You are on page 1of 8

Learning and Individual Differences 39 (2015) 8188

Contents lists available at ScienceDirect

Learning and Individual Differences

journal homepage:

Are balanced groups better? Belbin roles in collaborative

learning groups
Nicoleta Meslec a,, Petru Lucian Cureu b,c,1
Department of Organisation Studies, Tilburg University, PO Box 90153, 5000 LE Tilburg, The Netherlands
Department of Psychology, Babe-Bolyai University, Cluj-Napoca, Romania
Department of Organisation Studies, Tilburg University, PO Box 90153, 5000 LE Tilburg, The Netherlands

a r t i c l e i n f o a b s t r a c t

Article history: In a sample of 459 students organized in 84 groups this study tests the impact of group role balance on teamwork
Received 6 March 2013 quality and three performance indicators in collaborative learning groups (group cognitive complexity, perceived
Received in revised form 26 May 2014 performance and objective performance). The results show that group role balance positively predicts group
Accepted 29 March 2015
performance in preliminary phases of the group project but not in later phases. Moreover, group role balance
positively predicts group cognitive complexity and is negatively related to teamwork quality. The results hold
Group role balance
only when role balance is conceptualized as a congural property of groups instead of a sum of individual
Performance roles. The ndings of the study have implications for the design of collaborative learning groups.
Collaborative learning groups 2015 Published by Elsevier Inc.
Group cognitive complexity

1. Introduction adopt specic patterns of behavior in interpersonal interactions and

these stable individual differences can congure in various ways within
Modern organizations use groups to perform a variety of complex groups (Belbin, 1981). The way in which individual role preferences com-
tasks (Tannenbaum, Mathieu, Salas, & Cohen, 2012), therefore, next to bine within groups is a congural group property (Klein & Kozlowski,
job-related knowledge and expertise, teamwork skills become important 2000) that impacts on group dynamics and effectiveness. One of the
elements in personnel selection across a variety of organizational elds major claims of Belbin's role theory is that balanced groups (in which
(Burch & Anderson, 2004; O'Neil, Allerd, & Baker, 1997; Stevens & all nine roles are present) perform better than unbalanced groups, in
Campion, 1999; Zedeck & Goldstein, 2000). As a consequence, education- which existing roles are duplicating each other (Belbin, 1981). Although
al programs extensively use collaborative learning to help students devel- the claim of balanced groups is being extensively used in organizational
op teamwork skills (Cureu, Janssen, & Raab, 2012) and acquire specic consultancy, the empirical evidence supporting its validity is not conclu-
curricular knowledge (Cureu, 2011; Haugwitz, Nesbit, & Sandmann, sive (Blenkinsop & Maddison, 2007; Jackson, 2002; Partington & Harris,
2010; McCune & Entwistle, 2011). It becomes therefore important to 1999; Senior, 1997; van de Water, Ahaus, & Rozier, 2008). The aim of
identify group design features that increase the effectiveness of individual the current paper is to test the role balance claim in an educational setting
and collaborative learning in student groups (Cureu & Pluut, 2013). given that there is evidence that Belbin's group role theory can be applied
One of the most extensively used design tool for groups is based on to non-managerial personnel as well (Fisher, Hunter, & Macrosson, 1998).
the group role preferences described by Meredith Belbin (1981). Group A comprehensive approach is being employed in which various group
role preferences are dened as group members' predispositions to balance indices are used to predict a wide array of outcomes in collabora-
tive learning groups: teamwork quality, group cognitive complexity, and
group performance, across time.

The authors would like to thank Miel A. P. Vugts for helping with database manage-
ment and Marius T.H. Meeus for his useful comments on previous versions of the manu-
2. Theoretical underpinnings
script. Petru L. Curseu was supported by a grant of the Romanian National Authority for
Scientic Research, CNCS UEFISCDI, project number PN-II-ID-PCE-2011-3-0482. The
funders had no role in study design, data collection and analysis, decision to publish, or Apart from their functional role (prescribed through design), group
preparation of the manuscript. members have the tendency to display particular behavioral patterns
Corresponding author at: Department of Organization Studies, Tilburg University, in interpersonal interactions that inuence the progress of the group
Warandelaan 2, PO box 90153, 5000 LE Tilburg, The Netherlands. Tel.: +31 13466 8299.
E-mail addresses: (N. Meslec),
towards specic task achievement. These stable individual differences
(P.L. Cureu). are captured by the group role preferences (Belbin, 1981). The roots of
Tel.: +31 13 4662153; fax: +31 13 466 3002. group roles were considered to lie in a person's generic personality traits
1041-6080/ 2015 Published by Elsevier Inc.
82 N. Meslec, P.L. Cureu / Learning and Individual Differences 39 (2015) 8188

