5.1 Introduction
For individuals who work or game together online, the difference between a successful collaboration and disappointment often hinges on the quality of a virtual team's communication. Over time, teams develop communication norms of conversational interactivity, openness, and satisfaction. These norms are often unstated, but become part of the team's routine ways of communicating, including participants’ ideas about appropriate topics of conversation, expectations of feedback, appropriate turn-taking behaviors, openness to self-disclosure, and satisfaction with the ongoing interaction among group members.
In virtual teams, developing and maintaining effective communication norms for conversational interactivity, openness, and satisfaction are often difficult. Several aspects of participant computer-mediated communication (CMC) competence, such as skill, efficacy, and confidence, influence the development of these norms (Sherblom, Withers, & Leonard, Reference Sherblom, Withers and Leonard2013; Sundararajan, Reference Sundararajan2009, Reference Sundararajan2010). Relational attributes, such as trust, presence, and identity also have an effect on virtual team communication (Leonard et al., Reference Leonard, Sherblom, Withers and Smith2015). Managing these various influences is essential to effective virtual team collaboration.
Developing an effective collaboration within a virtual environment is particularly important to teams whose members are geographically dispersed. The ability to share ideas freely, strategize together, be understood by other team members, and feel connected to other participants through their avatar representations is important to constructing a shared virtual team relational space in which to interact (Leonard & Withers, Reference Leonard and Withers2009). Developing this shared virtual space in which members feel a sense of togetherness builds team cohesion and facilitates effective communication.
The purpose of the present study is to examine two types of influences on virtual team collaboration. We analyze these influences by surveying participants at universities in the United States and eastern Europe after they have engaged in team problem-solving projects in the virtual environment of Second Life®. Our survey measures the personal competence of participants in using the CMC medium, the relational attributes among team members, and the norms of team conversation that develop. The study then models the influences of personal competence and relational attributes on these team conversational norms. We predict that personal competence, as measured through each individual's level of skill, efficacy, and confidence in using the CMC medium, is necessary but not sufficient to facilitate interactive, open, and satisfying virtual team communication. In addition to this personal competence, team members must develop the relational attributes of trust, presence, and identity to facilitate interactive, open, and satisfying communication.
5.2 Background
This study is grounded in three eras of CMC theory development that inform the understanding of computer-mediated communication (Houtman Makos, & Meacock, Reference Houtman, Makos and Meacock2014; Oztok & Brett, Reference Oztok and Brett2011). The first era focuses on the influence of the medium. The second describes the communication competencies of the people engaging in CMC. The third analyzes the relationships that develop among those people in their interpersonal relationships, virtual teams, and communities.
Theories representing the first era, such as media richness, emphasize the role of the medium as a major influence on communication, in a way that is sometimes referred to as technological determinism (Daft & Lengel, Reference Daft and Lengel1986; Lengel & Daft, Reference Lengel and Daft1988; Trevino, Daft, & Lengel, Reference Trevino, Daft, Lengel, Fulk and Steinfield1990). Walther and his colleagues provide reviews and critiques of these early perspectives (Walther, Reference Walther2004; Walther, Loh, & Granka, Reference Walther, Loh and Granka2005; Walther & Parks, Reference Walther, Parks, Knapp and Daly2002). The more recent theories of the second era, such as social information processing and media naturalness, focus on how communicators develop personal competence in using the medium (Kock, Reference Kock2004, Reference Kock2005; Walther, Reference Walther2009, Reference Walther, Berger, Roloff and Roskos-Ewoldsen2010). This personal competence is often measured in an individual's perceived skill, efficacy, and confidence in communicating with others through the medium. The third era is represented in theories such as the hyperpersonal perspective and theories of virtual community that analyze the relational influences that shape CMC and the development of relationships in virtual environments (Walther, Reference Walther, Berger, Roloff and Roskos-Ewoldsen2010; Willson, Reference Willson2006).
Taking this third perspective, Leonard et al. (Reference Leonard, Sherblom, Withers and Smith2015) argue that personal competence with a CMC medium is necessary, but not sufficient, for effective virtual team communication. They suggest that competence with the medium is required but, in addition, virtual team members must develop a sense of relational trust, presence, and identity for the team to achieve interactivity, openness, and member satisfaction in its conversations. Only with the development of relational trust, presence, and identity, in addition to personal skill, efficacy, and confidence, can a virtual team achieve conversational interactivity, openness, and satisfaction. Their study shows that these relational attributes can be developed through participant training programs, but they do not directly test the relationship of these relational attributes to team member conversational participation.
The present study compares two path analysis models. The first model examines the influence of personal competence, as measured in an individual's skill, efficacy, and confidence, on virtual team conversational norms. The second model includes the relational influences of trust, presence, and identity on the virtual team's conversational interactivity, openness, and satisfaction. The goal of the study is to test whether these relational attributes contribute substantially to the personal competence influences described by Leonard et al. (Reference Leonard, Sherblom, Withers and Smith2015).
