Vous êtes sur la page 1sur 19

Journal of Occupational and Organizational Psychology {\991), 70, 241-258 1997 The British Psychological Society

Printed in Great Britain

24l

Team roles and team performance: Is there 'really' a link?


Barbara Senior*
New College of Higher Education. Park Campus. Northampton NN2 1 AL. UK

Belbin's (1981, 1993) team role theories are evaluated in terms of the claim that high team performance is associated with teams which are balanced in terms of the team roles represented amongst team members. Eleven teams from a mixture of private and public organizations were surveyed to test the proposition that a balanced team will be a high performing team and vice vetsa. Issues of measurement of team balance and team petformance are addressed. The research is believed to be unique in including a number of criteria tor team role balance rather than rhe single, simplistic measure frequently used. An innovative measure of team performance is used. The results of the investigation give some support to the link Belbin makes between team role balance and team performance. These are discussed in the light of the paucity of rigorous tesearch in this area to date.

The performance of an organization in seeking to achieve organizational goals depends on many factors such as strategy, structure, technology, people employed and management style. Of importance amongst these is the 'people' factor, that is the behaviour of individual employees and the contribution this makes to performance at individual, group and organization level. Some writers (e.g. Burns & Stalker, 1966; Morgan, 199.3; Weber, 1964) relate the organizarions to the structure they adopt whilst others (e.g. Herzberg, 1968; McClelland, 1988;Maslow, 1943; Roethlisberger & Dickson, 1964; Taylor, 1947}concentrate on what motivates the individual worker. However, current pressure to 'downsize' organizations, resulting in a reduction in the number of levels with consequent development of more flexible forms of working (Drucker, 1988; Tjosvold, 1991), arguments about the empowerment of workers (Conjer, 199.3; Moss-Kanter, 1993), and rhe interdependence between individuals which is said ro characterize most modern semi-auromated work (Smith, Peterson & Misumi, 1994), has brought about an emphasis on the work of the group or on the more recent development of that concept, the work of the team. Indeed, a recent article (Brown, 1995) emphasizes this trend by pointing out the rapid growth in group based rewards as opposed to individualistic based reward schemes. Likert s (1961) linking pin theory sees its expression in these ideas that everyone is part of one or more teams, whether production or service oriented or part of the management
*Requtsrs tor reprints.

242

Barbara Senior

structure of the organizations. The performance of teams within organizations is, therefore, an important variable in the performance of the organization as a whole. Crucial to the performance of teams are the abilities and behaviours of their members. Relevant to this are the roles that team members play. These can be thought of in two ways. The first is in terms of functional roles, that is those roles which relate to a person's job role and function in the organization; for instance, marketing manager, lecturer, typist, head of department. People are often chosen to be members of teams on the basis of their functional roles, these being considered most appropriate to the task{s) which the team has to perform. However, people's functional roles, though fitting them in terms of experience and expertise for the task in hand, will not necessarily help when it comes to the process through which a team of people makes decisions and implements them. They do not help in matters such as the way different team members approach a problem or task, the way team members interact with one another, and their style of behaviour in general. Consequently, writers such as Belbin (1981, 1993), Davis, Millburn, Murphy & Woodhouse (1992), Margerison & McCann (1990), Parker (1990), Spencer & Pruss (1992) and Woodcock (1989) have proposed the notion of team roles, or in the case of Parker, team-player styles. Thus, individuals will not only bring the characteristics of their functional roles to their activities as members of teams, but will also, naturally, take up one or more team roles. For instance, a person might naturally be imaginative-a good ideas person. Another might be good at gathering information, whilst another might take the role of coordmator of other people's contributions. A number of different team roles can be identified and these vary according to different writers. From those referred to above, the largest number of team roles is fifteen (Davis etal., 1992) and the least four (Parker, 1990). There is some overlap between different sets of roles, but also some roles which seem to be unique to a particular writer. The identification of team roles is sometimes claimed to be based on different personality characteristics (e.g. Margerison & McCann, 1990; Parker, 1990). However, all proponents of the team role concept claim to have observed the behaviouts typical of each team role, in a wide variery of occupational teams in many different types of organizations. As an early proponent of the team role concept, the work of Belbin (1981, 1993) is particularly interesting as, arguably, the basis upon which other team role frameworks have been built. Beibin's development of the team role concept arose ftom observing, over a period of nine years, teams ot managers on training courses playing management games where team performance was measured in terms of winning or losing (Dulewicz, 1993). Belbin claimed to be able to predict the performance of a team through knowledge of each team member's team role. Given team role profiles for each team member, where all team roles were strongly represented across the profiles (that is the team was assumed to be 'balanced'), the team was predicted to be high performing. Where some team roles were absent, he maintained that the team would have a lower success rate. However, the issue of what constitutes 'success' or high performance in real teams in real organizations, rather than artificially constituted management game playing teams, is more complex. Winning or losing can rarely be measured. Other measures must, therefore, be looked for. A small survey of the literature (Evenden & Anderson, 1992; King, 1988; Margerison & McCann, 1990; Prince, 1989) reveals at least 33 concepts associated with high performing teams. In other cases (Cohen & Ledford, 1994), team performance is measured on