and mental abilities (Aritzeta, Swailes, & Senior, 2005; Belbin, 1981) as described above. Therefore, the impact of group role balance on group dy-
well as the structure of environment (Arroba & Wedgwood- namics and outcomes is tested by using several balance indices (which
Oppenheim, 1994; Fisher & Macrosson, 1995; Yuwei & Tang, 1997). In rely on different assumptions) as well as a variety of group dynamics
the context of a 9-year research project developed by Belbin, behavioral and performance indicators (group processes, group cognitive complexi-
observations as well as personality and mental abilities of group members ty, perceived and actual group performance).
were recorded and used to develop a taxonomy of group role preferences.
Matching of these measurements resulted in the identication of eight 3. The impact of group role balance on teamwork quality
group roles: the coordinator (CO co-ordinates and controls the activi-
ties of the group), the resource-investigator (RI-extrovert, makes outside Teamwork quality (TWQ) is a multidimensional construct that
contacts and develops ideas), the teamworker (TW-person oriented, reects the quality of interpersonal interactions inside the group.
communicates well with the others), the plant (PL creative and imagi- It consists of several dimensions that reect both group processes
native), the monitor-evaluator (ME prudent and analytical), the imple- (communication, coordination and planning) and group emergent
menter (IM practical and task-oriented), the completernisher (CF states (cohesion, perceived performance and potency) (Cureu & Pluut,
attentive to details, nishes things), and the shaper (SH dynamic and 2013; Cureu, Schalk, & Schruijer, 2010; Hoegl & Gemuenden, 2001).
challenging). Later on, a ninth role was added to this taxonomy, the spe- Teamwork expertise and synergistic interactions within groups in
cialist (SP with high technical skills and in-depth knowledge for the educational settings have several advantages. First, organizations are
task) (see Belbin, 1981, 2009 for an extensive overview). often employing groups and teamwork as a exible way of organizing
One of the most important claims in Belbin's work is that balanced work and therefore are looking for candidates that already acquired
groups (with regard to their members' role preferences) have superior teamwork skills during their educational trajectory (Chen, Donahue, &
performance than unbalanced groups. In other words, it is useful to Klimoski, 2004). Thus, teamwork expertise increases workforce readi-
have members that possess particular strengths without duplicating ness given that students develop during their studies specic teamwork
the ones already present in the group (Belbin, 1981; van de Water, knowledge, skills and abilities. Second, when groups manage to develop
van de Water, & Bukman, 2007). A perfectly balanced group would be synergistic interactions (good teamwork quality) they generate com-
a group in which all nine roles are present in a high or very high level plex group-level knowledge structures (Cureu & Pluut, 2013) as well
while a perfectly unbalanced group would be one in which all the as increased group performance (in innovative projects) and at an indi-
group members report the same individual role preference. The concept vidual level are able to learn more (in terms of knowledge and skills)
of role balance is therefore a congural property of groups (Klein & and be more satised with their work (Hoegl & Gemuenden, 2001).
Kozlowski, 2000), and the conguration of roles are expected to be pre- Given the benets of teamwork quality, substantial effort was devoted
dictive for group dynamics and performance (Belbin, 1981). According to investigate how the quality of interpersonal interactions within groups
to the InputProcessOutput model of group effectiveness (Ilgen, can be improved in educational settings (Chen et al., 2004; Cureu &
Hollenbeck, Johnson, & Jundt, 2005), group role congurations can Pluut, 2013; Cureu et al., 2012). Solutions such as the development of
be considered inputs that predict the group processes and the qual- university courses in which students specically learn about teamwork
ity of interpersonal interactions, which in turn inuence group per- (e.g. The Psychology of Working in Groups and in Teams) (Chen et al.,
formance. However, the results of the studies investigating the impact of 2004) or personality student-group interventions in which group mem-
group role balance on performance are not conclusive. While some bers learn about each other's personalities and how to manage individual
studies identied little or no relation between the two (Blenkinsop differences (Clinebell & Stecher, 2003) were proposed.
& Maddison, 2007; Jackson, 2002; Partington & Harris, 1999; Senior, A more straightforward group design solution is the use of design
1997; van de Water et al., 2008), some others found evidence for principles that generate the most effective group congurations.
group balance as a valid predictor of group performance (Higgss, However, the simple placement of students in groups does not al-
Plewnia, & Ploch, 2005; Prichard & Stanton, 1999). ways guarantee the development of teamwork skills (Hansen, 2006;
The lack of converging results can be attributed to several factors. Johnson & Johnson, 1990). Student groups often experience unclear
First, in most of the studies there is no control for group size and gender goals, mismanagement, conicts or unequal participation (Cox &
diversity although previous research reports gender differences regard- Bobrowski, 2000; McCorkle et al., 1999; McKendall, 2000; Rau &
ing group role preferences (Anderson & Sleap, 2004). Next to the gender Heyl, 1990). The aim of the current study is to investigate whether com-
issue, some of the group balance indices are also sensitive to the group posing groups by using the role balance assumption leads to better group
size (van de Water et al., 2007). Therefore, gender and group size should interactions and thus higher teamwork quality.
be accounted for when analyzing the relation between balance and According to Belbin roles theory, group role balance, as a congural
performance. Second, little specications were given with respect to group property should have a positive impact on teamwork quality. A cru-
how group role balance can be computed. Therefore, there is little overlap cial role here is played by social roles such as coordinator and teamworker.
in the formulas currently used in computing group role balance. Group Due to their association with the extraversion dimension of personality
balance was computed by taking into account the behavioral or environ- (Davies & Kanaki, 2006), such roles facilitate communication and coordi-
mental focus of the roles (Higgss et al., 2005), the weights of the top three nation processes inside the group. The compromising style of conict
roles an individual in a group has (the so-called primary, secondary and management (Aritzeta et al., 2005) associated with these roles also buffers
tertiary roles) (van de Water et al., 2008) or how much the group deviates emergent relationship conicts and leads to positive emergent states such
from an ideally balanced group (Partington & Harris, 1999). The different as group cohesion. The resource-investigator, characterized by over-
assumptions underlying group balance formulas could stand as an expla- optimism not only contributes to the group's belief in their own strengths
nation for the non-conclusive results for the group balance-group perfor- but, also helps in collecting and bridging among different ideas, including
mance relation. Finally, most of the studies used just one group outcome those of introverted group members (e.g. plants or specialists).
indicator, which often differed across studies. For instance, performance Effective teamwork requires a balance between task related and in-
was measured in terms of group processes such as group organization terpersonal knowledge as research on shared mental models argues
and communication (Blenkinsop & Maddison, 2007; Prichard & Stanton, that members of effective groups need to share both task related as
1999) subjective reports of managers (Higgss et al., 2005) or group well as teamwork related knowledge (CannonBowers & Salas, 2001).
success in simulation games (Partington & Harris, 1999; van de Water The convergence of teamwork and taskwork mental models is condu-
et al., 2008). cive to effective teamwork processes which in turn impact on group
In the current study the concept of group role balance and its impact performance (Mathieu, Heffner, Goodwin, Salas, & Cannon-Bowers,
on performance is being reconsidered in order to tackle the limits 2000). Moreover, meta-analytical evidence also suggests that shared
N. Meslec, P.L. Cureu / Learning and Individual Differences 39 (2015) 8188 83

