Assessing the Influence of Virtuality on theEffectiveness of Engineering Project Networks:

“Big Five Theory” PerspectiveM. Reza Hosseini, M.ASCE1; Igor Martek2; Nicholas Chileshe3;Edmundas Kazimieras Zavadskas4; and Mehrdad Arashpour5

Abstract: Engineering project networks (EPNs), as digitally mediated units whose team members are scattered across differentorganizations/locations, have become the norm for construction project teams. In EPNs, virtuality captures various forms of boundariesamong team members and influences major dimensions of team functioning and outcomes. Despite this importance, investigating the roleof virtuality on the functional performance of EPNs has remained an unexplored area. Through a multidisciplinary review of literature, atheoretical model has been created to conceptualize the fabric of links between virtuality and effectiveness in EPNs. The theoretical basis ofthe model is the Big Five theory of team effectiveness. The model has been evaluated using empirical data collected from 285 completedquestionnaires, and analyzed through the use of partial least squares structural equation modeling (PLS-SEM) technique. Quantification ofassociations reveals that virtuality significantly affects effectiveness in teams by manipulating several mediators. However, the level of in-fluence of virtuality on these mediators, and eventually effectiveness, is much lower than previously assumed by current literature. This studygoes beyond existing studies, presenting the first quantified model of virtuality applicable to any type of EPN in the construction context.DOI: 10.1061/(ASCE)CO.1943-7862.0001494. © 2018 American Society of Civil Engineers.

Author keywords: Engineering project networks; Virtual teams; Virtuality; Big Five theory; Effectiveness; Construction projects.

Introduction

Construction projects are collaborative ventures that are designedand constructed by construction project teams (Son and Rojas2011). As defined by Buvik and Rolfsen (2015), constructionproject teams are “ : : : a group of people responsible for complextasks over a limited period, and are typically cross-functional, con-sisting of members who have complementary skills and come fromdifferent disciplines and functional areas in the organization.”For such teams, face-to-face collaborations are considered expen-sive, disruptive, and inefficient (Oraee et al. 2017). Consequently,the work is more often executed by geographically dispersed, dig-itally mediated teams of knowledge specialists who are organizedinto engineering project networks (EPNs) (Franz et al. 2017;

Zelkowicz et al. 2015). Conducting business in the constructioncontext relies on EPNs (Lobo and Whyte 2017; Nitithamyongand Skibniewski 2011). Hence, interactions among membersand the way such interactions affect different features in respectof the functioning of EPNs have become areas of particular interest(Bassanino et al. 2014), especially in which the phenomenon ofvirtuality plays a pivotal role (Hosseini et al. 2016).

Virtuality is a key factor affecting every aspect of a team’s func-tions and members’ interactions (Foster et al. 2015; Penarroja et al.2013). It is dictated by several considerations that reflect the impactof boundaries, such as cultural, nationality and time zone differen-ces, geographical distance, working for different organizations, andsimilar dividing characteristics present in the work context thatteam members have to bridge in order to interact and communicate(Zelkowicz et al. 2015).

Investigators such as Hosseini et al. (2016) identified perceiveddistance, task nature, difference in context, team tenure and matu-rity, dissimilarity of knowledge and skills, and team size as the fac-tors that affect virtuality in EPNs. Yet, the weight and strength oftheir impact on virtuality remain unknown (Schaubroeck and Yu2017). Moreover, findings of previous studies with regard to theimpacts of virtuality on team functioning and outcomes are mixed(Marlow et al. 2017; Merschbrock and Munkvold 2015). In view ofthese considerations, investigating the impacts of virtuality onEPNs and the weight and strength of factors that affect virtualityare areas yet to be explored (Hosseini et al. 2016; Merschbrockand Munkvold 2014). Through the lens of the Big Five theory, thisstudy aims at developing a model for virtuality for EPNs thatanswers two questions:1. What are the weights and levels of influence of factors that affect

virtuality in EPNs?2. What are the major impacts of virtuality and its level of influ-

ence on the functioning of EPNs?

1Lecturer, School of Architecture and Built Environment, DeakinUniv., Geelong 3220, Australia (corresponding author). ORCID:https://orcid.org/0000-0001-8675-736X. Email: Reza.Hosseini@deakin.edu.au

2Lecturer, School of Architecture and Built Environment, Deakin Univ.,Geelong 3220, Australia. Email: Igor.Martek@deakin.edu.au

3Associate Professor, School of Natural and Built Environments,Univ. of South Australia, Adelaide 5001, Australia. Email: Nicholas.Chileshe@unisa.edu.au

4Professor, Institute of Sustainable Construction, Faculty of CivilEngineering, Vilnius Gediminas Technical Univ., Sauletekio Ave. 11,Vilnius 10223, Lithuania. Email: Edmundas.Zavadskas@vgtu.lt

5Senior Lecturer, Dept. of Civil Engineering, Monash Univ., 23 CollegeWalk (B60), Clayton 3800, VIC, Australia. Email: mehrdad.arashpour@monash.edu

Note. This manuscript was submitted on April 5, 2017; approved onMarch 9, 2018; published online on May 12, 2018. Discussion period openuntil October 12, 2018; separate discussions must be submitted for indivi-dual papers. This paper is part of the Journal of Construction Engineeringand Management, © ASCE, ISSN 0733-9364.

© ASCE 04018059-1 J. Constr. Eng. Manage.

J. Constr. Eng. Manage., 2018, 144(7): 04018059

Dow

nloa

ded

from

asc

elib

rary

.org

by

Mon

ash

Uni

vers

ity o

n 11

/15/

18. C

opyr

ight

ASC

E. F

or p

erso

nal u

se o

nly;

all

righ

ts r

eser

ved.

Background

The current manner in which teams work in the constructionindustry entails linking project members in EPNs electronically,as well as transferring and processing project data to and from thesedispersed participants (Alin et al. 2013; Becerik-Gerber et al. 2012;Oraee et al. 2017). The use of EPNs is on the rise for severalreasons. First, the business environment is changing while the con-struction workforce is becoming increasingly globalized (Iorio andTaylor 2014; Nitithamyong and Skibniewski 2011; Zelkowicz et al.2015). In addition, economic fluctuations have resulted in fiercecompetition, which necessitates higher performance, access toglobal best talents, and international partnering (Chinowsky andSonger 2011; Zelkowicz et al. 2015). Moreover, due to the adventof technological innovations, including building informationmodeling (BIM) alongside virtual and augmented reality, virtualcollaboration forms, such as EPNs are becoming mainstream(Oraee et al. 2017). Accelerated by technical advancements inWeb and cloud computing (Petri et al. 2017), these have resultedin a rapid growth of EPNs on construction projects (Alreshidi et al.2018; Merschbrock and Munkvold 2015; Mignone et al. 2016).

Virtuality

Members in EPNs have to cross several boundaries, such as organi-zational culture and practices, as well as geographic distance, tocommunicate and interact (Iorio and Taylor 2014; Mignoneet al. 2016; Zhang and Ng 2013). These boundaries can be collec-tively conceptualized under the term virtuality as an attribute ofsuch networks (Hosseini et al. 2016; Penarroja et al. 2013;Schaubroeck and Yu 2017). Virtuality is the measure of proximityalong a continuum in which one extreme is the traditional co-location of persons in a singular physical place, and the otherextreme is where interacting persons do so from wholly mutuallyremote locations (Martins and Schilpzand 2011). As dichotomousas such a linear representation may appear to be—people are eitherphysically near or not physically near—those elements that sepa-rate people or bring people together are in fact complex andmultifaceted. Office arrangements and layouts; connective andcommunicative technologies; meeting, reporting, and work culturepractices; among many other factors, all combine to enhance ordetract from the effects of proximate work experience. Conse-quently, an examination of the singular concept of virtuality neces-sitates the actual examination of the multidimensional phenomenon(Dulebohn and Hoch 2017; Hosseini et al. 2015).

