LIT REVIEW

Diversity in team composition, relationship conflict and team

leader support on globally distributed virtual software

development team performance Vathsala Wickramasinghe and Sahan Nandula

Department of Management of Technology, University of Moratuwa, Moratuwa, Sri Lanka

Abstract Purpose – This study aims to investigate whether diversity in team composition leads to relationship conflict, and, consequently, relationship conflict leads to team performance, and whether team leader support moderates the negative effects of relationship conflict on team performance. Design/methodology/approach – For the study, 216 team members working in globally distributed virtual software development projects responded. To examine the hypothesized relationships, structural equation modeling with maximum likelihood estimation was performed. Findings – It was found that diversity in team composition leads to relationship conflict, relationship conflict leads to team performance and team leader support moderates the latter relationship. Practical implications – The findings suggest the role of team leaders in reducing the harmful effect of relationship conflict on team performance. The findings imply the need of providing training to team leaders to create cohesive teams that deliver on project goals. Originality/value – Empirical studies on globally distributed virtual teams could provide new insights into challenges and issues associated with team composition, relationship conflict and team leader support in achieving higher levels of team performance.

Keywords Knowledge-based systems, Emerging markets, Virtual teams, Economic and social systems, Global outsourcing

Paper type Research paper

Introduction Owing to competitive pressures, organizations are forced to find more flexible and versatile business structures (Bell and Kozlowski, 2002; Peters and Karren, 2009). Team-based virtual work structures are identified as one of the best means to achieve this flexibility and versatility (Mishra and Mahanty, 2014; Peters and Karren, 2009; Saxena and Burmann, 2014). According to Curşeu and Wessel (2005, p. 271), “a virtual team is a collection of individuals who are geographically, organizationally or otherwise dispersed and who collaborate, using varying degrees of communication and information technologies in order to accomplish a specific goal”. Innovations in information and communication technologies (ICT) have accelerated the growth of globally distributed virtual software development (GDVSD) or offshore outsourcing of

The current issue and full text archive of this journal is available on Emerald Insight at: www.emeraldinsight.com/1753-8297.htm

SO 8,2/3

138

Received 8 February 2015 Revised 1 April 2015 15 May 2015 Accepted 27 May 2015

Strategic Outsourcing: An International Journal Vol. 8 No. 2/3, 2015 pp. 138-155 © Emerald Group Publishing Limited 1753-8297 DOI 10.1108/SO-02-2015-0007

software development during the past two decades (Kotlarsky et al., 2007; Mishra and Mahanty, 2014).

A GDVSD project team is a network of dispersed group of people who work together toward a common goal and who share together resources, know-how, services and by-products (Bourgault et al., 2008; Reed and Knight, 2010). For the success of project teams, organizations expect a higher level of team performance that leads to increased team effectiveness (such as work quality and ability to meet project goals) and efficiency (such as adherence to time schedules and budget) as well as psychosocial outcomes (such as the degree of experienced friendliness and support) (Faraj and Sproull, 2000; Pinto and Pinto, 1990; Saxena and Burmann, 2014). However, there are several factors in virtual team environment that operate as barriers for effective team performance. Temporarily assigned members of globally distributed virtual teams (GDVTs) typically come from different national, ethnic, functional and educational backgrounds; reside in different countries or continents with different time zones; and rarely or never see each other in-person (Cusumano, 2008; Kankanhalli et al., 2006; Kotlarsky and Oshri, 2005; Reed and Knight, 2010). Yet, they work together interdependently and use ICT to communicate and coordinate their time-constrained project tasks (Bourgault et al., 2008; Peters and Karren, 2009). In general, team diversity and space–time dispersion in GDVSD project teams are celebrated for stimulating creativity and allowing a variety of skills to be brought to bear on problems at hand (Kankanhalli et al., 2006). Yet, such characteristics may also cause problems in achieving successful collaboration by reducing team cohesion and increasing team conflict (Furumo, 2009; Kotlarsky et al., 2007; Pazos, 2012). The present study is confined to investigate causes and consequences of relationship conflict.

The relationship conflict arises due to disagreements between team members that are characterized by feelings of anger, hostility, frustration and distrust among team members (Furumo, 2009; Hinds and Bailey, 2003; Jehn and Mannix, 2001). The majority of previous studies provided evidence that relationship conflict decrease team performance and the intent of team members to remain in the team, while some others provided less conclusive evidence (Furumo, 2009; Jehn and Mannix, 2001). Hence, on the one hand, factors influencing virtual team performance are not well-understood. On the other hand, the dynamics of team leader support are not well-researched and understood (Thomas and Bostrom, 2010; Wakefield et al., 2008; Zakaria et al., 2004). Although team leader support is vital in alleviating team conflict and enhancing team performance, little empirical research has specifically examined the role of team leader support in GDVTs (Hertel et al., 2005; Thomas and Bostrom, 2010; Wakefield et al., 2008). Further, the majority of previous empirical studies were conducted on student project teams (Kankanhalli et al., 2006; Van Dick et al., 2008; Zhang and Wang, 2011) compared to the number of studies conducted on actual virtual teams (Peters and Karren, 2009; Saxena and Burmann, 2014; Thomas and Bostrom, 2010; Wakefield et al., 2008). Therefore, empirical studies examining team composition, relationship conflict and team leader support in achieving higher levels of team performance in virtual teams could make a significant contribution to the literature.

In the above context, the purpose of the study was to investigate whether diversity in team composition leads to relationship conflict, and, consequently, relationship conflict leads to team performance, and whether team leader support reduces (moderates) the negative effects of relationship conflict on team performance. For the present study, 216

139

Diversity in team

composition

team members working in GDVSD projects responded by revealing their actual project experiences. Consistent with the objectives, the rest of the article reviews, briefly, the theoretical background of the study. This is followed by the methodology adopted. Thereafter, the main findings are presented and discussed. The article concludes with a discussion on managerial implications of the findings and research areas for further inquiry and understanding.