5.3 Personal Competence: Skill, Efficacy, and Confidence
Several lines of research describe the influence of CMC competence on virtual team communication. One line identifies participant skill with using the CMC medium as a primary influence (Bubas, Radosevic, & Hutinski, Reference Bubas, Radosevic and Hutinski2003). A second focuses on how a person's sense of efficacy in using the medium affects communication (Kelly et al., Reference Kelly, Keaten, Hazel and Williams2010; Wrench & Punyanunt-Carter, Reference Wrench and Punyanunt-Carter2007). A third examines the confidence of participants in being able to express themselves appropriately, interpret the meanings of others, engage in smooth virtual conversations, and develop social relationships as a major influence (Spitzberg, Reference Spitzberg2006). These three lines of research indicate that skill, efficacy, and confidence each play an influential role and each competency must be developed over time, through individual effort and experience with the medium, for effective team communication to occur.
Media naturalness theory predicts that virtual communication initially requires more time, cognitive effort, and experience to develop these skills. It argues that humans have become adapted through evolution with neurologically optimized brains designed to function most efficiently in synchronous face-to-face communication, with its auditory and visual cues that assist in the interpretation of another person's meaning (Kock, Reference Kock2004, Reference Kock2005). The less a communication medium incorporates colocated, face-to-face, synchronous speech, the greater the cognitive effort required for a person to both convey and understand meaning. Over time, people can develop the cognitive schemas and social skills needed to communicate effectively in a virtual environment, but developing these schemas and learning the appropriate skills requires substantial cognitive effort and practice (DeRosa et al., Reference DeRosa, Hantula, Kock and D'Arcy2004; Kock, Reference Kock and Roberts2008; Kock, Verville, & Garza, Reference Kock, Verville and Garza2007). Once people develop the cognitive schema and social skill to communicate effectively through a medium, however, the virtual environment begins to feel more natural and participants experience greater cognitive ease in coordinating and managing their meanings with others (Kock, Reference Kock2004, Reference Kock2005). Hence, personal skill, efficacy, and confidence are important to communicating effectively in a CMC medium.
5.3.1 Skill
Skill with communicating in a CMC medium develops through repeated interaction with others and is a major influence on a person's ability to be expressive, attentive to others, and engage in smooth conversational coordination (Bubas et al., Reference Bubas, Radosevic and Hutinski2003). Expressiveness recognizes the ability to create messages that seem alive and animated to others. Attentiveness shows in a person's interest, concern, affection, and adaptability to others. Coordination means being able to achieve smooth conversational transitions, timing, topic initiation, and conversational repairs as needed (Bubas et al., Reference Bubas, Radosevic and Hutinski2003).
5.3.2 Efficacy
Efficacy describes the belief among participants that they have the necessary cognitive schema and social abilities to communicate effectively through a medium. Some communicators experience CMC reticence, apprehension, anxiety, and inhibition, which can reduce their ability to express their meanings and emotions effectively (Kelly et al., Reference Kelly, Keaten, Hazel and Williams2010). Hence, anxiety, apprehension, and reticence affect a person's ability to communicate effectively in a virtual medium (Wrench & Punyanunt-Carter, Reference Wrench and Punyanunt-Carter2007).
5.3.3 Confidence
Confidence is the sense that one has the social knowledge and skill to communicate, and the ability to accurately interpret the meanings of others (Spitzberg, Reference Spitzberg2006). A virtual environment has particular communication constraints and affordances, and a participant must develop an ability to communicate effectively within them (Erhardt et al., Reference Erhardt, Martin-Rios, Gibbs and Sherblom2016). A person's confidence in having the ability to perform competently within a virtual environment affects the willingness and motivation to participate. Confidence builds over time as a person experiences successful virtual conversations and develops relationships with others (Spitzberg, Reference Spitzberg2006). These positive experiences orient a person toward increased attentiveness, expressiveness, and willingness to engage in future conversations. This confidence facilitates conversational interactivity, openness, and the likelihood of satisfaction (Spitzberg, Reference Spitzberg2006).
5.3.4 Importance of Skill, Efficacy, and Confidence
Sherblom et al. (Reference Sherblom, Withers and Leonard2013) use regression analysis to examine the influence of skill, efficacy, and confidence on virtual team communication. They conclude that increased skill, efficacy, and confidence, as represented in a lack of apprehension, positively affect virtual team conversations. They do not, however, investigate the potential of relational influences.
Leonard et al. (Reference Leonard, Sherblom, Withers and Smith2015) do examine the relational influences of trust, presence, and identity on virtual team conversation. Their qualitative analysis indicates that relational trust, sense of presence, and development of an online identity all affect virtual team interactivity, openness, and satisfaction (Leonard et al., Reference Leonard, Sherblom, Withers and Smith2015). Building on their findings, the present study examines these influences: first by modeling the effects of personal skill, confidence, and efficacy, and then by modeling the additional influences of relational trust, presence, and identity.