Factors Affecting Virtuality

Distance among members and the consequent increase in technol-ogy use has been treated as the key dimensions of virtuality(Dulebohn and Hoch 2017; Foster et al. 2015; Gilson et al. 2015).Yet, as asserted by Schaubroeck and Yu (2017), a variety of otherfactors, apart from distance, are influential. Six broad categoriesidentified include temporal, cultural, spatial, organizational, workgroup, and relationship (Gilson et al. 2015). Indeed, a review ofliterature reveals that there is no agreement on what factors affectvirtuality (Dulebohn and Hoch 2017). That is because virtuality is acontext-specific phenomenon, hence factors affecting virtuality dif-fer according to the context at hand (Ibrahim et al. 2013; Maynardet al. 2012).

In response to this disparity, Hosseini et al. (2016) identified themain factors that affect virtuality and customized them for the con-text of EPNs. Key factors were task nature, the number of teammembers (team size), and the skills and knowledge of these mem-bers. Other factors identified were disparity of context among

members (termed as degree of dispersion), history of team collabo-ration (prior ties), and maturity of a team in supply chain relation-ships management. Hosseini et al. (2016), however, deployed aqualitative approach, thus the level of influence and strength of im-pacts of these factors on virtuality were not quantified. In fact,measuring the weight and strength of the factors that affect virtual-ity has remained an underresearched area in the literature(Schaubroeck and Yu 2017).

Impacts of Virtuality on Team Effectiveness (Big FiveTheory)

Team effectiveness provides a holistic insight into how a team per-forms (i.e., how the team completed the team’s task), together withhow the team interacted in achieving the final outcome (Franz et al.2017; Salas et al. 2005). With this in mind, the Big Five theory ofteam effectiveness was used in the present study to conceptualizethe impacts of virtuality on team functioning. The Big Five theory,as proposed by Salas et al. (2005), provides an appropriatetheoretical lens to examine teamwork effectiveness (Kay et al.2006). It was utilized in the present study because, according toFransen et al. (2011), the Big Five theory is “[t]he best knownframework for teamwork.”

The input-mediator-outcome (IMO) model incorporates alltypes of factors and mechanisms that link inputs to the effectivenessfor teams (Mathieu et al. 2008) (Fig. 1). According to IMO, inputsrepresent the composition of a team, its structure, the organizationaldesign, and the nature of the problem which provides the focus ofactivities for the team (Bosch-Sijtsema et al. 2011; Mathieu et al.2008). The activities and interactions among team members, in-cluding their motivation, cognition, and affect make up the medi-ators that enable them to utilize the team’s resources synergistically(Bosch-Sijtsema et al. 2011). According to the Big Five theory, fivekey factors together with three supporting processes have activeroles as mediators of team effectiveness. The five factors are lead-ership, backup behavior, mutual performance monitoring, teamorientation, and adaptability. The three supporting processes aremutual trust, communication, and shared mental models (Salaset al. 2005). The Big Five theory is a theory applicable to all typesof teams; however, the generic form of the Big Five theory is to becustomized for specific contexts or particular types of teams(Fransen et al. 2011). Through a qualitative approach, Hosseini(unpublished data, 2016) provided a revised version of the Big Fivetheory for construction teams. According to this model, factorsapplicable to construction project teams were backup behavior,leadership, team orientation, and accountability. Supporting proc-esses comprised communication, mutual trust, social interactions,and shared mental models.

Outcomes are the results expected of teams and cover a widerange of factors that are encapsulated in the concept of team effec-tiveness (Kozlowski and Ilgen 2006). The effectiveness of a team ismeasured by indicators such as (1) task-related indicators, namely,team performance (i.e., quantity, quality, timeliness, and customersatisfaction); and (2) affect-related indicators, such as team membersatisfaction (Kozlowski et al. 2015). In the same vein, in theirauthoritative work, Mathieu et al. (2008) maintained that the teameffectiveness framework should include “performance (e.g., qualityand quantity) and members’ affective reactions (e.g., satisfaction,commitment, and viability).”

As a real-life example provided by Franz et al. (2017), establish-ing a common culture in a team can be seen as the input. Develop-ment of this team into a single working unit, however, is mediatedthrough the level of integration achieved on the project (as a media-tor). These contribute to final completion of the construction

© ASCE 04018059-2 J. Constr. Eng. Manage.

J. Constr. Eng. Manage., 2018, 144(7): 04018059

Dow

nloa

ded

from

asc

elib

rary

.org

by

Mon

ash

Uni

vers

ity o

n 11

/15/

18. C

opyr

ight

ASC

E. F

or p

erso

nal u

se o

nly;

all

righ

ts r

eser

ved.

project on budget, schedule, and according to expected quality(overall performance, as an element of effectiveness).

Conceptual Model

The major factors affecting virtuality in EPNs were drawn from thestudy by Hosseini et al. (2016). As for the mediators (factors andsupporting processes), the revised version of the Big Five theory forconstruction project teams (Hosseini, unpublished data, 2016) wasutilized as the source. There were four factors: (1) leadership,(2) backup behavior, (3) team orientation, and (4) accountability.

Similarly, there were four supporting processes: (1) mutual trust,(2) communication, (3) shared mental models, and (4) social inter-actions. Moreover, indicators of team effectiveness were encapsu-lated into two factors: performance and satisfaction of members, asrecommended by Mathieu et al. (2008) (Fig. 2).

Research Methods

Factors that affect virtuality in teams are to be seen and measured assubjective phenomena (Siebdrat et al. 2014). As an example,deploying objective measures of distance as a factor affecting

Fig. 2. Conceptual model of the study.

Fig. 1. Conceptualizing the impacts of virtuality on team effectiveness.

© ASCE 04018059-3 J. Constr. Eng. Manage.

J. Constr. Eng. Manage., 2018, 144(7): 04018059

Dow

nloa

ded

from

asc

elib

rary

.org

by

Mon

ash

Uni

vers

ity o

n 11

/15/

18. C

opyr

ight

ASC

E. F

or p

erso

nal u

se o

nly;

all

righ

ts r

eser

ved.

virtuality leads to misleading results (Siebdrat et al. 2014). That isbecause, “treating distance in purely objective terms provides anincomplete view of how people experience it” (Wilson et al.2005). In the same vein, regarding the impacts of virtuality on teameffectiveness, Lurey and Raisinghani (2001) proposed that “teammember perceptions can be extremely valid predictors of the team’seffectiveness since team members are central to the work, and thus,they directly influence the team’s productivity and satisfaction.”Consequently, the aggregation of shared perceptions amongteam members is the major indicator for factors affecting virtualityand is the predictor of team outcomes and effectiveness (Siebdratet al. 2014).

This distinction is not a trivial matter. If the distance of virtualityto be measured is a subjective one, then the method by which it ismeasured must correspond. In this regard, questionnaires, ratherthan direct quantitative approaches, are suitable. Here, the use ofquestionnaires is not merely a resort of practical expediency,but, rather, the most appropriate research tool for this study.

Survey Design

Survey is a robust, ubiquitous method used in the research field forthe collection of perceptions from a population of interest. Thus, aquestionnaire was deemed the best approach for testing the validityof the conceptual model. It comprised two parts, as illustrated in theSupplemental Data. The first collected demographic attributes ofthe respondents, while the second assessed factors affecting virtual-ity in EPNs along with impacts on virtuality that are associated withthe effectiveness of EPNs.

Questions used were modeled on previous studies which, as in-dicated by Rowley (2014), is an efficient strategy for researchersattempting to measure complex and multidimensional concepts.A 7-point Likert scale (from 1 = very strongly disagree to 7 = verystrongly agree) was used. This reduces skewness, improves normal-ity, makes the data closer to interval-level scaling, while creating anadequate amount of variance (Blunch 2013). Pretesting was con-ducted with a group of 17 experts, and on the strength of the feed-back received, the clarity of questions was strengthened withduplications eliminated; reducing the number of questions from63 to 53.