mental models have a positive inuence on the quality of interpersonal resulted from the constraints imposed by the number of students enrolled
processes within groups (DeChurch & Mesmer-Magnus, 2010). A for each of the workshops. Groups had to work together throughout the
balanced role composition secures that both task related (e.g. semester in order to deliver three group assignments that consisted of
specialist, plant) as well as teamwork related (e.g. teamworker, re- three case studies that covered 40% of their individual grades. A cross-
source investigator) orientations are simultaneously present and lagged design was used, data being collected across a seven-week period
as such, balanced groups are expected to have more effective inter- which was indicative for the duration of the course. Data collection in-
actions than unbalanced groups. Therefore, the ne adjustment of volved groups belonging to three different cohorts of students which
one role to another in a balanced group should lead to synergistic were taking the course in three distinct academic years. Group roles (as
teamwork processes. We therefore hypothesize that: a congural property of groups) were assessed in the rst week of the se-
mester. Close to the end of the term (week 7), teamwork quality, per-
H1. Group role balance positively predicts teamwork quality. ceived performance and group cognitive complexity were assessed.
Objective performance was measured with three different group assign-
4. The impact of group role balance on group outcomes ments at three different time points (weeks 3, 5 and 7 of the course).
The separation of measurements in time has the advantage of reducing
Next to perceived group performance and objective performance the common method bias (Podsakoff, MacKenzie, Lee, & Podsakoff, 2003).
group cognitive complexity (GCC) is also used as an outcome variable.
Group cognitive complexity (GCC) was dened as the richness of the 5.2. Measures
collective knowledge structures that emerge as a group-level phenom-
enon from the integration of individual specialized knowledge through 5.2.1. Belbin roles questionnaire and role balance indices
interpersonal interactions (Cureu et al., 2010). The higher the cog- Group roles were assessed using an adaptation of the role self-
nitive complexity, the stronger the group's capabilities to absorb the assessment instrument developed by Belbin (BTRSPI) (Belbin, 1981).
variety of representations held by group members in relation to the The questionnaire (Belbin's team role questionnaire) was part of the
task or the environment (Neill & Rose, 2006) and the higher group per- student's handbook and contained items referring to all nine Belbin
formance (Cureu et al., 2010). roles, including the specialist role which was not present in the 1981
Research to date investigated how GCC and performance can be stim- version of BTRSPI (Huczynsky & Buchanan, 2001). The instrument has
ulated in collaborative learning groups. Factors such as the composition seven sections and each section contains nine items, one for each of
of groups with respect to gender, personal values, teamwork quality the nine roles. For each section, participants had to distribute ten points
and conict were found to play a key-role for group's cognitive complex- across the nine items, considering the behavior that best described him/
ity and performance (Cureu & Pluut, 2013; Cureu et al., 2012; Glew, her in a group. A nal score was computed for each group member for
2009). In line with the claim developed by Belbin (1981), this study in- each of the nine roles. Each role score was computed by adding up the
vestigates whether role balance (as a congural group property) leads points allocated to the seven items that were referring to that particular
to higher cognitive complexity and performance in groups. According role. Theoretically, each role score can take values between 0 and 70.
to Belbin (1981), it is benecial to create groups in which individual char- In order to test the claim that balanced groups perform better, we
acteristics and role preferences are not redundant. A pure group (in reviewed all the balance indices proposed in the literature and used
which all group members detain the same role) will develop a style of the ones that were presented detailed enough in order to be replicated.
its own, focused only on task or relational domains and such group role In addition, a new balance index was added up and also other (thematic)
congurations are not conducive to performance (e.g. a group of CW role patterns indices proposed in the literature. The balance indices (TB)
will be well organized but with a lack of real ideas and inexibility). and role thematic patterns (TP) are described below.
Therefore, a balanced group in which all nine roles are present will The rst balance index (TB1) relies on the assumption that in a bal-
contain all the characteristics needed for group performance. In line anced group the aggregate score from all members will be evenly
with Belbin's balance claim, we hypothesize the following: spread across the nine roles (Partington & Harris, 1999). TB1 is comput-
ed with the following formula: TB1 = 9/[( R 7.77) + 9] 100,
H2. Group role balance will positively predict group outcomes (GCC, where R represents the deviation of each of the nine roles, summed
perceived and objective group performance). up at a group level. Thus, TB1 considers the amount of roles directly at
a group level, disregarding their original congurations at an individual
level. A perfectly balanced group would have an R on each role equal to
5. Method
7.77. This score represents the 70 points to be distributed divided by the
nine roles. If the total deviation of the role mean from 7.77 is 0, then
5.1. Sample and procedure
TB1 = 100%.
The premise for the second balance index (TB2 and TB2a) is that a
The sample consisted of 459 students (151 female) enrolled in an
balanced group would have at least one person scoring high or very
Organizational Behavior course at a Dutch university. Participants
high in as many as possible of the nine roles (Partington & Harris,
were asked to form groups at the onset of the course. Group size ranged
1999). For each role, norms for dening high and very high scores
from three to seven members, resulting in 84 groups. The vast majority of
were computed. TB2 uses the norms dened by Belbin (1981, p.158)
the groups had ve or six members. For a detailed overview of demo-
while TB2a uses the norms constructed by using the sample of this par-
graphics see Table 1. The more extreme group sizes (three and seven)
ticular study. For each role, all the individual scores are ordered from the
smallest score to the highest. The scores are split in four quartiles, the
Table 1
highest two quartiles (from 66 to 100%) containing the range of scores
Sample demographics.
which qualify as high and very high scores for that particular role. The
Group size Number of women/group Total norms for this study were similar with the ones identied by Belbin
0 1 2 3 4 5 (1981, p.158) and Partington and Harris (1999, p. 700). Each group
3 1 0 0 1 0 0 2 received a maximum of one point for each role represented in a
4 0 4 1 0 0 0 5 high or very high level and two points if the role was not represent-
5 4 8 15 7 0 0 34 ed in a high or very high level. The following formula was used
6 4 6 17 6 3 1 37 TB2 = (9/ HR) 100%, where HR represents the sum of points
7 1 1 2 1 0 0 5
given for each role represented in a high or very high level.
84 N. Meslec, P.L. Cureu / Learning and Individual Differences 39 (2015) 8188

