Exploring the Concepts which Constitute the Perception of Trustworthiness in Virtual Sponsored Communities

Authors: Xiaohang Chen Olle Falkenäng Supervisor: Mikael Gidhagen

Program: Ekonomi Kandidatprogram Department of Business Studies Level: Bachelor-level

Semester: VT 2012

6/2/2012

Uppsala University

Foreword

This journey has been very interesting. To gain a deeper understanding on how the interaction in virtual communities works has been greatly satisfying. Even though it has been hard during some periods, the paper has developed to something we are very proud of.

We want to thank our mentor Mikael Gidhagen for his guidance and help in

times of need. We also want to thank the gamers of Blizzard´s community that

made this paper possible. Most important of all, we want to honor the team work

and the friendship that has developed during this journey.

Abstract

The development of virtual sponsored communities has created great possibilities for people and companies to interact on shared platforms. To create interaction in the community, it is essential to create the perception of trustworthiness of the community. The purpose of this study is to examine which concepts of trust-building constitute the perception of trustworthiness of virtual sponsored communities. A theoretical framework based on previous studies on trust-building in virtual communities is used to provide the basis for our empirical studies. The paper examines the virtual sponsored community of an online video game, World of Warcraft, with both quantitative surveys and qualitative netnographical observations to determine the important concepts which affect the perception of trustworthiness of the community from the perspectives of the sponsor and the gamers.

Our findings show that the concepts of quality content, privacy policy, interaction, knowledge growth, satisfaction, embeddedness, shared value and perceived responsiveness constitute the perception of trustworthiness of the virtual sponsored community. The findings also show implications for causal relationships between the concepts affecting the perception of trustworthiness for future studies.

Key Words: Concepts of trust-building, Internet forums, Relationship marketing,

T

Table of Contents

1.0 Introduction ... - 1 -

1.1 Research question ... - 2 -

1.2 Purpose ... - 2 -

1.3 Disposition ... - 3 -

2.0 Theoretical Framework ... - 4 -

2.1 A model of trust-building ... - 4 -

2.1.1 Participant trust ... - 4 -

2.1.2 Institutional trust ... - 4 -

2.1.3 Trusting belief ... - 5 -

2.1.4 Outcomes of trust ... - 5 -

2.1.5 Implications of the trust-building model on our study ... - 5 -

2.2 Concepts affecting trust-building ... - 6 -

2.2.1 Shared value ... - 6 -

2.2.2 Privacy Policy ... - 6 -

2.2.4 Knowledge growth ... - 7 -

2.2.5 Perceived responsiveness ... - 7 -

2.2.6 Quality content ... - 7 -

2.2.3 Satisfaction ... - 7 -

2.2.7 Embeddedness ... - 8 -

2.2.8 Interaction ... - 8 -

2.3 Trust and trustworthiness ... - 8 -

2.4 Operationalization of theories ... - 9 -

3.0 Methodology ... - 11 -

3.1 The subject of this study ... - 11 -

3.2 Disposition ... - 12 -

3.3 The quantitative study ... - 12 -

3.4 The qualitative study... - 13 -

T

4.0 Empirical findings ... - 16 -

4.1 Quantitative study ... - 16 -

4.2 Qualitative study... - 19 -

5.0 Analysis ... - 22 -

5.1 Quantitative study ... - 22 -

5.2 Qualitative study... - 24 -

6.0 Conclusion... - 27 -

7.0 Implications ... - 28 -

8.0 Limitations ... - 30 -

9.0 References ... - 31 -

Appendix I: Table of Wu et al.’s measure variables ... - 33 -

Appendix II: The Questionnaire form ... - 34 -

Appendix III: Results from the questionnaire ... - 36 -

Appendix IV: Correlation Matrixes ... - 42 -

Appendix V: Results for the combined variables for each concept affecting the perception of trustworthiness ... - 44 -

I

- 1 -

1.0 Introduction

The internet has evolved into one of the most important marketing channels for companies; it gives the opportunity for companies to communicate with the customers among other things.

Virtual communities, out of all internet marketing channels, have given the possibilities for companies to interact directly with the customers to best create customer value (Newell 2000:184). A virtual community is a platform for customers to share brand related information and interests; it allows interactions both among customers and between customers and the company, which can benefit the brand by generating loyal behaviors through word-of- mouth, knowledge sharing and commitment (Hur et al. 2011:1206). By exchanging information and values about the brand, the interactions on the virtual community can help the companies to identify customer needs more accurately, and enable the companies to manage long-term relationships (Stokburger-Sauer 2010). However, in order for companies to benefit from information shared by members on the virtual community, members must have trust towards the virtual community and the company sponsoring it. Therefore, long-term customer relationships require the members of a virtual community to have trust (Doney & Cannon 1997). Trust can be seen as the tendency of the customer to believe a brand keeps its promises and is trustworthy (Hur et al. 2011:1198). Trust directly influences the possibility for the members to share information with other members of the virtual community. The direct correlation between trust and information sharing shows that the more trust a member has towards a virtual community the more likely he/she will visit the community and/or share information with other members and the company (Porter & Donthu 2008). Through building trust, members of a community participate more. It is through this participation information can be obtain to build business strategies, which ultimately leads to increased brand loyalty (Hur et al. 2011:1196). But generating strong brand loyalty can only be achieved if the customers see the company and the community as trustworthy (Hur et al. 2011:1198).

Research has been conducted to study how to create trust on a virtual community towards the

company and among customers. Past research focused on determining concepts that affect

trust-building (Hsu et al. 2011) on virtual communities from either the members’ perspective

or that from the sponsor’s (company’s) perspective. For example, Hsu et al. (2011) emphasize

on concepts of shared values, policies and satisfaction to determine important concepts of

building trust among the members in a virtual community, while Porter and Donthu (2008)

emphasize on the importance for sponsors of virtual communities to provide efforts of

content, embeddedness and interaction on the community to generate trust towards the

sponsor.