Sampling

Survey Monkey was used as the online service for conducting thesurvey. Professional associations of construction practitioners inAustralia (e.g., Master Builders, Australian Institute of Building,Civil Contractors Federation, and Australian Institute of Architects)were approached to distribute the questionnaire link. Similarly,building companies were contacted directly from website addressesand regulatory body lists. The kind of sampling chosen wasstratified random sampling, with the stratified groups comprisingarchitects, engineers, design consultants, contractors, facility man-agers, and project managers. As postulated by Robson (2002),“stratified random sampling can be more efficient than simplerandom sampling, in the sense that, for a given sample size, themeans of stratified samples are likely to be closer to the populationmean.”

The accumulated lists used for distributing the questionnairesincluded the contact details of 3,683 construction practitioners.The contacts belonged to 728 architectural companies; 1,308 con-tractors; 852 designing firms; and 795 construction managemententerprises. In due course, 537 questionnaires were completedby the end of March 2015, with 475 of these completed to a suffi-ciently high level of 85%. Of the 475, 285 claimed previous

involvement with EPNs, 76.5% of whom had more than 11 yearsof experience in the industry. Thus, these residual 285 question-naires, showing a response rate of around 15%, formed the basisof the data set for the subsequent stages of data analyses. This wasdeemed acceptable given the novelty of the topic. Even so, responserates as low as 10–12% are not untypical of contemporary construc-tion management research (Bing et al. 2005).

The professional areas in which respondents worked are illus-trated in Fig. 3. This shows that the survey respondents covered awide range of professions within the Australian construction indus-try. Thus, the views put forward are properly reflective of the gen-eral perceptions of practitioners working in construction projectteams across the construction industry.

As can be seen, the percentages are divided fairly equallybetween design practitioners (i.e., architects, civil/structural engi-neers, and quantity surveyors) and practitioners directly involvedin on-site activities (contractors and building services).

Structural Equation Modeling

According to the test selection grid developed by Ho (2006), incircumstances in which the hypothesis falls within the categoryof relationships in which the nature of data could be regarded asinterval, the method of analysis should be based on multivariateregression. On the other hand, in the present study, the main con-struct in the hypotheses was virtuality, which is an unobservedtheoretical phenomenon that represents the degree of virtualityin teams, in which that virtuality is considered a latent variable(Ho 2006). Structured equation modeling (SEM) is capable of han-dling latent variables in multivariate regression analysis (Blunch2013; Ho 2006) and is adopted in this study.

Model Analysis

Constructing Parcels

A common approach used to form SEM models involves aggregat-ing two or more indicators into a parcel and using the parcel in lieuof individual indicators. First, forming parcels is a viable solution tothe problems of both nonnormality when combined with the cat-egorical nature of the data collated through Likert-scale question-naires (Ho 2006). Second, using parcels instead of individualindicators for latent variables reduces the complexity of the model,facilitating greater stability of the parameter estimation (Blunch2013). Thus, parceling is used, however, their reliability must beconfirmed.

Reliability TestThe reliability of an instrument is associated with its consistency,which in turn has two major aspects, consistency over timeand internal consistency (Ho 2006; Punch 2005). The presentstudy meets the three key reliability requirements in creatingparcels:• Cronbach’s alpha coefficient value for the battery of measures

for each concept is higher than 0.70 (Blunch 2013; Spector1992).

• Corrected item-total correlation (the simple correlation betweenthe item and the sum of the rest) is higher than 0.40 (Blunch2013; Spector 1992).

• Squared multiple correlation [the coefficient of determination(R2) when the item is regressed on the other items] is higherthan 0.30 (Cronk 2014).

© ASCE 04018059-4 J. Constr. Eng. Manage.

J. Constr. Eng. Manage., 2018, 144(7): 04018059

Dow

nloa

ded

from

asc

elib

rary

.org

by

Mon

ash

Uni

vers

ity o

n 11

/15/

18. C

opyr

ight

ASC

E. F

or p

erso

nal u

se o

nly;

all

righ

ts r

eser

ved.

Validity TestExploratory factor analysis (EFA) is used to discover if a conceptof interest is unidimensional or multidimensional (Spector 1992).Indicators were submitted to SPSS for analysis. Principal compo-nent analysis (PCA) with Varimax rotation was utilized to explorethe underlying structure of each construct. All but two measuresloaded cleared the minimum recommended (loading ≥ 0.5)(Spector 1992). Still, two of the eight measures for the conceptof communications failed this test, and were, therefore, deletedfrom the battery of measures for the construct (Spector 1992).These were items measuring bidirectionality and sense of presence.The measures retained were considered sufficiently valid to formthe basis for constructing the parcels in the present study.

Constructed Parcels

As mentioned, items were assigned to parcels on the basis of theo-retical intended content of the parcel. In this approach, parcels re-present concepts and their scores are calculated by summing theitems assigned to that specific parcel (Spector 1992). Parcels wereassembled using Eq. (1)

Parcel scorei ¼Xn

k¼1

itemki ð1Þ

where i = observation number (i ¼ 1–285); and k = number ofitems assigned to each parcel after validation of the items for thatparcel. There is a strong underlying assumption regarding themultivariate normality of variables for SEM, with any deviationfrom this assumption resulting in incorrect interpretations. To test

univariate normality in a data set, values for the skew index (SI)and kurtosis index (KI), with their standard errors, should be cal-culated. The ratio of the values of the SI and KI over the corre-sponding standard errors is a good measure as a z-test of thenull hypothesis (Rose et al. 2015). Considering this, results of thetest of normality for variables in the data set are presented inTable 1.

Values for SI/Standard error and KI/Standard error beyond�1.96 suggest that distribution of a particular variable is not normal(Rose et al. 2015). As shown in Table 1, values above�2.0 indicatethat the variable distribution is very likely skewed, or has excesskurtosis, with few variables meeting the requirements of normality.The outcome of the analysis underscored the necessity of consid-ering SEM techniques that are applicable to nonnormal categori-cal data.

Partial Least Squares Structural Equation Modeling

The two broad methods for conducting SEM are covariance based(CB-SEM) and partial least squares (PLS-SEM) (Hair et al. 2014).PLS-SEM is the most appropriate method when the nature of thestudy is more exploratory than confirmatory, with a less developedtheoretical background. Therefore, PLS-SEM was used to test themodel in the present study. A range of software packages forPLS-SEM are available, with SmartPLS, being one of the mostcommon (Hair et al. 2014), and this package was used.

Model SpecificationThe model for SEM (Fig. 4) was developed according to the detailsshown in Fig. 2. Unlike software packages based on CB-SEM,

Fig. 3. Professional area of respondents.

© ASCE 04018059-5 J. Constr. Eng. Manage.

J. Constr. Eng. Manage., 2018, 144(7): 04018059

Dow

nloa

ded

from

asc

elib

rary

.org

by

Mon

ash

Uni

vers

ity o

n 11

/15/

18. C

opyr

ight

ASC

E. F

or p

erso

nal u

se o

nly;

all

righ

ts r

eser

ved.

SmartPLS is able to analyze models in both formative and reflectiveforms. Since measures in the reflective form are viewed asrepresentative samples of the possible items available in the domainof the associated concept, with the intention being that all measuresare caused by the same concept, this approach was taken as themost appropriate. The assessment of PLS-SEM model generated

follows a two-stage sequential procedure. This requires separateassessments of measurement models (indicators–constructs), andan assessment of the structural model, that is, between the con-structs (Hair et al. 2014). Researchers then proceed to assessingthe structural model after establishing the goodness of fit (GoF)of constructs of the specified model.