The third balance score (TB3) was computed on the premise that dimensionality of the scale, with the rst factor having an eigenvalue
ideally roles should not be duplicated (Belbin, 1981; Partington & of 3.33 and explaining for 33.30% of the total variance while the other
Harris, 1999). The formula used is similar with the one for TB2, two factors identied explain only 14.13 and respectively 11.49 of the
TB3 = (9/HR) 100%. The difference lies in the number of points at- total variance. This comes in line with previous studies indicating that
tributed to each role. In the case of TB3, each role represented in a high although teamwork quality has multiple facets its underlying structure
or very high level counts for the number of points given. pertains to one unique factor (Eby et al., 1999; LePine, Piccolo, Jackson,
Finally, the last balance index (TB4) considers group roles as a Mathieu, & Saul, 2008). The Alpha Cronbach for teamwork quality was
qualitative property of the groups, in which both the richness of roles .75, which indicates a good reliability of the scale. The teamwork quality
represented in the group and the abundance of the roles making up the scale was administered at the end of the course.
richness are considered. Each group received one point for each role rep-
resented in a high or very high level and 0 points for roles not represented Group cognitive complexity (GCC). Group cognitive complexity
as such. Group balance was computed using Simpson's index of diversity was measured through the cognitive map technique that is a group
1 - D = n (n 1)/N(N 1), where n is the total number of high and task (Cureu et al., 2010). Groups received 20 cards on which major con-
very high scores of a particular role and N is the total number of high and cepts from the course domain were written. Their task was to organize
very high scores of all roles (Simpson, 1949). the concepts in a way that reects their understanding as a group of the
To summarize, four group balance indices were used, each of them relations between the concepts. Once agreed on the nal chart, they had
relying on different theoretical considerations. The rst balance index to glue the cards on a paper and draw the relations between the con-
TB1 accounts for Belbin roles directly at a group level while TB2 and cepts as well as to indicate the type of relations between the concepts.
TB2a consider roles at an individual level being afterwards congured The nal maps were assessed while using three criteria: total number
at a group level. TB3 is similar to TB2, bringing in addition a correction of concepts used (NoC), total number of connections established
for roles that duplicate each other. Finally, TB4 attempts to account for between the concepts (CMC) and the number of distinct relations
Belbin roles in a more comprehensive way, while considering both the established (CMD). These three dimensions are indicative for group
variety of roles represented in the group as well as their elevation. cognitive complexity as the complexity of the conceptual systems in-
Next to the group role balance, Belbin classies the proposed roles cluding on the one hand the number of dimensions belonging to the
in three thematic patterns (TP): relationship (CO, RI, TW), thinking system and one the other hand the nature and the extent of rules elicit-
(PL, ME and SP) and action roles (IM, CF, SH) (Belbin, 2009). This last ed for integrating these dimensions (Calori, Johnson, & Sarnin, 1994;
distinction was also included in our analyses. For each thematic catego- Cureu et al., 2010). The type of relations was rated using seven different
ry a group score was computed by averaging the individual scores to categories: causal, association, equivalence, topological, structural,
the group level. For instance, for the relationship thematic pattern, chronological, and hierarchical (Cureu et al., 2010; Gmez, Moreno,
TPrelationship = ( COIndScores/GroupSize + RIIndScores/ Pazos, & Sierra-Alonso, 2000). Group cognitive complexity was comput-
GroupSize + TWIndScores/GroupSize)/3, where IndScores represents ed using the following formula CMCo = (CMCxCMD)/NoC (Cureu
the individual score for that particular role. et al., 2010).

5.2.2. Teamwork quality and group outcomes Group performance. Perceived performance was measured with 2
items (Cureu et al., 2010) on a 5-point Likert scale. One example of an Teamwork quality (TWQ). Groups had to rate their own item is: I think that the present group performance is. For objective
teamwork quality on a 5-point scale questionnaire in which they evaluat- performance the grades obtained by the groups at three different as-
ed different dimensions of teamwork: communication, cohesion, coordi- signments were used. Groups had to deliver three assignments in
nation, planning and potency (Carron, Widmeyer, & Brawley, 1985; weeks 3, 5 and 7, which represented 40% of their nal grade. In the
Cureu et al., 2010; Eby, Meade, Parisi, & Douthitt, 1999; Guzzo, Yost, three assignments, groups had to analyze and solve three case studies
Campbell, & Shea, 1993). All items and their respective factor loadings while using topics studied during the course (motivation at work, lead-
are presented in Table 2. Data was collected directly at a group level, ership and group dynamics). The maximum grade for the rst assign-
group members being asked to reect on their own teamwork processes ment was 10 points while for the other two assignments they could
during the semester and decide on a group rating for each of the have earned a maximum of 15 points.
teamwork items. Data collection through group agreement was
argued to be superior to individual data collection aggregated sub- 5.3. Control variables and data analysis strategies
sequently at a group level (Kirkman, Tesluk, & Rosen, 2001).The average
score of all the scales was used as an indicator for TWQ. An exploratory In order to test the hypotheses several Ordinary Least Squares (OLS)
factor analysis conducted with the scale items supports the one- regression models were run with group balance indices and thematic
patterns as predictors and teamwork quality, group cognitive complex-
ity, perceived performance and objective performance (at three differ-
Table 2
Items and factor loadings for teamwork quality scale. ent time points) as dependent variables. A separate OLS regression for
each of the dependent variables was run. Empirical evidence shows
Items Factor
that there are signicant differences between men and women in
group roles assumed (Anderson & Sleap, 2004). In order to control for
It was easy for all team members to express their ideas during team .375 the potential inuence of gender-based role preferences on our results,
Our group is united in trying to make high quality assignments .732
in the OLS we used the percentage of females in each of the groups as a
The group members feel they belong to this group .463 control variable. Given that some group balance indices are sensitive to
The group members share with each other the information they have .599 the number of members a group has (van de Water et al., 2007), group
regarding the projects size was used as a control variable.
The group members have synchronized their actions in order to reach .489
the group goals
Each group member was completely aware of his/her own tasks .414 6. Results
The group has clearly dened performance standards .490
Each member is completely aware of the group's goals and objectives .750 Perceived performance correlates positively with teamwork quality
This team has condence in itself .656 r(84) = .63, p = .00. and with objective performance at time 1, r
This team believes it can be very productive .662
(84) = .45, p = .00 (Tables 3 and 4). Group size is not associated with
N. Meslec, P.L. Cureu / Learning and Individual Differences 39 (2015) 8188 85

Table 3
Bivariate correlations at group level.

Variable 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

1. Group size
2. Percent females .11
3. TB1 .06 .05
4. TB2 .46 .00 .03
5. TB2a .26 .11 .06 .39
6. TB3 .32 .09 .01 .13 .02
7. TB4 .32 .06 .08 .03 .53 .33
8. Relationship roles .19 .00 .36 .08 .16 .19 .21
9. Acting roles .09 .06 .19 .21 .05 .16 .05 .47
10. Thinking roles .07 .07 .02 .14 .09 .02 .25 .41 .59
11. TWQ .11 .17 .20 .02 .10 .04 .17 .22 .06 .25
12. Perceived performance .23 .08 .05 .09 .01 .07 .07 .06 .18 .11 .63
13. GCC .15 .25 .28 .16 .04 .08 .15 .04 .08 .14 .14 .13
14. Performance 1 .22 .12 .28 .10 .13 .01 .06 .24 .22 .05 .18 .45 .00
15. Performance 2 .03 .00 .03 .02 .11 .00 .02 .04 .06 .12 .12 .17 .09 .19
16. Performance 3 .01 .02 .06 .03 .00 .02 .11 .00 .00 .02 .08 .12 .16 .20 .61
p b .05.
p b .01.