I

- 2 -

However, there has been less focus on virtual sponsored communities (Porter & Donthu 2008). A virtual sponsored community differs from a virtual non-sponsored community in the sense that a virtual sponsored community is sponsored, monitored and moderated by the brand company, and allows for interaction between the consumers and the company, whereas a virtual non-sponsored community has no official intervention from the company and only allows members to interact with each other. Past studies on trust-building are only based on either the perspective of the company or the members, these studies are only one-sided and cannot offer a more thorough understanding on how trust is developed to generate brand loyalty; additionally, it does not make a clear distinction between sponsored and non- sponsored communities. This can be a reason why many companies have failed to create value after launching their virtual sponsored communities (Porter & Donthu 2008:113).

Furthermore, the concepts of trust proposed from both perspectives are partially different with each other, but yet they all lead to trust-building. This inconsistency in the concepts which gives the same results leads our research to find the actual concepts affecting trust-building regardless of the perspective.

In a virtual sponsored community, members and sponsors interact together and form a single entity – the community; therefore their trust is towards the community. Ennew & Sekhon (2007) argue that trust differs from trustworthiness because trust is attributed to consumers while trustworthiness is the attribute of the medium. The medium in this case is the community. Members and sponsors of a virtual sponsored community are ultimately affected by their perceptions of the trustworthiness of the community as they rely on the community for knowledge and information. The platform of a virtual sponsored community is important for building trust for both actors. Therefore instead of focusing on concepts affecting trust- building in a virtual sponsored community, it is more relevant to focus on concepts affecting the trustworthiness of the community. Consequently, it is relevant to examine the concepts that affect the perception of trustworthiness of virtual sponsored communities from the dual perspective of the members and of the sponsors.

1.1 Research question

Which concepts of trust-building constitute the perception of trustworthiness of virtual sponsored communities?

1.2 Purpose

The purpose of this study is to examine how the perception of trustworthiness of virtual

sponsored communities is constituted. We will do so by examining several virtual community

I

- 3 -

trust theories from both the perspective of the members and the sponsor, and hereafter operationalize them to test them through quantitative surveys and qualitative observations on an online-game virtual sponsored community.

1.3 Disposition

This paper first explores the theoretical backgrounds of concepts affecting trust-building from

both the perspective of the gamers (members) and that of the sponsor of a virtual sponsored

community. After examining the concepts, we will explain the difference between trust and

trustworthiness. Consequently we operationalize the concepts by combining them to

determine how trustworthiness of a virtual sponsored community can be perceived. Then we

will empirically test our operationalization through quantitative surveys and qualitative

observations to determine the concepts of trust-building which constitute the perception of

trustworthiness of virtual sponsored communities. Lastly, this paper will discuss the

similarities and differences in the empirical test to explore theoretical and managerial

implications.

T

- 4 -

2.0 Theoretical Framework

As mentioned earlier, a virtual sponsored community is the combined platform with interactions from both members and sponsors; it is therefore more relevant to examine the concepts which affect trust-building from early literatures regardless of the perspective to suit this study.

To understand how trust affects the members of a virtual community, it is important to first understand how trust is built. In this part of the paper, theories are brought up to first examine a trust-building model of a virtual community. Based on the model, relevant concepts affecting trust-building from the perspectives of both the members and the sponsors of a virtual community are examined. After various concepts from both perspectives are examined and compared, the differences between trust and trustworthiness are explained. Lastly, we introduce our operationalization with an illustration of concepts affecting the perception of trustworthiness of virtual sponsored communities based on previous studies.

2.1 A model of trust-building

Wu et al. (2008) propose that trust in a virtual community should consist of three stages:

building, process and outcome. Based on this, they anticipate their model of trust-building with four constructs of trust: participant trust, institutional trust, trusting belief and outcomes of trust. Participant trust and institutional trust facilitate cooperation and information sharing among parties. Trusting beliefs are dimensions of trust towards individuals. Outcomes of trust are relevant measures to examine the actual effect of trust. The four constructs of trust are explained below (Hsu et al. 2011:588)

2.1.1 Participant trust

“Participant trust” is the trust among members of a virtual community. It consists of three variables – benefit attraction, tendency to trust, and share value. Benefit attraction suggests that trust is more likely to be forged when two parties have overlapping benefits. Tendency to trust suggests that individuals have different tendencies to trust, which directly influences the degree of trust a member has towards other members and the community. Shared value is the concept that trust is more likely to be developed when the parties have a shared sense of value (e.g. behavior, goals, culture, interest, etc.). (Wu et al. 2008)

2.1.2 Institutional trust

“Institutional trust” is the members’ trust towards the platform of the virtual community; this

includes the trust towards the information accuracy, security, fairness, and control of the

T

- 5 -

community. It consists of two variables – monitoring and privacy policy. Monitoring is the management of the community, where the members’ activities are checked and regulated to fit a general purpose. It allows for the smooth operations of the website and promotes responsible behaviors among the members. Privacy policy has an important role on affecting the trust of members; privacy policy allows the members to know the intention of the forum so that they can have protected personal information. Privacy policies strengthen the trust members have on the virtual community. (Wu et al. 2008)

2.1.3 Trusting belief

“Trusting belief” refers to “individuals believing other individuals have characteristics from which they can benefit from.” (Wu et al. 2008:118). There are four main trust beliefs which constitute the trust members have towards the virtual community, and they are (Wu et al.

2008:118):

 Competence – Believe that others have the capability to help them.

 Benevolence – Believe that others care about their needs; expresses kindness and

altruism which reduce uncertainty and opportunism.

 Integrity – Believe that others are truthful and follow certain set principles and rules;

reduces uncertainty.

 Predictability – Believe that actions of others are within expectations; reduces

uncertainty and risk.

2.1.4 Outcomes of trust

Members have different trust beliefs. These trusting beliefs ultimately lead to two outcomes:

stickiness and sharing information. Stickiness refers to the ability for the virtual community to keep the members. Sharing information is the act for members to be engaged and actively participate in sharing information to the community. To be noted is the fact that the first three constructs of trust generally leads to this construct. But yet the effects and correlations have varied based on different findings (McKnight et al. 2002; Wu et al. 2008). This could be due to the fact that the researches were conducted on virtual communities of product of different categories. Nevertheless it could be applied in this study as a basis for our study of perception of trustworthiness.