Teajit roles and team performance

243

a small number of scales by reference to team managers' judgments and/or that of their superiors. Cohen & Ledford show how performance can also be measured by objective data such as customer complaints, team members' accidents, illness and absenteeism. However, where no such objective criteria exist, for instance in management teams, some form of subjective criteria has to be used. Even so, a survey of the literature reveals little agreement on what criteria to set or how judgments should be made as to whether these criteria have been met. It may be that the 'difficulty of measuring salient, ecologically valid and reliable, teamdependent outcome variables' (Furnham, Steele & Pendleton, 1993, p. 245) has led to the paucity of academic studies which attempt, in any systematic way, to test team role theories in the context of real teams in real organizations. What is without doubt is the continuing and widespread use of the team role concept and the idea of a balanced team. This is particularly the case with trainers and consultants in their pursuit of helping teams improve their team performance. Given the extensive application of team role theories, it is surprising that there is little, other than anecdotal, evidence to support the original premise from which these activities are derived. What evidence there is concentrates on the psychometric properties of different ways of measuring individuals' team roles (Furnham et al.. 1993), and the relationship between self-perception measures and those used by observers who know the participants well (Parkinson, 1995). Thus, considering the widespread use of team role theory in recruitment of members to teams and in training associated with team building, it seems timely that the theories which form the basis of team theory should be tested in a more systematic way. Aims of the study Using real teams in real organizations, the study sets out to evaluate Belbin's team role theories in terms of their ability to predict team performance. The means of doing this, however, raise a number of methodological issues which must be addressed. These are: (a) how to identify a person's 'natural' team role; (b) how to measure whether a team is balanced or not; and (c) how to measure team performance. These are discussed below, prior to a more detailed description of the methodology used and discussion of the results. As a preface to this, Belbin's team role theory is described in more detail. Belbin's team role theory In his book, published in 1981, Belbin claimed to have identified eight team roles as a result of his original research using management teams playing management games. In his later book, published in 1993, he renamed some of the team roles and added a ninth role as a result of further experience. Table 1 identifies the nine team roles together with a short description of their strengths and weaknesses. The roles are described in more detail in Belbin's book (1993) and in the notes associated with Belbin Associate's software program (Interplace IV) which has been designed to measure people's Belbin team roles. The identification of people's team roles is, however, ot more than academic interest.

a. a _ -5 o
a> ^i? c o c

.2^

J -

-a c
u.

c t;

u O

-a
T3 C

e
E o

pond

D.. "^

ega

S o
^ s

i?

c U

cQ

n X

. "^

QO

nj

l-l

un

-0

c
Ul

.S E

UJ

~o c 3

-^ a 5
I-

S "c E g - 8

'legac , dyn;

1- ^ F ,c o -0

optio ns.

c o o -z

[iJ

l-l

u
ul

f
E u
o

3
1

-0 c

l-l

c o - -a ao c E .;
ij C I r-

2 n

OJ

'L'

s -e

2 -S ,1 J

f^

:? &
Q .H a i;
c^-^
.M ft rt

U
3
Cl- C/!)

5 ^ a,^
a c

.
c s 3 S u

,2 O

a.^

C Q
a

-a

CL

DS

-a
=
o

Team roles and team performance

245

Belbin maintains that, not only do team roles exist as behaviours and thinking styles, but individuals will tend to have distinctive preferences, or 'natural' roles which will be assumed on most occasions. For instance. Table 1 shows the profile of someone whose natural roles (that is those scoring at 70 or above) are plant and completet-finisher. Thus, as well as contributing their professional skills or expertise to a team, team members will also contribute particular behavioural characteristics which can be identified with their natural team roles. Team role theory, then, finds its expression in the assumption that for a team to be high performing, there needs to be a balance, or spread, of naturally occurring team roles across the team. This idea is encapsulated in two statements made by Belbin. These are:
Each team needs an optinriunn balance in both functional roles and team roles. The ideal blend will depend on the goals and tasks the team faces,

and
A team can deploy its technical resources to best advantage only when it has the requisite range of team roles to ensure efficient teamwork (Belbin, 1981, pp, 132133).

Identification of team roles


Two instruments designed by Belbin exist for the purpose of identifying a person's natural' team roles. The first is Belbin's Self-Perception Inventory (SPI), an ipsative instrument which is completed by each team member. The second is an observer checklist which is completed by team members' colleagues who know their team member behaviour well. There have been criticisms (Furnham et al.. 1993) of the psychometric properties of the SPI, but these were of the original eight-role version. No published tests of the psychometric properties of the current nine-role version have been found. However, a more recent study of 218 SP!s completed by a variety of practising managers (Swailes & Senior, 1996) provides some statistical evidence for the independence of each team role. Little attention has been given to the use of the observer checklists other than in the study by Parkinson (1995) and comments in the Intetplace manual (not dated). The statement in the latter reads, "In groups of executives and senior managers tested there was found to be a degree of agreement between self-percept ion and observed perception, and between the observers themselves that was significantly better than would occur by chance' (p. 50). Parkinson compared the nine-role version of the SPI with results from observers and found a limited association between the two. Drawing on this, logic suggests that a person's team roles could, most accurately, be identified by using a combination of the SPI and observer checklist results. It was decided, therefore, to use this method in this study. Defining team role balance The issue of balance is an important one, and Belbin's linking of a balanced team with high performance is no different from other team role theorists on this issue. However, whilst logically, the idea of balance is appealing, ways of measuring balance ate contentious.