TB1, r(84) = .06, p = .56, and positively associated with TB2, r(84) = to teamwork quality ( = .25, t(81) = 2.26, p = .02) and GCC ( = .24,
.46, p = .00, and TB2a, r(84) = .26, p = .00 while negatively associated t(81) =2.24, p = .02).
with TB3, r(84) = .32, p = .00, and TB4, r(84) = .32, p = .00. Per-
centage of women is positively associated to group cognitive complexi-
ty, r(84) = .25, p = .02. Finally, some signicant correlations between 7. Discussions
the group balance indices were identied: TB2 and TB2a, r(84) = .39,
p = .00, TB2a and TB4, r(84) = .53, p = .00, TB3 and TB4, r(84) = .33, In the current study one of the major claims of Belbin's role theory
p = .00. was tested, namely that balanced groups (in which all nine Belbin
In order to test the proposed hypotheses a separate OLS regression roles are present) are associated with several group outcomes (group
analysis was conducted for each of the dependent variables. Group cognitive complexity, perceived and objective performance) and team-
size and percentage of women were used as control variables in the work quality. Among the team balance indicators, TB1 appears to be the
rst step of the regression analysis. Team balance indices were added only to predict objective performance (measured at time 1) GCC and
in the second step of regression analysis. Due to multicolinearity issues, teamwork quality (measured at the end of the group project). What dif-
the thematic patterns could not be included in one single OLS regres- ferentiates TB1 from the other balance indices is that roles are not con-
sion. Thus separate regression models were conducted for each of the sidered at an individual level rst and then summed up at a group level
three thematic patterns. but, rather as congurations of roles directly at a group level. For example,
The results indicate that TB1 positively predicts GCC ( = 0.28, for the coordinator role all the points belonging to all group members
t(81) = 2.65, p = .01) and objective performance at time 1 ( = 0.24, (whether they are high or low) are being summed up and included in
t(81) = 2.30, p = .02), and negatively predicts teamwork quality the balance score. The rest of the formulas start from the assumption
( = .21, t(81) = 2.02, p = .04). All the other team balance indices that one particular role (e.g. coordinator) should be represented in a
are not signicantly related with any of the other team outcomes high or very high level at one particular group member and that should
(Tables 5 and 6). The three thematic congurations also do not predict be considered for the group balance. The other low individual scores on
group perceived performance, objective performance or GCC. Neverthe- that particular role are not being accounted for.
less, thinking roles negatively predict teamwork quality ( = .25, TB1 was found to positively predict objective performance, but only
t(81) = 2.36, p = .02) while relationship roles positively predict at time 1 and to be negatively related to teamwork quality as measured
teamwork quality ( = .26, t(81) = 2.47, p = .01) (Tables 7 and 8). Fi- at the end of the group project. We originally hypothesized that accord-
nally, the percentage of women as a control variable is positively related ing to Belbin's balance claim, the ne adjustment of one role to another
in a balanced group would lead to synergistic teamwork processes and
increased team performance. Furthermore, as research on shared men-
Table 4 tal models argues, effective groups share in a balanced way both task as
Descriptive statistics.

Variable Mean SD Min Max

Table 5
1. Group size 5.46 .79 3 7 Impact of group balance on teamwork quality, perceived performance and cognitive
2. Percent females 33.66 19.88 0 100 complexity.
3. TB1 99.99 3.71 88.21 116.33
4. TB2 79.61 8.61 61.53 100 Step Variable Teamwork Perceived Cognitive
5. TB2a 81.87 6.37 66.67 88.89 quality performance complexity
6. TB3 54.76 7.38 38.10 72.73
R2 R2 R2
7. TB4 .90 .02 .85 .94
8. Relationship roles 8.03 .87 6.13 12.08 1 Group size .25 .02 .32 .04 .02 .06
9. Acting roles 8.81 .95 6.60 12.17 Percent females .25 .13 .20
10. Thinking roles 6.47 .93 3.17 8.80 2 TB1 .21 .07 .08 0 .28 .11
11. TWQ 3.87 .41 2.80 4.80 TB2 .06 .01 .11
12. Perceived performance 3.86 .59 2.50 5.00 TB2a .11 .12 .07
13. GCC 2.64 .89 .60 5.70 TB3 .18 .01 .00
14. Performance 1 7.95 1.06 5 10 TB4 .08 .08 .22
15. Performance 2 12.19 2.18 5 15
p b .10.
16. Performance 3 11.79 1.77 7 15
p b .05.
86 N. Meslec, P.L. Cureu / Learning and Individual Differences 39 (2015) 8188