2.1.5 Implications of the trust-building model on our study

To summarize, Wu et al. (2008) proposed a model of trust-building of virtual communities

consisting of four constructs – participant trust, institutional trust, trusting belief and

outcomes of trust (See Appendix I). Each one of the constructs provide basis for the concepts

T

- 6 -

which affects trust-building. The constructs in Wu et al.’s (2008) model are closely related to the concepts, which therefore can be used to analyze the effect of the concepts in relation to trust-building. This leads to the next part of this theoretical framework, namely concepts affecting trust-building.

2.2 Concepts affecting trust-building

Trust can be created not only between individuals of a community, but also towards the sponsor of the virtual community. It is therefore necessary both to study concepts which affect trust-building among members and to study how sponsors can create value from virtual communities via cultivating trust. In this part, the concepts which affects trust-building of a virtual community from previous studies are examined. Each concept is explained in relation to the trust-building model in 2.1 to provide a deeper understanding for their effects on the dimensions of trust, which later will be used as a part of the analysis of our empirical study.

2.2.1 Shared value

One of the concepts which affect trust-building for the members of a virtual community is shared value. Shared value refers to common social principles, norms and goals guiding the behavior pattern of members of a virtual community. Shared values connect diverse individuals to a common logic system and enable group members to interpret the communication in a similar manner” (Wu et al. 2010:1027). According to Wu et al. (2010), shared values increase competence, benevolence and predictability of a virtual community in the “trusting belief” construct of trust-building model. Similar concept is identified by Hsu et al. (2011), in which the concept of shared vision is explained. It refers to the perception of values, goals and vision about knowledge sharing a member has with other members of the community. Shared value and shared vision are related as they both emphasis on the sharing of similar qualities among members, therefore they are categorized under shared value in this study.

2.2.2 Privacy Policy

Privacy policy refers to the ability for individuals to control, manage and selectively reveal

personal information, and is one of the concepts affecting trust-building for members of

virtual communities (Wu et al. 2010). Privacy policy of a virtual community increases the

trusting belief of integrity and predictability. It is also related to institutional trust, in which it

can affect the trust towards the platform of the virtual community, and in turn affect the

perception of its trustworthiness.

T

- 7 - 2.2.4 Knowledge growth

Knowledge growth refers to the degree members believe their participation on a community will lead to increased expertise, and it increases competence, benevolence, and predictability.

(Hsu et al., 2011:592).

2.2.5 Perceived responsiveness

Hsu et al. (2011) also identifies perceived responsiveness as a concept which affects trust- building for members in a virtual community. Perceived responsiveness refers to the degree of likelihood for response a member perceives he/she will receive from other members of the community, and it increases competence, benevolence, integrity and predictability (Hsu et al.

2011:592).

2.2.6 Quality content

Porter & Donthu (2008) conceptualized the concept of quality content by referring it to credibility, accessibility, relevancy and importance of the information provided by the sponsor of the community. By providing access to quality content the members will respond favorably and gain positive attitudes towards the sponsor due to the signs of shared values and respect to the members. The access of new information will also motivate continuing participation, as well as viewing the sponsor´s behavior as less opportunistic and more trustworthy (Ridings et al. 2002). Hsu et al. (2011)’s study on trust-building among members of virtual communities identify a similar concept, namely knowledge quality. Knowledge quality is a member’s perception of the quality, relevance, comprehensibility, completeness, applicability and timeliness of the knowledge shared on a community.

Despite the fact that quality content and knowledge quality derive from theories of trust- building from two different perspectives, they are similar in definition, and therefore they are combined into one single concept in this study, namely quality content. Quality content increases competence, benevolence, integrity and predictability (Porter & Donthu 2008).

2.2.3 Satisfaction

Satisfaction refers to the positive discrepancy between individual expectations and actual experience with the community or services obtained; it is not a one-time experience but rather an accumulated outcome of past experiences. Satisfaction increases competence, benevolence, integrity and predictability of a virtual community. (Wu et al. 2010)

Satisfaction and quality content can be similar based on their definitions if the members

emphasize on the quality of information shared on the platform of the community to reach

T

- 8 -

satisfaction; but since there are other aspects than just knowledge quality for members to be satisfied (e.g. website layout), they are considered as two different concepts in this study.

2.2.7 Embeddedness

The sponsor can create trust by emphasizing on embeddedness within the virtual community network. Embeddedness refers to the members´ feeling of being organizational insiders, which is the members´ belief on feeling part of the sponsor´s organization (Balasubramanian

& Mahajan 2001, Bhattacharaya & Sen 2003). Porter & Donthu (2008) propose that by managing the policies of the virtual sponsored community, the sponsors can foster embeddedness by giving the members the possibility to contact the representatives of the sponsoring company (Porter & Donthu 2008:115). This would allow the members to influence the policies within the virtual sponsored community and provide the members with legitimate roles (McAlexander et al 2002, Preece 2001). Porter and Donthu (2008) explain that these efforts signal respect for the members and align the sponsor´s values with the members´. This view of creating stronger embeddedness within the virtual sponsored community will also signal that the sponsor is not acting opportunistically (Porter & Donthu 2008:117). Embeddedness increases competence, benevolence, integrity and predictability (Porter & Donthu 2008:117).

2.2.8 Interaction

According to Porter & Donthu (2008:116), facilitating interaction is essential for the sponsor to creating trust within the virtual sponsored community. By managing the members through interaction, the sponsor can manage and affect the participation of the members by moderating the discussions. The sponsor´s efforts to facilitate interaction demonstrates that the company act less opportunistic and signal respect for the members and respect for their needs (Porter & Donthu 2008:117). However, interaction among members of a virtual sponsored community is also an important aspect of trust-building. Interaction is essential in a virtual community, and “trusting beliefs” all emphasize on the trust for other actors through interaction (Wu et al. 2008). Therefore interaction is considered a relevant concept, and it increases competence, benevolence, integrity and predictability (Porter & Donthu 2008:117).