246

Barbara Senior

The most straightforward and most used method of determining team balance is to establish whether team members' profiles, collectively, have all nine team roles represented at the natural level, that is at a score of 70 or above. If all nine team roles are represented in this way, then, according to Belbin, the team is said to be balanced, with the implication that a balanced team will be high performing. However, Belbin acknowledges that, where some team roles are not naturally' present, if they are at the able to be assumed' level (that is scoring between 30 and 60see Table 1), this may not be too detrimental to the team's performance. What seems to be implied here, is that there is a subjective judgment to be made as to the extent of balance of team roles in any team. Looked at from the other direction, a team would definitely be unbalanced where a team role was not represented, either 'naturally' ot at the 'able to be assumed' level. The measure of balance discussed above relies on examining individual team members' most naturally occurring roles. One could also argue that team role balance occurs when the team's average scores on each team role (taking all the team role scores into account) are found to be similar, although this measure is not commonly found in the literature. Thus, although one or more naturally occurring roles may be absent from any of the individual team members' profiles, the fact that, on average, those team role characteristics appear in similar quantity to any other may suggest a degree of balance. It was decided, therefore, to use both measures of balance in this study.
Team roles and a team's ikey stage of activity

The concept of balance tied to a particular level at which team roles occur across team member profiles is relatively clear. In addition, it is easy to set a measure of similarity for team role average scores across team roles. However, in his 1993 book, Belbin also stresses the link between the stages of a team's project or activities and the need for different team roles to be dominant at different stages. Table 2 lists the six stages with Belbin's comments on the team roles deemed relevant to each stage. Given the importance of the role of any team leader, it could be argued that his or her natural roles should be congruent with at least one of the roles deemed necessary for a specific stage of team activity and this was argued at the 1996 Interplace Users Conference. In addition, there is a case to be put for these particular team role characteristics to be present on average when all the team role scores of *?// the team members are taken into account. Both these measures are used in this study. Definition of team performance The central tenet of Belbin's theories is that the more balanced a team is in terms of the spread of naturally occurring team roles, the greater the propensity for it to be high performing. In other words, there is a cause and effect relationship between team role balance and team performance. As discussed above, in Belbin's original work (Belbin, 1981; Dulewicz, 1995), team performance was measured in terms of winning and losing. However, difficulties occur when there are no objective measures with which to judge team performance. This is a particular problem for management teams, or in Katzenbach & Smith's (1991) terms.

Team roles and team performance Table 2. Team roles relevant to different stages of a team's project or activities
Key stages of team activity 1. Identifying needs 2. Finding ideas 3. Formulating plans

247

Team roles relevant to particular stages Key figures ar this stage are individuals with a strong goal awareness. Shapers and coordinators make their mark here. Once an objecrive is set, the means of achieving it are required. Here plants and resource investigators have a crucial role to play. Two activities help ideas turn into plans. One weighing up the options, the second making good use of all relevant experience and knowledge to ensure a good decision. Monitor evaluacors make especially good long-term planners and specialists have a key role to play at this stage. People must be persuaded that an improvement is in prospect. Champions of the plans and cheer leaders must be found. This is an activity where resource investigators are in their element. However, to appease disturbed groups, a team worker is required. Plans need turning into protedures, methods and working practices to become routines. Implementers are the people required here. These routines, however, need people to make them work. Getting people to fit the system is what coordinators are good at. Too many assumptions are made that all will work out well in the end. Good follow-through benefits trom rhe acrentions of complecers, Implementers, too, pull their weight in this area, for they pride themselves on being efficient in anything they undertake.

4. Making contacts

3. Establishing the organization

6. Following through

Based on Bdbin (1993. pp. lUO-lOl),

teams classified as "recommending things' or running things' rather than those which 'make ot do things'. Methods which could be used include observation of the team's behaviour, tracking the effects of its decision making or the administration of pre-prepared questionnaires to team members and other relevant persons. However, all these methods rely on the researcher's own perception of what makes any team effective or not and do not take into account the different goals and purposes of different teams (particularly when operating in different organizations). Neither do they take account of team members' own ideas on what constitutes 'performance' for them. One method which attempts to meet these criticisms is the use of repertory grid (Stewart & Stewart, 1981), first, to generate the constructs which define team performance, for the team in question, and, second, to rate the team on these constructs to obtain a measure of its performance. Thus, each team sets its own performance criteria and rates its performance accordingly. This process is recommended by writers such as Galpin (I994, p, 245) who says, '. . . asking teams and individuals to rate themselves on whatever factors are determined to be important is a good way to approach "immeasurables" like customer service, teamwork and communication skills'. The particular process used for this study is based on this view. It is detailed further in the methodology section below and reported fully in a separate article (Senior, 1996a).