Table 6 Table 8
Impact of group balance on objective performance at 3 different time points. Impact of role thematic congurations on objective performance at three different time
Step Variable Performance 1 Performance 2 Performance 3
Step Variable Performance 1 Performance Performance
R2 R2 R2
2 3
1 Group size .33 .05 .18 0 .15 0
R2 R2 R2
Percent females .17 .05 .00
2 TB1 .24 .08 .04 0 .09 0 1 Group size .22 .05 .04 0 .01 0
TB2 .04 .02 .02 Percent females .14 .01 .01
TB2a .05 .24 .18 2 Action roles .19 .08 .06 0 0 0
TB3 .06 .03 .02 1 Group size .24 .05 .03 0 .02 0
TB4 .15 .16 .28 Percent females .17 .01 .02
2 Thinking roles .06 .04 .12 0 .02 0
p b .05.
1 Group size .20 .05 .02 0 .01 0
Percent females .15 .01 .01
2 Relationship roles .20 .08 .04 0 .01 0
p b .10.
well as teamwork related knowledge (CannonBowers & Salas, 2001). p b .05.
Our results indicate that although balanced teams perform signicantly
better in initial phases of group development, this effect disappears to- accurately predict group objective performance or GCC. Nevertheless,
wards the end of the group project. This could be explained by other fac- thinking roles negatively predict teamwork quality while relationship
tors that have not been accounted for in the study. For example, the roles positively predict teamwork quality. Having teams composed most-
deterioration of interpersonal interactions in originally balanced groups ly of thinking oriented roles (plant, monitor evaluator and specialist)
could have buffered the positive relation between group role balance might interfere with the quality of interactions developed within groups.
and group performance at time points 2 and 3. The negative relation be- This comes in line with previous studies indicating that thinking roles are
tween TB1 and teamwork quality (measured at the end of the group characterized by a tendency towards domination in conict, the use of
project) comes to support this claim. Studies indicate that initial group Machaivelian behaviors and fewer integrative cognitive styles (Aritzeta,
conict is benecial for group performance as it stimulates information Swailes, & Senior, 2007; Aritzeta et al., 2005; Macrosson & Hemphill,
exchange but when it escalate in more profound relationship conicts it 2001). On the other hand relationship oriented roles (coordinator, re-
becomes detrimental for the group (Yang & Mossholder, 2004). It could source investigator and teamworker) might facilitate the quality of inter-
be that balanced groups, engaged in constructive task conicts at the actions within groups. Relationship roles were shown to correlate with
onset of the group project, yet in time, task conict evolved into rela- the extraversion dimension of personality (Davies & Kanaki, 2006) and
tionship frictions that damaged the quality of interpersonal interactions bridging and adaptive cognitive styles (Aritzeta et al., 2007; Fisher et al.,
and ultimately group performance. Thus, the disappearing effect of 1998). Further on, group members adopting relationship roles are more
team balance on group performance from time 1 to time 2 and 3 could likely to adopt avoiding and compromising conict management styles
be due to process losses experienced by the group members. Further (Aritzeta et al., 2005). These characteristics of relationship roles might fa-
studies could investigate the mechanisms at play here. Interestingly, al- cilitate the quality of interpersonal interactions within groups.
though balanced groups do not manage to maintain the same levels of Percentage of women was found to be signicantly related to team-
performance at later stages, they do develop a higher level of cognitive work quality and GCC. One possible explanation for this nding can be
complexity than unbalanced groups. TB1 positively predicts the level of that women display a stronger preference for relationship roles than
GCC (as measured at the end of the group project). GCC was shown pre- men. This claim is in line with previous empirical evidence that
viously to be a stable group emergent state and for this reason might be women have a stronger preference for the teamworker role than men
more resilient to changes or negative interaction patterns (Cureu et al., (Anderson & Sleap, 2004; Sommerville & Dalziel, 1998). The same pref-
2010). Although balanced groups might be able to consider a higher num- erence was also found in our study t = 2.09, p b .05 (mean scores for
ber of relations between relevant knowledge structures and thus to have women = 9.24 and mean scores for men = 8.24). An alternative expla-
a richer representation of the task they might still have difculties in nation can be linked to the development of collective emotional intelli-
translating this into performance due to negative interactions developed gence in groups. Recent studies indicate that the percentage of women
at later stages of group development. in the group fosters the emergence of collective emotional intelligence
Next to team balance indices, the impact of the three thematic cong- that further increases social integration (high cohesion and reduced
urations (action, relationship and task related roles) on group outcomes relationship conict) within groups (Cureu, Pluut, Boro, & Meslec, in
was also tested. Our results indicate that thematic congurations do not press).
Finally, a few correlational ndings warrant attention. Perceived
Table 7 performance was found to signicantly correlate with both teamwork
Impact of role thematic congurations on teamwork quality, perceived performance and quality and objective performance at time 1. This comes in line with
cognitive complexity.
previous studies indicating that teamwork processes are positively as-
Step Variable Teamwork Perceived Cognitive sociated with both group performance and members' satisfaction
quality performance complexity (LePine et al., 2008). Group size is not associated with TB1 and positively
R2 R2 R2 associated with TB2 and TB2a. This indicates that larger groups have a
higher chance of being role-balanced, which involves the representa-
1 Group size .13 .02 .22 .04 .14 .06
Percent females .18 .10 .24 tion of all nine roles at a high or very high level. However, this reasoning
2 Action roles .03 .01 .15 .05 .04 .05 does not hold for TB1, an index in which roles are counted for directly at
1 Group size .15 .02 .25 .04 .18 .06 a group level. Thus, group size is not as relevant for TB1 as it is for TB2
Percent females .17 .10 .23
and TB2a. Further on, group size is negatively correlated to TB3 and
2 Thinking roles .25 .07 .13 .04 .13 .07
1 Group size .18 .02 .24 .04 .16 .06 TB4. TB3 and TB4 although similar to TB2 in counting the roles which
Percent females .19 .11 .25 are represented at a high or very high level, also include corrections
2 Relationship roles .26 .08 .01 .03 .07 .05 for roles that are duplicating each other and respectively for the intensi-
p b .10. ty of roles. These corrections could explain why the link with group size
p b .05. is negative. TB1 does not correlate with any of the other balance indices,
N. Meslec, P.L. Cureu / Learning and Individual Differences 39 (2015) 8188 87