2.3 Trust and trustworthiness

Trust differs from trustworthiness. In the case of virtual sponsored communities, trust refers to

the willingness to depend and is attributed to members who depend on the community for

knowledge, while trustworthiness refers to the extent trust is perceived by members on the

T

- 9 -

community (Gidhagen & Persson, 2011 (a); Ennew & Sekhon, 2007). Concepts affecting trust-building mentioned earlier could affect trustworthiness of a virtual sponsored community. Therefore trustworthiness of a virtual sponsored community can be seen from both the perspective of the members and the perspective of the company – concepts of trust- building affect members´ perception of the trustworthiness of the virtual community, and companies can emphasis on the concepts of trust-building to affect members’ perceptions of the trustworthiness of the community. Consequently concepts affecting trust-building also affect the perception of trustworthiness of virtual sponsored communities. Additionally, trustworthiness is an attribute of the platform for the community to interact, and therefore it differs from trust from one party towards the other.

2.4 Operationalization of theories

Previously we identify different concepts of trust-building from both the members’

perspective and the sponsors’ perspective. We find eight concepts from earlier literatures that could affect the perception of trustworthiness of a virtual sponsored community, and they are:

shared value, privacy policy, knowledge growth, perceived responsiveness, quality

content, satisfaction, embeddedness, and interaction. Despite these concepts derive from

studies on different perspectives, they are applicable in this study because of the purpose of

this study, namely trustworthiness. Since our study is focused on trustworthiness of a virtual

sponsored community, which is the combination of the members and the sponsor, it is

therefore relevant to examine the trustworthiness from the different concepts rather than from

different perspectives. To simplify our theoretical framework, we combined the concepts

affecting the perception of trustworthiness into an illustration to create a framework for our

empirical study (See Illustration 2.1).

T

- 10 -

Illustration 2.1: Our illustration of the perception of trustworthiness of a virtual sponsored community.

Illustration 2.1 shows the eight concepts affecting the perception of trustworthiness – privacy policy, satisfaction, shared values, perceived responsiveness, knowledge growth, quality content, embeddedness and interaction. The first five concepts are derived from the users’

perspective, while the latter two are from the sponsor’s perspective. The remaining concept,

quality content, is derived from both perspectives as mentioned above.

M

- 11 -

3.0 Methodology

In this section of the paper, we will explain the methodology chosen for answering our research question. We will account for reasons and motivations for the chosen methods and how they have been implemented.

3.1 The subject of this study

We focused on the virtual sponsored community of an online computer game. The computer game industry is suitable in this case because it is advanced in consumer relationship marketing, both in the production of the products and on managing virtual communities. Due to the homogenous products within the industry, emphasizing on virtual sponsored communities is necessary to be competitive. One genre, the massive multiplayer online role playing game (MMORPG), focus greatly on virtual sponsored communities. The consumers are highly involved both with the actual game content and with the discussions on the virtual communities; hence the virtual community is a big part of the total gaming experience (Gidhagen et al. 2011 (b); Dupuis, & Ramsey 2011:2486). The virtual sponsored community is there to offer help and create discussions concerning the game. The virtual community is an important competitive advantage for the developers of MMORPG, and an important tool to create greater consumer values and brand loyalty. Therefore it is important that the virtual community must be viewed as trustworthy to be successful by both the members and the sponsors, or in other words, the gamers and the moderators (Porter & Donthu 2008).

We focused on the virtual sponsored community of one MMORPG game, World of Warcraft (WoW), which is created by Blizzard Entertainment. WoW is one of the biggest MMORPG game in the world with over 12 million gamers (Leffler 2011). Due to its popularity, WoW has many fan sites and discussion forums. It also has its own official virtual sponsored community which we have focused on in this study. WoW´s virtual sponsored community is interesting to study because it has existed for many years with many gamers interacting in the community so it can be expected that both the sponsor and the gamers have knowledge about what constitutes trustworthiness in virtual sponsored communities. Because it is a virtual sponsored community, there is also interaction from the sponsor which makes it possible to study the trustworthiness of the virtual sponsored community from both perspectives. Also the size of the community can compensate the need of observing several communities.

Without being a gamer of the game, one has no access to joining the discussions in WoW´s

virtual sponsored community. This helps to limit the discussions to relevant members. The

M

- 12 -

community is divided into many sub-forums such as “gameplay”, “support”, “development”

etc. Representatives from Blizzard join discussions with gamers about the gameplay, but mostly use the community to both help and provide information. The identification of Blizzard representatives and normal gamers is rather simple – representatives distinct themselves by having blue names and text, and have nameplates for their positions, while gamers have standardized colors for their names and texts.

3.2 Disposition

To able to examine our purpose we conducted both a qualitative and a quantitative study. The reason for this was that we wanted to understand which concepts of trust-building constitute the perception of trustworthiness on a virtual sponsored community from the perspective of the members and the sponsors. The quantitative approach was based on a questionnaire aimed for the gamers of the community. The qualitative analysis was based on a netnographical observation of the sponsor´s (representatives from Blizzard Entertainment) participation on the virtual community (Saunders et al. 2009:297). By using multiple methods for examining our research question, we could enhance the credibility and reliability by canceling out the

"method effect" of using only one type of procedure (Saunders et al. 2009: 154).

3.3 The quantitative study

The questionnaire consisted of 24 questions and was answered by 231 gamers (out of 289 views). A questionnaire was beneficial because we could reach a big amount of respondents which we saw as a necessity to represent the population of gamers in a credible way, and it was time- and cost-efficient (Bryman & Bell 2007:241). The questionnaire was in the form of self-completion questionnaire (See Appendix II), which was announced in a general section of the WoW´s virtual sponsored community to avoid the risk of being removed by Blizzard due to creation of an unrelated subject in the sub-forums. Because the questionnaire would reach people that were familiar and active in the community, we expected the respondents to have a good perception of what was important to constitute trustworthiness in communities.