248

Barbara Senior Method

Process
Two determinants of team performance have been identified. The first concerns whether a ream is balanced or not in terms of team roles; the second concerns the presence or absence of those team roles particularly relevant to a team's ,specific stage of project development or ream activities. Given these, the following formed the basis on which the data collection and analysis were carried out in order to meet the aims of the study. 1. Through an identification of team members' team roles, and taking account of the stage of activity of each team (a) assess the degree of balance of each team and, on chis basis {b) predict its likely level of performance, 2. Test these predictions against the actual perceived performance of the team. 3. Conclude the extent to which Belbin's team role theories can usefully be used in actual organizational settings.

Sample
The participants of interest to this study were 11 management teams of between four and nine members; that is groupings of people who meet together and work as teams which involve decision making related to the whole or part of their organizations. The teams were distributed amongst public and private sectors as follows: social services, 3 teams; borough councils, 4 teams; hospital trusts, 1 team; benefit agency, 1 team; financial institution, 1 team; brewery, 1 team. All members of all teams completed the Belbin nine-role version of the Self Perception Inventory (SPI) and participated in interviews to collect data on certain characteristics of their teams, the stage ot team activity and their teams' performances, A total of 67 team members took part in the research.

Measures
Identification of team roles. Belbin's nine-role version of the SPI was used, in conjunction with results from team members' colleagues completing observer sheets, Belbin Associates' Interplace IV computer software was used to analyse rhe SPI and observers' scores resulting in the production of a team role ranking for each team member (see Appendix 1). Ideally, results from at least four observer sheets are required tor each team member In some cases, the observers' results were not acceptable owing to a lack of discrimination in their completion. These were returned once to the individual concerned requesting another attempt. Where this also failed or was not forthcoming, the identification of the team role profile was done on the results of rhe SPI and fewer than four observers. Only three teams included members who did not have a full set of observer results. The extent of the shortfall was judged not to affect significantly the overall result. Measures of team role balance. The Interplace IV program uses the results of the analysis of team members' team roles to produce a 'Team Role Combinations Report', This is shown in Fig, 1 as a graph of the scores of t,a) the individuals with the highest team rote scores, and {b) the ream's average score for each team role. The definition of team role balance uses the two measures described above. First, that all individual team roles should score 70 or above, a measure which indicates a 'natural' role (the left hand bars in Fig, 1), Second, that the average team role stores (the right hand bars in Fig. 1) should show no more than a 20 per cent variation one from another. The figure of 20 per cent is arbitrary, but thought to be reasonable given no guidance in any of the literature nor trom Belbin himself on this measure. General team characteristics and stage of team activity. Team members were asked six open-ended questions relating to general characteristics of their teams and the stage of team activity at the point of the research according to the stages identified by Belbin and listed in Table 2. The questions related to: the purpose and tasks of the team; che organizational functions represented by the members; the length ot time the team had existed; the degree ot stability of the team's membership; the

Team roles and team performance

249

Figure 1. Team role combinations report. For each pair of bars, the bar on the left represents the individual with the highest overall score in that team role. The bar on the right represents the team's average score. team's stage of activity; the type of internal and external environment experienced by the ream. Explanations of Beibin's key stages of team activity were given by the researcher to help categorize the teams accordingly. From the answers given, a pen picture of each team was assembled. Presence of team rules appropriate to team's stage of development. A team leader was judged to have team roles relevant to his/her team's particular stage of activity ifthe roles identified in Table 2 matched the strongest roles in the team leader's profile. Where a team was clearly at a single activity stage, a match was measured by the team leader's strongest two team roles matching those required for that activity stage. Where a team was moving between activity stages, a match was measured by the team leader's strongest three roles being amongst che three or four required for those activity stages. In addition, the tliaracceriscics of the desired roles should be amongst the cop four average team role scores, Agam this latter measure is arbitrary, there being no guidance in the literature nor from Belbin himself on this issue. Team performance. Each team member participated in a one-co-one interview with the researcher to use the repertory grid technique chosen to determine each team's performance. The technique, used in this study, required each participant to choose real examples of the following seven elementsa 'good' team, a 'bad' team, an "okay' team, a well acted play, a badly acted play, the team of interest (i.e. the research team), and another team with which the participant was personally involved. The inclusion ot elements such as the well and badly acted play is a recognized part of the technique, the argument being chat elements somewhat different from those of direct interest help che creaciveness of the chinking in che 'compare and contrxst' process discussed below. Participants were then asked to compare and contrast different groups of three elements in terms of their pertormance. By use of this methodknown as che 'triad' methodof comparing elements, each team member identified a range of constructs, all relating to team performance. The method used is suth thac all constructs have two poles allowing the rating of each of the seven elements on each construct, using a six-poinc scale. An exception is that of the firsc ceam involved which rated on a five-point scale. Appendix 2 is an example ot a completed construct and rating sheet for a single team member. The GAP PREFAN program located at the Manchester Computing Centre was used to produce group analyses of che repercory grid data for each team. The program includes a principal component analysis and identifies the loadings of each element and each construct on each principal component. This allows che production of cognitive maps showing the positive or negative positioning of a team in relation to che team per-