Theoretical background and hypotheses Team composition Temporary, geographically and organizationally dispersed and electronically communicating work teams have been brought about to get projects completed in a minimum time while incorporating a wide range of cross-functional knowledge and expertise possessed by individual members into a collective body for effective problem-solving (Andres, 2002; Garrison et al., 2010; Van Knippenberg et al., 2004). Therefore, it is expected that the very nature of GDVT would bring different and complementary knowledge and expertise enhancing creativity among team members (Zakaria et al., 2004).

However, many GDVTs fail to meet expected project results (Zakaria et al., 2004). The literature suggests that although team composition offers potential benefits, it also presents major challenges in GDVTs (Peters and Karren, 2009; Saxena and Burmann, 2014; Zakaria et al., 2004). First, the temporary nature of virtual teams suggests that team membership is fluid, as it evolves according to changing project requirements influencing how they work – members do not share past history and they may not work together in the future (Kanawattanachai and Yoo, 2007). When team members have a short history together, it may negatively influence how they work together in a team (Gibson and Gibbs, 2006). Therefore, traditional coordination and control mechanisms are less effective in GDVTs due to the complexity of team dynamics.

Second, geographical dispersion implies that team members rarely, if ever, meet in a face-to-face setting (Kanawattanachai and Yoo, 2007; Saxena and Burmann, 2014; Schiller and Mandviwalla, 2007). Although they may rely on a combination of ICT, these may reduce non-verbal cues about interpersonal affections such as tone, warmth and attentiveness that could in turn reduce message clarity, delay in feedback and the interpretation of feedback (Kanawattanachai and Yoo, 2007; Kotlarsky and Oshri, 2005). In addition, when team members are residing in different time zones in various parts of the world, it may reduce opportunities for real-time collaboration (Pauleen and Yoong, 2001; Sarker and Sahay, 2004; Zhou et al., 2014). Therefore, time and distance are important boundaries that GDVT should cross in building rapport and developing long-term relationships.

Third, national and linguistic diversity also create challenges for effective GDVT collaboration (Kankanhalli et al., 2006; Kotlarsky and Oshri, 2005; Pauleen and Yoong, 2001; Zhou et al., 2014). Virtual teams often comprise members from a wide range of organizations, countries and continents, and may impact on identification in virtual teams, leading to unhealthy racial and national stereotypes (Au and Marks, 2012). When team members are situated across national and organizational boundaries, they may have differing attitudes toward hierarchy and authority which may influence how they operate as a team (Kanawattanachai and Yoo, 2007; Zhou et al., 2014). Further, national diversity can be identified as a barrier if team members of one nationality have negative

SO 8,2/3

140

feelings toward other nationalities believing that one’s own nationality is superior to others (Kankanhalli et al., 2006). Furthermore, distributed team work is also a linguistically bounded concept (Au and Marks, 2012; Zakaria et al., 2004; Zhou et al., 2014). Language will become a barrier if misunderstanding and frustration occur due to accents or lack of fluency (Duckworth, 2008). Such situations impact on daily communication and coordination between team members (Au and Marks, 2012). Au and Marks (2012, p. 282) conclude “linguistic differences can lead to the loss of information and communication problems as team members attempt to decipher their colleagues’ communications through their own cultural perspectives”. Therefore, team diversity presents enormous challenges to team members, team processes and, ultimately, team outcomes, although ICT makes it easier than ever to form GDVT consisting of members from different countries with varying time zones and cultures (Malhotra et al., 2007; Pauleen and Yoong, 2001; Zhou et al., 2014). Fourth, the literature identifies team member differences in terms of sex, age or tenure as demographic diversity, which could act as a barrier for effective GDVT (Kankanhalli et al., 2006). Demographic differences are, in general, found to encompass negative consequences for team processes such as communication and social integration (Homan et al., 2008).

Overall, GDVT can typically be conceived as a team of people working toward a common goal but separated by a number of boundaries, such as those of distance, time, organizational affiliation, nationality and language. Hence, team composition of GDVT could provide numerous implications for academics and practitioners.

Relationship conflict According to De Dreu and Weingart (2003), conflicts emerge from team members’ tension for real or perceived differences. Intra-team conflict is commonly defined by Jehn and Mannix (2001, p. 238) as “an awareness on the part of the parties involved of discrepancies, incompatible wishes, or irreconcilable desires”. The literature identifies three types of intra-team conflict, namely, relationship conflict, task conflict and process conflict (De Dreu and Weingart, 2003; Jehn and Mannix, 2001). Relationship conflict relates to interpersonal frictions and disagreements among team members on personal issues; task conflict relates to disagreements among team members about the team task at hand; process conflict relates to disagreements among team members on the way in which the task should be achieved (Jehn and Mannix, 2001). The three types of conflict have differential impact on team performance (De Jong et al., 2008). Relationship conflict distracts team members from the task; task conflict might increase team performance if it leads to better understanding of task issues; process conflict might increase time spent on deciding how to achieve the task (Martínez-Moreno et al., 2009). As mentioned in the Introduction, the present study is confined to investigate relationship conflict.

Relationship conflict (also known as affective, personalized or emotional conflict) may evoke due to interpersonal incompatibilities and frictions between team members resulting in tension, annoyance and animosity (Furumo, 2009; Hinds and Bailey, 2003; Jehn and Mannix, 2001) and interpersonal disagreements between team members on matters that are not directly related to project tasks (Thomas and Bostrom, 2010; Wakefield et al., 2008). For instance, interpersonal misunderstandings, attitude problems and damaged relations between team members could lead to relationship inadequacies (Lim and Suh, 2014; Thomas and Bostrom, 2010). In other words, relationship conflict is characterized by interpersonal issues involving mutual dislike,

141

Diversity in team

composition

personality clashes, the feelings of anger and frustration and distrust between team members (Hinds and Bailey, 2003; Jehn and Mannix, 2001; Thomas and Bostrom, 2010; Wakefield et al., 2008).