5.4 Relational Attributes: Trust, Presence, and Identity
5.4.1 Trust
Trust is a complex relational attribute that influences both face-to-face and virtual team communication (Feng, Lazar, & Preece, Reference Feng, Lazar and Preece2004; Henderson & Gilding, Reference Henderson and Gilding2004; Himelboim et al., Reference Himelboim, Lariscy, Tinkham, Kaye and Sweetser2012; Sarker et al., Reference Sarker, Ahuja, Sarker and Kirkeby2011). Trust between communicators builds over time. It is assessed as reputations develop among participants who judge each other to be trustworthy, or not, based on patterns of behavior that fulfill or disappoint expectations. Trustworthiness is the perception that a person possesses a set of relational communication characteristics such as responsiveness, benevolence, cooperation, and integrity that are beneficial to the team (Beldad, de Jong, & Steehouder, Reference Beldad, de Jong and Steehouder2010; Cheng & Macaulay, Reference Cheng and Macaulay2013; Henderson & Gilding, Reference Henderson and Gilding2004; Jarvenpaa & Leidner, Reference Jarvenpaa and Leidner1998; Morrison, Cegielski, & Rainer, Reference Morrison, Cegielski and Rainer2012; Schiller, Mennecke, Nah, & Luse, Reference Schiller, Mennecke, Nah and Luse2014).
Some characteristics of a virtual environment, such as the reduced social cues, perception of anonymity, and asynchronous communication, can create interpersonal uncertainty that hinders the development of trust and assessment of trustworthiness (Beldad et al., Reference Beldad, de Jong and Steehouder2010; Henderson & Gilding, Reference Henderson and Gilding2004; Jarvenpaa & Leidner, Reference Jarvenpaa and Leidner1998; Nam, Reference Nam2014; Turilli, Vaccaro, & Taddeo, Reference Turilli, Vaccaro and Taddeo2010). It is possible, however, for communicators to compensate for these reduced social cues, increased anonymity, and asynchronous communication through clear, specific, frequent verbal statements and development of a social presence (Jarvenpaa & Leidner, Reference Jarvenpaa and Leidner1998). Communicating with others through avatars provides adequate social information for participants to both express and assess trustworthiness over time (Henderson & Gilding, Reference Henderson and Gilding2004; Turilli et al., Reference Turilli, Vaccaro and Taddeo2010). Additionally, online relational trust builds as a participant's sense of presence. With that presence, participant feelings of connectedness, openness, and commitment to the group task increase (Beldad et al., Reference Beldad, de Jong and Steehouder2010; Green-Hamann & Sherblom, Reference Green-Hamann and Sherblom2014; Hains, Reference Hains2014; Himelboim et al., Reference Himelboim, Lariscy, Tinkham, Kaye and Sweetser2012; Morrison et al., Reference Morrison, Cegielski and Rainer2012).
5.4.2 Presence
Presence represents a person's subjective psychological state in which the individual does not perceive the technology as mediating the interpretation of sensory stimuli (Aymerich-Franch, Reference Aymerich-Franch2010). Instead, the individual overlooks the technology as an influence on the interpretation of that experience (International Society for Presence Research, 2000). This conception of presence has moved away from a focus on the media richness or capacity as affecting a person's thoughts and feelings about the ongoing social, relational, and contextual processes in a virtual environment. Recent conceptualizations of presence focus more on how virtual community participants use the affordances of the medium to build relationships with other team members (Erhardt et al., Reference Erhardt, Martin-Rios, Gibbs and Sherblom2016; Houtman et al., Reference Houtman, Makos and Meacock2014; Kehrwald, Reference Kehrwald2008; Sherblom, Reference Sherblom2010; Tu & McIsaac, Reference Tu and McIsaac2002).
Thus, presence is a subjective, relational phenomenon that is dependent on the ability and willingness of participants to achieve interpersonal relationships with others. It represents a dynamic phenomenon rather than a simple learned behavior or skill (Kehrwald, Reference Kehrwald2008). Individuals must participate in an ongoing performance of presence that demonstrates to others that they have a willingness and ability to communicate within the virtual environment and to engage in interpersonal relationships (Kehrwald, Reference Kehrwald2008). This idea of presence places agency in the individual, rather than in the medium, and adds a relationally performative component. Participants must negotiate the medium, establish their presence through self-disclosure, and actively demonstrate that presence through ongoing relational cues that show an attentiveness, trust, empathy, rapport, and emotional expressiveness (Kehrwald, Reference Kehrwald2008; Sherblom, Reference Sherblom2010). The ability of individuals to effectively negotiate the virtual environment and construct this performative sense of presence in their conversations influences their ability to actively participate in the conversations of a virtual team and build personal satisfaction with those interactions (Houtman et al., Reference Houtman, Makos and Meacock2014; Kehrwald, Reference Kehrwald2008).
The realism of the virtual environment, sophistication of avatar designs, and avatar-mediated communication facilitate this sense of presence (Jin & Bolebruch, Reference Jin and Bolebruch2010). Virtual team members frequently report experiencing high levels of presence in the vivid, immersive, three-dimensional spaces of virtual environments (Aymerich-Franch, Reference Aymerich-Franch2010; Biocca, Kim, & Choi, Reference Biocca, Kim and Choi2001; Jin & Bolebruch, Reference Jin and Bolebruch2009; Leonard et al., Reference Leonard, Sherblom, Withers and Smith2015). Within this virtual environment, presence is associated with feelings of connectivity, immediacy, intimacy, warmth, mutual social awareness, involvement, and emotional accessibility (Green-Hamann, Eichhorn, & Sherblom, Reference Green-Hamann, Eichhorn and Sherblom2011; Jin & Bolebruch, Reference Jin and Bolebruch2009; Lee, Reference Lee2004; Nowak, Reference Nowak2001; Sivunen & Nordbäck, Reference Sivunen and Nordbäck2015). These feelings produce important influences on a team's communication in the virtual environment, affecting team member interactivity, openness, and satisfaction.
