http://oss.sagepub.com
Organization Studies
DOI: 10.1177/0170840607076003
2007; 28; 347 Organization Studies
Caroline Wiertz and Ko de Ruyter
Online Communities
Beyond the Call of Duty: Why Customers Contribute to Firm-hosted Commercial
http://oss.sagepub.com/cgi/content/abstract/28/3/347
The online version of this article can be found at:
Published by:
http://www.sagepublications.com
On behalf of:
European Group for Organizational Studies
can be found at:Organization Studies Additional services and information for
http://oss.sagepub.com/cgi/alerts Email Alerts:
http://oss.sagepub.com/subscriptions Subscriptions:
http://www.sagepub.com/journalsReprints.navReprints:
http://www.sagepub.co.uk/journalsPermissions.navPermissions:
http://oss.sagepub.com/cgi/content/refs/28/3/347 Citations
at SAGE Publications on May 21, 2009 http://oss.sagepub.comDownloaded from
Beyond the Call of Duty: Why Customers
Contribute to Firm-hosted Commercial Online
Communities
Caroline Wiertz and Ko de Ruyter
Abstract
Firm-hosted commercial online communities, in which customers interact to solve each
other’s service problems, represent a fascinating context to study the motivations of col-
lective action in the form of knowledge contribution to the community. We extend a
model of social capital based on Wasko and Faraj (2005) to incorporate and contrast the
direct impact of commitment to both the online community and the host firm, as well as
reciprocity, on quality and quantity of knowledge contribution. In addition, we examine
the moderating influence of three individual attributes that are particularly relevant to the
firm-hosted community context: perceived informational value, sportsmanship, and
online interaction propensity. We empirically test our framework using self-reported and
objective data from 203 members of a firm-hosted technical support community. In addi-
tion to several interesting moderating effects, we find that a customer’s online interac-
tion propensity, commitment to the community, and the informational value s/he
perceives in the community are the strongest drivers of knowledge contribution.
Keywords: collective action, social capital, firm-hosted commercial online communities,
online interaction propensity
Introduction
Online communities have an unparalleled ability to facilitate the collective action
of knowledge contribution, as evidenced for example in the open source move-
ment (e.g. Lakhani and von Hippel 2003; von Hippel and von Krogh 2003). Even
though knowledge, which is the main resource exchanged in online communities,
possesses the quality of a public good that can be consumed by anyone (regard-
less of whether this individual has contributed to its production), a puzzling
number of people forgo the economically rational tendency to free-ride, and
rather share their knowledge for the good of the collective (Wasko et al. 2004).
An increasing number of firms are now attempting to exploit this phenome-
non by hosting online communities for commercial purposes, such as building
relationships with their customers, getting their feedback, strengthening the
brand, and reducing customer service costs by enabling peer-to-peer problem
solving (e.g. Moon and Sproull 2001). The success of these firm-hosted com-
mercial online communities entirely depends on the willingness of customers to
spend time and effort responding to each other’s requests for help. Given that not
article title
Organization
Studies
28(03): 347–376
ISSN 0170–8406
Copyright © 2007
SAGE Publications
(Los Angeles,
London, New Delhi
and Singapore)
Caroline Wiertz
Cass Business
School, City
University, UK
Ko de Ruyter
University of
Maastricht, the
Netherlands
www.egosnet.org/os DOI: 10.1177/0170840607076003
at SAGE Publications on May 21, 2009 http://oss.sagepub.comDownloaded from
only fellow customers, but ultimately also the host firm, benefit from the knowl-
edge exchanged in firm-hosted commercial online communities, it is surprising
that customers are willing to answer ‘the call of duty’ and actively engage in
knowledge contribution. Empirical evidence with respect to the predictive abil-
ity of antecedents of knowledge contribution has remained equivocal (e.g. Wasko
and Faraj 2005). Furthermore, it is not clear whether empirical findings in the
emerging literature on open source and practice-oriented online communities
(e.g. von Hippel and von Krogh 2003) can be extended to the context of firm-
hosted commercial online communities. Thus, there seems to be both a manage-
rial and theoretical need for a more in-depth understanding of the factors that
predict customers’ contribution to knowledge resources in this context.
Following Wasko and Faraj (2005), we advance theories of collective action
and social capital as a theoretical basis for gaining such an understanding. These
authors have tested a theoretical framework incorporating individual motivations
and social capital to explain voluntary behavior in computer-mediated knowl-
edge exchange networks. As the very existence of companies depends on sound
relationships with their customers (e.g. Morgan and Hunt 1994), our focus is on
the relational dimension of social capital within the context of firm-hosted com-
mercial online communities. In accordance with Wasko and Faraj (2005), we
examine the role of commitment and reciprocity as predictors of knowledge
contribution. However, in the case of firm-hosted commercial communities, it
remains unclear which object of commitment is most important in driving
knowledge contribution behavior. Commitment may not only be directed at the
online community, but also at the firm hosting the community. Therefore, we
assess how both types of commitment affect the quality and quantity of knowl-
edge contribution.
In addition to the aforementioned relational capital dimensions, Wasko and
Faraj (2005) argue that individual attributes of community members impact
knowledge contribution in electronic networks of practice. For the context of
firm-hosted commercial online communities, we propose three individual
attributes pertaining to the message, the medium, and the messenger: (1) infor-
mational value, (2) sportsmanship, and (3) online interaction propensity. First,
people contribute knowledge in online communities as they expect that ‘some
new value will be created’ (Wasko and Faraj 2005: 39). We propose that the
availability of valuable content will stimulate community members to con-
tribute. Second, while many companies are experimenting with online commu-
nities as an additional service support channel, there is little theoretical and
practical guidance on how to develop, manage, and improve this online chan-
nel. In such a learning-by-doing environment, it seems important that commu-
nity members display a goodwill tolerating less than ideal circumstances and a
willingness to face inconveniences and tackle challenges. This is reflected in
sportsmanship, a form of citizenship behavior. There is accumulating evidence
(e.g. Bell and Menguc 2002; Yoon and Suh 2003) that sportsmanship induces
customer commitment, compliance to service standards, and cooperative behav-
ior in service delivery operations. Third, it has been observed that members of
online communities strongly differ in their inclination to interact (Burnett
2000). Often a distinction is made between active contributors and lurkers
348 Organization Studies 28(03)
at SAGE Publications on May 21, 2009 http://oss.sagepub.comDownloaded from
(Cothrel and Williams 1999). Therefore, we examine the impact of online inter-
action propensity, or an individual’s general tendency to engage in online inter-
actions, on knowledge contribution.
Wasko and Faraj’s (2005) study on antecedents of knowledge contribution in
electronic networks of practice did not yield a consistent pattern between pre-
dictor and criterion variables. Recent theorizing on online behavior has sug-
gested that it may be conceptually relevant to investigate the moderating impact
of individual attributes, rather than direct effects (Dabholkar and Bagozzi 2002).
This line of reasoning has been substantiated empirically by recent research that
shows that people’s online behavior may differ as a result of cross-sectional het-
erogeneity between persons and different commercial contexts (Bucklin and
Sismeiro 2003). Therefore, it seems necessary to refine our understanding of dri-
vers of knowledge contribution in firm-hosted online communities by examining
the interaction effects between relational capital and individual attributes.
The Collective Action Problem in Firm-hosted Online
Communities
Online communities originally began to form as social entities with the aim of
bringing back a sense of belonging that was lost during the shift from community
to society (Fischer et al. 1996). More and more private individuals clustered online
with similar others to anchor themselves, support each other, and exchange infor-
mation (Bressler and Grantham 2000). By the mid-1990s, the commercial poten-
tial of such online groups was strongly propagated in the popular management
literature (e.g. Hagel and Armstrong 1997), with the result that numerous organi-
zations started to explore the opportunities for building their own online commu-
nity. These firm-hosted commercial online communities of customers are the
research context of this paper. We define commercial online communities as firm-
hosted online aggregations of customers who collectively co-produce and consume
content about a commercial activity that is central to their interest by exchanging
intangible resources. These intangible resources can take the form of information,
knowledge, socio-emotional support, and the like (Butler et al. forthcoming).
One of the most common types of firm-hosted commercial online communi-
ties is the community for service support (Rainie and Horrigan 2005). As
opposed to online brand communities, the main purpose of which is the cele-
bration of the brand and the affiliation with other brand enthusiasts (e.g.
Algesheimer et al. 2005), online communities for service support focus on peer-
to-peer problem solving and information exchange. In contrast to telephone and
on-site support, communities are commonly a free-of-charge support channel
that the host firm offers to its customers. Notable examples of firm-hosted ser-
vice support communities are the Dell, HP, Adobe, and iPod (Apple) user com-
munities for technical support, the Lonely Planet and Fodors communities for
knowledge exchange concerning all travel-related questions, or Ensemble
Studio’s Age of Empire community for strategy advice on online gaming.
It is worth noting that firm-hosted commercial online communities are con-
ceptually very different from open source communities, which have received
Wiertz and Ruyter: Beyond the Call of Duty 349
at SAGE Publications on May 21, 2009 http://oss.sagepub.comDownloaded from
recent research attention in the information systems literature (e.g. Lakhani and
von Hippel 2003; von Hippel and von Krogh 2003). Most importantly, open
source communities are not explicitly sponsored by companies, but are set up
by individuals or small groups as independent software development projects
(Moon and Sproull 2000). Hence, open source project participants can reap
direct benefits from their voluntary contributions to the project. Two of the most
cited reasons for participating in this type of community are a personal need for
the product being developed and the enhancement of career opportunities
(Lakhani and von Hippel 2003; von Hippel and von Krogh 2003). Members in
commercial online communities, in contrast, are customers of the host company
who have paid for ownership of the company’s products. Traditionally, one
would expect the company to provide a support service to its customers, either
free of charge or as part of a service contract. In firm-hosted commercial online
communities, however, customers not only seek this support service from other
customers, they even invest their own time and effort solving fellow customers’
problems. As such, members of firm-hosted commercial online communities
take over service functions traditionally provided by the host company, often-
times without getting any monetary compensation or other direct rewards.
