Return to Iconoclastic Papers Home Page


Return to Index of Iconoclastic Papers


TEAM ROLES AND TEAM BUILDING:

SELF-PERCEPTION IS NO BASIS ON WHICH TO BUILD A TEAM



Barbara Senior


Barbara Senior is a Chartered Occupational Psychologist and Principal Lecturer and Consultant in Management at Nene College. She specialises in the areas of organisational behaviour and change, creative problem solving and cross cultural studies.

Research interests are in team-working, particularly ways in which the characteristics of team members influence team performance; organisation and change; cross cultural studies Her published work includes a range of papers and two books.





There has been much debate on the importance of the work of teams in achieving organisational tasks. For instance West begins his book Effective Teamwork, by asking, and attempting to answer, the question are teams more effective than individuals working alone? (1). The answer can be summed up in the words of West and Slater: "The research evidence is consistent in suggesting that the quality of group decision making generally equals but does not exceed the quality of decision making of the average member" (2). However Katzenbach and Smith believe that "teams will become the primary unit of performance in high-performance organizations" and that "every company faces specific performance challenges for which teams are the most practical and powerful vehicle at top management's disposal" (3).


Even though individuals working alone may achieve better results then groups of people working together in some situations, the fact that occupational teams are a common and increasing characteristic of organisational life places a responsibility upon all involved with them to ensure they work as effectively as possible. There are a number of factors which contribute to the performance of teams; for instance, the organisational structure within which the team works, the type of task to be accomplished, resources available and the characteristics of the team and the team members. The last, the characteristics of the team members, is the subject of this paper.


TEAM ROLES AND TEAM PERFORMANCE


It is generally accepted that people are chosen for their membership of teams because of the job and task skills they possess; in other words, because of the functional role they perform. However, for some fifty years or so, it has been recognised that members of groups play roles additional to those which gained them admission to the group in the first place. Thus, Benne and Sheats proposed a number of roles such as 'energiser', 'opinion seeker', initiator-contributor' 'harmoniser', 'encourager' and so on (4). Bales differentiated between task-oriented behaviours and socio-emotional behaviours, the latter being more concerned with a group's processes and the former more concerned with the task (5). More recently, various academics, consultants and others have applied the notion of behavioural roles to teams and claim to have identified sets of roles which they term 'team roles'. Thus, the early work of Belbin identified eight team roles (6), to which he later added a ninth (7). Davis, Millburn, Murphy and Woodhouse identified five team roles, which they subdivided into fifteen (8). Margerison and McCann found nine roles (9); Spencer and Pruss ten roles (10), and Woodcock twelve roles (11).


Different team roles indicate different types of behaviour which are not necessarily linked to job and task skills. For instance, a person might be naturally imaginative - a 'good ideas' person. Another might be good at checking details to make sure everything has been covered. Yet another might be the person to make sure decisions are implemented and the task carried through to completion. Even though these team roles are not associated with particular job and task skills, they are considered crucial to task and goal achievement in that their presence or absence is said to influence significantly the work and achievements of teams. Consequently, most team role exponents maintain that, for a team to be high performing, it should be 'balanced'; that is, there should exist amongst the typical behaviours of members, the full range of team roles.


Consultants and trainers have built upon these theories and applied in putting them together occupational teams and in expanding the activities of existing teams. A range of questionnaire type instruments have been developed to identify the natural team role of individuals - each team role proponent having devised his or her own instrument. Some have developed computer programmes to produce, for each team member, team role profiles and accompanying prose explanations. In addition, the programmes frequently give summaries of a team's expected performance on the basis of the team members' range of identified team roles.


Two main issues, crucial to the application of team role theory, arise from this work. The first is the method for identifying an individual's team roles and the question of whether he or she can make this judgement through self-perception only, or whether other people's judgements of them are also required. The second is the concept of balance. How balanced must a team be to be judged balanced enough to ensure high performance? The research reported in this paper arises out of these issues. They are related, specifically, to Belbin's work in this field, work which has resulted in one of the most widely applied set of team role theories. The remainder of this paper, therefore, starts with a discussion of these two issues. This is followed by a description of a piece of research which aims to identify different measures of team role identification, and using these, to test their validity in determining team role balance as a determinant of team performance.