By announcing the questionnaire in the same platform as the qualitative method the both

research methods would be able to complement each other and make it easier to compare

them to enhance the quality of the paper. We chose to have a short questionnaire because this

tends to achieve better response rate than longer ones and make the respondents more

interested to answer thoroughly (Bryman & Bell 2007:261). The majority of the questions

M

- 13 -

consisted of closed questions with a 7-Likertscale to enhance the comparability between the answers, and made it easier to analyze the opinions and attitudes of the gamers. In order to make it possible to examine which concepts that affect the perception of trustworthiness, the different questions were divided to represent each concept of trust-building. Each concept had between 1-3 questions depending on if we believed a certain concept needed several dimensions to make sure the questionnaire would have high validity. To evaluate if a certain concept needed several dimensions of questions, we used methodology in previous studies that have been done about virtual communities and trustworthiness. By adapting the questionnaire from studies with similar subject and respondents which already have been empirically tested and validated, we could more efficient maintain high reliability and validity (Saunders et al. 2009: 382). To make sure the different dimensions to each concept had high level of statistical significance, we conducted correlation matrix on the dimensions to determine if there were linear relationships between them (Newbold et al 2007).

To achieve high reliability and validity we used clear wording with terms that probably would be familiar and easy understood by the respondents (Saunders et al. 2009:374). We chose to begin the questionnaire with easy questions about the respondents so the respondents got the chance to get an open mind set before the questions based on their opinions and values. At the end we used open questions to give the respondents the chance to reflect about the subject to obtain interesting replies and thoughts that otherwise would not have been covered (Bryman

& Bell, 2007:261). The results from our questionnaire could then be analyzed to see the importance of each concept on the perception of trustworthiness from the gamers’

perspective.

3.4 The qualitative study

The qualitative study was made by a netnographical observation of the interaction of Blizzard

Entertainment´s representatives on the virtual community. Netnography is a research method

that is adapted to study cultures and communities that emerges through computer-mediated

communication (Kozinets 2002). The data was collected by observation of the interaction of

the different parties on the community (Kozinets 2002:5). Netnography was beneficial for our

case because it was less time consuming than traditional ethnography due to its downloadable

content and was totally unobtrusive because it was not fabricated or impacted by the

researcher (Kozinets 2002:2-3). It also gave us the chance to focus on a small part of the

M

- 14 -

virtual community which could limit the observation time and the threats that can affect reliability.

The observation was conducted by measuring how the representatives of Blizzard interacted in the virtual community by looking at what they wrote and how they responded to the interaction of the gamers. The data from the observation was then used to evaluate on which concepts affecting the perception of trustworthiness that the representatives of Blizzard might emphasize on in the Analysis section. The evaluation of concepts of trustworthiness was based on the theoretical framework which had been developed in the theory (Saunder et al 296). It was important to take into account the context and events of the interaction between all parties on the threads and not only determining the individual activity of the sponsor to attain a good understanding (Bryman & Bell 2007:458; Kozinets 2002:7). The number of posts from the representatives of Blizzard was also counted in relation to the posts of the gamers to determine the level of interaction from the sponsor.

We identified two types of representatives of Blizzard – moderators and Most Valuable Posters (MVPs). Moderators were identified by their blue text colors which are hired by Blizzard to manage the community to provide answers and create discussions. MVPs were gamers on the other hand. MVPs had green text color and were chosen by Blizzard if gamers

“post constructively, answer questions about the game, and encourage discussion throughout the forums” (Forum Guidelines, Blizzard, 2012). MVPs were created to “free up Blizzard representatives” (Forum Guidelines, Blizzard, 2012). Even though MVPs did not have the full extensive power to moderate as official Blizzard moderators did, they essentially have the same function as moderators, and therefore viewed as the same as Blizzard moderators in this study.

Our qualitative observation was divided into two parts. We first observed how the forums generally work and then limited ourselves to five different “cases”, or “threads”

: two in

“Gameplay – General” sub-forum, and three in “Mist of Pandaria Beta” sub-forum. By having used five threads from two sections of the virtual community, we could generate a wide spectrum of data without having to observe too many threads that would have been time consuming. Also by having used two sections, the behavior from the sponsor within every thread would probably be consistent within each section due to similarities in subjects. The

1 A “thread” is a common name referring to the concentrated line of posts about a specific subject where different parties can participate.

M

- 15 -

choice of threads was made by selecting threads which contained at least four posts from the sponsor with the latest activity in the month of April 2012. Kozinets (2002:6) explains that on the community there are messages which are off-topic or on-topic from the view of the research question. To only observe the posts that were interesting for the research question, we did an active evaluation to make sure off-topic messages were excluded from the analysis.

We recorded continuously field notes regarding conditions, subtexts and our own personal emotions during the observation to develop good insights for the analysis (Kozinets 2002:6).

Due to the research method there was risk of observer bias, but Saunders et al (2009:297)

explains that observer bias is something that cannot be avoided. But because we could

observe the behaviors in a natural context we could generate findings with great quality. To

avoid observer bias and observer error because of different interpretations, we developed a

clear theoretical framework which we used during the observation. To maintain high

reliability it was crucial to learn as much as possible about the virtual community, the groups

and the participants (Kozinets 2002:4). Therefore we thoroughly investigated the virtual

community before conducting the observation to have a good understanding of how the

virtual community works. Furthermore, we were aware of the fact that we already had a pre-

defined mind set because we were conducting an observation to find concepts that have

effects on trustworthiness. But as a part of netnography, one must take into consideration of

the situation in reality and then connect to the theories; after the connection, it was then

possible to determine how and which variables should be observed. Since the variables in this

case were hard to measure, netnography provided the ability for us to determine the variables

from a general perspective (Kozinets 2002).

E

- 16 -

4.0 Empirical findings

In this section of the paper, the empirical findings will be presented based on our quantitative and qualitative studies. We will present the findings in relation to the concepts which affect the perception of trustworthiness in our model to provide basis for analysis later on.

4.1 Quantitative study

The results from the questionnaire are categorized into different concepts which affect the perception of trustworthiness. We conducted a correlation matrix on the results to determine the significance between the relationships of the questions within each concept category (See Appendix III). The correlation matrix shows strong positive linear relationships between the different questions for all of the concepts, with the correlation at 0.000 level (2-tailed) for all of them. This give us the possibility to combine the questions (variables) within every concept together by calculate the average on each respondent´s answers. It is these combined results we use as the foundation in the quantitative findings (See Appendix V). We have translated the 7-Likertscale into different categories to provide better understanding of the results. Aside from the original scales of "not important at all" to "extremely important", we also divide the scale between the categories; the interval of 1-2 is “not important”, of 3-5 is “neutral”, and of 6-7 is “very important”.