While the host company obviously benefits from reduced service costs and
other valuable by-products of this customer-to-customer problem solving (e.g.
rich customer feedback, relationship building potential, etc.), it is less clear why
the host firm’s customers are willing to contribute knowledge to these commer-
cial online communities in the absence of obvious direct rewards. After all, the
community’s resources, which result from the knowledge contribution of its
members, have the quality of public goods. Public goods are defined by two char-
acteristics: nonexcludability and nonrivalry (Samuelson 1954). Once made avail-
able to one person, public goods can be consumed by all others at no additional
marginal costs, without being ‘used up’ (Olson 1965). In addition, individuals
cannot be excluded from consuming the public good, even though they might not
contribute to its production. Similarly, the knowledge contributed by one online
community member is visible and accessible to all other members, regardless of
whether they have ever actively contributed anything themselves (Wasko and
Faraj 2000, 2005; Wasko et al. 2004). Thus, there is ample opportunity for
members to lurk and free-ride instead of participating in the creation of the com-
munity’s resources, potentially resulting in underprovision of the public good.
This is known as the collective action problem (Olson 1965; Ostrom 2000). Firm-
hosted commercial online communities aggravate the collective action problem
and therefore represent a particularly interesting context to study knowledge con-
tribution behavior and its drivers. This aggravation stems from the fact that not
only fellow customers, but also the host firm can benefit from the public good (i.e.
free-of-charge service provision, knowledge about problems with its products and
services, and associated solutions) created by its customers. Hence, it is our aim
in this paper to further our understanding of why customers forgo their apparent
inclination to act out of self-interest and contribute knowledge to the firm-hosted
commercial online community rather than free-ride on the efforts of others. Our
point of departure for gaining such an understanding will be the concept of social
capital (e.g. Bourdieu 1986; Coleman 1988; Putnam 1993).
350 Organization Studies 28(03)
at SAGE Publications on May 21, 2009 http://oss.sagepub.comDownloaded from
Social Capital and Knowledge Contribution in
Firm-hosted Online Communities
Social capital is an elastic term with a variety of definitions. Generally, it is con-
ceptualized as an intangible resource of support that emanates from member-
ship of a social group which can be mobilized in times of need (Adler and
Kwon 2002; Bourdieu 1986; Coleman 1988). Traditional examples of such a
resource of support include babysitting clubs and neighborhood watches. The
basic premise of social capital is that investment in social relations results in
benefits (Coleman 1988). While investments are made by individuals in the col-
lective, the benefits accrue to both the collective as a whole and the individual
members (Lin 2001). Consequently, the promise of social capital accumulation
enables participants to act together and pursue shared objectives (Putnam
1993). In the case of the neighborhood watch, for example, the entire neigh-
borhood benefits every night from increased security, while each neighbor has
to participate only once a week. Thus, social capital operates on the assumption
that the total is more than the sum of its parts. However, beyond the basic agree-
ment that social capital resides in social relations, there is considerable debate
about whether it stems from the formal structure of these social relations or their
content (Adler and Kwon 2002). Whereas the former approach focuses on
social network analysis and measures such as centrality, the latter concentrates
on the quality of social relations based on trust, norms, and commitment
(Nahapiet and Ghoshal 1998; Wasko and Faraj 2005). In addition to this struc-
tural and relational dimension of social capital, Nahapiet and Ghoshal (1998)
have identified a third, cognitive dimension, which refers to a shared system of
meaning within a group, such as a common jargon.
These three dimensions of social capital have been put forward as main drivers
of knowledge sharing within organizations (Nahapiet and Ghoshal 1998), based
on the reasoning that they create supportive conditions for exchange. Wasko and
Faraj (2005) have adopted the same reasoning in an attempt to explain knowledge
contribution to electronic networks of practice. However, as opposed to the net-
work level of analysis employed by Nahapiet and Ghoshal (1998), they developed
their research model on the individual level, arguing that members of electronic
networks not only build a relationship with the network as a whole, but also with
individuals within the network. As a result, these individual relationships are
important sources of social capital and determine how individual members behave
in relation to others, for example with regard to knowledge sharing.
Since this is applicable to firm-hosted online communities as well, we adopt
the same approach and develop our theorizing on the individual level of analy-
sis. However, we focus primarily on the relational dimension of social capital as
a predictor of knowledge contribution, based on the following rationale. First, we
have no reasons to suppose that the structural or cognitive dimension would dif-
fer in the case of firm-hosted online communities. A recent study by Tsai (2006)
indeed confirms the importance of structural capital in virtual communities.
Second, Wasko and Faraj (2005: 51) did not find a consistent positive effect of
relational social capital on knowledge sharing between network members, and
concluded that ‘relational capital may not develop in electronic networks due to
Wiertz and Ruyter: Beyond the Call of Duty 351
at SAGE Publications on May 21, 2009 http://oss.sagepub.comDownloaded from
a lack of shared history, high interdependence, frequent interaction, and co-
presence’. This result is in contrast to Granovetter’s (1973) theory of weak ties,
which predicts that knowledge can also be exchanged in loosely knit network
structures, and that the information stemming from weak ties might in fact be
more valuable. Constant et al. (1996), for example, proposed that electronic
weak ties can lead to improved knowledge exchange because more people are
reached, their knowledge repositories are more diverse, and they possess more
resources. Hence, it seems theoretically relevant to re-examine the influence of
relational social capital on knowledge contribution in firm-hosted online com-
munities. Finally, since the main focus of these communities is to develop and
maintain relationships with customers, it seems managerially relevant to exam-
ine relational capital as a driver of knowledge contribution.
Relational social capital refers to the affective nature of social relationships
within a collective (Wasko and Faraj 2005) and has been identified as an important
facilitator of an individual’s actions within the collective (Coleman 1990). As such,
the relational dimension of social capital is expected to have a strong influence on
individual member behavior, such as knowledge contribution (Nahapiet and
Ghoshal 1998). There are two main aspects of relational social capital. The first is
mutual trust that help provided will be returned. This mutual trust arises within the
context of regular, cooperative behavior based on commonly shared norms
(Paldam and Svendsen 2001). In an online community, this trust facilitates the ease
of cooperation by reducing the risk of a one-way knowledge flow in which the
knowledge provider would be taken advantage of. The norm of reciprocity speci-
fies that people should help those who have helped them by returning equivalent
benefits. Recipients of beneficial actions feel a sense of indebtedness, which leads
to a motivation to alleviate this indebtedness through reciprocation (Gouldner
1960). Previous research has found that the reciprocity norm operates in online set-
tings and is able to motivate knowledge sharing (Wasko and Faraj 2000). Thus,
when the individual members of an online community perceive that a strong norm
of reciprocity governs the exchanges within the community, they trust that their
valuable knowledge contribution will be reciprocated at some point in the future.
In line with Wasko and Faraj (2005), we differentiate between the quality and the
quantity of knowledge contribution, and derive the following hypotheses:
H1a: An individual’s perception of the norm of reciprocity has a positive impact on the
quality of her/his knowledge contribution.
H1b: An individual’s perception of the norm of reciprocity has a positive impact on the
quantity of her/his knowledge contribution.
The second aspect of relational social capital is commitment to the collective,
which results in a perceived duty to help fellow members of the same collective
through knowledge contribution. As members have repeated positive exchange
experiences, the importance of the relationship with the community as a whole
increases accordingly and members become committed. Kollock (1999) posits
that it is this commitment that motivates members to contribute content. When
commitment to the community increases, members feel a sense of responsibil-
ity to assist others in the collective by sharing their valuable knowledge (Wasko
and Faraj 2005). This leads us to propose the following:
352 Organization Studies 28(03)
at SAGE Publications on May 21, 2009 http://oss.sagepub.comDownloaded from
H2a: An individual’s commitment to the online community has a positive impact on the
quality of her/his knowledge contribution.
H2b: An individual’s commitment to the online community has a positive impact on the
quantity of her/his knowledge contribution.
A unique feature of firm-hosted online communities is that its members not
only act out of concern for each other, but potentially also out of concern for the
host company. Members of firm-hosted commercial online communities are usu-
ally customers of the host firm, which means that they might have established a
bond with the firm that is independent of the online community. In most cases,
this commitment is motivated by the repeated purchase of and enthusiasm for the
firm’s products or services. In the literature on brand communities, it has been
established that members of these types of communities are not only dedicated to
their specific online group, but also the brand and underlying firm (Algesheimer
et al. 2005). Hence, it is possible that customers contribute their valuable knowl-
edge to the online community’s resources not only out of concern for the good of
the community itself, but because they also treasure their bond with the host firm.
Thus, in firm-hosted commercial online communities, an additional driver of
knowledge contribution may be the members’ commitment to the host firm:
H3a: An individual’s commitment to the host firm has a positive impact on the quality
of her/his knowledge contribution.
H3b: An individual’s commitment to the host firm has a positive impact on the quantity
of her/his knowledge contribution.
In addition to social capital, previous research proposes that knowledge con-
tribution is also influenced by individual attributes of network participants (e.g.
Nahapiet and Ghoshal 1998; Wasko et al. 2004; Wasko and Faraj 2005). In the
next section, we will elaborate on the potential impact of three individual vari-
ables that are particularly important in the context of firm-hosted online com-
munities: perceived informational value, sportsmanship, and online interaction
propensity. Our full conceptual model is summarized in Figure 1.
Individual Attributes and Knowledge Contribution in Firm-hosted
Online Communities
Social capital researchers have proposed that one important reason why some
individuals build up more social capital and engage more willingly in collective
action than others are individual attributes, such as motivations and abilities
(Adler and Kwon 2002; Coleman 1990; Lakhani and von Hippel 2003; Nahapiet
and Ghoshal 1998; Putnam 1993). Specifically, we propose that three individ-
ual attributes pertaining to the message, the medium, and the messenger will
influence knowledge contribution in firm-hosted online communities. These
attributes are (1) the perceived informational value of the message, i.e. the
knowledge exchanged in the community, (2) an individual’s level of tolerance
of imperfections in the medium, i.e. the online community, expressed by sports-
manship, and (3) online interaction propensity, the messenger’s tendency to
engage in online interactions.
Wiertz and Ruyter: Beyond the Call of Duty 353
at SAGE Publications on May 21, 2009 http://oss.sagepub.comDownloaded from
To our knowledge, studies on knowledge sharing to date have tended to study
direct effects of individual attributes (e.g. Constant et al. 1996; Nahapiet and
Ghoshal 1998; Wasko and Faraj 2005). Despite convincing hypotheses, several
of these direct effects have been found to be insignificant, much to the surprise
of the researchers. For example, Wasko and Faraj (2005) did not find a signifi-
cant relationship between ‘enjoying helping’ and knowledge contribution, and
Lakhani and von Hippel (2003) observed that the empirical evidence of a direct
relationship between reputational gains and knowledge contribution is frag-
mented. A possible reason for these inconsistent findings might be that individ-
ual attributes exert a moderating rather than direct effect. Nahapiet and Ghoshal
(1998: 251) suggest that the process that leads to knowledge contribution is inter-
related and complex, and that their own focus on direct effects ‘limits the rich-
ness of the present exploration and identifies an important area for future work’.