Scholars used similarity attraction theory and social identity theory to explain relationship conflict in GDVT (Kankanhalli et al., 2006). Similarity attraction theory suggests that people prefer to interact with similar rather than dissimilar people; social identity theory suggests that people like to be affiliated with others belonging to their own social category with some emotional and value significance (Kankanhalli et al., 2006). By using these theories, previous studies found that relationship conflict could occur when people belonging to diverse groups work together (Hong, 2010; Kankanhalli et al., 2006; Williams and O’Reilly, 1998). For instance, Hong (2010) states that individuals feel comfortable in interacting with those who are similar to them; cultural differences among team members may influence how they attach to and identify with one another within the virtual team. Williams and O’Reilly (1998) found that diverse teams are more likely to be less integrated and less communicated and have more conflict. Hence, as mentioned in the section on Team composition, relationship conflict could occur in GDVT when members work across cultural, national, linguistic, geographical and time boundaries.

According to the available literature, the effect of team composition on relationship conflict is not well-understood (Kankanhalli et al., 2006; Van Knippenberg et al., 2004; Williams and O’Reilly, 1998). Some previous studies provide evidence that teams composed with individuals from diverse backgrounds could be a liability to team performance (Kankanhalli et al., 2006; Van Knippenberg et al., 2004), while some other studies failed to provide conclusive evidence (Mortensen and Hinds, 2001). Therefore, empirical evidence for the effect of team composition on relationship conflict in GDVSD is not very clear, and this demands further investigations. Based on the literature reviewed above, it is proposed for the study that diversity in team composition leads to relationship conflict. Therefore, it is hypothesized:

H1. Teams composed with members from diverse backgrounds experience higher levels of relationship conflict.

Team performance Previous researchers conceptualized the dimensions of team performance under several broad areas. For instance, Faraj and Sproull (2000) classified team performance into two broad areas of team effectiveness (e.g. work quality and ability to meet project goals) and efficiency (e.g. adherence to time schedules and budgets). Pinto and Pinto (1990) classified team performance into two broad areas of task outcomes (e.g. adherence to the estimated schedule and budget) and psychosocial outcomes (e.g. degree of experienced friendliness and support).

In operatinalizing team performance, some researchers such as Huckman et al. (2007) and Henderson and Lee (1992) proposed objective measures (e.g. delivered source code instructions per man-hour and number of defects in acceptance testing), while some others such as Bourgault et al. (2008) and Staples and Webster (2007) proposed subjective measures (e.g. team autonomy and ability to cope). In this regard, Faraj and Sproull (2000) state that either objective measures are often unavailable or they are subject to manipulation and may reflect the specific accounting practices of a particular project. Therefore, Faraj and Sproull (2000) suggest that relying on subjective measures

SO 8,2/3

142

is a better source for team performance data. However, there are no agreed-upon measures to evaluate team performance in the literature. Table I summarizes some of the frequently used objective and subjective dimensions of team performance. It is apparent from Table I that some of the frequently used dimensions to evaluate team performance include multiple parameters such as adherence to time schedules (Huckman et al., 2007), adherence to budget (Faraj and Sproull 2000) and positive work experiences for team members (Staples and Webster, 2007). Of the previous research shown in Table I, the studies of Saxena and Burmann (2014), Huckman et al. (2007), Faraj and Sproull (2000) and Henderson and Lee (1992) addressed virtual software development setting.

Previous studies provide evidence for the negative effect of relationship conflict on team performance (Au and Marks, 2012; De Dreu and Weingart, 2003; Gibson and Gibbs, 2006; Wakefield et al., 2008). For instance, Au and Marks (2012) argued that language barriers due to cultural diversity coupled with poor communication via technology could reduce cooperative behavior of team members, which in turn reduce performance. Further, when relationship conflict is perceived to increase in a team, members wish to distance themselves from each other and cooperate with one another only to deal with just the task at hand (Au and Marks, 2012; Stark et al., 2014). Overall, the findings of these studies suggest that teams with unresolved relationship conflict lower cohesion, the effectiveness of information exchange and the intention of members to remain in the team and increase the time and energy consumption of team members associated with emotional disagreements. As a consequence, relationship conflict is detrimental to team performance. Therefore it is hypothesized:

H2. Relationship conflict lowers team performance.

Team leader support In GDVSD teams, members are located in various countries, participating in a project where their contributions are independent of each other, but the parts contributed by each member make a complete project (Ahuja, 2002). In this regard, Lacity and Willcocks (2014) provide evidence for assigning right leadership in the context of business process outsourcing for dynamic innovation. Previous studies provide evidence that team leaders play a crucial role in effective GDVT management by coordinating tasks, motivating team members, monitoring and/or facilitating collaboration and resolving conflict (Anh et al., 2012; Zakaria et al., 2004; Zhang and Wang, 2011). Further, some previous studies emphasize the importance of team leaders in social facilitation in the virtual team context to increase team unity and to create a cohesive work unit (Duckworth, 2008; Kargar and An, 2011; Zakaria et al., 2004). Furthermore, Wakefield et al. (2008) found a negative relationship between facilitating role of team leader and the level of relational conflict perceived by team members.

Therefore, literature suggests the importance of relational coordination or horizontal coordination for virtual teams (Anh et al., 2012; Duckworth, 2008; Kargar and An, 2011; Zakaria et al., 2004; Zhang and Wang, 2011). In this regard, Gittell (2000) states that work processes in distributed workplaces require reciprocal, iterative interactions among team members rather than sequential hand-offs performed by operators on a production line. According to Gittell (2000), relational coordination is characterized by frequent and timely problem-solving communication, helpfulness, shared goals, shared knowledge and mutual respect. Gittell (2000) found that supervisors, or in the case of GDVSD, team leaders, can support this relational coordination. In a similar vein, Barge

143

Diversity in team

composition

Table I. Team performance measures

A ut

ho rs

M ea

su re

H en

de rs

on an

d le

e (1

99 2)

F ar

aj an

d sp

ro ul

l (2

00 0)

H in

ds an

d ba

ile y

(2 00

3)

H in

ds an

d m

or te

ns en

(2 00

5)

H er

te l

et al

. (2

00 5)

H uc

km an

et al

. (2

00 7)

St ap

le s

an d

w eb

st er

(2 00

7)

B ou

rg au

lt et

al .