5.4.3 Identity
The use of a screen name, or pseudonym, and consistently recognizable avatar appearance establishes a person's identity within a virtual community and helps develop an online reputation (Leonard & Toller, Reference Leonard and Toller2012). This identity contains both internal (personal) and external (social) components. The internal component is a person's self-belief, such as perceiving oneself to be honorable, competent, or funny. The external aspect represents how a person talks to and connects with others, and the types of social interactions that person participates in, such as belonging to and identifying with a particular group or maintaining friendships with certain types of individuals (Cheek & Briggs, Reference Cheek and Briggs1982; Leary, Wheeler, & Jenkins, Reference Leary, Wheeler and Jenkins1986). “Through their avatars and associated profiles, virtual world residents can establish their virtual identities, which can be molded according to their desires and expectations” (Nagy & Koles, Reference Nagy and Koles2014, p. 279). These virtual identities often represent a duality of self, similar to the dichotomy described by Cheek and Briggs (Reference Cheek and Briggs1982) and Leary et al. (Reference Leary, Wheeler and Jenkins1986). That is, a person's virtual identity embodies a self-representation made by the participant and an external representation that marks the choices made when communicating with others in the virtual environment (Seung-A, Reference Seung-A2012). These internal and external choices are influenced by the person's knowledge of that virtual environment and affect the participation within it (DeGrove, Courtois, & Van Looy, Reference DeGrove, Courtois and Van Looy2015; Leonard, Withers, & Sherblom, Reference Sherblom2010).
The relational attributes of trust, presence, and identity are essential to the development of successful virtual team communication. These attributes, along with personal competence, as measured in skill, efficacy, and confidence, influence communication norms that a team develops. Three measures of these norms are participant interactivity, openness, and satisfaction.
5.5 Team Conversational Norms: Interactivity, Openness, and Satisfaction
5.5.1 Interactivity
Interactivity includes the communication rate, feedback, turn-taking, timeliness, responsiveness, immediacy, and synchronicity that participants experience in a conversation (Karimi, Ramenzoni, & Holme, Reference Karimi, Ramenzoni and Holme2014; Tu & McIsaac, Reference Tu and McIsaac2002). This interactivity is generally greater in a synchronous communication medium, such as in face-to-face discussion. In these face-to-face discussions the conversations are often more dialogic, open, and immediate, and the rate of information exchange is typically greater (Tu & McIsaac, Reference Tu and McIsaac2002).
5.5.2 Openness
Openness describes an individual's level of comfort and ease with expressing personal thoughts, opinions, ideas, and emotions to others (Ayoko, Reference Ayoko2007; Nam, Reference Nam2014). This openness is reflected in a participant's willingness to self-disclose to the group (Goffman, Reference Goffman1959; Jourard, Reference Jourard1971). This self-disclosure is encouraged or discouraged by the interactivity norms established and implicitly maintained in the group. In a virtual team these openness cues may be communicated explicitly in words or demonstrated implicitly through such nonverbal cues as a person's avatar appearance, the topics discussed, or the verbal style of other participants (Gottschalk, Reference Gottschalk2010).
5.5.3 Satisfaction
Satisfaction is an affective response that represents an enjoyable, fulfilling experience (Hecht, Reference Hecht1978). Several factors affect an individual's level of satisfaction, including the communication climate of the group, amount of personal feedback received, and sociocultural expectations (Diener, Reference Diener2000; Downs & Hazen, Reference Downs and Hazen1977). Participants who engage in group activities in a virtual environment often express satisfaction with the richness of the virtual environment, number of social cues available, sense of presence with others, and ease of participating (Hazel, Crandall, & Caputo, Reference Hazel, Crandall and Caputo2015; Nowak, Watt, & Walther, Reference Walther2009; Sherblom, Reference Sherblom2010; Simon, Reference Simon2010; Walther & Bazarova, Reference Walther and Bazarova2008). The amount of feedback, complexity of information exchanged, and participant skill in using the medium all affect satisfaction (Walther & Bazarova, Reference Walther and Bazarova2008). Adequate training and technical support positively influence participant satisfaction with a new CMC medium (Hiltz & Johnson, Reference Hiltz and Johnson1990; Leonard et al., Reference Leonard, Sherblom, Withers and Smith2015; Sherblom et al., Reference Sherblom, Withers and Leonard2013). The self-disclosure of other participants within that medium can increase a participant's communication satisfaction, as well (Morry, Reference Morry2005).