In general, it has been suggested that focusing on direct effects of individual
attributes may be somewhat ‘redundant and obvious’ (Dabholkar and Bagozzi
2002: 186), and that the investigation of moderating effects is much more mean-
ingful (e.g. Ajzen et al. 1982; Baron and Kenny 1986; James and Brett 1984).
Therefore, we investigate the impact of interactions between an individual’s rela-
tional social capital and her/his attributes on the level of knowledge contribution.
The relevance and moderating effects of perceived informational value, sports-
manship, and online interaction propensity are discussed next.
Moderating Effects of Perceived Informational Value
An important observation that has emerged from social capital research in the
context of organizations is that collective action is often driven by instrumental
motivations of the individual, such as career advancement (e.g. Lin et al. 1981).
Indeed, also in the context of electronic networks of practice and open source
354 Organization Studies 28(03)
Informational
value
Sportsmanship
Online
interaction
propensity
Individual attributes
Relational social
capital
Commitment
to community
Commitment
to host firm
Reciprocity
Quantity
Knowledge
contribution
Quality
Figure 1.
Conceptual Model
at SAGE Publications on May 21, 2009 http://oss.sagepub.comDownloaded from
communities, prior research indicates that individuals participate out of instru-
mental reasons: access to high-quality information and the opportunity to
exchange ideas and solutions (Lakhani and von Hippel 2003; Wasko and Faraj
2000; Wasko et al. 2004). Nahapiet and Ghoshal (1998) already suggested that
individuals will only contribute knowledge if they expect that this action will
create value for the collective, with the anticipation of personally benefiting
from this value in the future. In a professional environment such as a firm-
hosted online community, the main source of value that accumulates in the col-
lective and accrues to the individual is information. Customers visit the
firm-hosted community first and foremost because they have an information
need and hope to get answers from fellow customers. Thus, the perceived level
of the informational value that the community provides is an important individ-
ual attribute that clearly will have an effect on the relationship between rela-
tional social capital and knowledge contribution.
More specifically, we expect that higher levels of perceived informational
value will strengthen the relationship between reciprocity and knowledge con-
tribution. Even though the norm of reciprocity is universal (i.e. prevalent in all
groups to which an individual belongs), it is not unconditional (Gouldner,
1960). Rather, the intensity of the norm is ‘contingent upon the imputed value
of the benefit previously received’ (Gouldner 1960: 171). Hence, if a member
of a firm-hosted online community perceives the information provided to
her/him by other community members to be valuable, the indebtedness towards
the community and consequently her/his desire to reciprocate and contribute
knowledge will increase. Therefore, we propose the following hypotheses:
H4a: If an individual’s perception of the community’s informational value increases,
the relationship between reciprocity and the quality of her/his knowledge contri-
bution will be strengthened.
H4b: If an individual’s perception of the community’s informational value increases,
the relationship between reciprocity and the quantity of her/his knowledge contri-
bution will be strengthened.
In addition, perceived informational value may also have an impact on the
relationship between an individual’s commitment to the online community and
knowledge contribution. A committed community member feels a sense of
responsibility towards the collective and therefore assists other members.
Commitment builds over repeated interactions with others (Coleman 1990;
Nahapiet and Ghoshal 1998; Wasko and Faraj 2005). We expect that the value
of these repeated interactions will have an effect on how obliged an individual
feels to ‘pay back’ the collective by contributing her/his knowledge:
H4c: If an individual’s perception of the community’s informational value increases,
the relationship between commitment to the community and the quality of her/his
knowledge contribution will be strengthened.
H4d: If an individual’s perception of the community’s informational value increases,
the relationship between commitment to the community and the quantity of
her/his knowledge contribution will be strengthened.
As stated before, members of firm-hosted online communities might not only
be committed to the community itself, but also to the underlying company that
Wiertz and Ruyter: Beyond the Call of Duty 355
at SAGE Publications on May 21, 2009 http://oss.sagepub.comDownloaded from
hosts it, based on previous experiences with its products and services. There is
evidence that the impact of commitment on customer behavioural intentions is
enhanced by value perceptions (Pura 2005). Since the members of such a com-
munity are customers of the host firm who come to the community with prob-
lems related to those products and services, the online community in fact acts
as service support channel. If this support channel provides valuable informa-
tion to the customer — and thus a valuable service — s/he might feel that the
host firm has fulfilled its obligations and it is now her/his turn to contribute to
the continued success of the online community as a service channel by sharing
her/his knowledge. This leads to the following hypotheses:
H4e: If an individual’s perception of the community’s informational value increases,
the relationship between commitment to the host firm and the quality of her/his
knowledge contribution will be strengthened.
H4f: If an individual’s perception of the community’s informational value increases,
the relationship between commitment to the host firm and the quantity of her/his
knowledge contribution will be strengthened.
Moderating Effects of Sportsmanship
An attribute that is of particular importance in the context of firm-hosted online
communities is an individual’s level of sportsmanship. Sportsmanship has its
roots in organizational theory and is one of the original dimensions of organi-
zational citizenship behavior (Organ 1988). It is defined as ‘the willingness to
tolerate less than ideal circumstance without complaining’ (Podsakoff et al.
1997: 263), and has been demonstrated to have a beneficial effect on coopera-
tive behavior (e.g. Bell and Menguc 2002). In firm-hosted online communities,
‘less than ideal circumstances’ might arise from two different sources. On the
one hand, since online communities are computer-mediated and depend entirely
on information technologies, technical problems, for example with the web
server, as well as design and functionality issues might cause annoyance to
community members. On the other hand, the community members themselves
might create imperfections, for example by behaving inappropriately, or unin-
tentionally providing incorrect answers to service queries.
Higher levels of sportsmanship provide a shield against the potential negative
consequences of such imperfections, and as such strengthen the effects of rela-
tional social capital on knowledge contribution. Even if a community member
desires to reciprocate help received, a technical breakdown or inappropriate
member behavior might deter her/him from participating if s/he has a low level
of tolerance for such problems. If, however, the member displays high sports-
manship, s/he might be more willing to overlook these problems or see them as
improvement opportunities and continue trying to make helpful contributions.
Consequently, we suggest the following hypotheses:
H5a: If an individual’s level of sportsmanship increases, the relationship between reci-
procity and the quality of her/his knowledge contribution will be strengthened.
H5b: If an individual’s level of sportsmanship increases, the relationship between reci-
procity and the quantity of her/his knowledge contribution will be strengthened.
356 Organization Studies 28(03)
at SAGE Publications on May 21, 2009 http://oss.sagepub.comDownloaded from
A similar reasoning applies to the effect of sportsmanship on the relationship
between commitment to the online community as well as the host firm and
knowledge contribution. If the repeated interactions between online community
members are continuously disrupted or otherwise negatively impacted by tech-
nical problems or inappropriate member behavior, the feeling of responsibility
towards the community and consequently knowledge contribution would decline.
Sportsmanship provides a defense mechanism against such a chain of events.
Community members who display high sportsmanship are more willing to tol-
erate these negative aspects of the community and continue to assist fellow
members. Thus, we hypothesize:
H5c: If an individual’s level of sportsmanship increases, the relationship between com-
mitment to the community and the quality of her/his knowledge contribution will
be strengthened.
H5d: If an individual’s level of sportsmanship increases, the relationship between com-
mitment to the community and the quantity of her/his knowledge contribution will
be strengthened.
In relation to the interaction between commitment to the company and
sportsmanship, Mattila (2004) demonstrates that the negative impact of service
failures on customer-perceived commitment is less pronounced for those cus-
tomers who exhibit higher levels of failure tolerance. Since a firm-hosted
online community has the function of a service channel, its members might be
less forgiving of imperfections than they would be in other contexts. After all,
they are customers who have paid money for ownership of the firm’s products
or services that lie at the heart of the community. Often, the online community
is the only free-of-charge support option available to them, and they might not
easily forgive problems with elements that the host firm controls, such as
design, functionality, and other technical issues. If the quality of the service
support provided by the community suffers as a result of such imperfections,
the customer’s bond with the host firm might be negatively affected, hand in
hand with her/his willingness to continue investing in the community through
knowledge contribution. As a result, the level of sportsmanship will be crucial
in determining the amount of damage caused. More specifically, we expect the
following:
H5e: If an individual’s level of sportsmanship increases, the relationship between com-
mitment to the host firm and the quality of her/his knowledge contribution will be
strengthened.
H5f: If an individual’s level of sportsmanship increases, the relationship between com-
mitment to the host firm and the quantity of her/his knowledge contribution will
be strengthened.
Moderating Effects of Online Interaction Propensity
Interaction is a precondition for the development and maintenance of dense
social capital (Bourdieu 1986), and in firm-hosted commercial online commu-
nities, interaction takes place through asynchronous computer-mediated com-
munication. Knowledge contribution should therefore be strongly influenced by
Wiertz and Ruyter: Beyond the Call of Duty 357
at SAGE Publications on May 21, 2009 http://oss.sagepub.comDownloaded from
an individual’s willingness to engage in such online interactions. However,
despite its apparent importance, this crucial individual attribute has, to our
knowledge, received no previous research attention.
It is a common observation in online communities that members strongly dif-
fer in their interaction frequency (e.g. Burnett 2000; Hammond 2000). For
example, Hammond (2000) concludes from his qualitative work that there are
two types of membership in online groups: communicative membership, in
which individuals interact frequently, articulate concerns, and respond to mes-
sages; and quiet membership, in which individuals read messages but rarely
send/post messages of their own. Based on a review of the communication lit-
erature, we propose that the type of membership that an individual displays may
be explained in part by the presence (or lack) of a general disposition to engage
in online interactions. In the traditional communication and psychology disci-
plines, which deal with face-to-face interactions, it is well known that individ-
uals have different propensities to communicate with others (Liu 2003). This
personality trait has been called ‘(un)willingness to communicate’ and
describes a general tendency to approach or avoid communication (Burgoon
1976; McCroskey and Richmond 1985). But as all interaction on the Internet is
mediated by technology, it is profoundly different from face-to-face communi-
cation (e.g. Hoffman and Novak 1996). Online interactions are mostly asyn-
chronous, text-based, and lack both verbal and especially non-verbal cues. In
addition, owing to the truly global nature of the Internet, a large percentage of
online interactions occur between relative strangers. Many people who com-
municate with each other online have never personally met, and postings on
newsgroups, discussion boards, and in online communities can have potentially
global audiences.