(2 00

8)

Sa xe

na an

d bu

rm an

n (2

01 4)

A bi

lit y

to co

pe (b

y te

am m

em be

rs )

✓ A

bi lit

y to

m ee

t pr

oj ec

t go

al s

✓ A

dh er

en ce

to ti

m e

sc he

du le

s ✓

✓ ✓

✓ ✓

A dh

er en

ce to

bu dg

et s

✓ ✓

✓ ✓

✓ A

dh er

en ce

to en

d- us

er sp

ec ifi

ca ti

on s

D el

iv er

ed so

ur ce

co de

in st

ru ct

io ns

pe r

m an

-h ou

r ✓

✓ E

xt en

t of

m ee

ti ng

de si

gn ob

je ct

iv es

✓ F

or m

al iz

at io

n of

de ci

si on

pr oc

es se

s ✓

In it

ia ti

ve of

th e

te am

✓ In

te nt

io n

to re

m ai

n on

th e

te am

✓ N

um be

r of

de fe

ct s

in ac

ce pt

an ce

te st

in g

✓ P

er ce

iv ed

in di

vi du

al pe

rf or

m an

ce ✓

P os

it iv

e w

or k

ex pe

ri en

ce s

fo r

te am

m em

be rs

/s at

is fa

ct io

n w

it h

th e

te am

✓ Q

ua lit

y of

de ci

si on

-m ak

in g

pr oc

es s

✓ ✓

R ep

ut at

io n

of w

or k

ex ce

lle nc

e/ w

or k

ex ce

lle nc

e ✓

✓ T

ea m

au to

no m

y ✓

T ea

m op

er at

io ns

/t ea

m ef

fe ct

iv en

es s

in pe

rf or

m in

g ta

sk s/

pe rc

ei ve

d te

am pe

rf or

m an

ce ✓

✓ ✓

✓ T

ea m

ef fi

ca cy

T ea

m ef

fi ci

en cy

in pe

rf or

m in

g ta

sk s

✓ ✓

✓ T

ec hn

ic al

in no

va ti

on ✓

U se

r fr

ie nd

lin es

s W

or k

qu al

it y/

qu al

it y

of th

e te

am w

or k

re su

lt s

✓ ✓

✓

N o te

: T

ab le

de ve

lo pe

d by

th e

au th

or s

SO 8,2/3

144

(1996) emphasized the importance of relational management, to which Barge (1996) refers to as the ability of leaders to develop interpersonal relations that foster a workable balance of cohesion and unity in the team. Therefore, team leaders in virtual environments are expected to project a wider degree of behavioral repertoires such as facilitating, mentoring, understanding, empathy and relationship building (Anh et al., 2012; Duckworth, 2008; Wakefield et al., 2008).

Therefore, some previous studies suggest that the managerial purpose of team leader changes from one of direct supervision to one of enabler or relationship moderator in GDVTs (Anh et al., 2012; Wakefield et al., 2008). In other words, diversity has potential value for teams because diverse teams may enhance team performance, while diversity may disrupt team processes and performance. Therefore, it is very difficult to apprehend the effect of diversity without taking moderators into account. In this regard, Ehrlich and Cataldo (2014) state that although team leaders are important for the smooth functioning of teams, their effect on team performance is not well-established. Wakefield et al. (2008) found that team leader’s ability to assume leadership roles to manage conflict before the conflict negatively impacts team outcomes is not statistically significant. In our study, we propose that the effect of relationship conflict on team performance will be reduced (moderated) by team leader support. Therefore, it is hypothesized:

H3. The negative effect of relationship conflict on team performance is moderated by team leader support.

The framework developed for the study is shown in Figure 1.

Methodology Sample and method of data collection A random sample of employees working in GDVSD project teams were selected from software development firms operating in Sri Lanka. In doing so, a contact person was identified at each firm. The contact persons together with one of the authors of this paper randomly selected software development project team members who have had completed a work assignment in a GDVSD project during the past 12-month period. Further, not more than three people were identified from the same development project. The contact persons distributed the link to on-line self-administered survey questionnaire via e-mail. A total of 216 usable responses resulted in 54 per cent response rate. Respondents were involved in software development work for client firms located in the USA, the EU and the Asia-Pacific region. Some of the broad job categories of respondents are software architect, principal software engineer, senior software engineer, software engineer, quality assurance engineer and business system analyst. The demographics of the sample are shown in Table II.

Diversity in team composition

Relationship conflict

Team performance

Team leader support

Figure 1. Research model

145

Diversity in team

composition

Table II. Sample characteristics

About respondents (%)

Gender Male 71 Female 29

Marital statues Single 49 Married 51

Age (in years) Mean 28.72 SD 3.43

Team size Mean 12.13 SD 3.94

Tenure in virtual teams (in years) Mean 4.17 SD 2.64

Highest level of education Degree or equivalent 79 Postgraduate degree 21 About other team members

Countries of other team membersa

Sweden 51 Norway 28 England 18 USA 17 India 14 Denmark 11 Bangladesh 8 China 8 The Netherlands 8 Pakistan 7 Austria 7 Germany 7 Arabic countries (various) 7 Canada 6 Indonesia 6 Japan 6 Thailand 5 Singapore 5 Spain 5 Finland 3 Australia 2 Mexico 2 African countries (various) 2 Russia 1

(continued)

SO 8,2/3

146

Measures Team composition was measured using a four-item scale developed based on the literature reviewed earlier. Relationship conflict was measured using a four-item scale adopted from Kankanhalli et al. (2006). Team leader support was measured using a three-item scale developed based on the literature reviewed earlier. Building on previous studies (Bourgault et al., 2008), team performance was conceptualized as team members’ perception of their own performance. A five-item scale was developed based on the literature reviewed earlier. Responses for team composition and relationship conflict were on a five-point Likert scale ranging from (1) never to (5) always. Responses for team leader support and team performance were on a five-point Likert scale ranging from (1) strongly disagree to (5) strongly agree.