given that the index is constructed by using a different assumption, explicitly compose groups based on role preferences and investigate
namely that group roles are considered directly at a group level. Per- the impact of group role balance on group outcomes.
centage of women positively correlates with group cognitive complexi-
ty. This comes in line with previous ndings indicating the important 9. Conclusions
role of gender for groups and group performance (Cureu et al., in
press; Woolley, Chabris, Pentland, Hashmi, & Malone, 2010). Belbin group roles are widely used in organizations, although the
initially proposed claims have created many scientic controversies
8. Implications, limitations and directions for further research (Belbin, 1993; Furnham, Steele, & Pendleton, 1993). Our study indicates
that when group role balance is conceptualized as a congural group
Our study has important implications for group design literature as property instead of a sum of individual roles, it positively predicts
well as for collaborative learning literature. First, we report the results group performance at least at initial phases of the group work and it
of a comprehensive test of various group congurations with respect positively predict group cognitive complexity. Nevertheless it links neg-
to group role preferences on various collaborative learning outcomes. atively with teamwork quality and it does not predict performance at
We nd that team balance is relevant at least for initial levels of group later points in time. Further studies could investigate the dynamic ef-
performance when it is considered as a group congural property rather fects of role balance over time on group performance. Next to this,
than a sum of individual roles. Further on, balanced groups are associated other group diversity indicators such as the percentage of females in
with higher levels of cognitive complexity. This indicates that forming the group appear to be highly predictive for teamwork quality and
groups based on Belbin role balance assumption might enhance group group cognitive complexity and thus should warrant further research.
performance, at least in initial phases and have persistent effects on the
level of group cognitive complexity. Further research could explore here References
other intervening processes such as group conict that might explain
the drop in performance from time 1 to time 2 and 3. Additionally, we Anderson, N., & Sleap, S. (2004). An evaluation of gender differences on the Belbin team
role self-perception inventory. Journal of Occupational and Organizational Psychology,
nd that the percentage of women positively predicts teamwork quality 77, 429437.
and GCC. We admit that using the percentage of women as a design prin- Aritzeta, A., Swailes, S., & Senior, B. (2005). Team roles: Psychometric evidence, construct
ciple when forming student groups in educational settings has limitations, validity and team building. Research Memorandum.
Aritzeta, A., Swailes, S., & Senior, B. (2007). Belbin's team role model: Development, valid-
yet a straightforward implication of our results is that educators should at
ity and applications for team building. Journal of Management Studies, 44(1), 96118.
least attempt to form mixed gender groups. This is in line with previous Arroba, T., & Wedgwood-Oppenheim, F. (1994). Do senior managers differ in the public
results showing that gender composition has positive inuence on GCC and private sectors? Journal of Managerial Psychology, 9(1), 1316.
Belbin, M. (1981). Management teams, why they succeed or fail. London: Heinemann.
(Cureu, Schruijer, & Boro, 2007) and group rationality (Cureu, Jansen
Belbin, R.M. (1993). A reply to Belbin team role self-perception inventory by Furnham,
& Chappin, 2013). When gender composition is not a feature open to ma- Steele and Pendleton. Journal of Occupational and Organizational Psychology, 66,
nipulation, educators could focus on short trainings to increase interper- 259260.
sonal awareness and social sensitivity or use normative interventions to Belbin, M. (2009). Method, reliability & validity, statistics and research: A comprehensive
review of Belbin team roles. Retrieved from www.
inuence directly the quality and nature of interpersonal interactions in Blenkinsop, N., & Maddison, A. (2007). Team roles and team performance in defence ac-
collaborative learning groups (Cureu & Schruijer, 2012). quisition. Journal of Management Development, 26(7), 667682.
Future research could explore possible explanatory mechanisms and Burch, G.S.J., & Anderson, N. (2004). Measuring personteam t: Development and
validation of the team selection inventory. Journal of Managerial Psychology, 19,
our study points out that the quality of interpersonal interactions needs 406426.
further attention in this respect. Finally, our study uses data collected Calori, R., Johnson, G., & Sarnin, P. (1994). CEOs' cognitive maps and the scope of the
from multiple sources (e.g., self-rated performance, objective perfor- organization. Strategic Management Journal, 15, 437457.
CannonBowers, J.A., & Salas, E. (2001). Reections on shared cognition. Journal of
mance indicators and group cognitive complexity rated by external Organizational Behavior, 22(2), 195202.
evaluators) and measures are separated in time, therefore the likely im- Carron, A.V., Widmeyer, W.N., & Brawley, L.R. (1985). The development of an instrument
pact of common method bias is low. Given the strengths of our design, to assess cohesion in sport teams: The Group Environment Questionnaire. Journal of
Sport Psychology, 7, 244266.
the substantial sample size and the variety of outcomes considered we Chen, G., Donahue, L.M., & Klimoski, R.J. (2004). Training undergraduates to work in orga-
can conclude that our ndings are robust, yet future research is wel- nizational teams. Academy of Management Learning and Education, 3, 2740.
comed in replicating them in other organizational contexts. Clinebell, S., & Stecher, M. (2003). Teaching teams to be teams: An exercise using the
Myers-Briggs type indicator and the ve-factor personality traits. Journal of
Next to these strengths, our study has also certain limitations. First,
Management Education, 27(3), 362383.
group roles were measured at the beginning of the workgroup as indi- Cox, P.L., & Bobrowski, P.E. (2000). The team chapter assignment: Improving the effec-
vidual preferences for particular behavioral styles. However, we did tiveness of classroom teams. Journal of Behavioral and Applied Management, 1,
not check whether these were indeed the roles assumed by the group 92103.
Cureu, P.L. (2011). Need for cognition and active information search in small student
members during their work. Although group roles are in principle con- groups. Learning and Individual Differences, 21(4), 415418.
sidered to be individual preferences that are stable over time, it could be Cureu, P. L., Jansen, R. J., & Chappin, M. M. (2013). Decision rules and group rationality:
the case that due to dynamics unfolded in the group, members take cognitive gain or standstill? PloS one, 8(2), e56454.
Cureu, P.L., Janssen, S.E.A., & Raab, J. (2012). Connecting the dots: Social network struc-
other roles (e.g. secondary or tertiary roles) than the ones they have ture, conict, and group cognitive complexity. Higher Education, 63(5), 621629.
their rst preference for. Second, the self-reported nature of the team- Cureu, P.L., & Pluut, H. (2013). Student groups as learning entities: The effect of group di-
work measure might be a limitation of the study. Nevertheless, in versity and teamwork quality on groups' cognitive complexity. Studies in Higher
Education, 38(1), 87103.
using this scale we do align with previously proposed measures for Cureu, P.L., Pluut, H., Boro, S., & Meslec, N. (2015). The magic of collective emotional
teamwork quality (Eby et al., 1999; LePine et al., 2008). Third, our intelligence in learning groups: No guys needed for the spell! British Journal of
groups differed in size and this might be problematic especially for the Psychology (in press).
Cureu, P.L., Schalk, R., & Schruijer, S. (2010). The use of cognitive mapping in eliciting and
computation of group role balance that involves the presence in a high evaluating group cognitions. Journal of Applied Social Psychology, 40(5), 12581291.
or very high level the presence of all nine roles. Although we tried to Cureu, P., & Schruijer, S. (2012). Normative interventions, emergent cognition and deci-
overcome this limit by controlling for group size in our analyses and sion rationality in ad-hoc and established groups. Management Decision, 50(6),
by using different group balance indices, we do acknowledge this as a
Cureu, P.L., Schruijer, S.G.L., & Boro, S. (2007). The effects of groups' variety and disparity
limitation of our study. Further studies could try to replicate these on groups' cognitive complexity. Group Dynamics: Theory, Research and Practice,
results in a more controlled setting in which all groups are composed 11(3), 187206.
with the same size and group roles preference stability is accounted Davies, M., & Kanaki, E. (2006). Interpersonal characteristics associated with different
team roles in work groups. Journal of Managerial Psychology, 21(7), 638650.
for. Finally, in this study group role balance was not manipulated and DeChurch, L.A., & Mesmer-Magnus, J.R. (2010). The cognitive underpinnings of effective
thus causal inferences are not warranted. Further studies could try to teamwork: A meta-analysis. Journal of Applied Psychology, 95(1), 32.
88 N. Meslec, P.L. Cureu / Learning and Individual Differences 39 (2015) 8188