Owing to these fundamental differences between offline and online commu-
nication, it is not suitable to simply transfer an offline communication trait and
apply it to the online context. Rather, it seems necessary to investigate online
interactions separately. Most research on the Internet has overlooked the exis-
tence of individual differences in online interaction preference (Liu 2003). In
order to overcome this shortcoming, we propose a new behavioral disposition —
online interaction propensity — that we define as a prevailing tendency of an
individual to interact with relative strangers (i.e. people they have never met
offline) in an online environment. It is this behavioral disposition, rooted in per-
sonality, that explains why one person will engage in online interaction and
another will not under identical circumstances.
The literature on norm theory alludes to the fact that the strength of the rela-
tionship between norms and resulting action is influenced by the presence of
certain conditions that are conducive to the activation of norms (Schwartz
1977). Schwartz (1977) proposes that the norm–behavior link is often influ-
enced by personality moderators. Likewise, Ostrom (2000) argues that the
development of a theory of collective knowledge sharing must take personality
traits into account. An important condition for the activation of norms is the
inclination of members to communicate with other members. By engaging in
community discussions and knowledge sharing, the communicative members
358 Organization Studies 28(03)
at SAGE Publications on May 21, 2009 http://oss.sagepub.comDownloaded from
become more aware of the norms that govern the community, making them in
turn a more powerful predictor of knowledge contribution. Thus, we put for-
ward the following hypotheses:
H6a: If an individual’s level of online interaction propensity increases, the relationship
between reciprocity and the quality of her/his knowledge contribution will be
strengthened.
H6b: If an individual’s level of online interaction propensity increases, the relationship
between reciprocity and the quantity of her/his knowledge contribution will be
strengthened.
In addition, we also expect a moderating effect of online interaction propensity
on the relationship between commitment to the online community and knowl-
edge contribution. As mentioned before, commitment to a collective is built
through repeated interactions with its members (Nahapiet and Ghoshal 1998).
Since interactions in online communities are computer-mediated, their fre-
quency will be impacted by an individual’s tendency to engage in such online
interactions. An online interaction-prone individual will communicate more fre-
quently and build stronger relationships with her/his fellow community
members and the collective as a whole. Moreover, these frequent interactions
will likely strengthen her/his feelings of obligation to provide help to fellow
members by contributing knowledge. We therefore propose:
H6c: If an individual’s level of online interaction propensity increases, the relationship
between commitment to the community and the quality of her/his knowledge con-
tribution will be strengthened.
H6d: If an individual’s level of online interaction propensity increases, the relationship
between commitment to the community and the quantity of her/his knowledge
contribution will be strengthened.
It has been argued that the communicative interplay between companies and
their customers plays a central role in forging relational bonds (Ford 2001).
Repeated interactions between company representatives and customers, for
instance, result in more favorable behavioral intentions towards the firm on the
part of the customer (Gutek 1995). A firm-hosted online community is often
managed by employees of the firm, and customers interacting with other cus-
tomers in the community know that they are being ‘watched’ by the firm. Thus,
they might perceive customer-to-customer interactions as an indirect communi-
cation with the host firm. As community members who are more inclined to
interact online are more likely to experience repeated interactions with firm rep-
resentatives and other customers, we predict that the influence of commitment
to the host firm on knowledge contribution in online communities will increase
as a result of this disposition.
H6e: If an individual’s level of online interaction propensity increases, the relationship
between commitment to the host firm and the quality of her/his knowledge con-
tribution will be strengthened.
H6f: If an individual’s level of online interaction propensity increases, the relationship
between commitment to the host firm and the quantity of her/his knowledge con-
tribution will be strengthened.
Wiertz and Ruyter: Beyond the Call of Duty 359
at SAGE Publications on May 21, 2009 http://oss.sagepub.comDownloaded from
Methodology
Research Setting
In order to empirically test our hypotheses, we conducted a quantitative study
among the members of an online technical support community hosted and mod-
erated by a large computer hard- and software supplier. The primary commer-
cial benefit of this online community is service cost reduction, and to a lesser
extent brand building and ‘listening in’. The community is available on a world-
wide basis, and the language for content is English. Furthermore, it is entirely
based on asynchronous discussion boards. Hence, information exchange is not
‘real time’, although most posts receive a reaction within minutes. All member-
generated content is stored and, owing to a powerful search engine, easily
accessible at all times. With a free-of-charge technical support service, the com-
munity is accessible to all of the host firm’s customers, but its main users are IT
professionals. It is worth noting, though, that the community does not exclu-
sively operate in a business-to-business context. The host firm estimates that at
least 25% of the community’s members are non-IT professionals. The goal of
the community is to provide a platform where like-minded IT enthusiasts —
regardless of whether they are business customers or consumers — can interact
and engage in peer-to-peer technical support and knowledge sharing. Conse-
quently, the majority of member interaction focuses on information exchange
about technical problems. The value of this information exchange is demon-
strated by the hosting firm’s estimate that at least 35% of the problems posted
in this online community are completely solved by other members. In addition,
members also converse about social topics, which range from exchanging
system-administrator jokes to discussing the isolation and frustration that some
experience at their workplace. A unique feature of this community is that it pro-
vides a member-controlled point reward system. Points (ranging from 0 to 10)
can be assigned by the member who posted a question based on the quality of
the answers received. When a member has collected a certain number of points,
s/he receives a ‘hat’ that appears next to her/his username. There are six levels
of hats: ‘Pro’, ‘Graduate’, ‘Wizard’, ‘Royalty’, ‘Pharaoh’, and ‘Olympian’. As
such, these hats are an indication of the quality of the knowledge contributions
that an individual member provides.
As mentioned before, the online community is moderated by employees of
the host firm. These moderators are clearly visible since the company logo is
placed next to their name. They practise a ‘hands off’ approach and usually do
not answer technical questions. Their role is to facilitate member interactions
and, as such, they mainly guide new members and gather feedback from exist-
ing ones. They interfere in the discussions only if members display disruptive
behaviors, as outlined in a community ‘code of conduct’. In addition to the
moderators, also other employees of the host firm participate in the commu-
nity, but they do so on a completely voluntary basis and behave like other
members. While the host firm logo also appears next to their name, they do
collect points and associated hats like all other members. In order to use the
online technical support community, members have to register to the site by
360 Organization Studies 28(03)
at SAGE Publications on May 21, 2009 http://oss.sagepub.comDownloaded from
choosing a user name and password. In this process, an accurate email address
has to be provided, but the hosting firm commits to not use or disclose this
email address without the consent of the member. Hence, real names and
email addresses are not visible online unless the member chooses to disclose
them. While several thousand people have at some point registered for the
community, the host firm estimates that there are roughly 750 active members
who regularly spend several hours per week in the online community.
Questionnaire Development
All latent variables are measured using a multiple-item measurement scale.
These measures use a seven-point Likert type response format, with ‘strongly
disagree’ and ‘strongly agree’ as the anchors. Except for online interaction
propensity, we used scales that have been validated only by previous empirical
research. Items were selected and adapted to the specific characteristics of our
research setting on the basis of interviews with four members and four man-
agers of a different online community (which is part of a professional career site
and focuses on the exchange of advice concerning job search). The resulting
questionnaire was pretested quantitatively on a sample of 85 community
members. As a consequence, several items were reworded or deleted.
In addition, we had to develop a scale for the individual difference variable
‘online interaction propensity’. We followed the procedure proposed by
Churchill (1979) and conducted in total two qualitative and four quantitative
studies (each using a different sample) to develop and validate a scale. In the
two qualitative studies (conducted at two different firms), we extensively
interviewed in total 14 e-business managers and eight online community
members to establish the domain of our construct and generated an initial pool
of 54 items. We then asked five academic experts and three e-business man-
agers to rate how well each item represents online interaction propensity.
Only the 30 most representative items were retained for further quantitative
analysis. In the first quantitative study, using an offline student sample (n =
287), we reduced the scale to eight items by randomly splitting the sample in
half and using exploratory and confirmatory factor analysis. In the following
study, using a second offline student sample (n = 308), we evaluated discrim-
inant validity by assessing the final eight-item scale together with conceptu-
ally close constructs (i.e. extraversion, offline willingness to communicate,
and involvement with online communication) and controlling for social desir-
ability bias. Next, we established nomonological validity in the fifth study (n =
195) by demonstrating the ability of the online interaction propensity scale to
explain self-reported communication behavior in an online technical support
community setting. In the final study, we administered the online interaction
propensity scale to 50 members of an online movie review community and
independently collected the actual number of postings during a four-week
period before the administration of the scale, and found that online interaction
propensity is indeed significantly correlated with actual online interaction
behavior.
Wiertz and Ruyter: Beyond the Call of Duty 361
at SAGE Publications on May 21, 2009 http://oss.sagepub.comDownloaded from
Measures
We measured the study variables through both self-reported survey data and
objective participation data. Sportsmanship is measured with four items adapted
from Podsakoff et al. (1997). Commitment to the community and commitment to
the host firm are both measured with three items each, adapted from Morgan and
Hunt (1994). Informational value is measured by three items based on Okleshen
and Grossbart (1998), and reciprocity by three items based on de Ruyter and
Wetzels (2000) and Wasko and Faraj (2000). Finally, the online interaction propen-
sity scale consists of eight newly developed items. The complete set of items is
provided in the appendix.
The quality of the knowledge contributions provided by a respondent was
assessed by recording the ‘hat’ that s/he had earned (ranging from ‘No hat’ to
‘Olympian’) right after the survey data collection had been completed. Finally, the
quantity of knowledge contribution was measured by collecting the number of
messages that each respondent had posted during a one-month period prior to our
data collection. In line with Wasko and Faraj (2005), we define knowledge con-
tribution as a response to a question. To ensure that only these kinds of messages
were included in our data set, we conducted a content analysis on all messages
and classified them into questions of a social nature, answers of a social nature,
questions of a technical nature, and answers of a technical nature. In total, 3349
posts were analyzed, of which 0.2% were social questions, 6.9% were social
answers, 4.4% were technical questions, and 88.5% were technical answers. Only
the latter category (i.e. 2966 messages) was included in the data analysis. One
author analyzed all messages, while the second author independently coded a sub-
set of 200 messages. There was agreement on all 200 messages.