Methods of data analysis Self-administered survey questionnaire was used for the data collection. The questionnaire was pre-tested with a random sample that fitted with the intended sample of the study prior to the distribution. When self-report measures from a single source are used to evaluate variables, the literature highlights the problem of common method bias (Podsakoff et al., 2003; Spector, 1994). However, Spector (1994) suggests that despite the weaknesses of the cross-sectional self-report methodology, this design can be quite useful in providing a picture of and inter-correlations among people’s job environment and their reactions to jobs, which can be useful for deriving hypotheses about how people react to jobs. Therefore, procedural and statistical measures were taken to reduce common method bias in the present study. The anonymity of respondents was ensured to reduce evaluation apprehension. Factor analysis was used to conduct Harman’s single-factor test; neither single factor emerged from this analysis nor was there a general factor that could account for the majority of variance fulfilling the recommended guidelines of Podsakoff et al. (2003). Data were tested for appropriate:

• internal consistency reliability; • factor structure; • convergent validity; and • construct reliability to identify any issues with multicollinearity and response

bias.

Table II.

About respondents (%)

Religions of other team membersa

Catholic 63 Christian 59 Islam 33 Hindu 32 Buddhist 27 Other 4

Note: a Does not add up to 100 due to multiple answers

147

Diversity in team

composition

With regard to the procedure used for the analysis of data, mean and standard deviation were used to describe the data. Correlation analysis was used to understand the nature of relationship between variables. To examine the hypothesized relationships, structural equation modeling (SEM) with maximum likelihood estimation was performed using AMOS 16. In this regard, Kremelberg (2011, p. 360) states that SEM is a very new and more powerful method than methods such as regression. As we used already available item measures from the literature to measure our latent variables, the full model was tested with SEM (Kremelberg, 2011). Fit measures relating to absolute fit (CMIN/df, GFI), relative fit (CFI) and parsimonious fit (PRATIO) and fit measures based on the non-central chi-square distribution (RMSEA, PCLOSE) were used to evaluate the model– data fit.

Results and discussion The items, factor loadings and reliability statistics of the four measures are shown in Table III. Specifically, the measure on diversity in team composition had Cronbach’s alpha reliability of 0.734; principal component factor analysis yielded one factor. The measure on relationship conflict had Cronbach’s alpha reliability of 0.793; principal component factor analysis yielded one factor. The measure on team leader support had Cronbach’s alpha reliability of 0.811; principal component factor analysis yielded one factor. The measure on team performance had Cronbach’s alpha reliability of 0.789; principal component factor analysis yielded one factor.

Table IV shows means, standard deviations and zero-order correlations between the variables. As can be seen in Table IV, a significant positive correlation (Pearson r � 0.356, p � 0.01) between diversity in team composition and relationship conflict supports the arguments made in H1. A significant negative correlation (Pearson r � �0.148, p � 0.05) between relationship conflict and team performance supports the arguments made in H2. A significant positive correlation (Pearson r � 0.273, p � 0.01) between team leader support and team performance together with a significant negative correlation (Pearson r � �0.165, p � 0.05) between team leader support and relationship conflict support the arguments made in H3. However, correlation and its associated significance does not imply causality; therefore, relationships should be assessed using the coefficient of determination (Hole, 2015) in SEM. Overall, correlations between variables of interest suggest a better prediction in SEM.

The structural model achieved a good level of fit having CMIN/df � 2.583, GFI � 0.871, CFI � 0.914 and RMSEA � 0.068. The model satisfies the requirements of goodness-of-fit criteria (Byrne, 2001). The summary of analysis is provided in Table V. Standardized regression coefficient of 0.433 (p � 0.001) reveals that diversity in team composition significantly positively predicts relationship conflict. This supports H1. Standardized regression coefficient of �0.406 (p � 0.001) reveals that relationship conflict significantly negatively predicts team performance. This supports H2. As predicted in H3, the negative regression coefficient between relationship conflict and team performance (�0.406, p � 0.001) is reduced by the interaction term (�0.305, p � 0.001), suggesting team leader support as a moderator in the above relationship. This supports H3. Further, when the full structural model is taken into account, the coefficient of determination of 0.312 suggests that the variables of interest account for 31 per cent of the variation of team performance.

SO 8,2/3

148

Overall, the results of the study support the arguments made by us where diversity in team composition operates as an antecedent to relationship conflict, which in turn affects team performance; team leader support operates as a moderator on the latter relationship.

Table III. Summarized results:

validity and reliability

Items Estimate Cronbach’s alpha AVE

Construct reliability

Diversity in team composition 0.734 0.704 0.904 I collaborate with team members whose nationality is different to mine 0.872 I collaborate with team members whose native language is different to mine 0.866 I collaborate with team members who are located in different geographical regions with different time zones 0.811 I collaborate with team members who belong to different age groups than mine 0.805 Relationship conflict 0.793 0.682 0.896 My team members confronted each other on personal matters 0.862 My team members made negative remarks about each other 0.829 Some of my team members tended to ridicule others 0.820 The differences experienced by my team were interpersonal-related 0.791 Team leader support 0.811 0.732 0.891 My team leader had ability to communicate details about team members 0.867 My team leader had knowledge and understanding about the globally distributed project work environment 0.893 My team leader had ability to influence team members 0.805 Team performance 0.789 0.715 0.926 My team adhered to time schedules 0.872 My team adhered to budget 0.882 My team was able to meet project goals 0.807 My team was able to meet design objectives 0.861 My team has had reputation for work excellence 0.803

Table IV. Correlations

No. Variables Mean SD 1 2 3

1 Diversity in team composition 3.90 0.50 1 2 Relationship conflict 3.67 0.59 0.356** 1 3 Team leader support 2.17 0.69 0.132 �0.165* 1 4 Team performance 3.43 0.66 �0.180** �0.148* 0.273**

Notes: * Significant at the 0.05 level; ** significant at the 0.01 level

149

Diversity in team

composition

Conclusions and implications The available literature provides little empirical evidence on the relationship between team diversity, relationship conflict, team leader support and team performance as manifested in the attitudes and expectations of knowledge workers. Drawing upon a random sample of software developers engaged in globally distributed project teams in Sri Lanka, we empirically tested whether diversity in team composition positively related to relationship conflict, relationship conflict negatively related to team performance and team leader support moderates the harmful effects of relationship conflict on team performance. The novel character of this research is that it brought together and analyzed, in parallel, several factors which until recently had been analyzed separately.