Table 1. Details of concepts in the model (constructed parcels)

ConceptNumberof items Minimum Maximum Mean

Standarddeviation

Skewness Kurtosis SI KI

StatisticStandarderror Statistic

Standarderror

Standarderror

Standarderror

Degree of dispersion 3 3.00 21.00 13.17 3.44 −0.679 0.144 0.628 0.288 −5.57 1.97Maturity relationships 5 5.00 35.00 21.69 6.24 −0.278 0.144 −0.123 0.288 −2.17 −2.30Context disparity 5 5.00 33.15 22.60 5.02 −0.510 0.144 0.733 0.288 −3.07 0.67Task nature 2 2.00 14.00 9.158 2.12 −0.959 0.144 1.534 0.288 −2.34 −0.37Prior ties 2 2.00 14.00 8.77 2.51 −0.474 0.144 −0.008 0.288 −4.71 2.18KSAs 2 2.00 14.00 8.26 2.66 −0.214 0.144 −0.361 0.288 −1.92 −0.43Size of team 1 1 7 4.62 1.23 −0.803 0.144 0.566 0.288 −3.53 2.55Backup behavior 2 3.00 14.00 9.15 2.34 −0.426 0.144 −0.324 0.288 −6.64 5.33Leadership 1 1 7 4.42 1.24 −0.313 0.144 −0.660 0.288 −3.29 −0.03Team orientation 2 2.00 14.00 8.83 2.43 −0.284 0.144 −0.245 0.288 −1.48 −1.26Accountability 2 2.00 14.00 8.89 2.31 −0.295 0.144 0.035 0.288 −2.95 −1.13Communications 6 10.00 42.00 27.11 6.49 −0.293 0.144 −0.301 0.288 −1.97 −0.85Mutual trust 2 2.00 14.00 8.780 1.97 −0.069 0.144 0.165 0.288 −2.04 0.12Social interactions 1 1 7 4.80 1.18 −0.337 0.144 −0.107 0.288 −2.03 −1.05Shared mental models 2 3.00 14.00 9.42 2.12 −0.402 0.144 −0.305 0.288 −0.48 0.57Performance 4 4.00 28.00 16.96 4.32 −0.337 0.144 0.409 0.288 −2.79 −1.06Satisfaction of members 1 1 7 4.74 1.18 −0.444 0.144 0.193 0.288 −2.34 1.42

Note: SI = skewness statistic; KI = kurtosis statistic; KSA = knowledge, skills, and abilities.

Fig. 4. Specified model (based on model in Fig. 2) with initial estimation of loadings and path coefficients.

© ASCE 04018059-6 J. Constr. Eng. Manage.

J. Constr. Eng. Manage., 2018, 144(7): 04018059

Dow

nloa

ded

from

asc

elib

rary

.org

by

Mon

ash

Uni

vers

ity o

n 11

/15/

18. C

opyr

ight

ASC

E. F

or p

erso

nal u

se o

nly;

all

righ

ts r

eser

ved.

Assessment of Measurement ModelsAccording to Hair et al. (2014), the assessment of measurementmodels comprises an estimation of (1) reliability of individualindicators; (2) composite reliability; (3) average variance extracted(AVE); and (4) discriminant validity.

Indicators in the model that show an outer loading below 0.40should be eliminated from further analysis (Hair et al. 2014).Such indicators do not have much in common with other indicatorscaptured by their common construct. In view of the results (Fig. 4),task nature, degree of dispersion, and size of team, as highlighted inFig. 4, were eliminated from the battery of indicators for the vir-tuality construct, since the outer loadings were significantly lowerthan other indicators, and well below the threshold.

After the removal of these indicators, this process led to a re-vised model for the virtuality construct. The outer loadings for therevised model are presented in Table 2.

Prior ties also showed loadings below 0.40 in the initial model;however, the outer loading increased to 0.50 in the revised modeland was within the acceptable range (Table 2), hence its removalwas not necessary.

Due to the limitations of Cronbach’s alpha (sensitivity to thenumber of items involved), composite reliability (ρc) was usedto assess the internal consistency (Hair et al. 2014). Values above0.60 are acceptable (Hair et al. 2014), and since all constructs in themodel met this threshold, all were accepted. AVE is a commonmeasure used to establish the convergent validity of the constructsin the model with values greater than 0.50, indicating acceptableconvergent validity. Again, all constructs met this test. Finally,checking the cross-loadings of indicators on constructs is a measureused to establish discriminant validity. Discriminant validity is es-tablished when an indicator’s loading on a construct is higher thanall cross-loadings of the other constructs. This was established,with validity of all indicators confirmed (Table 2).

Assessment of the Structural ModelThe next stage involved investigating the structural model to showwhether it can be empirically confirmed, and to assess how well theempirical data supports the underlying theory of the study (Hairet al. 2014). The study’s structural model, including the relation-ships between the latent variables, was assessed using five features:(1) collinearity issues; (2) significance of path coefficients; (3) levelof R2 values; (4) the effect size (f2); and (5) the predictiverelevance (Q2).

In the model, illustrated in Fig. 4, the factors and supportingprocesses constructs are both predictors of the effectivenessconstruct. A multiple regression was run with effectiveness asthe dependent variable, and factors and supporting processes as

independent variables. Tolerance values below 0.2, or variance in-flation factor (VIF) values above 5, are indicative of unacceptablecollinearity among the predictors. Both values were within theacceptable range and, consequently, there was no issue withcollinearity.

The significance of path coefficients in PLS-SEM models is es-timated using the bootstrapping method. The number of bootstrap-ping samples was set at 5,000, as recommended by Hair et al.(2014) for PLS-SEM models in a two-tailed test. Because allt-values were well above the threshold (2.57), the associationsin the model all proved significant (significance level = 1%).

As recommended by Hair et al. (2014), adjusted R2 is a pre-ferred measure to assess the power of the model in predictingendogenous constructs, removing bias due to the complexity ofthe model. The adjusted R2 values for endogenous constructs, thatis, effectiveness, factors and supporting processes, were 0.71,0.045, and 0.059, respectively. Effectiveness is strong againstfactors and supporting processes, explaining at least 70% of thevariance in effectiveness in EPNs. However, virtuality merely pre-dicted 4 and 6% of the variations in factors and supporting proc-esses, respectively, although these relationships were found to besignificant.

The effect size, that is f2, is used to assess whether a particularexogenous construct in the model has a substantial impact onendogenous constructs. SmartPLS is able to provide f2 valuesfor exogenous constructs in the model. Values for factors and sup-porting processes on effectiveness were above 0.15. On the otherhand, values for virtuality on factors and supporting processes werearound 0.02. Thus, as asserted by Cohen (2013), the effect size offactors and supporting processes on effectiveness was moderate,while the effect size of virtuality on factors and supporting proc-esses was small.

Blindfolding analysis was carried out to assess the predictivesuitability of the model. The omission distance was D ¼ 7, withestimations provided by construct cross-validated redundancy, inaccordance with the approach prescribed by Hair et al. (2014).Values of Q2 larger than zero indicate that exogenous constructsin a model have predictive relevance for the endogenous constructsin that model. Because allQ2 values were above zero, this indicatedthe predictive relevance of the model. However, virtuality, as theexogenous construct, only had small predictive relevance for fac-tors and supporting processes, whereas factors and supportingprocesses had medium predictive relevance for effectiveness.

Model InterpretationThe loadings of the indicators (significant associations) in reflectivemeasurement models are indicative of the relative importance of

Table 2. Summary of the revised reflective measurement models

Construct Indicators Loadings Indicator reliability Composite reliability (ρc) Average Discriminant validity?

Virtuality Context disparity 0.70 0.49 0.835 0.570 YesKSAs 0.89 0.79

Maturity relationships 0.87 0.76Prior ties 0.50 0.25

Factors Accountability 0.83 0.69 0.917 0.735 YesBackup behavior 0.88 0.77

Leadership 0.83 0.69Team orientation 0.89 0.79

Supportingprocesses

Communications 0.83 0.69 0.851 0.593 YesMutual trust 0.57 0.32

Shared mental models 0.87 0.76Social interactions 0.78 0.61

Effectiveness Performance 0.91 0.83 0.897 0.814 YesSatisfaction of members 0.90 0.81

© ASCE 04018059-7 J. Constr. Eng. Manage.

J. Constr. Eng. Manage., 2018, 144(7): 04018059

Dow

nloa

ded

from

asc

elib

rary

.org

by

Mon

ash

Uni

vers

ity o

n 11

/15/

18. C

opyr

ight

ASC

E. F

or p

erso

nal u

se o

nly;

all

righ

ts r

eser

ved.

the indicators with regard to their common factor (Hair et al. 2014).As shown in Table 3, the indicator members’ knowledge, skills andabilities (KSAs), with a loading of 0.89 on virtuality, was found tohave the highest influence on virtuality. Contrariwise, the indicatorprior ties (loading ¼ 0.50) was the least important factor in asso-ciation with virtuality. By the same token, team orientation was themost important indicator for factors (loading = approximately 0.90)and leadership was the indicator with the least level of influence onfactors (loading ¼ 0.828) in the model.