Data Collection
Owing to the unavailability of personal email of the online community
members, a link to the online questionnaire was posted in a discussion thread in
the forum about ‘general matters’. We introduced ourselves as independent
researchers, explained the purpose of the study, and invited online community
members to participate. In order to stimulate response, we promised to make the
results of the survey available by posting them in the community. Employees of
the host firm were not invited to participate. In total, we received 216 usable
responses. As we do not know how many community members have read the
thread featuring our survey, but decided not to respond, we cannot estimate a
precise response rate. The only possible measure of response rate is the number
of completed surveys per number of unique clicks on the link to the question-
naire (Ridings et al. 2002). The rate of completions per unique visit to the ques-
tionnaire was 86%. In order to avoid double entries, date and time of
completion, as well as the remote user name were captured. In addition, respon-
dents were asked to voluntarily indicate their community user name and email
addresses. In total, 203 respondents were willing to share their user name,
which we then cross-checked with the existing registration profiles. As such, we
could ensure that these 203 respondents are indeed members of the online tech-
nical support community.
362 Organization Studies 28(03)
at SAGE Publications on May 21, 2009 http://oss.sagepub.comDownloaded from
Sample Profile
The 203 respondents we considered in our analysis are predominantly male
(92.5%) and relatively young (71.5% below 45 years), which is not surprising in
the IT industry. They live in 33 different countries, but most respondents come
from the USA (40%) and the UK (10%). The majority (68%) have been members
of this online community for more than one year, and respondents spend on
average 5.5 hours per week in the community. Eighty-nine respondents have
not yet earned a ‘hat’, 50 are ‘Pros’, 23 are ‘Graduates’, 27 are ‘Wizards’, nine
are ‘Royalties’, three are ‘Pharaohs’, and two are ‘Olympians’. They estimate that
67.2% of the problems they post in the community are completely solved, which
is much higher than estimated by the host firm. Finally, we were able to not only
capture very active participants in our sample, but also the so-called ‘lurkers’,
who only read the online community dialogue without contributing. Overall,
16.7% of our respondents self-reported that they have not yet posted anything on
the discussion boards, and our analysis of posting frequency indicates that 34%
did not post anything during the month prior to our data collection. The majority
of respondents (62.1%) posted less than ten technical answers during this month,
whereas 10% contributed more than 40 technical answers. The average number of
technical answer posts was 15.74, with a standard deviation of 28.21.
In order to ensure that we do not have a significant response bias in our sam-
ple, we randomly selected 100 members of the community who did not respond
to our survey and analysed their posting behaviour according to the same pro-
cedure used for the respondents. The average number of technical answer posts
in the same time period from these non-respondents was 16.10, with a standard
deviation of 29.46. Therefore, we can conclude that at least in terms of our focal
variable knowledge contribution, our sample is not significantly different from
the non-respondents.
Results
Validation of Measures
We initially examined the psychometric properties of the administered scales by
conducting exploratory factor analysis. The exploratory factor analysis (using
principal axis factoring with varimax rotation) found an eight-factor solution
that explains 73.1% of the total variance. The eight factors correspond almost
exactly to the eight constructs investigated in our study. After inspection of the
individual item loadings, we deleted three OIP items and one sportsmanship
item with loadings lower than .70 and one OIP item with significant cross-load-
ings from further analysis, as indicated in Table 1.
We employed partial least squares (PLS) path analysis as implemented in PLS-
Graph version 3.0 (Build 1060) to estimate the parameters in the structural and
measurement part of the structural model presented in Figure 1 (Chin, 1998). PLS
has been widely used in marketing (e.g. White et al. 2003), information systems
(e.g. Wasko and Faraj 2005), and strategic management (e.g. Hulland 1999). As
opposed to the covariance-based approach to structural equation modeling (e.g.
Wiertz and Ruyter: Beyond the Call of Duty 363
at SAGE Publications on May 21, 2009 http://oss.sagepub.comDownloaded from
LISREL, EQS, Mplus, etc.), PLS path modeling is component-based and there-
fore requires less stringent assumptions in terms of multivariate normality, mea-
surement levels of the manifest variables, and sample size (Chin 1998; Falk and
Miller 1992; Hulland 1999; Tenenhaus et al. 2005). Furthermore, Chin et al.
(2003) find that PLS path modeling might be superior to moderated regression
analysis and covariance-based methods for testing moderating hypotheses.
In our study, we specify reflective indicators (Chin 1998) for all our constructs
except for quality and quantity of knowledge contribution which are represented
by single indicators. To assess the psychometric properties of the measurement
instrument, we specify a null model with no structural relationships. All remain-
ing items have standardized loadings that exceed the recommended cut-off of .70
(Hulland, 1999). We evaluate reliability by means of composite scale reliability
(CR) and average variance extracted (AVE) (Chin 1998; Fornell and Larcker
1981). For all measures, the CR is well above the cut-off value of .70, and the
AVE exceeds the .50 cut-off value (Fornell and Larcker 1981).
Discriminant validity of the measures can be assessed using multiple meth-
ods. First of all, a construct should share more variance with its measures than
it shares with other constructs in the model (Chin 1998; Howell and Aviolo
364 Organization Studies 28(03)
Table 1. Exploratory Factor Analysis Results
Factor
Item
a
1 2345 678
OIP4 .886 .076 .073 .025 −.053 .045 .001 −.035
OIP 5 .862 .133 .098 .003 −.009 .052 .061 −.094
OIP 6 .840 .151 .090 .038 .024 .077 .038 −.129
OIP 3 .829 .191 .107 −.062 −.002 .037 .108 .147
OIP 2* .798* .068 .063 .006 −.014 .047 .092 .496**
OIP 8** .644** .224 .045 .032 .034 .090 .035 −.028
OIP 1** .622** .183 −.009 −.133 −.023 .018 −.026 .195
OIP 7** .618** .157 .117 .098 .042 −.033 .138 −.150
Quantity .551 .212 .245 −.073 −.012 .166 .400 .026
CCOM2 .333 .885 .133 .043 −.025 .019 .108 .004
CCOM3 .320 .878 .112 .024 −.054 .015 .054 .007
CCOM1 .344 .874 .112 .031 −.036 .015 .062 .006
REC2 .121 .096 .921 .026 .007 .122 −.006 .036
REC 3 .193 .077 .891 .019 .020 .058 .061 .022
REC 1 .101 .143 .846 −.073 −.040 .139 .023 −.060
CHOST2 −.016 .018 .026 .922 −.060 .043 .001 −.012
CHOST 3 −.037 −.048 −.022 .901 −.091 .042 −.059 −.021
CHOST 1 .049 .088 −.023 .753 −.105 −.009 −.016 .018
SP1 −.032 −.074 −.018 −.043 .800 −.004 −.006 .003
SP3 −.028 .011 −.043 .025 .759 .040 .007 .026
SP2 .120 .002 .020 −.091 .713 −.009 −.030 .011
SP4** −.041 −.015 .023 −.096 .487** .050 −.019 −.037
IV2 .168 −.041 .106 .059 −.005 .875 .037 −.038
IV1 .076 .042 .123 .054 −.016 .810 −.016 −.004
IV3 −.016 .028 .032 −.031 .111 .707 .065 .043
Quality .234 .130 .027 −.065 −.062 .069 .708 .004
*Item deleted because of high cross-loading (> 0.40). **Item deleted because of loading < .70.
a
OIP = Online interaction propensity; Quantity = Quantity of knowledge contribution; CCOM = Commitment to the community;
REC = Reciprocity; CHOST = Commitment to the host firm; SP = Sportsmanship; IV = Perceived informational value; Quality =
Quality of knowledge contribution.
at SAGE Publications on May 21, 2009 http://oss.sagepub.comDownloaded from
1993), so the square root of the AVE should exceed the intercorrelations of the
construct with the other constructs in the model (Fornell and Larcker 1981). In
our study, none of the intercorrelations of the constructs exceed the square root
of the AVE of the constructs. Moreover, we inspect the Theta matrix (Θ) and
confirm that no item cross-loads higher on another construct than it does on its
associated construct (Chin 1998) and that the correlations of the residual terms
across blocks do not exceed |.20| (Falk and Miller 1992). Consequently, we con-
clude that all constructs exhibit satisfactory discriminant validity. Table 2a pro-
vides the mean, standard deviation, range, CR, and AVE, while Table 2b
presents the correlations between all latent variables and the interaction terms.
Hypothesis Testing
We use PLS path modeling to estimate both the direct and the interaction effects
in our model (see Figure 1). To test the 18 moderating hypotheses, we resort to
the two-step score construction procedure (Chin et al. 2003). PLS allows for
explicit estimation of latent variable (LV) scores, and after saving the standard-
ized LV scores (cf. Tenenhaus et al. 2005), we calculate the interaction terms
and include them in the model. This method enables us to test for a relatively
large number of interaction effects while simultaneously correcting for mea-
surement error (Chin et al. 2003). To test the effects and statistical significance
of the parameters in the structural model, we use a bootstrapping procedure
with 1000 resamples and construct-level correction (Chin 1998). As suggested
by Chin et al. (2003), we employ a hierarchical approach to test our hypothe-
ses, in which we first estimate a model with the direct effects (Model M1 and
M2) only and then add the interaction effects in model M3. We obtain the esti-
mates that we report next from the final model that includes the interaction
effects. Table 3 summarizes our results.
At a significance level (α) of .05 (one-tailed), our results reveal that reciprocity
does not have a significant effect on either quality or quantity of knowledge con-
tribution. Hence, we do not find support for hypotheses 1a and 1b. As expected, we
do find a positive and significant effect for commitment to the community on both
quality (β=.131) and quantity (β=.141) of knowledge contribution, in support of
Wiertz and Ruyter: Beyond the Call of Duty 365
Construct
a
Mean Std Dev Range CR AVE Sqrt AVE
REC 4.79 1.19 1-7 .97 .90 .95
CCOM 5.03 1.79 2.33-7 .98 .95 .91
CHOST 6.26 .76 4-7 .94 .83 .97
OIP 4.98 1.04 1.71-7 .96 .85 .92
SP 5.99 .73 3.33-7 .88 .71 .84
IV 4.85 1.33 1-7 .91 .76 .87
Quality 2.18 1.39 1-7 na na na
Quantity 15.74 28.21 0-190 na na na
a
REC = Reciprocity; CCOM = Commitment to the community; CHOST = Commitment to the host
firm; OIP = Online interaction propensity; SP = Sportsmanship; IV = Informational value; Quality =
Quality of knowledge contribution measured by the ‘hat’ of the respondent; Quantity = Quantity of
knowledge contribution measured by the number of answers to technical questions per respondent.
Table 2a.