It was found that respondents work in virtual teams that comprised diverse set of team members. With regard to relationship conflict, a mean value of 3.67 suggests a considerably high level of conflict between team members. Further, a positive correlation was found between diversity in team composition and relationship conflict. This supports the findings of previous studies such as Peters and Karren (2009) and Zakaria et al. (2004). With regard to team leader support, a mean value of 2.17 suggests a low level of team leader support in terms of communicating details of team members, team leader’s knowledge and understanding about the project work environment and team leader’s ability to influence team members in creating a sense of shared space when team members are located apart. With regard to team performance, a mean value of 3.43 suggests a moderate level of team performance in terms of time schedules, budget, project goals, design objectives and teams’ reputation for work excellence. It was also found that diversity in team composition leads to relationship conflict, which in turn negatively impacts on their performance. This supports the findings of previous studies such as Hinds and Bailey (2003) and Kankanhalli et al. (2006). The findings also supported our prediction that team leader support moderates the link between relationship conflict and team performance. Specifically, team leader support could reduce the negative effect of relationship conflict on team performance.

The findings of the study have both theoretical and practical implications. First, both academics and practitioners need valid information to better understand the influence of team leader support on relationship conflict and team performance in the context of GDVSD. This becomes important because the popularity of GDVSD is relatively recent and knowledge workers engaged in GDVT are growing, while scholarly research in this area is presently lacking. In this context, our research investigated relationships between diversity in team composition, relationship conflict, team leader support and team performance. By doing so, the study produced new insight beyond prior studies to the literature. Such extended investigations become important, as the role of team leader

Table V. Summary of the results–structural model coefficients

Path Standardized regression estimate (significance)

Coefficient of determination

Diversity in team composition ¡ Relationship conflict 0.433 (p�0.001) 0.187 Relationship conflict ¡ Team performance �0.406 (p�0.001) 0.312 Team leader support ¡ Team performance 0.144 (p�0.001) Team leader support � Relationship conflict ¡ Team performance

�0.305 (p�0.001)

SO 8,2/3

150

support has not been investigated so far in relation to GDVT with the other three variables (relationship conflict, diversity in team composition and team performance). We have not come across previous survey-based studies that investigated the links proposed in this study to make straightforward comparisons. Therefore, there is a lack of research investigating these four essential elements in totality within an extended framework. However, considerable future research is necessary to validate the links suggested in this study. The present study represents a step in that direction. Second, the findings of our study imply the importance of relationship building between virtual team members to reduce relationship conflict; findings could help academics and practitioners to design interventions to manage causes and consequences of relationship conflict. Third, the findings imply that when selecting virtual team members, the demands of globally distributed work context itself should be acknowledged. For instance, in the GDVT context, most of the team communication may occur in written form. This excludes most of the body language, vocal inflection and pacing and other social cues available in face-to-face interactions. Further, for some team members, communications will also be occurring in a language other than their native tongue. Furthermore, to reduce occurrences of relationship conflict in globally distributed teams, members’ prior experience in working across borders and their cross-cultural competencies could be taken into granted.

Fourth, the findings also imply the role of team leaders in reducing the harmful effects of relationship conflict on team performance. Therefore, the role of team leader in identifying conflict situations and in reducing its effect on team performance is vital. The findings also imply the need of providing training to team leaders to create a cohesive team that delivers on project goals.

Limitations of the study and areas for future research The data were collected from a considerably large homogeneous sample of individuals engaged as team members in GDVSD project teams in Sri Lanka. The study relied on self-reported data. However, as mentioned in the section on Methods of data collection, a considerable attempt has been made to overcome problems that may occur due to common method variance. With regard to areas for future research, the validity of a measure cannot be truly established on the basis of a single cross-sectional study. The validation of a measure requires the assessment of measurement properties over a variety of samples in similar and different contexts. Hence, future research, in different samples and longitudinal studies, is necessary that complements questionnaire surveys with interviews. The study relied on experiences of respondents working in Sri Lanka on GDVSD projects. As a result, they revealed their experiences of their immediate project leaders in Sri Lanka. Future research could overcome this limitation by expanding the research to include global virtual team members in other locations. Future research could incorporate other variables of interest such as how global virtual team leaders exercise power to resolve relationship conflict, what role communication media plays in supporting team leaders and what skills and resources virtual team leaders need to moderate the link between relationship conflict and team performance. Our study was confined to investigating the relationship conflict, which is detrimental to team performance. Future studies could investigate other types of conflict such as task-related. Unlike relationship conflict, task conflict may have positive results by influencing internal competition leading to innovation.

151

Diversity in team

composition

References Ahuja, S. (2002), “Distributed teams in e-work organisations: the new issues”, Journal of Services

Research, Vol. 2 No. 2, pp. 79-95.

Andres, H.P. (2002), “A comparison of face-to-face and virtual software development teams”, Team Performance Management: An International Journal, Vol. 8 Nos 1/2, pp. 39-48.

Anh, N.D., Cruzes, D.S. and Conradi, R. (2012), “Dispersion, coordination and performance in global software teams: a systematic review”, ESEM’12, 19-20 September, Lund, pp. 129-139.

Au, Y. and Marks, A. (2012), “Virtual teams are literally and metaphorically invisible”, Employee Relations, Vol. 34 No. 3, pp. 271-287.

Barge, J.K. (1996), “Leadership skills and the dialectics of leadership in group decision making”, in Hirokawa, R.Y. and Poole, M.S. (Eds), Communication and Group Decision Making, Sage, Thousand Oaks, CA, pp. 301-342.

Bell, S.B. and Kozlowski, S.W. (2002), “A typology of virtual teams: implications for effective leadership”, Group & Organization Management, Vol. 27, pp. 14-49.