Out of the 231 respondents, 65 % of the group is between the ages of 13 and 25 with the most respondents between 19 and 25 years old. Majority of the respondents spend at least one hour every week on the virtual sponsored community of WoW, with the biggest group spending more than four hours. The respondents choose to go to “official forums” (30 %) and “fan- sites” (31%) when they need help with issues regarding WoW, but some choose to relay on

“in-game friends” among other choices as well.

Results for the combined variables for each concept

The results from combining the different variables show that a majority of the respondents (54

%) has a relative "neutral" opinion regarding shared values, but with tendencies towards seeing it as "important"; for example 23 % of the respondents selected 5 on the 7-Likertscale.

Some respondents (16 %) consider it as "not important", while some others (8 %) view it as

"very important".

E

- 17 - 2. Privacy Policy

A majority of 76 % of the respondent considers it "very important" to have their personal information protected on the virtual sponsored community.

3. Perceived responsiveness

The combined results show very mixed opinions regarding how important perceived responsiveness is based on the different variables. Approximately 21 % considers it "very important", but only 3 % perceives it as "extremely important". However 7 % considers it as

"not important at all". In general the spread is rather equal through the 7-Likertscale, with a tendency to have "neutral" opinions (44 %).

4. Satisfaction

The findings from the combined results show that a big part of the respondents considers the concept of satisfaction as ”very important" (27 %), but over 10 % still believes it is "not important at all". The majority of the respondents (37%) have a rather "neutral" opinion regarding the importance of the concept of satisfaction of their experience with the virtual sponsored community.

5. Knowledge Growth

The empirical findings show that most of the respondents (58 %) have "neutral" opinions regarding the importance of knowledge growth with a tendency towards seeing it as

"important". For example, a large portion of the respondents (27 %) selects 5 on the 7- Likertscale, which is the higher end of the “neutral” category. Approximately 30 % of the group sees it as "very important", but 9 % perceives it as "not important at all".

6. Embeddedness

The findings from combining the different variables show that most of the respondents (45 %) have a "neutral" opinion. But a large portion of the group (27%) considers this as "not important", with only 14 % seeing it as "very important".

Even though the correlation matrix shows strong positive linear relationships between the

questions of this concept, we choose to separate the findings between the questions because

they are largely different – the results show that it is not important for the gamers to feel as a

part of the virtual community (question 14), but yet it is important for them to get direct

contact with the company representatives and other gamers (question 15).

E

- 18 - 7. Interaction

The combined results show that a majority of the respondents believe that the interaction in the virtual sponsored community is important. Approximately 30 % see it as "very important", with a large share of the group with "neutral" opinions (43%) with a tendency towards seeing it as important. Yet circa 10% of the respondents consider it as "not important at all".

8. Quality content

Over 52% of the respondents believe the concept of quality content is "very important", with approximately 32 % considering it as "extremely important". In general most of the respondents view this quality content as something important to them.

9. The respondents´ opinions about World of Warcraft´s virtual sponsored community

When the respondents are asked how important it is to trust the people on the virtual community, this is generally considered important. A big group sees this as "very important"

(38%). However over half of the respondents (54%) have a "neutral" opinion when asked how

much they trust World of Warcraft´s official virtual sponsored community. Many respondents

explain that they trust what the representatives of Blizzard are saying in general. However, the

lack of trust towards the virtual sponsored community is because of some gamers who try to

ruin the discussions by making jokes and writing comments that are out of context. Many

respondents express opinion for the representatives of Blizzard to make more effort to remove

these gamers. To ensure the information is accurate, some respondents express their need of

confirmation from the majority or the company representatives. The lack of interaction from

the representatives from Blizzard is also a reason for many to not trust the virtual sponsored

community thoroughly.

E

- 19 -

4.2 Qualitative study

Findings from our qualitative observations are both general and specific. The general observation findings give an overview of WoW´s community and how Blizzard runs it, while our specific observation findings follow the interactions in five specific threads.

Generally, the community is divided in many sub-forums to give the gamers the possibility to interact with various gamers about common interests. The community also has specific pages that are directly aimed at providing information to the gamers about their privacy policies on the forums (Privacy Policy, Blizzard, 2010).

On the specific threads, we will first present the context for each thread, followed by the number of posts the representatives of Blizzard posted in the threads.

1. Policy Spotlight – Account Sharing from “Gameplay – General” forum

This thread is created by one of the moderators of Blizzard (Policy Spotlight – Account

awareness about policies. In this case it is for reminding about the rules of sharing your own game account of WoW with others. The moderator explains why it is not allowed to share your account and what the consequences can be. The moderator is primarily referring back to the Terms of Use which every member should have read. The moderator also explains that

“Policy Spotlight” is partly for feedback from the gamers, and he encourages the gamers to share questions, thoughts and feelings about the policies. Many gamers both criticize the policies and show their understanding regarding Blizzard´s reasoning. Some of the gamers explain that they actually share their accounts and wonder how to handle it. It is primarily the gamers that answer these questions by referring back to the Terms of Use, but also the moderators are interacting in the discussions. In two cases the moderators answers directly to the gamers’ individual questions. One such example is that one gamer ask the question regarding if a friend can use the account; the moderator says that he understands the member´s situation, and suggests solutions to both the member and the friend. The moderator is very active on the thread. The moderator post 10 replies out of 41 posts in this thread;

which is a participation rate of 24.4 %.

2. Community News, Week of April 2 from “Gameplay – General” forum

The thread is created by one of the moderators of Blizzard (Community News. Week of April

E

- 20 -

websites regarding the World of Warcraft game and its upcoming expansion

. Each piece of news is aimed for a specific day for the upcoming week. The moderator starts the post by saying "be a champion of the World of Warcraft community" which followed by encouraging the gamers to post suggestions for interesting and cool links connected to WoW which might be a part of the next publication of "Community News". During the thread the gamers post suggestions on links connected to WoW, but mostly ask questions or point out concerns to the moderator about the upcoming expansion. Beside the first post from the moderator, the moderator post once a day during the week, this is copied text from the first post about news for each specific day. The moderator did not answer any questions or participated in the discussion. The moderator has 6 replies out of 63 posts in this thread; which makes the moderator´s participation to be 9.5 %.