Mean, Standard
Deviation, Range,
Composite Reliability,
and Average Variance
Extracted
at SAGE Publications on May 21, 2009 http://oss.sagepub.comDownloaded from
366 Organization Studies 28(03)
Table 2b. Correlations of Latent Variables and Interaction Terms
Construct
a
1 2 3 4 5 6 7 8 910 1112 1314 15 1617
1 REC 1.00
2 CCOM .29 1.00
3 CHOST −.02 .05 1.00
4 OIP .29 .49 −.01 1.00
5SP −.04 −.09 −.12 .01 1.00
6 IV .21 .07 .05 .15 .03 1.00
7 Quality .12 .27 −.09 .31 −.06 .10 1.00
8 Quantity .41 .46 −.07 .63 −.01 .24 .50 1.00
9 IV*REC .26 .18 −.10 .09 −.13 .01 .19 .24 1.00
10 IV*CCOM .19 −.01 −.04 .07 .01 .11 .17 .18 .18 1.00
11 IV*CHOST .10 .04 .03 .08 −.09 .15 .15 .11 −.01 .20 1.00
12 SP*REC .08 −.02 −.04 −.10 .06 −.14 −.06 −.01 −.01 −.10 .06 1.00
13 SP*CCOM −.02 .06 .02 .02 −.12 .01 .01 .01 −.07 −.02 .02 .33 1.00
14 SP*CHOST .03 −.02 .03 −.09 −.31 −.10 .08 −.04 .04 .02 −.02 .03 .03 1.00
15 OIP*REC .29 .16 .10 .21 −.09 .09 .18 .42 .27 .21 .01 −.08 −.01 −.11 1.00
16 OIP*CCOM .22 .15 −.05 .08 .03 .09 .23 .41 .16 .14 .03 −.02 −.04 −.10 .27 1.00
17 OIP*CHOST −.09 .03 .14 .12 −.07 .07 .07 .14 −.01 −.01 −.11 −.07 −.03 .26 .35 −.11 1.00
a REC = Reciprocity; CCOM = Commitment to the community; CHOST = Commitment to the host firm; OIP = Online interaction propensity; SP =
Sportsmanship; IV = Informational value; Quality = Quality of knowledge contribution measured by the ‘hat’ of the respondent; Quantity = Quantity of knowledge
contribution measured by the number of answers to technical questions per respondent.
at SAGE Publications on May 21, 2009 http://oss.sagepub.comDownloaded from
Wiertz and Ruyter: Beyond the Call of Duty 367
Table 3. Results for Hierarchical Models
Testing Hierarchical Models
R
2
F Incremental R
2
Incremental F GoF
d
Model
b
Quality Quantity Quality Quantity Quality Quantity Quality Quantity
M1 .082 .305 5.955** 29.257** na na na na .414
M2 .130 .491 4.981** 32.155** .048 .186 3.678* 24.361** .505
M3 .216 .619 3.673** 21.662** .086 .128 2.438* 7.466** .586
Parameter Estimates for Model M3
Construct
c
β Bootstrapped SE t-statistic Hypotheses Findings
Quality Quantity Quality Quantity Quality Quantity Quality Quantity Quality Quantity
REC −.100 .069 .078 .064 −1.281 1.076 H1a H1b Not supported Not supported
CCOM .131* .141** .071 .056 1.846 2.503 H2a H2b Supported Supported
CHOST −.091* −.081 .053 .056 −1.717 −1.453 H3a H3b Not supported Not supported
OIP .225** .455** .084 .078 2.655 5.846 na na na na
SP .009 .025 .081 .049 .110 .505 na na na na
IV .051 .119* .069 .052 .733 2.226 na na na na
IV*REC .102 .049 .068 .060 1.483 .802 H4a H4b Not supported Not supported
IV*CCOM .085 .032 .067 .036 1.259 .878 H4c H4d Not supported Not supported
IV*CHOST .110 .029 .067 .055 1.626 .531 H4e H4f Not supported Not supported
SP*REC −.024 .072 .072 .054 −.331 1.332 H5a H5b Not supported Not supported
SP*CCOM .010 -.011 .070 .037 .138 −.298 H5c H5d Not supported Not supported
SP*CHOST* .138* .065 .074 .052 1.847 1.254 H5e H5f Supported Not supported
OIP*REC .064 .202** .097 .077 .662 2.613 H6a H6b Not supported Not supported
OIP*CCOM .155* .245** .073 .046 2.123 5.402 H6c H6d Supported Supported
OIP*CHOST .022 .012 .104 .095 .215 .124 H6e H6f Not supported Not supported
a
*p < .05 (one-tailed test), ** p < .01 (one-tailed test).
b
M1: predictor variables; M2: predictor and moderator variables; M3: predictor variables, moderator variables and
interaction terms.
c
REC = Reciprocity; CCOM = Commitment to the community; CHOST = Commitment to the host firm; OIP = Online interaction propensity; SP =
Sportsmanship; IV = Informational value.
d
Goodness-of-fit measure (Tenenhaus et al. 2005).
at SAGE Publications on May 21, 2009 http://oss.sagepub.comDownloaded from
hypotheses 2a and 2b. Surprisingly, commitment to the host firm has a relatively
weak but significant negative impact on the quality of knowledge contribution
(β=–.091), and no impact on quantity. So while the effect for hypothesis 3a is sig-
nificant, the direction of the effect does not correspond with the hypothesis. In
addition, hypothesis 3b is not supported.
With regard to the moderating hypotheses, we find that hypotheses 4a to 5f
are not supported, with the exception of hypothesis 5e. The relationship
between relational social capital and knowledge contribution does not seem to
be moderated by perceived informational value and sportsmanship, except for
the unexpected negative relationship between commitment to the host firm and
the quality of knowledge contribution, which is attenuated for higher levels of
sportsmanship (β=.138). However, perceived informational value has a direct
effect on the quantity of knowledge contribution (β=.119), which we did not
hypothesize. Furthermore, in support of hypotheses 6c and 6d, OIP strengthens
the relationship between commitment to the community and both quality (β=
.155) and quantity (β=.245) of knowledge contribution. Hypotheses 6a, 6b, 6e,
and 6f are not supported. However, online interaction propensity has strong un-
hypothesized direct effects on quality (β=.225) and quantity (β=.455) of knowl-
edge contribution. Finally, we also find — contrary to our hypothesizing — that
reciprocity acts as a moderator and strengthens the positive direct effect of OIP
on the quantity (β=.202) of knowledge contribution.
The R
2
for in the final model M3 (including both main and interaction
effects) is .216 for quality and .619 for quantity of knowledge contribution.
Using an incremental F test to test M3 versus the main effects model M2 (Chin
et al. 2003; Pedhazur 1997), we find that the ∆R
2
of .086 (F(9,187) = 2.438 [p
=.012]) for quality and ∆R
2
of .128 (F(9,187) = 7.466 [p = .001]) for quantity
of knowledge contribution is statistically significant. We calculate f
2
to assess
the effect size of the interaction terms in the final model (Chin et al. 2003;
Cohen 1988), and the results suggest a medium effect size for quality (f
2
= .099)
and quantity (f
2
= .251). Although PLS path modeling includes no proper single
goodness-of-fit measure, such as the χ
2
statistic and its derived measures for
covariance-based SEM, the R
2
values of the endogenous constructs can be used
to assess model fit (Chin 1998; Tenenhaus et al. 2005). In accordance with the
categorization of R
2
effect sizes by Cohen (1988; small: .02; medium: .13;
large: .26), we conclude that these effect sizes are medium for quality (M3: R
2
= .216) and large for quantity (M3: R
2
= .619) of knowledge contribution.
Finally, Tenenhaus et al. (2005) recently have developed a global fit measure
(GoF=√(average R
2
* average AVE)) for PLS, and the goodness of fit in model
M3 is .586 (M2: GoF = .505; see Table 3). Assuming a large average effect size
for R
2
(.26) and a cut-off value of average AVE of .70, we would obtain a com-
parison value for GoF of .42, which both the GoF for M2 and M3 exceed.
Discussion
The aim of this study was to identify the unique drivers of knowledge contribu-
tion by customers in firm-hosted commercial online communities. To that end,
we extended and empirically tested a model of social capital based on Wasko
368 Organization Studies 28(03)
at SAGE Publications on May 21, 2009 http://oss.sagepub.comDownloaded from
and Faraj (2005). Given our unique research context, we focused our model on
the relationship between the relational dimension of social capital and knowledge
contribution, and then investigated the moderating effects of individual attrib-
utes on that relationship. Our results clearly indicate that it is worthwhile to
consider these interaction effects, as evidenced by the significant improvement
in the R
2
of both quality and quantity of knowledge contribution when the inter-
action terms are added.
Contrary to our expectations, reciprocity did not have a significant effect on
quality or quantity of knowledge contribution. This finding is surprising, given the
reported strength of the reciprocity norm in face-to-face contexts (e.g. Gouldner
1960), and the fact that we did find a significant positive bivariate correlation
between reciprocity and the quantity of knowledge contribution. Since there is no
multicollinearity between the latent variables in our model, we carried out addi-
tional analyses that indicate that online interaction propensity acts as a suppres-
sor. We investigated the suppressor effect by adding independent variables to the
model and investigating whether each addition affected the strength of the path
from reciprocity to the dependent variable. We found that reciprocity does have a
significant impact on the quantity of knowledge contribution until online interac-
tion propensity is added to the model. One potential explanation for this finding
might be that interaction is the currency of reciprocity in online contexts. If a
community member wants to reciprocate help received, s/he has to interact with
others by responding to requests for support. Thus, the online interaction propen-
sity of this member might overpower her/his reciprocity intentions.
Furthermore, contrary to Wasko and Faraj’s (2005) findings, customers who
are committed to the firm-hosted online community contribute knowledge more
frequently and provide more helpful answers. This indicates that even though
members in firm-hosted online communities do not know each other offline, and
the community operates in a commercial context, strong relationships between
individual members and to the collective as a whole develop. As a result, cus-
tomers feel a relational bond with the community that encourages them to assist
fellow customers and to share their knowledge. This is even more the case if the
customer is online interaction prone, emphasizing again that commitment in
online communities is formed over repeated computer-mediated interactions.
Even though we did not hypothesize these relationships, online interaction propen-
sity (OIP) also has very strong direct effects on both quantity and quality of
knowledge contribution. We can, therefore, conclude that because knowledge
contributions in online communities have to be made via online interactions, OIP
is an important individual attribute that should be taken into account in assess-
ments of online community members’ participation behavior.
Another un-hypothesized result that deserves highlighting is the fact that the
direct link between OIP and the quantity of knowledge contribution is strength-
ened by higher levels of reciprocity. Thus, reciprocity does not have a direct but
a moderating effect on knowledge contribution. An individual who would gen-
erally prefer not to engage in online interactions might overcome this inclina-
tion and contribute knowledge to the community out of a strong feeling of
indebtedness and desire to reciprocate.