In terms of indicators associated with supporting processes, theindicator shared mental models was ranked first in terms of relativeimportance, while mutual trust was the least important of the fourindicators. The indicators, satisfaction of members and perfor-mance, were found to be similar in terms of their relative impor-tance, with loadings for both of close to 0.90.

A significant association between virtuality and effectivenesswas observed (significance level = 1% due to t-value >2.57); thus,an increase in virtuality is significantly and positively associatedwith a reduction in effectiveness. Nevertheless, the value of thepath coefficient was found to be close to 0.22, indicating thatthe association between virtuality and effectiveness was weak.Converting this path coefficient by the amount of explained vari-ance (R2 ¼ 0.222 ¼ 0.048) shows that virtuality is responsible lessthan 5% of the variance in effectiveness in EPNs. In essence, 95%of the variations in effectiveness stem from factors other thanvirtuality. This clearly points to the small role virtuality plays with

regard to changes in the effectiveness of EPNs. The total impacts ofvirtuality on the indicators of factors in the model are shown inTable 4.

Discussion of Findings

The present study tested a revised version of the Big Five theory forEPNs. In so doing, it offered significant insights on the role of vir-tuality with regard to EPN effectiveness.

Factors Affecting Virtuality in EPNs

As far as virtuality is concerned, the findings do not uphold thewidely held view that the degree of dispersion, size of team,and task nature strongly impact the virtuality of EPNs. This studythus indicates a shift in the work culture of EPNs to one that is moreaccommodating and adapted to its usage. Previous seminal studies,such as those by Watson-Manheim et al. (2002) and O’Leary andCummings (2007), have regarded the degree of dispersion as a pil-lar of virtuality for teams. The present study reveals rather that suchfindings have eroded with regard to EPNs. This is explicable inlight of evolving remote work practices and the concept of “deathof distance” (Yamin and Sinkovics 2006), in which, due to exten-sive experience with online internationalization, any form of dis-parity will likely erode. As a response to internationalizationpressures, the construction industry is moving toward delivering

Table 3. Relative importance of indicators in the model

Association Loadings t-statistics p-values Rank (relative importance)

KSAs—virtuality 0.888 11.543 0.000 1Maturity relationships—virtuality 0.866 12.324 0.000 2Context disparity—virtuality 0.704 6.013 0.000 3Prior ties—virtuality 0.497 3.301 0.001 4Team orientation—factors 0.891 53.556 0.000 1Backup behavior—factors 0.875 47.416 0.000 2Accountability—factors 0.834 34.937 0.000 3Leadership—factors 0.828 29.982 0.000 4Shared mental models—supporting processes 0.868 51.225 0.000 1Communications—supporting processes 0.825 31.453 0.000 2Social interactions—supporting processes 0.781 25.568 0.000 3Mutual trust—supporting processes 0.573 7.532 0.000 4Performance—effectiveness 0.906 76.832 0.000 1Satisfaction of members—effectiveness 0.898 67.373 0.000 2

Table 4. Total impacts of virtuality on indicators in the structural model

Structural model associationPath

coefficient (1) Measurement models’ associations Loadings (2) Total impacts’ associationTotal impacts’value ð1Þ × ð2Þ

Virtuality—effectiveness 0.206 Performance—effectiveness 0.906 Virtuality—performance 0.187Virtuality—effectiveness 0.206 Satisfaction of members—

effectiveness0.898 Virtuality—satisfaction of

members0.185

Virtuality—factors 0.219 Team orientation—factors 0.891 Virtuality—team orientation 0.195Virtuality—factors 0.219 Backup behavior—factors 0.875 Virtuality—backup behavior 0.192Virtuality—factors 0.219 Accountability—factors 0.834 Virtuality—accountability 0.183Virtuality—factors 0.219 Leadership—factors 0.828 Virtuality—leadership 0.181Virtuality—supportingprocesses

0.249 Shared mental models—supportingprocesses

0.868 Virtuality—shared mentalmodels

0.216

Virtuality—supportingprocesses

0.249 Communications—supportingprocesses

0.825 Virtuality—communications 0.205

Virtuality—supportingprocesses

0.249 Social interactions—supportingprocesses

0.781 Virtuality—socialinteractions

0.194

Virtuality—supportingprocesses

0.249 Mutual Trust—supporting processes 0.573 Virtuality—mutual trust 0.143

© ASCE 04018059-8 J. Constr. Eng. Manage.

J. Constr. Eng. Manage., 2018, 144(7): 04018059

Dow

nloa

ded

from

asc

elib

rary

.org

by

Mon

ash

Uni

vers

ity o

n 11

/15/

18. C

opyr

ight

ASC

E. F

or p

erso

nal u

se o

nly;

all

righ

ts r

eser

ved.

business in the form of e-commerce (Zelkowicz et al. 2015).This shift is associated with a diminution of the sense of dispersionfor teams and companies (Yamin and Sinkovics 2006).

Also indicated is that for generic tasks assigned to EPNs, thesize of a team does not impact perceptions of virtuality. Similarly,task nature also has limited impact. Complexity, the uncertainty oftasks, the involvement of a large number of organizations and dis-ciplines, and the necessity of having skilled personnel from diverselocations are inherent to EPNs (Mignone et al. 2016). Simply, thereis no perceived variance in task nature and size of teams associatedwith the level of virtuality among different EPNs.

Contrariwise, four factors were found to be influential on thelevel of virtuality in EPNs. As shown in Table 3, these are mem-bers’ KSAs; maturity of teams; context disparity; and prior ties. Afew conceptual studies, such as Martins et al. (2004), have pointedout the likelihood of a link between virtuality and members’ KSAs.However, the findings of the present study have established theexistence of a direct link between members’ KSAs and virtualityfor EPNs. Likewise, the maturity of a team in handing supply chainrelationships as a contributor to virtuality has been overlooked inthe extant literature. Nevertheless, the concept of “team resilience,”namely, the “ability to absorb high levels of change while display-ing minimal dysfunctional behavior” (Hoopes 1999), is a directoutcome of team maturity.

The findings of the case study in the construction context byMignone et al. (2016) identified the impacts of context disparityon a wide range of performance and effectiveness dimensions inEPNs. In this study, prior ties as a concept was found to be twodimensional, with its dimensions being (1) a history of collabora-tion prior to forming the current team, and (2) the time team mem-bers have spent together in the current team. Ortiz de Guinea et al.(2012) acknowledged the impacts of working in previous teams onvirtuality. However, prior ties in the present study is conceptualizedas the total amount of time that team members have worked to-gether, whether in a previous team or only the current team.

Impacts of Virtuality on Effectiveness of EPNs

The findings reveal a number of original insights into the associ-ation of virtuality with the effectiveness of teams, as mediatedthrough the Big Five theory factors and supporting processes.Principally, the findings spotlight that the negative impacts of vir-tuality found in previous studies are overemphasized. Even so,what negative impacts there are can be fully managed through al-terations in team configuration (Eubanks et al. 2016). Moreover, astechnology continues to improve and teams increasingly becomestaffed by those familiar with the use of technology, the negativeimpacts of virtuality are all the more reduced (Maynard et al. 2012).

Leadership, as a dimension of factors (from the Big Fivetheory), is highly influential on the performance of EPNs. Never-theless, change in the level of virtuality affects leadership onlyslightly. Self-leadership by far outweighs managers’ leadershipat higher levels of virtuality owing to the lower influence of leaderson members working at a distance (Andressen et al. 2012) includ-ing EPNs.

Accountability as an area influenced by virtuality has remainedunexplored within the extant literature. The present study’s findingsreveal that virtuality is indeed significant, albeit weakly associatedwith variations of accountability in EPNs. Lack of accountabilitymay be the result of diminished supervision, feedback, and control(West 2012).