5.6 Hypotheses
Following some early studies that focused mainly on media effects, a substantial amount of CMC research has shown the influence of personal competence, as measured in indices of skill, efficacy, and confidence, on virtual team conversational participation. In addition, much of this more recent CMC research suggests the influence of relational attributes such as trust, presence, and identity. At least one recent qualitative study of focus-group participants indicates that these relational influences affect the conversational participation of team members who meet in a virtual environment (Leonard et al., Reference Leonard, Sherblom, Withers and Smith2015). The present study examines whether these relational influences add substantially to the well-documented contributions of personal competence. The specific expectations of this study are stated in the following two hypotheses.
H1: The personal competence indices of participant skill, efficacy, and confidence predict virtual team conversational interactivity, openness, and satisfaction.
H2: The relational indices of trust, presence, and identity will add substantially to these personal competence indices in predicting virtual team conversational interactivity, openness, and satisfaction.
5.7 Method
5.7.1 Participants
The sample of 104 participants consists of seven relatively equal-size subsamples averaging 15 participants apiece. These subsamples were collected over the course of a year from three universities: an eastern US university, a midwestern US university, and an eastern European university. The total sample comprised 47 female and 57 male participants. Participant ages range from 18 to 40, with a mean age of 21.
5.7.2 Participant Training
The participants in this study were trained to use the Second Life® program over a period of several weeks. After an initial introduction to the use of the program, participants were given a set of team projects to complete before being invited to participate in the survey of their Second Life® communication experiences. These projects involved team discussion, decision-making, problem-solving, and report-writing assignments. Each team of four or five members had a specific meeting place in Second Life® to carry on these activities. For their first project, team members were required to interview Second Life® residents about their media use and develop a single group report of their findings. From this project participants learned about standards of media use in Second Life®, and faced the challenges of working and writing as a virtual team. For the second project, each team joined a virtual community in Second Life® and, after a period of observation and interviewing, reported on the community's values, goals, and ethics. These reports provided team members the opportunity to discuss and reflect on their assumptions about the purpose of virtual communities and the communication within them. In addition, writing together as a virtual team that met and communicated in Second Life® required group discussion, decision making, and problem solving to synthesize the results of each individual interview and experience into a single unified team report.
5.7.3 The Second Life® Virtual Environment
Second Life® is the largest three-dimensional, multiuser virtual environment created by users. There are 15.5 million registered participants in Second Life® and 900,000 of them are active each month. These users interact in a virtual space four times the size of New York City (Flowers, Gregson, & Trigilio, Reference Flowers, Gregson and Trigilio2009; Kingsley & Wankel, Reference Kingsley, Wankel, Wankel and Kingsley2009). Within this virtual space, participants communicate through symbolic visual representations of themselves called avatars, using both public and private forms of nearly synchronous text messaging (Sherblom, Reference Sherblom2010).
These Second Life® participants engage in social activities and carry on business transactions, and numerous for-profit and nonprofit organizations operate and hold meetings in Second Life® (Sherblom, Withers, & Leonard, Reference Sherblom, Withers, Leonard, Wankel and Kingsley2009). Among the most prominent of these companies participating in Second Life® are the American Cancer Society, Coca-Cola, Crescendo Design, Kraft Foods, IBM, Pepsi, Pizza Hut, and Starwood Hotels (Sherblom & Green-Hamann, Reference Sherblom, Green-Hamann, Al-Deen and Hendricks2013). The Second Life® currency, known as Linden dollars, has a real-world exchange rate and in 2015, Second Life® residents earned about 60 million US dollars through their business activities (Charara, Reference Charara2016).
5.7.4 Procedures
The research procedures and survey questions used in this study were approved by all three university institutional review boards. Before beginning the survey, each participant was provided with an informed consent form detailing the study. Then participants responded to the survey that appears in the appendix. This electronically distributed survey took participants approximately 20 minutes to complete and participants were provided a small monetary incentive in Linden dollars for completing the survey.
5.7.5 Measures
Survey scales measure self-perceptions of skill, efficacy, confidence, trust, presence, identity, interactivity, openness, and satisfaction. Questions about participant biological sex and age appear at the end of the survey. Each measure consists of a series of statements to which participants respond on a five-point Likert-type scale. Response choices range from 1 = strongly agree to 5 = strongly disagree. Several items are reverse coded to reduce the likelihood of response-set bias.
Skill is measured using nine items. These items assess a participant's ability to engage in turn-taking, prioritize conversational responses, be articulate and expressive, display appropriate emphasis, and engage in a clear communication style. A Cronbach's alpha coefficient (α = .85) shows an acceptable reliability for this measure.
Efficacy is measured with seven items. These items examine the extent to which respondents feel able to accomplish tasks, be productive, and be efficient in communicating with team members in Second Life®. Scales ask how much they use, rely on, and how useful they find this communication medium. A Cronbach's alpha coefficient (α = .79) shows acceptable reliability for the efficacy measure.
Confidence is measured using five items. These items ask respondents to reflect on how capable, confident, knowledgeable, or nervous (reverse-coded) they feel, and how quickly they developed their ability in using this communication medium. A Cronbach's alpha coefficient (α = .81) demonstrates acceptable reliability for confidence.
Trust is measured with seven items. These items ask participants how they present themselves to others, get to know others, and how much they trust others to be responsible, produce high-quality work, meet deadlines, and participate as productive team members. A Cronbach's alpha coefficient (α = .81) demonstrates an acceptable reliability for trust.