250

Barbara Senior

formance constructs. It also gives the statistical distances of elements from each ocher, thus allowing judgement of how far or near che ceam of interest is from the examples of a good team, a bad team, an okay team etc. Figure 2 illustrates chis wich data from one of che ceams. The map includes all the seven elements used co elicit the performance constructs. It includes the constructs loading most highly on the top two principal components; typically, 20 on component 1 and 15 on component 2. In no case did the cop two principal componencs account for less than 70 per cent of the variance. Inspection of the map enables the team of interest's (in this case TF) performance to be assessed by observing it in terms of: {a) its positive, borderline or negative position on che map in relation to che ceam performance constructs; {h) its nearness to the good team and co the okay team. In addition, the GAP program gives data on the discance of che team of interest from che bad team. A high performing team would be one which is typically positioned fwsitively with regard co the performance constructs and which is at least 55 per cent nearer than by chance to the good team. A low perform-

MoUvitKl to nulii thU>> hipptn MSkn

Figure 2. Team F cognitive map showing loadings of elements and conscruccs on component 1 (horizontal axis) and component 2 (vertical axis). Note: Distances of elements from each ocher are denoted as follows: 0.70 C 0.7y {2'y'X nearer chan by chance) O 0,60 C 0,69 (,^5% nearer than by chance) O 0.50 to 0,59 (45%' nearer than by chance) ^^=^^ 0.40 to 0.49 (55%: nearer than by chance) ^ ^ ^ ^ Key to elements: GT BT OKT OT WAP BAP TF = = ^ good team bad team okay team own team well acced play badly acced play Team F

Teavi roles and team performance

251

ing ceam would be one which is typically positioned negatively on che map and having no association with the good ceam wich probably a near association with the bad ceam.

Results
Table 3 lists che key stage of activity fot each team, the team roles identified by Belbin as being particularly useful for those stages, and whether {a) the team leaders' strongest roles matched the ones required; ib) the overall team average scores matched the ones required as defined in the measures section above. Table 4 lists the results of applying the team role balance measures, together with rhe degree of matching of the teams' key stages to {a) the team members' strongest individual role scores and to {b) the average role scores as defined above in the measures section. Where the requirement of a measure is met a 'yes' is recorded. Where this is not mer, a 'no' is recorded. 'Nearly' is recorded where the measure was almost met. The footnotes at the end of the table indicate the precise meaning of the 'nearly' recordings.

Table 3. Team roles marched to team key stage of activity


Key stage and associated team roles" organization/follow through Match of key stage roles to leader's strongest team roles no yes no nearly'' nearly* nearly*
no

Team

Match of key stage roles to strongest overall team average role scores

IMPCO/CFIMP
ideas/plans PL Rl/ME SP follow through CFIMP organization nearly* nearly* nearly* yes no nearly* no no nearly yes

IMP CO
organization

IMP CO
needs/ideas SH CO/PL Rl organization/follow through IMP CO/CF IMP ideas/plans PL RI/ME SP contacr/organization Rl TW/IMP CO plans/organization ME SP/IMP CO ideas/con tacts PL Rl/RI TW

nearly nearlv*

no
nearly' nearly'' nearly'

"Kt-y itajjes as described in Table I, In these cases, che result was one role shon ot che requirt-cJ number. Key. CF = tompleter-finisher; CO = coordinator; IMP = implemencer; ME = monicor Rl = resource invtsii^ator; SH = shapet; SP = specialise; TW = team worker.

evaiuacor;

PL = plane;

252

Barbara Senior

Table 4. Predicted team performances using team role balance measures


Balance of individuals" strongest Team roles Balance of overall ceam role averages nearly* no nearly ^ no no no no no no no no Matching of key stage roles to leader's strongest team roles yes nearly^ nearly^ no no no nearly^ nearly^ nearly'^ no no Match of key stage roles to strongest overall team average role scores nearly' yes nearly' no yes nearly* nearly^ no no nearly"" no Predicted performance on the basis of team balance very good good good moderate moderate moderate moderate moderate to poor moderate to poor moderate to poor poor

B
E

yes yes

A K
C D

yes
yes no nearly^ no no no no no

H F G I

"For Team C, eight of che nine coles are at rhe 'nacural' (score of 70 or above) level whilsc the ninth is at che 'able to be assumed' level. *In these cases the difference becween che hij^hesc and lowesc scocing toles was 25 points rather than the 20 points criteria sec foe a balanced team based on average role scores, I n chese tascs, the result was one role short of che required number.

The final column of Table 4 is an assessment, given the information in the previous columns, of the likely performance of each team. This was calculated in two stages. First, by rank ordering the teams according to the greatest number of'yes' counrs, followed by a mixture of'yes' and 'nearly' counrs descending to the lowest rating for the greatest number of 'no' counts. The teams are listed in rhis order in the table, from the one with the best performance prediction to rhe one with rhe worst. This was followed by a more subjective grouping of the teams on a five-point scale of 'very good', 'good', 'moderate', 'moderate to poor', 'poor'. Table 5 summarizes the results of the repertory grid analyses of each team's performance in terms of how near or far (statistically) each team is from the team members' ideas of good, bad and okay teams. It also says, for each ream, whether it was positioned in rhe positive or negarive zones of the relevant cognitive maps. The penultimate column makes an assessment of each team's actual performance. This was done in two steps. First by rank ordering the teams according to the following process: 1. The teams were grouped according to whether they fell in the positive, negative or borderline positions (with respect to the performance constructs) on the cognitive maps. 2. Within these three categories, the teams were ordered by nearness to the 'good' team. 3. Where tied rankings occurred, the teams were ordered further by distance from the 'bad' team.