The present study’s findings also show an association betweenvirtuality and the factors that reflect supporting processes. The di-rect association of shared mental models with the effectiveness of

teams has also been recognized in previous studies (Jonker et al.2011). The present study’s findings, however, reveal that the im-pacts of virtuality on shared mental models are higher than itsimpacts on communications. Contrary to the widespread perceptionthat communication is the supporting process most negatively af-fected by virtuality (Hosseini et al. 2015), shared mental models isthe process shown to have highest level of variation due to virtual-ity. Developing shared mental models thus requires extensiveinteractions between members, the exchange of rich information,combined with a history of cooperation between team members(Fransen et al. 2011).

Communication and interaction between members in EPNs arenegatively affected at higher levels of virtuality (Nayak and Taylor2009). Moreover, construction teams are inherently temporary(Rezgui 2007). The combined outcome of these points engendersshared mental models becoming the supporting process most af-fected by virtuality. Previous researchers (Martins et al. 2004;Penarroja et al. 2013) have suggested that an increase in virtualitysignificantly diminishes communication, leading to a reduction inperformance (Ferreira et al. 2012; Schweitzer and Duxbury 2010).Here, however, the impacts for EPNs are by far lower than previ-ously asserted.

Researchers from the construction industry have endorsed theview that virtuality has a clear negative impact on levels of mutualtrust (Penarroja et al. 2013), while Bierly et al. (2009) have chal-lenged this belief. This study confirms a strong association betweenvirtuality and mutual trust for EPNs.

Moreover, virtuality was found to impact satisfaction of mem-bers and ultimately the effectiveness for teams. However, theimpacts of virtuality on the satisfaction of members have re-mained unreported within the extant literature, particularly withinthe construction context. Few studies in the construction context,including that by Nayak and Taylor (2009), have been able to iden-tify significant frustration among members working in dispersedteams.

Implications for Practice

With EPN use growing exponentially, this study’s outcomescould benefit organizations, construction firms, and managers byproviding fresh insight into the nature of working in EPNs.Specifically, the findings provide guidelines to assess and developessential competencies for working in EPNs, and could promoteimproved interactions and performance in EPNs. As such, the prac-tical implications associated with the findings are summarizedsubsequently.

Increased Awareness of EPNs

The present study’s findings raise the level of awareness of man-agers and construction practitioners in dealing with potential issuesinherent to EPNs, and in devising plans to deal with such chal-lenges. Moreover, the quantification of associations provides a toolfor generating predictions on the basis of level of virtuality.

Implications for Training for EPNs

The findings of the study could assist in providing training on ef-fective styles of leadership for EPNs, backup behavior, and mea-sures to develop trust and maintain communications. These areasare a few examples of what could be included within the topics fortraining prior to EPN involvement.

© ASCE 04018059-9 J. Constr. Eng. Manage.

J. Constr. Eng. Manage., 2018, 144(7): 04018059

Dow

nloa

ded

from

asc

elib

rary

.org

by

Mon

ash

Uni

vers

ity o

n 11

/15/

18. C

opyr

ight

ASC

E. F

or p

erso

nal u

se o

nly;

all

righ

ts r

eser

ved.

Design and Structure of EPNs

The present study’s findings lay a solid foundation for constructionorganizations, managers, and companies to make informed deci-sions on the adaptation of EPNs within the overall organizationalstructure of construction projects. In designing and arrangingEPNs, the desired level of effectiveness could be achieved throughthe manipulation of influential variables.

Implications for Performance and Effectiveness inEPNs

The model of virtuality presented in this study illustrates the driversand inhibitors of effectiveness and performance when it comes toimplementing EPNs in construction projects. This leads to identi-fying the main variables and areas that require concentrated man-agement interventions needed to facilitate a supportive workplaceenvironment.

Limitations of the Study and Future Research

The findings are to be treated in light of the limitations of the study.The building blocks of the conceptual model deployed in thepresent study rely on the available literature on virtuality. This bodyof knowledge is, however, limited, being in its infancy and yet toreach maturity. This resulted in the inclusion of a number of con-cepts in the conceptual model (such as shared mental models andteam orientation) which may be difficult to operationalize. In for-mulating the questionnaire, semistructured interviews were used toconsolidate the content, and to fit the items to the construction set-ting. While the number of interviews was adequate for a qualitativestudy, the novelty of the concept of virtuality for industry practi-tioners may have interfered with informant responses.

A further limitation stems from the sampling strategy. In thisstudy, interviewees and respondents were practitioners workingin the Australian construction industry. Clearly, the perceptionsof construction practitioners from other countries and in other con-texts may differ. Therefore, generalizability of the findings is ten-tative. Nevertheless, further contexts and settings could serve as abasis for further research, with larger quantitative samples.

Of course, a further limitation lies with the reliance on datacollected from questionnaire surveys themselves. Although themethod is ubiquitous and indeed well suited to the collection ofsubjective perceptions of humans, as is the case here, they do attractthe criticism that results are variable when tests are replicated.Nevertheless, the return of 475 usable questionnaires, with 76%having more than 11 years’ relevant experience, provides a bench-mark against which further research into this area can be usefullycompared and contrasted.

Conclusion

The present study conducted an empirical investigation on virtual-ity in EPNs within the construction context. It offers the first quan-tified model linking virtuality with the effectiveness of EPNs.In this regard, the study developed theories that were based on find-ings imported from information technology (IT), and managementand business disciplines, modifying them through exposure toempirical data from the construction context. This facilitated “thetransferability and applicability” of existing theories from otherdisciplines to the specific setting of construction.

Both the Big Five theory and the IMO model are well-knowntheoretical frameworks in the literature. However, the present studyused virtuality as an input, connecting it with effectiveness in

hybrid teams by synthesizing the IMO model and the Big Fivetheory. This presents a new interpretation of the associations ofvirtuality with effectiveness.

Furthermore, the present study provides an original insight intothe phenomenon of virtuality through the use of contingency ap-proaches. Contingency approaches provide an original understand-ing on a topic by bringing to light “what processes and practicesapply in which contexts, what relationships hold or do not hold inwhich contexts and where do methods work and do not work orhow do they vary in different contexts” (Boer et al. 2015). Thisrefers to the findings of the study illustrating the specific factorsthat were not influential in the context of EPNs, as well as the itemsthat were overlooked within the existing literature, yet were re-vealed in the present study.

Data Availability Statement

Data generated by the authors or analyzed during the study,along with created models, are available at https://figshare.com/s/40c40d257edeb4aeb56f. Information about the Journal’s datasharing policy can be found here: http://ascelibrary.org/doi/10.1061/%28ASCE%29CO.1943-7862.0001263.

Supplemental Data

The questionnaire used in this study is available online in the ASCELibrary (www.ascelibrary.org).

References

Alin, P., J. Iorio, and J. E. Taylor. 2013. “Digital boundary objects as ne-gotiation facilitators: Spanning boundaries in virtual engineering projectnetworks.” Project Manage. J. 44 (3): 48–63. https://doi.org/10.1002/pmj.21339.

Alreshidi, E., M. Mourshed, and Y. Rezgui. 2018. “Requirements forcloud-based BIM governance solutions to facilitate team collaborationin construction projects.” Requirements Eng. 23 (1): 1–31. https://doi.org/10.1007/s00766-016-0254-6.

Andressen, P., U. Konradt, and C. P. Neck. 2012. “The relation betweenself-leadership and transformational leadership: Competing modelsand the moderating role of virtuality.” J. Leadersh. Organizational Stud.19 (1): 68–82. https://doi.org/10.1177/1548051811425047.

Bassanino, M., T. Fernando, and K.-C. Wu. 2014. “Can virtual workspacesenhance team communication and collaboration in design review meet-ings?” Archit. Eng. Des. Manage. 10 (3–4): 200–217. https://doi.org/10.1080/17452007.2013.775102.

Becerik-Gerber, B., K. Ku, and F. Jazizadeh. 2012. “BIM-enabled virtualand collaborative construction engineering and management.” J. Prof.Issues Eng. Educ. Pract. 1 (1): 234–245. https://doi.org/10.1061/(ASCE)EI.1943-5541.0000098.