Presence is measured using fourteen items. These ask how connected, immersed, and close participants feel to others when communicating through this medium. A Cronbach's alpha coefficient (α = .91) shows an acceptable reliability for presence.
Identity is measured with seven items. These items ask about whether a participant's avatar looks, acts, and represents the participant adequately, and whether the participant feels able to make a true representation of self in Second Life.® A Cronbach's alpha coefficient (α = .84) shows an acceptable reliability for identity.
Interactivity is measured using fourteen items. These items ask about the flow of the conversation, equality of participation among team members, ability to keep up with and follow the conversation, smoothness of turn-taking, and perceived willingness of other participants to respond quickly with relevant and on-topic contributions. A Cronbach's alpha coefficient (α = .91) provides an acceptable reliability for interactivity.
Openness is measured using ten items. These items ask about whether participants feel like they are able to openly share their thoughts, feelings, ideas, and opinions; say what they really think and feel; and comment honestly on the ideas of others. A Cronbach's alpha coefficient (α = .80) reveals an acceptable reliability for openness.
Satisfaction is measured with thirteen items. These items ask whether participants get what they want out of team interactions, achieve their goals, are able to express their ideas clearly, understand what others say, enjoy their interactions, believe their communication is effective, and feel good about their conversations. They also ask whether participants feel like they get to know people, can make friends easily, are pleased with their encounters, and are generally satisfied with their communication in this medium. A Cronbach's alpha coefficient (α = .94) demonstrates an acceptable reliability for satisfaction.
5.7.6 Statistical Analyses
Three types of statistical analyses describe the relationships among these measures. Correlations describe the relationships among them. Regression analysis compares the influence of personal skill, efficacy, and confidence on team conversational interactivity, openness, and satisfaction found in the present study to the earlier results reported by Sherblom et al. (Reference Sherblom, Green-Hamann, Al-Deen and Hendricks2013), who found that the personal competence measures of knowledge, skill, motivation, and lack of apprehension affect virtual team communication. Finally, structural equation modeling produces two path analysis models that provide tests of the research hypotheses. These two path analysis models more fully delineate the multiple direct and mediated regression relationships among personal skill, efficacy, and confidence, and of relational trust, presence, and identity, as influences on team conversational interactivity, openness, and satisfaction (Hayes, Reference Hayes2009).
The first path analysis model shows the influence of personal skill, efficacy, and confidence on team interactivity, openness, and satisfaction. The second path analysis model adds the relational attributes of trust, presence, and identity to the model along with these personal competence measures. To show an adequate fit to the data, a model must produce a nonsignificant χ2 and a χ2 to df ratio that is less than 5 (Marsh & Hocevar, Reference Marsh and Hocevar1985). In addition, an incremental fit measure such as the comparative fit index (CFI), which is relatively insensitive to model complexity, and a measure of absolute fit, such as the root mean square error of approximation (RMSEA), can be used to test and demonstrate an adequate model fit (Hair et al., Reference Hair, Black, Babin, Anderson and Tatham2006). The CFI is normed so that values closer to 1 indicate a better fit (Bentler & Bonett, Reference Bentler and Bonett1980). A CFI value of .95 or greater indicates a model that has a good fit (Hayes, Slater, & Snyder, Reference Hayes, Slater and Snyder2008). The RMSEA, with a built-in parsimony index that corrects for sample size and model complexity, provides an index of how well a model fits the population as well as the sample (Hair et al., Reference Hair, Black, Babin, Anderson and Tatham2006). An RMSEA value of .06 or less indicates a good model fit (Hayes et al., Reference Hayes, Slater and Snyder2008).
In sum, a nonsignificant χ2, χ2 to degrees of freedom ratio of less than 5, CFI value of .95 or greater, and RMSEA of less than .06 indicate a good model fit (Hayes et al., Reference Hayes, Slater and Snyder2008). See Tabachnick and Fidell (Reference Tabachnick and Fidell2013) for a fuller description of these indices and model fit criteria. In developing the path analysis model, weak links, that is, links producing a β value of less than .10, are dropped to produce a more parsimonious model. The examination of each model's fit statistics offers a good comparison for showing if the relational attributes of trust, presence, and identity provide a substantial contribution in addition to the personal skill, efficacy, and confidence effects on virtual team conversational interactivity, openness, and satisfaction.
5.8 Results and Discussion
Table 5.1 shows that skill has a weak correlation with efficacy (r = .19) and a moderate correlation with confidence (r = .52). Efficacy presents a weak correlation with confidence as well (r = .15). Trust has moderate correlations with both presence (r = .53) and identity (r = .38), and presence is moderately correlated with identity (r = .64). Interactivity is moderately correlated with both openness (r = .64) and satisfaction (r = .48), and openness is moderately correlated with satisfaction (r = .55). These correlations show that the measures are moderately correlated, but that each indexes a unique attribute.