Team roles and team performance

253

Table 5. Comparison of perceived actual team performances (using repertory grid results) with performances predicted on the basis of ream balance
Team J Distance from good team 55% nearer than by chance Distance from bad team 25% further than by chance Distance from okay team 75% nearer than by chance Actual perceived performance very good Predicted performance good

(Positioned in positive zone of cognitive map) E 35% nearer than by chance 25% further than by chance 35% nearer than by chance good good

(Positioned in posicive zone of cognitive map) F 25% nearer than by chance 25% further than by chance 25% nearer than by chance good moderate to poor

(Positioned in positive zone of cognitive map) D 25% nearer than by chance no different chan by chance 35%' nearer than by chance good moderate

(Positioned in positive zone of cognitive map) B 259? nearer than by chance no different than by chance 35% nearer than by chance good very good

(Positioned in positive zone of cognitive map) C 25f^ nearer than by chance no different than by chance no different rhan by chance no different than by chance no different than by chance no different than by chance 25% further than by chance no different than by chance no different than by chance no different than by chance 25%' nearer than by chance no different than by chance 35%'nearer than by chance 25%' nearer than by chance 25% nearer than by chance moderate moderate

(Positit)ned in positive zone of cognitive map) K moderate moderate

(Borderline position on cognitive map) H moderate to poor moderate to poor moderate to poor poor moderate to poor moderate

(Positioned in negarive zone of cognitive map) A

(Positioned in negative zone of cognitive map) 1 poor

(Positioned in negative zone of cognitive map) G no different than by chance 25% nearer than by chance 55% nearer than by chance moderate to poor

(Positioned in negative zone of cognitive map)


No/t, Teams rank-ordered by perceived actual performance. Rank I indicated (he highest performing; team, rank 11 the lowest performing team.

254

Barbara Senior

4. Where tied rankings still occurred, teams were further ordered by nearness to the okay team. The teams are listed in this order in the table with the best performing team ranked one and the lowest performing team ranked 11. This step was then followed by subjectively grouping the teams on the 'very good" ro 'poor' five-point scale as was done for the predicted performances. The final column of Table 5 gives the assessment for predicted performance, for comparison with actual performance. Finally, to test the degree of congruence of the rank ordering of predicted team performance with actual team performance, a Spearman's rank order correlation was carried out (see Table 6). The result shows a positive relationship between the two measures which is significant at the /> < . 10 level. Table 6. Comparison of predicted team performance with actual team performance Rank order of predicted performance 4.5
1

Team
A B

Rank order of actual performance 9-5 4.5 6 4.5 2

C D
E

6.5 6.5 2
8 8

F
G

3
U 8 9.5
1 7

H I

J
K

n 3
4.5

Ntite. Correlation significant at clie p < , 1 0 level, rwo-cailed,

Discusson The aims of this research were to evaluate Belbin's team role theories in the context of a range of management reams working within the public and private sectors. One of the issues explored was the idea that a team should be balanced in terms of members' team roles (both individually and averaged across the team) in order for the team to be effective and high performing. In addition, importance was attached to the stage of activity of the team and strength in the recommended ream roles for that stage sought. From the results relating to balance and the fit between stage of activity and leaders' and overall team average roles, predictions were made as to the likelihood of each team being high or low performing. These predictions were then compared ro the collective judgments of the team members, themselves, as to the actual performance of their teams. From the results reported above, four out of the 11 teams' actual performance confirmed the predicted performance when the five-point 'very good' to 'poor' scale was