Eby, L.T., Meade, A.W., Parisi, A.G., & Douthitt, S.S. (1999). The development of an individ- McCorkle, D.E., Reardon, J., Alexander, J.F., Kling, N.D., Harris, R.C., & Iyer, R.V. (1999). Un-
ual level teamwork expectations measure and the application of a within group dergraduate marketing students, group projects, and teamwork: The good, the bad,
agreement statistic to assess shared expectations for teamwork. Organizational and the ugly? Journal of Marketing Education, 21, 106117.
Research Methods, 2, 366394. McCune, V., & Entwistle, N. (2011). Cultivating the disposition to understand in 21st
Fisher, S.G., Hunter, T.A., & Macrosson, W.D.K. (1998). The structure of Belbin's team roles. century university education. Learning and Individual Differences, 21(3), 303310.
Journal of Organizational and Occupational Psychology, 71, 283288. McKendall, M. (2000). Teaching groups to become teams. Journal of Education for Business,
Fisher, S.G., & Macrosson, W.D.K. (1995). Early inuences on management team roles. 75, 277282.
Journal of Managerial Psychology, 10(7), 815. Neill, S., & Rose, G.M. (2006). The effect of strategic complexity on marketing strategy and
Furnham, A., Steele, H., & Pendleton, D. (1993). A response to Dr. Belbin's reply. Journal of organizational performance. Journal of Business Research, 59, 110.
Organizational and Occupational Psychology, 66, 261. O'Neil, H.F., Allerd, K., & Baker, E.L. (1997). Review of workforce readiness theoretical
Glew, D.J. (2009). Personal values and performance in teams: An individual and team- frameworks. In H.F. O'Neil (Ed.), Workforce readiness: Competencies and assessment
level analysis. Small Group Research, 40(6), 670693. (pp. 325). Mahwah, NJ: Erlbaum.
Gmez, A., Moreno, A., Pazos, J., & Sierra-Alonso, A. (2000). Knowledge maps: An essential Partington, D., & Harris, H. (1999). Team role balance and team performance: An empir-
technique for conceptualisation. Data and Knowledge Engineering, 33, 169190. ical study. Journal of Management Development, 18(8), 694701.
Guzzo, R.A., Yost, P.R., Campbell, R.J., & Shea, G.P. (1993). Potency in groups: Articulating a Podsakoff, P.M., MacKenzie, S.B., Lee, J.Y., & Podsakoff, N.P. (2003). Common method
construct. British Journal of Social Psychology, 32, 87106. biases in behavioral research: A critical review of the literature and recommended
Hansen, R.S. (2006). Benets and problems with student teams: Suggestions for improv- remedies. Journal of Applied Psychology, 88(5), 879.
ing team projects. Journal of Education for Business, 82(1), 1119. Prichard, J.S., & Stanton, N.A. (1999). Testing Belbin's team role theory of effective groups.
Haugwitz, M., Nesbit, J.C., & Sandmann, A. (2010). Cognitive ability and the instructional The Journal of Management Development, 18(8), 652665.
efcacy of collaborative concept mapping. Learning and Individual Differences, 20(5), Rau, W., & Heyl, B.S. (1990). Humanizing the college classroom: Collaborative learning
536543. and social organization among students. Teaching Sociology, 18, 141155.
Higgss, M., Plewnia, U., & Ploch, J. (2005). Inuence of team composition and task com- Senior, B. (1997). Team roles and team performance: Is there really a link? Journal of
plexity on team performance. Team Performance Management, 11, 227250. Occupational and Organizational Psychology, 70, 241258.
Hoegl, M., & Gemuenden, H.G. (2001). Teamwork quality and the success of innovative Simpson, E.H. (1949). Measurement of diversity. Nature, 163, 688.
projects: A theoretical concept and empirical evidence. Organization Science, 12(4), Sommerville, J., & Dalziel, S. (1998). Project teambuildingThe applicability of Belbin's
435449. team-role self-perception inventory. International Journal of Project Management,
Huczynsky, A., & Buchanan, D. (2001). Organizational behaviour: An introductory text (4th 16(3), 165171.
ed.). Essex: Pearson Education Limited. Stevens, M.A., & Campion, M.J. (1999). Stafng work teams: Development and validation
Ilgen, D.R., Hollenbeck, J.R., Johnson, M., & Jundt, D. (2005). Teams in organizations: From of a selection test for teamwork settings. Journal of Management, 25, 207228.
inputprocessoutput models to IMOI models. Annual Review of Psychology, 56, Tannenbaum, S.I., Mathieu, J.E., Salas, E., & Cohen, D. (2012). Teams are changing: Are
517543. research and practice evolving fast enough? Industrial and Organizational
Jackson, C. (2002). Predicting team performance from a learning process model. Journal of Psychology, 5, 224.
Managerial Psychology, 17(1), 613. van de Water, H., Ahaus, K., & Rozier, R. (2008). Team roles, team balance and perfor-
Johnson, D.W., & Johnson, R.T. (1990). Cooperative learning and achievement. In S. Sharan mance. Journal of Management Development, 27(5), 499512.
(Ed.), Cooperative Learning: Theory And Research (pp. 2337). New York: Praeger van de Water, T., van de Water, H., & Bukman, C. (2007). A balanced team generating
Press. model. European Journal of Operational Research, 180, 885906.
Kirkman, B.L., Tesluk, P.E., & Rosen, B. (2001). Assessing the incremental validity of team Woolley, A.W., Chabris, C.F., Pentland, A., Hashmi, N., & Malone, T.W. (2010). Evidence for
consensus rating over aggregation of individual-level data in predicting team effec- a collective intelligence factor in the performance of human groups. Science,
tiveness. Personnel Psychology, 54, 645667. 330(6004), 686688.
Klein, K.J., & Kozlowski, W.J. (2000). From micro to meso: Critical steps in conceptualizing Yang, J., & Mossholder, K.W. (2004). Decoupling task and relationship conict: The role of
and conducting multilevel research. Organizational Research Methods, 3(3), 211236. intragroup emotional processing. Journal of Organizational Behavior, 25(5), 589605.
LePine, J.A., Piccolo, R.F., Jackson, C.L., Mathieu, J.E., & Saul, J.R. (2008). A metaanalysis of Yuwei, S., & Tang, H.K. (1997). Team role behaviour and task environment: An explorato-
teamwork processes: Tests of a multidimensional model and relationships with team ry study of ve organizations and their managers. Journal of Managerial Psychology,
Effectiveness criteria. Personnel Psychology, 61(2), 273307. 12(2), 8594.
Macrosson, W.D., & Hemphill, D.J. (2001). Machiavellianism in Belbin team roles. Journal Zedeck, S., & Goldstein, I.L. (2000). The relationship between I/O psychology and public
of Managerial Psychology, 16(5), 355363. policy: A commentary. In J.F. Kehoe (Ed.), Managing selection in changing organizations:
Mathieu, J.E., Heffner, T.S., Goodwin, G.F., Salas, E., & Cannon-Bowers, J.A. (2000). The in- Human resource strategies (pp. 371396). San Francisco, CA: Jossey- Bass.
uence of shared mental models on team process and performance. Journal of Applied
Psychology, 85(2), 273.