3. Scarlet Monastery and Scholomance from “Mist of Pandaria Beta” forum.

A gamer begins the thread with complaints and suggestions about the upcoming game patch

, and is joined by other gamers with similar thoughts (Scarlet Monastery and Scholomance, Blizzard 2012). After a long discussion among the gamers, moderators join in and provide very short answers to some specific questions to give some insights in the discussion. Gamers are not satisfied with the answers and continue to discuss. Eventually moderators join in again to provide further thoughts from Blizzard’s perspective. Some comments from the moderators are found unsatisfactory by gamers, one example is this reply from a moderator “Aww, you didn’t like the updated Deadmines?”, which many gamers think is insulting because the response (especially “Aww”-part) does not answer their questions but instead treats the gamers like children. Outcries from gamers continue, but moderators do not participate further in the thread. In total, moderators and MVPs commented 4 times out of 99 replies in this thread, which makes the participation rate to be 4.0%

4. Feedback: Red Panda Male Tail from “Mist of Pandaria Beta” forum.

Gamers offer their thoughts and feedbacks on the upcoming patch after testing out the beta (Feedback: Red Panda Male Tail, Blizzard 2012). Other gamers join in and discuss the matter, and they explicitly expect moderators to interact and provide insights for the discussion. Moderators do so and provide information to answer some of the gamers’

2 “Expansion” refers to packaged additional content for the original game

3 “Patch” refers to a piece of software designed to fix problems with the game

E

- 21 -

questions. Gamers think that some answers diminish discussions and remove the meaning of feedback because Blizzard already made up their mind on certain issues. For example

“Essentially because we like it this way” is the answer the moderators provides when gamers ask about the reason behind one of Blizzard’s actions. Moderators and MVPs commented 4 times out of 164 total replies in this thread, which makes the participation rate to be 2.5%.

5. European MoP beta forums? from “Mist of Pandaria Beta” forum.

A gamer posts a suggestion for improvement regarding the forum in this thread (European

suggestions. Moderators also join in to answer and discuss the suggestions, including directed

answers to some individual questions. The answers are further discussed by gamers. In the

end, the discussions lead to a few specific questions in which the moderators then provide

answers for. The moderators and MVPs are very active in this thread in comparison to the

other threads, with 12 replies out of 91 total replies in this thread; which is a participation rate

of 13.2%.

A

- 22 -

5.0 Analysis

In this part of the paper, we will analyze our empirical findings with our theoretical framework. First we analyze our findings from the quantitative study to find the concept which affect the perception of trustworthiness which the gamers of WoW virtual community view as important. Then we analyze our findings from the qualitative study to find the concepts of trust-building WoW moderators are affecting the perception of trustworthiness of the virtual community. We use the deductive method to analyze our findings because the findings are based on the operationalization of the theories.

5.1 Quantitative study

The findings from the quantitative study show some interesting results regarding how the gamers of World of Warcraft´s official forum view the importance of different concepts which affects the perception of trustworthiness of a virtual sponsored community.

The gamers emphasize on having a virtual sponsored community that have good privacy policies which protects their personal information. The concept of “privacy policy” is a part of institutional trust, which is a construct of trust-building; therefore we can see that privacy policy is necessary in the constitution of trustworthiness in virtual sponsored communities.

Both the opinions of shared values within the community and to emphasize on embeddedness are considered rather neutral from the empirical findings. Trustworthiness can be enhanced by sharing common principles, norms and goals with other gamers. Even though there are tendencies towards seeing shared values as important, it does not seem to be crucial.

The findings regarding embeddedness follow the same trend. The respondents´ opinions

regarding the importance of embeddedness are mixed from not seeing it as important to be

relative neutral. Embeddedness and shared values are related because of the feeling of being a

part of something, also creates common norms and values which we saw earlier were not

crucial for the gamers. But it is important to stress that the need to create shared values with

others is not dependent on feeling as a part of the virtual sponsored community. However, it is

considered important within the variables of embeddedness to have direct contact with other

gamers and representatives of the company. As the theories explain that by participating and

by being present in the virtual sponsored community, the sponsor can generate respect from

the gamers and seem lees opportunistic, hence create a more trustworthy virtual sponsored

community.

A

- 23 -

Interaction is considered important to for members in a virtual sponsored community. Both the interaction from the gamers and the sponsor is necessary to create a dynamic platform where discussions can take place. As the theory explains, if the sponsor affects the interaction in the community, it creates the belief that other members cares about one’s needs and express kindness. It also reduces the uncertainty of the virtual sponsored community which affects the perception of the community being trustworthy.

The concepts of quality content and knowledge growth are important in affecting the perception of trustworthiness of a virtual sponsored community. By providing access to quality content, the gamers will gain positive attitudes towards the virtual community and towards the sponsors due to their signs of respect for the gamers. Because the gamers continuously participate in the virtual sponsored community to both attain better knowledge about the game and to share knowledge, it will create the necessary conditions to perceive the virtual sponsored community as trustworthy. By affecting the interactions in the community, the discussions generate greater knowledge for the gamers which would generate the perception of trustworthiness for the community. We can therefore see a connection between generating qualitative information and the conditions of interaction. The need of emphasizing for interaction from the sponsor can be seen both from the open questions and the concept of perceived responsiveness. The interaction of the gamers has to be rewarded and acknowledged by the sponsor and other parties to have trust towards the virtual sponsored community. Additionally, the respondents have a rather neutral point of the view regarding how much they trust WoW´s virtual community. This might be linked to the comments about how gamers ruining the discussion, which affects the quality of the information and the desire to interact in the community.

Most of the respondents consider the concept of satisfaction important in the virtual sponsored community. Satisfaction is the positive discrepancy between expectations and actual experiences. Hence we can see satisfaction as an outcome from how successful the expectations of the other concepts are fulfilled. Our findings show that the gamers value the positive discrepancy high which is important in affecting the perception of trustworthiness.