Contrary to our expectations, commitment to the host firm does not impact
the quantity of knowledge contribution, and negatively affects the quality of
Wiertz and Ruyter: Beyond the Call of Duty 369
at SAGE Publications on May 21, 2009 http://oss.sagepub.comDownloaded from
contributions. Since the firm-hosted online community in fact represents an
important service channel, we had expected that customers may be motivated to
contribute to the success of this channel through their participation because of
the bond they have established with the host firm. However, our results indicate
that the opposite seems to be the case: being committed to the host firm leads
customers to make lower-quality contributions. A possible explanation for this
unexpected finding pertains to the firm-hosted community’s function as a ser-
vice channel. Compared to other service alternatives, such as telephone or on-
site support, the community requires the customer to make more effort and
participate actively in the service provision. A customer who is very committed
to the host firm and its products and services might expect ‘better treatment’
than that, and in fact expects the host firm to provide the service, rather than fel-
low customers. As a result, s/he is not willing to make an effort in the commu-
nity and provide high-quality answers. However, the more tolerant the customer
is of imperfections in the community, the weaker this negative relationship
becomes. We can therefore conclude that if the customer displays high levels of
sportsmanship and does not pay attention to minor problems in the community,
s/he is also more willing to make an effort and contribute helpful answers.
It is also worth noting that sportsmanship does not affect the relationship
between commitment to the online community and knowledge contribution. It
seems clear that customers consider the community to be a social entity of its
own, and therefore do not attribute any problems in the community to that
entity, but rather to the host company. Therefore, their tolerance of imperfec-
tions in the community does not have an impact on their relationship with and
behavior in the community.
Finally, we had also expected moderating effects of perceived informational
value on the relationship between relational social capital and knowledge contri-
bution, based on the argument that the more valuable the information a customer
receives from the community, the stronger will be her/his feeling of indebtedness
and hence obligation to reciprocate this help and assist fellow members. However,
we only found a direct effect of perceived informational value on the quantity of
knowledge contribution. This finding highlights the importance of instrumental
motivations — such as information need — to participate in online communities
(e.g. Wasko and Faraj 2000; Wasko et al. 2004), and mirrors conclusions drawn
from open source communities (e.g. Lakhani and von Hippel 2003).
Our results have several interesting implications for firms that host commercial
online communities for their customers. Most importantly, firms need to under-
stand that their online communities must and will develop a life of their own. The
finding that commitment to the firm has no impact on the quantity of knowledge
contribution indicates that the social group dynamics of the community dominate
any other influences. The host firm needs to respect that and make sure that it
keeps interference to a minimum. Furthermore, it also has to be aware that the
community has its limitations as a service support channel. Customers who are
highly committed to the host firm in fact expect the host firm and not fellow cus-
tomers to provide them with service, resulting in an unwillingness to provide
quality contributions to the community. This finding re-emphasizes the need for
firms to offer a portfolio of service delivery channels that allows customers to
370 Organization Studies 28(03)
at SAGE Publications on May 21, 2009 http://oss.sagepub.comDownloaded from
choose a service option that is best suited to their individual requirements. Finally,
it is also important for firms hosting a commercial online community to realize
the importance of individual traits that affect customer behavior, such as online
interaction propensity. For example, a certain percentage of members, those who
have a low online interaction propensity, may never become active contributing
members. It might be worthwhile to attempt identifying the level of online inter-
action propensity upon registration to the online community in order to better
understand subsequent communication activity. Also, as reciprocity increases the
impact of OIP, an explicit code of conduct emphasizing this social norm may
stimulate members to contribute actively.
Our findings can only be interpreted in the light of certain limitations. While
we have focused on the relational dimension of social capital and three individ-
ual attributes that seem particularly important in the context of firm-hosted
online communities, the structural and cognitive dimensions of social capital and
other individual difference variables are clearly important in studying knowledge
contribution. For example, several researchers have alluded to the role of gener-
alized trust and identification with the group (e.g. Ridings et al. 2002) and exper-
tise and intrinsic enjoyment of the participant on her/his interaction behavior
(Wasko et al. 2004). Even though our study clearly underlines the importance of
studying moderating effects, we did not find a consistent pattern between inde-
pendent and moderator variables. The exact nature of the influence of social cap-
ital dimensions and individual attributes on knowledge contribution should
therefore be investigated in future research. Moreover, we focused our model on
only one type of collective action — knowledge contribution. While this is
arguably one of the most important drivers of the success of a firm-hosted com-
mercial online community, other types of collective action pertaining to socially
oriented goals might also be interesting to investigate. Furthermore, the general-
izability of our results may be limited to firm-hosted commercial online com-
munities whose main purpose is service support. Whereas we do think that the
findings from the technical support community that we investigated will essen-
tially apply to all customer-to-customer problem–solving communities, it might
be interesting to also study brand communities, the other important category of
firm-hosted commercial online communities. In addition, all concepts and rela-
tionships were only measured once, thus essentially from a static perspective.
Therefore, we cannot investigate any feedback loops. For example, we would
expect that knowledge contribution feeds back to relational social capital as well
as informational value but, given the cross-sectional nature of our data, we are
unable to test this. Finally, we should also note that we collected our dependent
variables prior to our independent variables but, given the short time that elapsed
between the two data collections, this should not have impacted our results.
In conclusion, we have investigated the influence of relational social capital
on knowledge contribution of customers in firm-hosted online communities,
and paid particular attention to the potential moderating effects of individual
attributes. The overall picture that emerges from our results is that knowledge
contribution is most strongly influenced by a customer’s online interaction
propensity, commitment to the community, and the informational value s/he
perceives in the community.
Wiertz and Ruyter: Beyond the Call of Duty 371
at SAGE Publications on May 21, 2009 http://oss.sagepub.comDownloaded from
Adler, Paul S., and Seok-Woo Kwon
2002 ‘Social capital: Prospects for a new
concept’. Academy of Management
Review 27/1: 17–40.
Ajzen, Icek, Christine Timko, and John
B. White
1982 ‘Self-monitoring and the
attitude–behavior relation’. Journal of
Personality and Social Psychology
42/3: 426–435.
Algesheimer, Rene, Uptal M. Dholakia, and
Andreas Hermann
2005 ‘The social influence of brand
community: Evidence from European
car clubs’. Journal of Marketing
69/July: 19–34.
Baron, Reuben M., and David A. Kenny
1986 ‘The moderator-mediator variable
distinction in social psychological
research: Conceptual, strategic, and
statistical considerations’. Journal of
Personality and Social Psychology
51/6: 1173–1182.
Bell, Simon J., and Bulent Menguc
2002 ‘The employee–organization
relationship, organizational
citizenship behavior, and superior
service quality’. Journal of Retailing
78/Summer: 131–146.
Bourdieu, Pierre
1986 ‘Forms of capital’ in Handbook of
theory and research for the
sociology of education. J. G.
Richardson (ed.). New York:
Greenwood.
Bressler, Stacey E., and Charles E. Grantham
Sr.
2000 Communities of commerce. San
Francisco: McGraw-Hill.
Bucklin, Randolph E., and Catarina Sismeiro
2003 ‘A model of web site browsing
behavior estimated on clickstream
data’. Journal of Marketing Research
August: 249–267.
Burgoon, Judee K.
1976 ‘The unwillingness-to-communicate
scale: Development and validation’.
Communication Monographs 43:
60–69.
Burnett, Gary
2000 ‘Information exchange in online
communities: A typology’.
Information Research 5/4.
Butler, Brian, Lee Sproull, Sara Kiesler, and
Rober Kraut
Forthcoming ‘Community effort in online
groups: Who does the work and
why?’ in Leadership at a distance. S.
Weisband and
L. Atwater (eds). Mahwah, NJ:
Lawrence Erlbaum.
Chin, Wynne W.
1998 ‘The partial least squares approach to
structural equation modeling’ in
Common problems/proper solutions:
Avoiding error in quantitative
research. G. A. Marcoulides (ed.).
Mahwah, NJ: Lawrence Erlbaum.
Chin, Wynne W., Barbara L. Marcolin, and
Peter R. Newsted,
2003 ‘A partial least squares latent variable
modeling approach for measuring
interaction effects: Results from a
Monte Carlo simulation study and an
electronic-mail emotion/adoption
study’. Information Systems Research
14/2: 189–217.
Churchill, Gilbert A. Jr.
1979 ‘A paradigm for developing better
measures of marketing constructs’.
Journal of Marketing Research
16/February: 64–73.
Cohen, Jacob
1988 Statistical power analysis for the
behavioral sciences. Hillsdale, NJ:
Lawrence Erlbaum.
Coleman, James
1988 ‘Social capital in the creation of
human capital’. American Journal of
Sociology 94: 95–120.
Coleman, James
1990 Foundations of social theory.
Cambridge, MA: Harvard University
Press.
Constant, David, Lee Sproull, and Sara
Kiesler
1996 ‘The kindness of strangers: The
usefulness of electronic weak ties for
372 Organization Studies 28(03)
The authors would like to thank Paul Dholakia, Charla Mathwick, Thorsten Hennig-Thurau,
Benedict Dellaert, and Tor Andreasson for their valuable feedback on earlier versions of this paper.
We are grateful to Sebastian Rubin for his help with the data collection and Martin Wetzels for help-
ing with the analysis. Finally, we would like to thank the editors and three anonymous reviewers for
their constructive criticism and helpful suggestions.
References
Note
at SAGE Publications on May 21, 2009 http://oss.sagepub.comDownloaded from
Wiertz and Ruyter: Beyond the Call of Duty 373
technical advice’. Organization
Science 7/2: 119–135.
Cothrel, Joseph P., and Ruth L. Williams
1999 ‘Online communities: Helping them
form and grow’. Journal of
Knowledge Management 3/1: 54–60.
Dabholkar, Pratibha A., and Richard P.
Bagozzi
2002 ‘An attitudinal model of technology-
based self-service: Moderating effects
of consumer traits and situational
factors’. Journal of the Academy of
Marketing Science 30/3: 184–201.
De Ruyter, Ko C., and Martin G. M. Wetzels
2000 ‘With a little help from my fans –
extending models of pro-social
behaviour to explain supporters’
intentions to buy soccer club shares’.
Journal of Economic Psychology 21:
387–409.
Falk, R. Frank., and Nancy B. Miller
1992 A Primer for Soft Modeling. Akron,
OH: University of Akron Press.
Fischer, Eileen, Julia Bristor, and Brenda
Gainer
1996 ‘Creating or escaping community?
An exploratory study of Internet
consumers’ behaviors’. Advances in
Consumer Research 23: 178–182.