Table 5.1 Correlations
| 1. | 2. | 3. | 4. | 5. | 6. | 7. | 8. | 9. | |
|---|---|---|---|---|---|---|---|---|---|
| 1. Skill | 1 | .19 | .52* | .46* | .52* | .34* | .50* | .50* | .57* |
| 2. Efficacy | .19 | 1 | .15 | .19 | .40* | .36* | .35* | .30* | .43* |
| 3. Confidence | .52* | .15 | 1 | .20* | .26* | .22* | .19 | .33* | .47* |
| 4. Trust | .46* | .19 | .20* | 1 | .53* | .38* | .62* | .55* | .38* |
| 5. Presence | .52* | .40* | .26* | .53* | 1 | .64* | .53* | .55* | .60* |
| 6. Identity | .34* | .36* | .22* | .38* | .64* | 1 | .40* | .52* | .50* |
| 7. Interactivity | .50* | .35* | .19* | .62* | .53* | .40* | 1 | .64* | .48* |
| 8. Openness | .50* | .30* | .33* | .55* | .55* | .52* | .64* | 1 | .55* |
| 9. Satisfaction | .57* | .43* | .47* | .38* | .60* | .50* | .48* | .55* | 1 |
N = 104;
* p < .05.
Hypothesis 1 states the expectation that participant skill, efficacy, and confidence, as indices of personal competence, will predict virtual team conversational norms of interactivity, openness, and satisfaction. The linear regression analysis demonstrates that skill, efficacy, and confidence do combine to account for 33% of the variance in interactivity (r2 = .33), 29% of the variance in openness (r2 = .29), and 46% of the variance in satisfaction (r2 = .46). These findings are similar to and corroborate the results reported by Sherblom et al. (Reference Sherblom, Green-Hamann, Al-Deen and Hendricks2013). They found that the personal competence measures of knowledge, skill, motivation, and lack of apprehension accounted for 24% (r2 = .24) of the variance in their virtual team communication.
Figure 5.1 provides a path analysis model that more carefully delineates these personal competence influences of skill, efficacy, and confidence on the virtual team conversational participation norms of interactivity, openness, and satisfaction. The model shows that skill predicts interactivity (β = .46) and satisfaction (β = .33). Efficacy predicts interactivity (β = .27) and satisfaction (β = .31). Confidence predicts openness (β = .26) and satisfaction (β = .32). In addition, interactivity also affects openness (β = .54).
Figure 5.1 Model of personal competence influences on team conversational norms, with no relational attributes included. χ2 = 14.32; df = 5; p = .014; χ2/df ratio = 2.84; CFI = .82; RMSEA = .13.
Similar to the linear regression results, this model shows that the personal competence indices of skill and efficacy combine to account for a substantial amount of the variance in interactivity (r2 = .32). Confidence, along with interactivity, accounts for variance in openness (r2 = .44), and all three indices combine to predict satisfaction (r2 = .48). This model, however, demonstrates a relatively poor fit to the data. This poor fit is shown in the fit indices: χ2 = 14.32; df = 5; p = .014; χ2/df ratio = 2.84; CFI = .82; RMSEA = .13. This relatively poor fit suggests that there are other predictors of these team conversational norms that are not recognized in this model. Hypothesis 2 predicts that the relational attributes of trust, presence, and identity will add substantially to the personal competence measures modelled in this first figure.
Figure 5.2 shows a path analysis model that includes these relational attributes of trust, presence, and identity as predictors of interactivity, openness, and satisfaction. The model fit indices of χ2 = 10.11; df = 15; p = .813; χ2/df ratio = .67; CFI = .99; RMSEA = .01 demonstrate that this model provides a good fit to the data. The χ2 value is nonsignificant, χ2/df ratio is less than 5, CFI is greater than .95, and RMSEA is less than .06. Each of these indices indicates a good model fit.
Figure 5.2 Model of personal competence and relational attribute influences on team conversational norms. χ2 = 10.11; df = 15; p = .813; χ2/df ratio = .67; CFI = .99; RMSEA = .01.
This model also shows that skill, efficacy, and confidence have both direct and indirect mediated effects on interactivity, openness, and satisfaction. Skill directly affects interactivity (β = .22) and satisfaction (β = .22). Efficacy influences satisfaction (β = .22). Confidence affects openness (β = .17) and satisfaction (β = .22). These effects show that the development of personal competence directly predict team conversational interactivity, openness, and satisfaction.
In addition, skill affects the relational attributes of trust (β = .25) and presence (β = .46). Trust affects openness (β = .19) and interactivity (β = .42), indicating that skill has an additional mediated effect, through trust, on team interactivity and openness. Efficacy shows an additional mediated connection to interactivity and satisfaction through presence (β = .31). Presence is associated with both interactivity (β = .12) and satisfaction (β = .34). Efficacy shows a mediated association with openness through its connection with identity (β = .13), as well. Identity (β = .25) and interactivity (β = .38) are both associated with greater openness, as well.
This second model, which includes trust, presence, and identity, provides a much fuller picture and more complex description of the influences on team conversational norms. This model shows the direct and mediated relationships of personal skill, efficacy, and confidence, and the additional contributions of relational trust, presence, and identity to developing team conversational norms of interactivity, openness, and satisfaction. Including the relational attributes of trust, presence, and identity in this model not only provides a substantially better fit to the data, but also increases the amount of variance accounted in the team conversational norms.