Team roles and team performance

255

used. The predictions for six of the remaining seven teams were within one point of rhe actual performance on the five-point scale. Only one team showed a two pomr difference. Finally, a positive correlation {p < .10) was found between the rank ordering of the teams according to their predicted performances and the rank ordering of the teams according to their actual performances. These findings give some support ro Belbin's team role theories which associate team balance with team performance. The issue of measuring team balance was addressed. In predicting the performance of the teams, it is usual to rely on one measure of team balance onlyhow many team roles are represented in individual team members' profiles at a score of 70 or above. This measure, whether derived from the use of the SPI only or combined SPI and observer results, does not appear to act as a good predictor (Senior, 1996/;). Therefore, the inclusion of other, albeit more arbitrary measures to help the prediction, is justified. An argument could be made that there was a degree of researcher judgment in allocating teams to different key stages and the use of the five-point 'very good' to 'poor' scale. However, not all measures can be easily quantified. What is more, relying only on those measures which can be objectively quantified unnecessarily restricts rhe study and leads to somewhat impoverished results. The study has thrown up a number of issues. First, with 11 reams, it was difficult to controt for other factors such as team purpose, status differences amongst team members, individual members' intelligence and skills, and those factors mentioned at the beginning of the papere.g. organizational structure, motivation and degree of autonomy, which could affect team performance regardless of the presence or absence of team roles. However, all of the teams appeared to be able to organize rheir own work and had powers of decision making over how they operated and, in some cases, over policy as well. In no case was there any reason ro believe thar status differentials between team members (where they existed) affected individuals' ability to contribute to their team's operation. Second, a sample of 11 teams is not large and it would have been useful to have had additional teams from the private sector. Third, whereas all the teams investigated thought of themselves as teams, there were clear differences in the degree of interaction between members in pursuit of ream goals. It should be remembered that the majority of the teams were what could loosely be termed 'management" teams who only came together at intervals, rather than teams who worked together all the time. Consequently, a large number of ream performance constructs elicited through rhe repertory grid process described team processes rather than team goals and outcomes. This is not to say the latter were absent, merely fewer in number. Fourth, it would have been useful to obtain a more rigorous understanding of each team's stage of development and activity. However, working with real teams puts constraints on how much members can be asked to do. As it was, each person gave up at least one and a half hours of time to work with the researcher. Future research could well improve this aspect of the investigation. Finally, the concept of a balanced team was operationalized to have a particular meaning in quantifiable terms. Ultimately, a team might be considered balanced if some team roles are only at the 'able to be assumed' (see Table 1) level and even if one or more is absentifthe circumstances in which the team is operating allow this without detriment ro rhe team's performance. Team F illustrates a situation where the predicted performance

256

Barbara Senior

was low yet Its members perceived the team to be effective in performance. This team had only four members and worked in a fairly closed environment. The weakest individual roles were monitor evaluator, plant and resource investigator. The plant and resource investigator roles are both associated with 'having ideas' and finding things out. The monitor evaluator is rhe intelligent checker of details and evaluator of options. In the environment in which this team worked, it is probable that bright ideas were not required in the regulated day to day work they had to do. Without further investigation, it is difficult to say whether the weak monitor evaluator role would be detrimental to the teams' performance, but on current knowledge, it is suspected not. Despite these limitations, the study has several major strengths. Primarily, the use of real teams in real organizations. The research goes beyond previous evaluations of ream role theories which are anecdotal in nature and frequently done by consultants or trainers who may have a vested interest in the results. As a result of using a repertory grid method for measuring team performance, there now exist around 650 constructs which team members have offered as descriptors of team performance. The possibility exists of distilling these into a validated questionnaire which can be used to measure team performance. Such a questionnaire would cut down the time spent interviewing team members and allow a greater number of teams to be surveyed. This study has concentrated on one of the most commonly used ream role frameworks (Belbin). Whilst there is a need to continue to test this framework in order to produce increased reliability for the results reported here, more studies are needed to evaluate other theories of team roles, with the outcome of debating the notion of team roles itself. Acknowledgements
The auchor would like co chank che members of all che ceams who cook pare in chis research. Everyone ,^ freely and uncomplainingly of cheir cime. Parcicularly pleasing has been che incecesc shown by che ceams in the resulcs of the research. Graceful chanks go also co Nene College for supporcing che research and Sandy McDonald for her extensive help wich daca colleccion and her encouragemenc chroughouc che period of che research. The commencs of my colleague Scephen Swailes have been helpful in arriving ac che tinal version of this paper.

References
Belbin Associates, interplace IV Human Resource Management System User's Manual. Cambricige, Belbin, R. M. (19H1). Management Teann: Why They Succeed or Fail. London; Heinemann, Belbin, R. M. (1993). Team Roles at Work. Oxford: Buccerworth-Heinemann, Brown, D. I. (1995), Team-based reward plans. Team Performance Management. 1, 2 3 - 3 1 . Burns, T. & Stalker, G. M. (1966). The Management of Innot-ation. London: Taviscock Cohen, S, G. & Ledford, G. E, Jr (1994), The effectiveness of self-managing teams: A quasi-experimenc. Human Relations. 47, 13-43. Conjer, J. A. (1993), Leadership: The art of empowering others. In J, R, Gordon (Ed,), A Diagnostic Approach to Organizational Behavior, pp. 420428, Boston, MA: Allyn & Bacon, Davis, J.. Millburn, P,, Murphy, T-& Woodhouse, M. (1992). Successful Team Building. How to Create Teams that Really Work. London: Kogan Page, Drucker, P. F. (19H8). The coming of the new organiziicion. Harvard Business Review. Jan-Feb, pp, 45-53, Dulewicz, V, (1995), A validacion of Belbin's ceam roles from 1 6PF and OPQ using bosses' ratings of competence, _/oarKij/ of Occupational and Organizational Psychology. 68, 1-18, Evenden, R. & Anderson, G. (1992). Making the Most of People. Cambridge, MA; Addison-Wesley.