Bierly, P. E., E. M. Stark, and E. H. Kessler. 2009. “The moderating effectsof virtuality on the antecedents and outcome of NPD team trust.”J. Prod. Innovation Manage. 26 (5): 551–565. https://doi.org/10.1111/j.1540-5885.2009.00680.x.

Bing, L., A. Akintoye, P. J. Edwards, and C. Hardcastle. 2005. “Theallocation of risk in PPP/PFI construction projects in the UK.” Int.J. Project Manage. 23 (1): 25–35. https://doi.org/10.1016/j.ijproman.2004.04.006.

Blunch, N. J. 2013. Introduction to structural equation modeling usingIBM SPSS statistics and AMOS. Thousand Oaks, CA: SAGEPublications.

Boer, H., M. Holweg, M. Kilduff, M. Pagell, R. Schmenner, and C. Voss.2015. “Making a meaningful contribution to theory.” Int. J. Oper. Prod.Manage. 35 (9): 1231–1252. https://doi.org/10.1108/IJOPM-03-2015-0119.

© ASCE 04018059-10 J. Constr. Eng. Manage.

J. Constr. Eng. Manage., 2018, 144(7): 04018059

Dow

nloa

ded

from

asc

elib

rary

.org

by

Mon

ash

Uni

vers

ity o

n 11

/15/

18. C

opyr

ight

ASC

E. F

or p

erso

nal u

se o

nly;

all

righ

ts r

eser

ved.

Bosch-Sijtsema, P. M., R. Fruchter, M. Vartiainen, and V. Ruohomäki.2011. “A framework to analyze knowledge work in distributed teams.”Group Organ. Manage. 36 (3): 275–307. https://doi.org/10.1177/1059601111403625.

Buvik, M. P., and M. Rolfsen. 2015. “Prior ties and trust developmentin project teams: A case study from the construction industry.” Int.J. Project Manage. 33 (7): 1484–1494. https://doi.org/10.1016/j.ijproman.2015.06.002.

Chinowsky, P., and A. Songer. 2011. “Introduction.” In Organizationmanagement in construction, edited by P. Chinowsky and A. Songer.New York: Spon Press.

Cohen, J. 2013. Statistical power analysis for the behavioral sciences.Burlington, VT: Elsevier Science.

Cronk, B. C. 2014. How to use SPSS: A step-by-step guide to analysis andinterpretation. Glendale, CA: Pyrczak Publishing.

Dulebohn, J. H., and J. E. Hoch. 2017. “Virtual teams in organizations.”Hum. Resour. Manage. Rev. 27 (4): 569. https://doi.org/10.1016/j.hrmr.2016.12.004.

Eubanks, D. L., M. Palanski, J. Olabisi, A. Joinson, and J. Dove. 2016.“Team dynamics in virtual, partially distributed teams: Optimal role ful-fillment.” Comput. Hum. Behav. 61: 556–568. https://doi.org/10.1016/j.chb.2016.03.035.

Ferreira, P. G. S., E. P. D. Lima, and S. E. G. da Costa. 2012. “Perception ofvirtual team’s performance: A multinational exercise.” Int. J. Prod.Econ. 140 (1): 416–430. https://doi.org/10.1016/j.ijpe.2012.06.025.

Foster, M. K., A. Abbey, M. A. Callow, X. Zu, and A. D. Wilbon. 2015.“Rethinking virtuality and its impact on teams.” Small Group Res.46 (3): 267–299. https://doi.org/10.1177/1046496415573795.

Fransen, J., P. A. Kirschner, and G. Erkens. 2011. “Mediating team effec-tiveness in the context of collaborative learning: The importance of teamand task awareness.” Comput. Hum. Behav. 27 (3): 1103–1113. https://doi.org/10.1016/j.chb.2010.05.017.

Franz, B., R. Leicht, K. Molenaar, and J. Messner. 2017. “Impact of teamintegration and group cohesion on project delivery performance.”J. Constr. Eng. Manage. 143 (1): 0401608. https://doi.org/10.1061/(ASCE)CO.1943-7862.0001219.

Gilson, L. L., M. T. Maynard, N. C. Jones Young, M. Vartiainen, and M.Hakonen. 2015. “Virtual teams research: 10 years, 10 themes, and 10opportunities.” J. Manage. 41 (5): 1313–1337. https://doi.org/10.1177/0149206314559946.

Hair, J. F., G. T. M. Hult, C. Ringle, and M. Sarstedt. 2014. A primeron partial least squares structural equation modeling (PLS-SEM).Thousand Oaks, CA: SAGE Publications.

Ho, R. 2006. Handbook of univariate and multivariate data analysis andinterpretation with SPSS. Boca Raton, FL: CRC Press.

Hoopes, L. 1999. “Team resilience absorb change without creating dys-function.” Executive Excellence 16 (5): 16.

Hosseini, M. R., N. Chileshe, B. Baroudi, J. Zuo, and A. Mills. 2016.“Factors affecting perceived level of virtuality in hybrid constructionproject teams (HCPTs): A qualitative study.” Constr. Innovation16 (4): 460–482. https://doi.org/10.1108/CI-05-2015-0029.

Hosseini, M. R., J. Zuo, N. Chileshe, and B. Baroudi. 2015. “Evaluatingvirtuality in teams: A conceptual model.” Technol. Anal. StrategicManage. 27 (4): 385–404. https://doi.org/10.1080/09537325.2014.1003206.

Ibrahim, K. I., S. B. Costello, and S. Wilkinson. 2013. “Key practice indica-tors of team integration in construction projects: A review.” TeamPerform. Manage. Int. J. 19 (3/4): 132–152. https://doi.org/10.1108/TPM-10-2012-0033.

Iorio, J., and J. E. Taylor. 2014. “Boundary object efficacy: The mediatingrole of boundary objects on task conflict in global virtual project net-works.” Int. J. Project Manage. 32 (1): 7–17. https://doi.org/10.1016/j.ijproman.2013.04.001.

Jonker, C. M., M. B. van Riemsdijk, and B. Vermeulen. 2011. “Sharedmental models.” In Coordination, organizations, institutions, andnorms in agent systems VI, 132–151. Berlin: Springer.

Kay, J., N. Maisonneuve, K. Yacef, and P. Reimann. 2006. “The bigfive and visualisations of team work activity.” In Proc., IntelligentTutoring Systems: 8th Int. Conf., edited by M. Ikeda, K. D. Ashley,and T.-W. Chan, 197–206. Berlin: Springer.

Kozlowski, S. W. J., J. A. Grand, S. K. Baard, and M. Pearce. 2015.“Teams, teamwork, and team effectiveness: Implications for human sys-tems integration.” In APA handbook of human systems integration,edited by D. A. Boehm-Davis, F. T. Durso, J. D. Lee, D. A.Boehm-Davis, F. T. Durso, and J. D. Lee, 555–571. Washington,DC: American Psychological Association.

Kozlowski, S. W. J., and D. R. Ilgen. 2006. “Enhancing the effectiveness ofwork groups and teams.” Psychol. Sci. Public Interest 7 (3): 77–124.https://doi.org/10.1111/j.1529-1006.2006.00030.x.

Lobo, S., and J. Whyte. 2017. “Aligning and reconciling: Building projectcapabilities for digital delivery.” Res. Policy 46 (1): 93–107. https://doi.org/10.1016/j.respol.2016.10.005.

Lurey, J. S., and M. S. Raisinghani. 2001. “An empirical study of best prac-tices in virtual teams.” Inf. Manage. 38 (8): 523–544. https://doi.org/10.1016/S0378-7206(01)00074-X.

Marlow, S. L., C. N. Lacerenza, and E. Salas. 2017. “Communication invirtual teams: A conceptual framework and research agenda.” Hum. Re-sour. Manage. Rev. 27 (4): 575–589. https://doi.org/10.1016/j.hrmr.2016.12.005.