Bourgault, M., Drouin, N. and Hamel, E. (2008), “Decision making within distributed project teams: an exploration of formalization and autonomy as determinants of success”, Project Management Journal, Vol. 39, pp. S97-S110.

Byrne, B.M. (2001), Structural Equation Modelling with AMOS: Basic Concepts, Applications and Programming, Lawrence Erlbaum, London.

Curşeu, P.L. and Wessel, I. (2005), “Information processing in virtual teams: implications for virtual team effectiveness”, in Curşeu, P.L. (Ed.), Complexity within Organizations, Pearson Education, Harlow.

Cusumano, M.A. (2008), “Managing software development in globally distributed teams”, Communications of the ACM, Vol. 51 No. 2, pp. 15-17.

De Dreu, C.K.W. and Weingart, L.R. (2003), “Task versus relationship conflict, team performance, and team member satisfaction: a meta-analysis”, Journal of Applied Psychology, Vol. 88 No. 4, pp. 741-749.

De Jong, R., Schalk, R. and Curşeu, P.L. (2008), “Virtual communicating, conflicts and performance in teams”, Team Performance Management: An International Journal, Vol. 14 Nos 7/8, pp. 364-380.

Duckworth, H. (2008), “How TRW automotive helps global virtual teams perform at the top of their game”, Global Business and Organizational Excellence, November/December, pp. 6-16. doi: 10.1002/joe.20237.

Ehrlich, K. and Cataldo, M. (2014), “The communication patterns of technical leaders: impact on product development team performance”, CSCW’14, 15-19 February, Baltimore, MD, pp. 733-744.

Faraj, S. and Sproull, L. (2000), “Coordinating expertise in software development teams”, Management Science, Vol. 46, pp. 1554-1568.

Furumo, K. (2009), “The impact of conflict and conflict management style on deadbeats and deserters in virtual teams”, Journal of Computer Information Systems, Vol. 49 No. 4, pp. 66-73.

Garrison, G., Wakefield, R.L., Xu, X. and Kim, S.H. (2010), “Globally distributed teams: the effect of diversity on trust, cohesion and individual performance”, The DATA BASE for Advances in Information Systems, Vol. 41 No. 3, pp. 27-48.

SO 8,2/3

152

Gibson, C.B. and Gibbs, J.L. (2006), “Unpacking the concept of virtuality: the effects of geographic dispersion, electronic dependence, dynamic structure, and national diversity on team innovation”, Administrative Science Quarterly, Vol. 51 No. 3, pp. 451-495.

Gittell, J.H. (2000), “Organizing work to support relational coordination”, International Journal of Human Resource Management, Vol. 11 No. 3, pp. 517-539.

Henderson, J.C. and Lee, S. (1992), “Managing I/S design teams: a control theories perspective”, Management Science, Vol. 18 No. 6, pp. 757-777.

Hertel, G., Geister, S. and Konradt, U. (2005), “Managing virtual teams: a review of current empirical research”, Human Resource Management Review, Vol. 15, pp. 69-95.

Hinds, P.J. and Bailey, D.E. (2003), “Out of sight, out of sync: understanding conflict in distributed teams”, Organization Science, Vol. 14 No. 6, pp. 615-632.

Hinds, P.J. and Mortensen, M. (2005), “Understanding conflict in geographically distributed teams: the moderating effects of shared identity, shared context, and spontaneous communication”, Organization Science, Vol. 16 No. 3, pp. 290-307.

Hole, G. (2015), Research Skills One, Correlation Interpretation, University of Sussex, available at: www.sussex.ac.uk/Users/grahamh/RM1web/Eight%20things%20you%20need%20to% 20know%20about%20interpreting%20correlations.pdf (accessed 10 March 2015).

Homan, A.C., Hollenbeck, J.R., Humphrey, S.E., Van Knippenberg, D., Ilgen, D.R. and Van Kleef, G.A. (2008), “Facing differences with an open mind: openness to experience, salience of intragroup differences, and performance of diverse work groups”, Academy of Management Journal, Vol. 51 No. 6, pp. 1204-1222.

Hong, H. (2010), “Bicultural competence and its impact on team effectiveness”, International Journal of Cross Cultural Management, Vol. 10 No. 1, pp. 93-120.

Huckman, R.S., Staats, B.R. and Upton, D.M. (2007), “Team familiarity, role experience, and performance: evidence from Indian software services”, Working Paper, available at: www. hbs.edu/research/pdf/08-019.pdf (accessed 26 September 2011).

Jehn, K.A. and Mannix, E.A. (2001), “The dynamic nature of conflict: a longitudinal study of intragroup conflict and group performance”, Academy of Management Journal, Vol. 44, pp. 238-251.

Kanawattanachai, P. and Yoo, Y. (2007), “The impact of knowledge coordination on virtual team performance over time”, MIS Quarterly, Vol. 31 No. 4, pp. 783-808.

Kankanhalli, A., Tan, B.C.Y. and Wei, K.-K. (2006), “Conflict and performance in global virtual teams”, Journal of Management Information Systems, Vol. 23 No. 3, pp. 237-274.

Kargar, M. and An, A. (2011), “Discovering Top-k teams of experts with/without a leader in social networks”, CIKM’11, 24-28 October, Glasgow, pp. 985-994.

Kotlarsky, J. and Oshri, I. (2005), “Social ties, knowledge sharing and successful collaboration in globally distributed system development projects”, European Journal of Information Systems, Vol. 14, pp. 37-48.

Kotlarsky, J., Oshri, I., Kumar, K. and Van Hillegersberg, J. (2007), “Globally distributed component-based software development: an exploratory study of knowledge management and work division”, Journal of Information Technology, Vol. 22 No. 2, pp. 161-173.

Kremelberg, D. (2011), Practical Statistics – A Quick and Easy Guide to IBM SPSS Statistics, STATA, and Other Statistical Software, Sage Publications, Thousand Oaks, CA.

Lacity, M. and Willcocks, L. (2014), “Business process outsourcing and dynamic innovation”, Strategic Outsourcing: An International Journal, Vol. 7 No. 1, pp. 66-92.