Team roles and team performance

257

Furnham, A,, Sceele, H, & Pendlecon, D. (1993). A psychomecrit assessmenc of the Belbin Team-Role SelfPerception Xnwentoty. Journal of Occupational and Organizational Psychology. 66, 245-257. Galpin, T, (1994), How co manage human performance. Employment Relations Today. Summer, 207225. Herzberg, F. W, (1968). One more time: How do you motivate employees? Harvard Business Review. 46,

53-62,
Katzenbach, J, R. & Smich, K, (1991), The discipline of teams. Harvard Business Revieu: MarchApril, 111-120. King, D. (1988). Team excellence. Management Solutions. October, 25-28, Likert, R, (1961), New Patterns of Management. New York: McGraw-Hill, McClelland, D, C. (1988). Human Motivation. Cambridge: Cambridge Universicy Press. Margerison, C. & McCann, D. (1990). Team Management. London: W. H. Allen. Maslow, A, H. (1943), A theory of human motivation. Psychological Revieu: 50, 370396. Morgan, G, (1993), Imaginization: The Art oj Creative Management. London: Sage, Moss-Kanter, E, (1993). Power failure in managemenc circuics. In D, S, Pugh (Ed,), Organisation Theory: Selected Readings. London: Penguin, Parker, G. M. (1990). Team Players and Teamwork: The New Competitive Business Strategy. Oxford: Jossey-Bass, Parkinson, R. (1995), A silk purse out of a sow's ear. Organisations and People. 2, 2225. Prince, G. (1989), Recognismg genuine teamwork. Supervisory Management, April, 25-31, Roechlisberger, F, J, & Dickson, W, J. (1964). Management and the Worker. New York: Wiley, Senior, B, (1996^), Team performance: Using repercory grid technique to gain a view from che inside. Journal
of Managerial Psychology. 1 1 , issue 3,

Senior, B, (1996i), Team roles and ceam building: Self-perception is no basis on which to build a ceam. Iconoclastic Papers. 1(1), December, Internet address: hctp://www,solent.ac,uk/busmgmt/iconcM;lastic/ mdex,htm! Smith, P, B,, Peterson, M, F, & Misumi, J, (1994). Event management and work ceam effectiveness in Japan, Britain and USA. Journal of Occupational and Organizational Psychology, (si., 33-^,^Spencer, J, & Pruss, A. (1992), Managing Your Team. London: Piatkus. Stewart, V, & Stewart, A, (1981), Business Applications of Repertory Grid. London: McGraw-Hill, Swailes, S, & Senior, B. (1996), Team role and learning scyle: A correlacional study. Nene College School of Management Working Paper. Norchampton: Nene College of Higher Education, Taylor, F, W, (1947), Scientifk Management. New York: Harper & Row, Tjosvold, D. (1991), Team Organization. An Enduring Competitive Advantage. Chichester: Wiley. Weber, M. (1964), The Theory of Social and Economic Organizations. New York: Collier Macmiltan. Woodcock, M, (1989), Team Development Manual. 2nd ed, Brookfield, VT: Gower.
Received 2 November 1994: final version received 16 April 1996

Appendix 1 Assessment results in ranked order


Candidate: Date: Team roles Assessment Self perception Observer A Observer B Observer C Observer D Overall ranking 1 MF SH CF SH PL SH 2 CF CF SH CF CO CF

3
IMP

4
PL Rl SP CO SP PL

5
SH SP TW Rl CF TW

6
TW TW Rl TW TW Rl

7
Rl PL PL ME ME IMP

8 CO IMP IMP PL Rl CO

9
SP CO CO SP IMP SP

ME ME IMP SH ME

258

Barbara Senior Appendix 2 Elements scored on each construct: A single team member's results
Elements

Scale I Clear objeccives Positive link becween individual differences and ceam objeccives Many curf issues Lot of openness and trusc Very process aware Unclear roles and responsibilities Good individual performances Locs of ciienc feedback Good skills misc. Inadequate communicacions between team and others Thorough planning operationally Supporcive ceam leader Gcx>d idea generation Has a lot of fun Conscruccive conflicc managemenc Good prioricy secting

GT 2 2

BT OKT WAP BAP 5 6 4 4 2 1 4 5

T2 1DWN(T)

Scale 6 Unclear objeccives No link/negative link between individual differences and ceam objeccives Few curf issues Secrecy and scepcicism Very cask awate Clear roles and responsibilities Bad individual performances Lictle client feedback Insufficienc skills amongsc members Good communicacions becween ceam and ochers Ad hoc operacional planning Inflexible leadership style Poor idea generation Takes it all too seriously Descruccive conflicc management No idea of priorities

3
2

4
1

6 3 3 5
1 1 2

2 6 3 4 3 4 4 1

4
2 2

4
2 2

3 6 3
1

5
1 1

4
2

4 3 3 4 2 3

3 5 5
2

5 3 2 2 6

4
2

3
2

5
2

5
i

4
1 2 1

6 5 3 5 6 6

1 2

6 5 5
5

3 3
2

3
2 5

3 3

3 3 3 4

4
2 2 2

4
2

5 4

3
2

For each characceriscit, giveeath t-lemenc a score between 1 and (i. The left hand end of each stale scores 1, che rijrhc hand end scorts 6. Key. GT = ^ood teamjBT = bad ceam; OKT = okay team; WAP - well acted play; BAP = badly atred play;T2 = ceam of inceresc (rtsearch team); OWN(T) = addiciotial ceam.

Vous aimerez peut-être aussi