Martins, L. L., L. L. Gilson, and M. T. Maynard. 2004. “Virtual teams:What do we know and where do we go from here?” J. Manage.30 (6): 805–835. https://doi.org/10.1016/j.jm.2004.05.002.

Martins, L. L., and M. C. Schilpzand. 2011. “Global virtual teams: Keydevelopments, research gaps, and future directions.” In Research in per-sonnel and human resources management, edited by A. Joshi, H. Liao,and J. J. Martocchio. Bingley, UK: Emerald Group Publishing Limited.

Mathieu, J., M. T. Maynard, T. Rapp, and L. Gilson. 2008. “Team effec-tiveness 1997–2007: A review of recent advancements and a glimpseinto the future.” J. Manage. 34 (3): 410–476. https://doi.org/10.1177/0149206308316061.

Maynard, M. T., J. E. Mathieu, T. L. Rapp, and L. L. Gilson. 2012. “Some-thing(s) old and something(s) new: Modeling drivers of global virtualteam effectiveness.” J. Organizational Behav. 33 (3): 342–365. https://doi.org/10.1002/job.1772.

Merschbrock, C., and B. E. Munkvold. 2014. “Succeeding with buildinginformation modeling: A case study of BIM diffusion in a healthcareconstruction project.” In Proc., 47th Hawaii Int. Conf. on SystemSciences, 3959–3968. New York: IEEE.

Merschbrock, C., and B. E. Munkvold. 2015. “Effective digital collabora-tion in the construction industry—A case study of BIM deployment in ahospital construction project.” Comput. Ind. 73: 1–7. https://doi.org/10.1016/j.compind.2015.07.003.

Mignone, G., M. R. Hosseini, N. Chileshe, and M. Arashpour. 2016.“Enhancing collaboration in BIM-based construction networks throughorganisational discontinuity theory: A case study of the new RoyalAdelaide Hospital.” Archit. Eng. Des. Manage. 12 (5): 333–352.https://doi.org/10.1080/17452007.2016.1169987.

Nayak, N. V., and J. E. Taylor. 2009. “Offshore outsourcing in global de-sign networks.” J. Manage. Eng. 25 (4): 177–184. https://doi.org/10.1061/(ASCE)0742-597X(2009)25:4(177).

Nitithamyong, P., and M. Skibniewski. 2011. “Success factors for the im-plementation of web-based construction project management systems:A cross-case analysis.” Constr. Innovations 11 (1): 14–42. https://doi.org/10.1108/14714171111104619.

O’Leary, M. B., and J. N. Cummings. 2007. “The spatial, temporal, andconfigurational characteristics of geographic dispersion in teams.”MIS Q. 31 (3): 433–452. https://doi.org/10.2307/25148802.

Oraee, M., M. R. Hosseini, E. Papadonikolaki, R. Palliyaguru, and M.Arashpour. 2017. “Collaboration in BIM-based construction networks:A bibliometric-qualitative literature review.” Int. J. Project Manage.35 (7): 1288–1301. https://doi.org/10.1016/j.ijproman.2017.07.001.

Ortiz de Guinea, A., J. Webster, and D. S. Staples. 2012. “A meta-analysisof the consequences of virtualness on team functioning.” Inf. Manage.49 (6): 301–308. https://doi.org/10.1016/j.im.2012.08.003.

Penarroja, V., V. Orengo, A. Zornoza, and A. Hernández. 2013. “Theeffects of virtuality level on task-related collaborative behaviors:The mediating role of team trust.” Comput. Hum. Behav. 29 (3):967–974. https://doi.org/10.1016/j.chb.2012.12.020.

© ASCE 04018059-11 J. Constr. Eng. Manage.

J. Constr. Eng. Manage., 2018, 144(7): 04018059

Dow

nloa

ded

from

asc

elib

rary

.org

by

Mon

ash

Uni

vers

ity o

n 11

/15/

18. C

opyr

ight

ASC

E. F

or p

erso

nal u

se o

nly;

all

righ

ts r

eser

ved.

Petri, I., T. Beach, O. F. Rana, and Y. Rezgui. 2017. “Coordinating multi-site construction projects using federated clouds.” Autom. Constr.83: 273–284. https://doi.org/10.1016/j.autcon.2017.08.011.

Punch, K. F. 2005. Introduction to social research: Quantitative and quali-tative approaches. Thousand Oaks, CA: SAGE Publications.

Rezgui, Y. 2007. “Exploring virtual team-working effectiveness in the con-struction sector.” Interact. Comput. 19 (1): 96–112. https://doi.org/10.1016/j.intcom.2006.07.002.

Robson, C. 2002. Real world research. Oxford, UK: Blackwell Publishers.Rose, S., A. I. Canhoto, and N. Spinks. 2015. Management research:

Applying the principles. Hoboken, NJ: Taylor & Francis.Rowley, J. 2014. “Designing and using research questionnaires.”

Manage. Res. Rev. 37 (3): 308–330. https://doi.org/10.1108/MRR-02-2013-0027.

Salas, E., D. E. Sims, and C. S. Burke. 2005. “Is there a ‘big five’ in Team-work?” Small Group Res. 36 (5): 555–599. https://doi.org/10.1177/1046496405277134.

Schaubroeck, J. M., and A. Yu. 2017. “When does virtuality help or hinderteams? Core team characteristics as contingency factors.” Hum. Resour.Manage. Rev. 27 (4): 635–647. https://doi.org/10.1016/j.hrmr.2016.12.009.

Schweitzer, L., and L. Duxbury. 2010. “Conceptualizing and measuring thevirtuality of teams.” Inf. Syst. J. 20 (3): 267–295. https://doi.org/10.1111/j.1365-2575.2009.00326.x.

Siebdrat, F., M. Hoegl, and H. Ernst. 2014. “Subjective distance and teamcollaboration in distributed teams.” J. Prod. Innovation Manage. 31 (4):765–779. https://doi.org/10.1111/jpim.12122.

Son, J., and E. M. Rojas. 2011. “Evolution of collaboration in temporaryproject teams: An agent-based modeling and simulation approach.”J. Constr. Eng. Manage. 137 (8): 619–628. https://doi.org/10.1061/(ASCE)CO.1943-7862.0000331.

Spector, P. E. 1992. Summated rating scale construction: An introduction.Newbury Park, CA: SAGE Publications.

Watson-Manheim, M. B., K. M. Chudoba, and K. Crowston. 2002.“Discontinuities and continuities: A new way to understand virtualwork.” Inf. Technol. People 15 (3): 191–209. https://doi.org/10.1108/09593840210444746.

West, M. A. 2012. Effective teamwork: Practical lessons from organiza-tional research. 3rd ed. Hoboken, NJ: Wiley.

Wilson, J. M., M. B. O’Leary, A. Metiu, and Q. Jett. 2005. “Subjective dis-tance in teams.” Accessed February 2, 2015. http://www.insead.edu/facultyresearch/research/doc.cfm?did=1619.

Yamin, M., and R. R. Sinkovics. 2006. “Online internationalisation, psy-chic distance reduction and the virtuality trap.” Int. Bus. Rev. 15 (4):339–360. https://doi.org/10.1016/j.ibusrev.2006.03.002.

Zelkowicz, A., J. Iorio, and J. E. Taylor. 2015. “Exploring the roleof cultural boundary spanners at complex boundaries in globalvirtual AEC networks.” J. Inf. Technol. Constr. (ITcon) 20 (24):385–398.

Zhang, P., and F. F. Ng. 2013. “Explaining knowledge-sharing in-tention in construction teams in Hong Kong.” J. Constr. Eng. Man-age. 139 (3): 280–293. https://doi.org/10.1061/(ASCE)CO.1943-7862.0000607.

© ASCE 04018059-12 J. Constr. Eng. Manage.

J. Constr. Eng. Manage., 2018, 144(7): 04018059

Dow

nloa

ded

from

asc

elib

rary

.org

by

Mon

ash

Uni

vers

ity o

n 11

/15/

18. C

opyr

ight

ASC

E. F

or p

erso

nal u

se o

nly;

all

righ

ts r

eser

ved.