The findings from the surveys also show that most of the concepts that constitute the

perception of trustworthiness are relative important, but a few of them appear to be more

significant than others. Quality content, knowledge growth, interaction and privacy policy

seemed to be most important concepts

The concept of quality content as well as knowledge

growth have their main functions to generate qualitative information sharing, which is

A

- 24 -

dependent on creating the conditions where interaction can take place

As the findings show, the quality of the information is highly dependent on participation from the sponsor. Even though the creation of interaction between the members is important, the findings show that it is less crucial to focus on shared values and embeddedness to create a trustworthy virtual community. Privacy policy, which is clearly important for gamers, can be explained by the fact that the gamers need to feel secure and perceive the virtual sponsored community as trustworthy before participating. The importance of the concept of satisfaction is harder to understand because the total experiences in the virtual community for the gamers are highly dependent on how the remaining concepts affect the perception of trustworthiness of the community.

5.2 Qualitative study

Based on our findings from the five threads on the official World of Warcraft forum, we manage to connect relevant points to the theories on concepts affecting the perception of trustworthiness. Generally, seven out of eight concepts of trust-building are presented here, and they are: quality content, interaction, shared values, privacy policy, knowledge growth, satisfaction, embeddedness. The concept perceived responsiveness is not presented in the findings because there is no evidence that the moderators are affecting this concept.

According to the theories, shared values such as goals and visions connect individuals in a virtual community. Our findings show that the moderators in the threads operate in accordance to this concept. The moderators post in the threads to either provide information or to answer gamers’ questions; this shows that the moderators share the same value as the gamers. The existence of the different forums itself is even a proof of this, as the sub-forums categorizes the threads so that relevant ones are found within proximity of each other.

However, thread 3 and 4 shows that even though the moderators can share value with the gamers by participating in the threads, their participation can have a negative effect on the trustworthiness of the community because the things they write in the threads do not show shared value. Therefore our findings show that shared values have a direct effect on the perception of trustworthiness of the virtual sponsored community.

The concept privacy policy is only prevalent in thread 1 because the thread is created for the

purpose of bringing awareness to privacy policies in the game and in the forum. However, by

having dedicated webpages on privacy policies show that the sponsors are taking this concept

A

- 25 -

into consideration when trying to affect the perception of trustworthiness of the virtual sponsored community.

The moderators’ active participation in the threads is trying to promote satisfaction through providing answers to gamers’ questions. However, failures to create satisfaction for gamers in thread 3 and 4 have resulted in outcries from unsatisfied gamers.

The moderators’ participation increases the knowledge of the gamers by providing “insider”

information. The knowledge growth leads to both positive and negative perceptions of trustworthiness of the virtual community. In thread 1, 2 and 5, the moderators offer detailed responses to gamers which affect the perception of trustworthiness in a positive way; in thread 3 and 4, however, the gamers’ perception of trustworthiness is affected in a negative way despite the knowledge growth the moderators try to affect because the responses are short and vague.

The quality of content of the moderators’ participation in the threads varies. The moderators provide relevant and credible information in thread 1, 2 and 5, which increase the perception of trustworthiness. In thread 3 and 4, the quality of the content is not up to the expectations of the gamers; this includes frequent off-topic answers, which not only avoid the gamers’

questions but also cause the gamers to be unsatisfied. Therefore quality content is a concept that affects the perception of trustworthiness on virtual sponsored communities.

Interaction and embeddedness are two closely related concepts. The moderators interact directly with gamers to provide individual solutions and answers, which give the sense of embeddedness to gamers as the gamers feel empowered of being organizational insiders. One exception is thread 2, in which there is no interaction between the moderators and the gamers.

The moderator only posts links to the relevant news each day, and completely ignores gamers’

questions. This could be because this thread’s sole purpose is to provide relevant news to increase embeddedness, and that interaction is not viewed as a part of this process.

The moderators’ management in thread 3 and 4 fails to create discussions, which leads to the

feeling of insult felt by gamers (e.g. gamers’ reactions to “Aww” in thread 3), which does not

necessarily increase trustworthiness. Nevertheless, the dissatisfying interactions from the

moderators cause gamers to distrust the sponsor. Therefore simply emphasizing on interaction

is not enough to increase embeddedness and perception of trustworthiness of the virtual

community. It should rather be done in a substantial manner so that the gamers are satisfied

with the moderators’ inputs. Otherwise the interactions create the opposite effect and decrease

A

- 26 -

the trustworthiness as can be seen in these two threads. Nevertheless, the participation from the moderators to create interaction is necessary in promoting the perception of trustworthiness of the virtual sponsored community. This can further be deduced from the negative perception of trustworthiness among the gamers when the moderators show lack of interest to interact in thread 4. The failures in interaction also result in the failure to create the sense of embeddedness.

By analyzing the five threads, we have come to the conclusion that the perception of trustworthiness of a virtual sponsored community is constituted by seven concepts - quality content, interaction, shared values, privacy policy, knowledge growth, satisfaction, embeddedness. Generally, most of the threads on the forums do not have inputs from the moderators. The threads that have moderators’ interactions that this study examines either increase or decrease the trustworthiness of the virtual sponsored community. The perception of trustworthiness is dependent on the concepts of trust-building – if the moderators act in accordance to the concepts (e.g. the moderators provide detailed information to the gamers and enhance interaction), the trustworthiness of the community increases; otherwise the trustworthiness decreases. Therefore the occurrences of the concepts of trust-building do not affect the trustworthiness of the forums, but rather the occurrences of the concepts together with the positive reinforcement which leads to positive reception that result in increased trustworthiness.

Another general trend seems to be that the higher the moderators’ participation rate, the better

the effect is on increasing trustworthiness of the forum. Thread 4 and 5 are dissatisfying for

the gamers, and the participation rate of moderators are 4% and 2.5% respectively. The other

threads which increase trustworthiness have moderator participation rate of 24.4%, 9.5%, and

13.2% respectively. An explanation to this is that the increased participation rate offers more

opportunities for the moderators to interact on the threads, which can in turn provide more

opportunities for interactions which are affected by the concepts of trust-building, and

consequently lead to better perception of trustworthiness of the virtual sponsored community.