153

Diversity in team

composition

Lim, H. and Suh, B. (2014), “Now here or nowhere: conflict resolution strategies for intimate relationship in diverse geographical contexts”, CSCW’14 Companion, 15-19 February, Baltimore, MD, pp. 197-200.

Malhotra, A., Majchrzak, A. and Rosen, B. (2007), “Leading virtual teams”, Academy of Management Perspectives, Vol. 21 No. 1, pp. 60-70.

Martínez-Moreno, E., González-Navarro, P., Zornoza, A. and Ripoll, P. (2009), “Relationship, task and process conflicts on team performance”, International Journal of Conflict Management, Vol. 20 No. 3, pp. 251-268.

Mishra, D. and Mahanty, B. (2014), “The effect of onsite-offshore work division on project cost, schedule, and quality for re-engineering projects in Indian outsourcing software industry”, Strategic Outsourcing: An International Journal, Vol. 7 No. 3, pp. 198-225.

Mortensen, M. and Hinds, P.J. (2001), “Conflict and shared identity in geographically distributed teams”, International Journal of Conflict Management, Vol. 12 No. 3, pp. 212-238.

Pauleen, D.J. and Yoong, P. (2001), “Relationship building and the use of ICT in boundary-crossing virtual teams: a facilitator’s perspective”, Journal of Information Technology, Vol. 16, pp. 205-220.

Pazos, P. (2012), “Conflict management and effectiveness in virtual teams”, Team Performance Management: An International Journal, Vol. 18 Nos 7/8, pp. 401-417.

Peters, L. and Karren, R.J. (2009), “An examination of the roles of trust and functional diversity on virtual team performance ratings”, Group & Organization Management, Vol. 34 No. 4, pp. 479-504.

Pinto, J.K. and Pinto, M.B. (1990), “Project team communication and cross-functional cooperation in new program development”, Journal of Product Innovation and Management, Vol. 7 No. 3, pp. 200-212.

Podsakoff, P., Mackenzie, S., Lee, J.-Y. and Podsakoff, N. (2003), “Common method biases in behavioral research: a critical review of the literature and recommended remedies”, Journal of Applied Psychology, Vol. 88 No. 5, pp. 879-903.

Reed, A.H. and Knight, L.V. (2010), “Project risk differences between virtual and co-located teams”, Journal of Computer Information Systems, pp. 19-30.

Sarker, S. and Sahay, S. (2004), “Implications of space and time for distributed work: an interpretive study of US-Norwegian system development teams”, European Journal of Information Systems, Vol. 13 No. 1, pp. 3-20.

Saxena, A. and Burmann, J. (2014), “Factors affecting team performance in globally distributed setting”, SIGMIS-CPR’14, 29-31 May, Singapore, pp. 25-33.

Schiller, S.Z. and Mandviwalla, M. (2007), “Virtual team research – an analysis of theory use and a framework for theory appropriation”, Small Group Research, Vol. 38 No. 1, pp. 12-59.

Spector, P.E. (1994), “Using self-report questionnaires in OB research: a comment on the use of a controversial method”, Journal of Organizational Behavior, Vol. 15 No. 5, pp. 385-392.

Staples, D.S. and Webster, J. (2007), “Exploring traditional and virtual team members’ ‘best practices’ – a social cognitive theory perspective”, Small Group Research, Vol. 38 No. 1, pp. 60-97.

Stark, E., Bierly, P. and Harper, S.R. (2014), “The interactive influences of conflict, task interdependence and cooperation on perceptions of virtualness in co-located teams”, Team Performance Management, Vol. 20 Nos 5/6, pp. 221-241.

Thomas, D.M. and Bostrom, R.P. (2010), “Vital signs for virtual teams: an empirically developed trigger model for technology adaptation interventions”, MIS Quarterly, Vol. 34 No. 1, pp. 115-142.

SO 8,2/3

154

Van Dick, R., Van Knippenberg, D., Hägele, S., Guillaume, Y.R.F. and Brodbeck, F.C. (2008), “Group diversity and group identification: the moderating role of diversity beliefs”, Human Relations, Vol. 61 No. 10, pp. 1463-1492.

Van Knippenberg, D., De Dreu, C.K.W. and Homan, A.C. (2004), “Work group diversity and group performance: an integrative model and research agenda”, Journal of Applied Psychology, Vol. 89, pp. 1008-1022.

Wakefield, R.L., Leidner, D.E. and Garrison, G. (2008), “A model of conflict, leadership, and performance in virtual teams”, Information Systems Research, Vol. 19 No. 4, pp. 434-455.

Williams, K.Y. and O’reilly, C.A. (1998), “Demography and diversity in organizations: a review of 40 years of research”, in Staw, B.M. and Cummings, L.L. (Eds), Research in Organizational Behavior, Vol. 20, JAI Press, Greenwich, CT, pp. 77-140.

Zakaria, N., Amelinckx, A. and Wilemon, D. (2004), “Working together apart? Building a knowledge-sharing culture for global virtual teams”, Creativity and Innovation Management, Vol. 13 No. 1, pp. 15-29.

Zhang, C. and Wang, J.A. (2011), “Effects of communication, leadership, and team performance on successful it capstone projects: a case study”, SIGITE’11, 20-22 October, West Point, NY, pp. 281-286.

Zhou, N., Gifford, W.M., Ratakonda, K., Westerwick, G.H. and Engel, C. (2014), “On the quantification of global team performance and profitability”, SCC’14 Proceedings of the 2014 IEEE International Conference on Services Computing, IEEE Computer Society Washington, DC, pp. 378-385. doi. 10.1109/SCC.2014.57.

Further reading Scholarios, D. and Marks, A. (2004), “Work-life balance and the software worker”, Human

Resource Management Journal, Vol. 14 No. 2, pp. 54-74.

Corresponding author Vathsala Wickramasinghe can be contacted at: [email protected]

For instructions on how to order reprints of this article, please visit our website: www.emeraldgrouppublishing.com/licensing/reprints.htm Or contact us for further details: [email protected]

155

Diversity in team

composition

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.