Ford, William
2001 ‘Customer expectations for
interaction with service providers:
Relationship versus encounter
orientation and personalized service
communications’. Journal of Applied
Communication Research 29/1: 1–29.
Fornell, Claes, and David F. Larcker
1981 ‘Evaluating structural equation
models with unobservable variables
and measurement error’. Journal of
Marketing Research 18: 19–50.
Gouldner, Alwin W.
1960 ‘The norm of reciprocity’. American
Sociological Review 25: 165–178.
Granovetter, Mark S.
1973 ‘The strength of weak ties’. American
Journal of Sociology 78/6: 1360–1380.
Gutek, Barbara
1995 The Dynamics of Service Reflections
on the Changing Nature of
Customer/Provider Interactions. San
Francisco: Jossey-Bass.
Hagel, John III, and Arthur G. Armstrong
1997 Net gain: Expanding markets through
online communities. Boston, MA:
Harvard Business School Press.
Hammond, Michael
2000 ‘Communication within on-line
forums: The opportunities, the
constraints and the value of a
communicative approach’. Computers
& Education 35/4: 251–262.
Hoffmann, Donna L., and Thomas P. Novak
1996 ‘Marketing in hypermedia computer-
mediated environments: Conceptual
foundations’. Journal of Marketing
60/July: 50–68.
Howell, Jane M., and Bruce J. Aviolo
1993 ‘Transformational leadership,
transactional leadership, locus of
control, and support for innovation:
Key predictors of consolidated-
business-unit-performance’. Journal
of Applied Psychology 78/6:
891–902.
Hulland, John
1999 ‘Use of partial least squares (PLS) in
strategic management research: A
review of four recent studies’.
Strategic Management Journal 20:
195–204.
James, Lawrence R., and Jeanne M. Brett
1984 ‘Mediators, moderators, and tests for
mediation’. Journal of Applied
Psychology 69/2: 307–321.
Kollock, Peter
1999 ‘The economics of online
cooperation: Gifts and public goods
in cyberspace’ in Communities in
Cyberspace
. M.A. Smith and P.
Kollock (eds). London: Routledge.
Lakhani, Karim R., and Eric von Hippel
2003 ‘How open source software works:
“Free” user-to-user assistance’.
Research Policy 32: 923–943.
Lin, Nan
2001 Social Capital. Cambridge :
Cambridge University Press.
Lin, Nan, Walter M. Ensel, and John C.
Vaughn
1981 ‘Social resources and strength of ties:
Structural factors in occupational
status attainment’. American
Sociological Review 46: 393–405.
Liu, Yuping
2003 ‘Generating value through online
interaction: Individual and situational
differences’. Presented at the 2003
Academy of Marketing Science
Annual Conference: Washington, DC.
Mattila, Anna S.
2004 ‘The impact of service failures on
customer loyalty’. International
at SAGE Publications on May 21, 2009 http://oss.sagepub.comDownloaded from
374 Organization Studies 28(03)
Journal of Service Industry
Management 15/2: 134–149.
McCroskey, James C., and Virginia P.
Richmond
1985 ‘Willingness to communicate and
interpersonal communication’.
Presented at the West Virginia
Symposium on Personality and
Interpersonal Communication:
Morgantown, WV.
Moon, Jae Yun, and Lee Sproull
2000 ‘Essence of distributed work: The
case of Linux Kernel’. First Monday
5/1: available online at http://
firstmonday.org/issues/issue5_7/
moon/index.html.
Moon, Jae Yun, and Lee Sproull
2001 ‘Turning love into money: How some
firms may profit from voluntary
electronic customer communities’.
Forthcoming in Electronic Commerce
Handbook: Issues, Technology and
Society. P. Lowry, J. Cherrington and
R. Watson (eds). Portland: CRC
Press.
Morgan, Robert M., and Shelby D. Hunt
1994 ‘The commitment–trust theory of
relationship marketing’. Journal of
Marketing 58/July: 20–38.
Nahapiet, Janine, and Sumantra Ghoshal
1998 ‘Social capital, intellectual capital,
and the organizational advantage’.
Academy of Management Review
23/2: 242–266.
Okleshen, Cara, and Sanford Grossbart
1998 ‘Usenet groups, online community
and consumer behaviors’. Advances
in Consumer Research 25: 276–282.
Olson, Mancur
1965 The Logic of Collective Action.
Cambridge, MA: Harvard University
Press
Organ, Dennis W.
1988 Organizational Citizenship Behavior:
The Good Soldier Syndrome.
Lexington, MA: Lexington Books.
Ostrom, Elinor
2000 ‘Collective action and the evolution
of social norms’. Journal of
Economic Perspectives 14/3:
137–158.
Paldam, Martin, and G. Tinggaard Svendsen
2001 ‘Missing social capital and the
transition in Eastern Europe’. Journal
for Institutional Innovation,
Development and Transition 5:
21–33.
Pedhazur, Eleazar
1997 Multiple Regression in Behavioral
Research: Explanation and
Prediction. Singapore:
Wadsworth/Thomson Learning.
Podsakoff, Philip M., Michael Ahearne, and
Scott B. MacKenzie
1997 ‘Organizational citizenship
behavior and the quantity and
quality of work group performance’.
Journal of Applied Psychology 82/2:
262–270.
Pura, Minna
2005 ‘Linking perceived value and loyalty
in location-based mobile services’.
Managing Service Quality 15/6:
509–538.
Putnam, Robert D.
1993 Making Democracy Work: Civic
Traditions in Modern Italy.
Princeton, NJ: Princeton University
Press.
Rainie, Lee, and John Horrigan
2005 ‘How the Internet has woven itself
into American life’. The Pew Internet
& American Life Project: available
online at
http://www.pewinternet.org/reports.
Ridings, Catherine M., David Gefen, and
Bay Arinze
2002 ‘Some antecedents and effects of
trust in online communities’. Journal
of Strategic Information Systems 11:
271–295.
Samuelson, Paul A.
1954 ‘The pure theory of public
expenditure’. Review of Economics
and Statistics 36/4: 387–389.
Schwartz, Shalom H.
1977 ‘Normative influences on altruism’.
Advances in Experimental Social
Psychology 10: 222–280.
Tenenhaus, Michel, Vincenzo Esposito Vinzi,
Yves-Marie Chatelin, and Carlo Lauro
2005 ‘PLS path modeling’. Computational
Statistics & Data Analysis 48/1:
159–205.
Tsai, Yi-Ching
2006 ‘Effect of social capital and
absorptive capability on innovation in
internet marketing’. International
at SAGE Publications on May 21, 2009 http://oss.sagepub.comDownloaded from
Wiertz and Ruyter: Beyond the Call of Duty 375
Journal of Management 23/1:
157–166.
Von Hippel, Eric, and Georg von Krogh
2003 ‘Open source software and the
“private-collective” innovation
model: Issues for organization
science’. Organization Science 14/2:
209–223.
Wasko, Molly, and Samer Faraj
2000 ‘It is what one does: Why people
participate and help others in
electronic communities of practice’.
Journal of Strategic Information
Systems 9: 155–173.
Wasko, Molly, and Samer Faraj
2005 ‘Why should I share? Examining
social capital and knowledge
contribution in electronic networks of
practice’. MIS Quarterly 29/1: 35–57.
Wasko, Molly, Samer Faraj, and Robin
Teigland
2004 ‘Collective action and knowledge
contribution in electronic networks of
practice’. Journal of the Association
for Information Systems 5/11–12:
493–513.
Yoon, Mahn Hee, and Jaebeom Suh
2003 ‘Organizational citizenship behaviors
and service quality as external
effectiveness of contact employees’.
Journal of Business Research 56:
597–611.
White, Chris J., P. Rajan Varadarajan, and
Peter A. Dacin
2003 ‘Market situation interpretation and
response: The role of cognitive style,
organizational culture, and
information use’. Journal of
Marketing 67/3: 63–79.
Caroline Wiertz is an Assistant Professor of Marketing at Cass Business School, City
University London. Before joining Cass, she completed her PhD on customer behavior
in commercial virtual communities at the University of Maastricht, the Netherlands. Her
research interests are in the areas of online consumer behavior, customer participation in
service delivery, commercial online communities, and social capital theory.
Address: Cass Business School, 106 Bunhill Row, London ECIY 8T2, UK.
Email: sa545@city.ac.uk
Ko de Ruyter is a Professor of Interactive Marketing and International Service Research
at the University of Maastricht, the Netherlands, and Chairman of the Department of
Marketing. He received his PhD in Management Science from the University of Twente
and has been a visiting professor at Purdue University. He serves on the editorial boards
of various international academic journals, including the Journal of Service Research
and the International Journal of Service Industry Management. His research interests
concern international service management, e-commerce, and customer satisfaction and
dissatisfaction.
Address: Maastricht University, P. O. Box 616, 6200 MD Maastrict, the Netherlands.
Email: K.deRuyter@mw.unimaas.nl
Caroline Wiertz
Ko de Ruyter
at SAGE Publications on May 21, 2009 http://oss.sagepub.comDownloaded from
376 Organization Studies 28(03)
Appendix
Questionnaire Items
Construct Wording of the items
Reciprocity Members should return favors when the X community is in need.
When I receive help, I feel it is only right to give back and
help others.
The principle of give and take is important in the X community.
Sportsmanship I tolerate minor imperfections in the X community.
I overlook the negative details of the X community and focus on
the positive ones instead.
I accept that not every answer to my questions in the X
community is perfect.
I do not spend a lot of time complaining about trivial matters in
the X community.*
Commitment to The relationship I have with the X community is something
community to which I am very committed.
The relationship I have with the X community deserves my effort
to maintain.
The relationship I have with the X community is one I intend to
maintain indefinitely.
Commitment to host firm The relationship I have with the X community host is something
to which I am very committed.
The relationship I have with the X community host deserves my
effort to maintain.
The relationship I have with the X community host is one I intend
to maintain indefinitely.
Online interaction propensity In general, I like to get involved in online discussions.
I am someone who enjoys interacting with like-minded
others online.
I am someone who likes actively participating in online
discussions.
In general, I thoroughly enjoy exchanging ideas with other
people online.
I find the idea of belonging to an online discussion group
pleasant.*
I am someone who likes interacting with like-minded others
online.*
In general, I am someone who enjoys initiating a dialogue
online.*
In general, I am someone who, given the chance, seeks contact
with others online.*
Informational value The information provided by the X community is useful.
The information provided by the X community is valuable.
The X community is a great way to get answers to X-related
questions.
*Item deleted from final analysis
at SAGE Publications on May 21, 2009 http://oss.sagepub.comDownloaded from