Skill and presence combine to account for 33% of the variance in trust (r2= .33). Skill, efficacy, presence, and trust account for 50% of the variance in interactivity (r2= .50). Skill, efficacy, confidence, and presence account for 55% of the variance in satisfaction (r2= .55). Confidence, trust, interactivity, and identity account for 54% of the variance in openness (r2= .54). That is, when the relational attributes are included in the model the variance accounted for in interactivity increases from 32% to 50%, in openness from 44% to 54%, and in satisfaction from 48% to 55%.
The superior fit and increased variance accounted for by this model shown in Figure 5.2 provide support for hypothesis 2. The relational attributes add substantially to the measures of personal competence in predicting virtual team conversational interactivity, openness, and satisfaction. Table 5.2 shows a comparison of the fit statistics for the two path analysis models. This comparison demonstrates that the model that includes the relational attributes of trust, presence, and identity provides a better fit to the data and a fuller picture of the relationships than the model that includes only the measures of personal competence.
Table 5.2 Model Fit Criteria Comparison
| Criteria: | χ2 | df | p of χ2 | χ2/ df ratio | CFI | RMSEA |
|---|---|---|---|---|---|---|
| > .05 | < 5 | > .95 | < .06 | |||
| Model (no relational attributes) | 14.32 | 5 | .02 | 2.84 | .82 | .13 |
| Model (with relational attributes) | 10.11 | 15 | .81 | .67 | .99 | .01 |
In addition, a comparison of the residual correlation matrices associated with the two models shows a significant difference, χ2 = 410.03; df = 42; p < .001. The model that includes the relational attributes contains a smaller set of residual correlations (M = .0008, SD = .09) than the model without the relational attributes (M = .032, SD = .14). This indicates that the model containing the relational attributes of trust, presence, and identity does a better job explaining the observed relationships than the model containing no relational attributes. That is, including the relational attributes leaves less unaccounted correlations in the residual matrix (Anderson & Williams, Reference Anderson and Williams1992; Veit & Ware, Reference Veit and Ware1983).
5.9 Conclusion
The literature review highlights two potential influences on virtual team communication. The first influence is each individual's ability to communicate competently within the medium. We measured this personal competence through indices of skill, efficacy, and confidence in the using the communication medium. Recent literature indicates that differences in the relational attributes of trust, presence, and identity also affect the virtual team communication. Our path analysis models examine and compare these two types of influences on virtual team conversational norms. Results of this comparison indicate that the relational attributes of trust, presence, and identity are associated with personal competence as measured in skill, efficacy, and competence, and contribute substantially to team conversational interactivity, openness, and satisfaction.
Virtual teams establish and enforce conversational norms of interactivity, openness, and satisfaction as they work together. The present results indicate that having team members who are competent in communicating in a virtual environment is necessary but not adequate to achieve effective norms of team conversational interactivity, openness, and satisfaction. To develop these norms requires relational trust, presence, and identity. Training programs designed to build effective virtual team communication should focus some attention on developing these relational attributes within the team, in addition to building the personal competence needed to use the medium. Media naturalness theory suggests that participants can develop these relational skills through mindful attention and cognitive effort.
Relational trust is built through the verbal and nonverbal communication in which one learns to trust another person and establishes that person's trust in oneself. Engaging in relational talk as well as task-oriented business, that is, offering personal information about oneself such as one's thoughts and feelings toward the project, and asking the other person for ideas and opinions can, over time, establish rapport and build relational trust. This trust-building process is not different than that engaged in by teams who meet face-to-face, but it may take longer and require more effort from virtual team members.
Presence is an ongoing dynamic process of being with the other person within the communication medium. It is built upon one's skill, efficacy, and confidence in understanding and be understood by the other person. Developing this relational presence with others through the medium comes with practice and experience, and forms the basis for an interactive, open, and satisfying task-oriented virtual team conversation.
Identity builds upon this sense of presence to define the participant's role expectations of others and of self within the group. This identity becomes an enduring attribute of a participant in the virtual team. It affects the roles a participant is given by the team, and the ones a person accepts and expects (Brandon & Hollingshead, Reference Brandon and Hollingshead2004; Hollingshead, Reference Hollingshead2001). A participant's identity within the team affects conversational interactivity, openness, and satisfaction.
The model presented in Figure 5.2 shows that including these relational attributes of trust, presence, and identity provides a better model fit and a more complete description of the influences affecting a virtual team's conversational interactivity, openness, and satisfaction. Personal skill, efficacy, and confidence are necessary, but not sufficient, to generate conversational interactivity, openness, and satisfaction in a virtual team. In addition, team members must be able to form relationships that exhibit trust, presence, and identity with each other. These relational attributes of trust, presence, and identity build upon the personal competence influences of skill, efficacy, and confidence to affect the interactivity, openness, and satisfaction norms of the virtual team.
The present study shows the relationship between these relational attributes and a virtual team's conversational interactivity, openness, and satisfaction. A virtual team's conversational interactivity, openness, and satisfaction are associated with these relational attributes. Without relational trust, presence, and identity even team members who possess personal competence with the medium will have difficulty engaging in interactive, open, and satisfying conversations in a virtual environment.