IDENTIFICATION OF TEAM ROLES


The most popular method for identifying a person's Belbin team role is to ask for completion of a Belbin Team Role Self-Perception Inventory (SPI), which gives a score between 0 and 100 for each team role. From these scores, a team role profile for each team member can be plotted. Based on Belbin's nine role framework, Figure 1 illustrates such a profile with a description of each team role.


SELF PERCEPTION TEAM ROLE PROFILE


Name:............................................................


ROLES BEST AVOIDED

ROLES ABLE TO BE ASSUMED

NATURAL ROLES

ROLES AND DESCRIPTIONS TEAM-ROLE CONTRIBUTION

ALLOWABLE WEAKNESSES

0 10 20

30 40 50 60

70 80 90 100



. ..

. . PL . .

. . x .

PLANT: Creative, imaginative, unorthodox.

Solves difficult problems

Ignores incidentals. Too preoccupied to communicate effectively.

. ..

. . RI x .

. . . .

RESOURCE INVESTIGATOR: Extrovert, enthusiastic, communicative. Explores opportunities, Develops contacts.

Over-optimistic. Loses interest once initial enthusiasm has passed.

x ..

. . CO . .

. . . .

CO-ORDINATOR: Mature, confident, a good chairperson. Clarifies goals, promotes decision-making, delegates well.

Can be seen as manipulative.

Offloads personal work.

. ..

. . SH . x

. . . .

SHAPER: Challenging, dynamic,

thrives on pressure. The drive and courage to overcome obstacles

Prone to provocation. Offends people's feelings

x ..

. . ME . .

. . . .

MONITOR EVALUATOR: Sober, strategic and discerning. Sees all options. Judges accurately.

Lacks drive and ability to inspire others.

. ..

. x TW . .

. . . .

TEAMWORKER: Co-operative, mild, perceptive and diplomatic. Listens, builds, averts friction.

Indecisive in crunch situations.

. ..

. . IMP x .

. . . .

IMPLEMENTER: Disciplined, reliable, conservative and efficient. Turns ideas into practical actions.

Somewhat inflexible. Slow to respond to new possibilities.

. ..

. . CF . .

. x . .

COMPLETER: Painstaking, conscientious, anxious. Searches out errors and omissions. Delivers on time.

Inclined to worry unduly. Reluctant to delegate.

. ..

x . SP . .

. . . .

SPECIALIST: Single-minded, self-starting, dedicated. Provides knowledge and skills in rare supply.

Contributes on only a narrow front.

Dwells on technicalities.


Figure 1: Example of an individual's Belbin team role profile with descriptions of the nine different team roles.

Note there are two team roles which are 'natural' roles, both scoring 70 or above.

(Based on Belbin Associates Interplace IV computer programme materials)


An examination of Figure 1 shows that team roles are classified into those which are the natural roles for that person, those which are able to be assumed by that person, and those which he or she should avoid. Roles at score 70 or above are considered to be the roles that person would naturally assume given no pressures to act otherwise.


There have been criticisms of Belbin's eight role version of the SPI (which is similar to the one used here) in terms of its psychometric properties, which give rise to questions as to the use of 70 as the determining score for a 'natural' role (12). However, as important is the fact that the results rely solely on self-perception, taking no account of other people's views. Consequently, when Belbin Associates developed computer software (Interplace) to score the SPI, they also designed an observer adjective checklist for use by close work colleagues of the subject (Belbin advised at least four observations per team member). Therefore, in addition to using the results of the SPI on its own to identify a person's team role, it is also possible to use data from a combination of the SPI and the observer checklists to achieve a combined self and other-perception team role score.


However, although these two measures of team roles exist, and logic suggests the combined measure (SPI plus the perceptions of others) as the more rounded team role description, it is the SPI measure, on its own, which is predominantly used. For instance, only one account has been found which reports any other than the use of the SPI, and this claims systematic discrepancies between the SPI scores and the other-perception scores (13). To the author's knowledge, the majority of consultancy work done, based on the use of the team role concept, uses only the SPI results to identify a person's team roles. The dilemma remains, therefore, as to which measure to use for identifying someone's natural team roles as a means of building a balanced team. This leads to the issue of defining team role balance.


TEAM ROLE BALANCE


Belbin makes a number of statements about balanced teams; for example "In a perfectly balanced team there is always someone who can deal naturally with any set of responsibilities" (14). This means that at least one person in the team should have at least one of the team roles in their profile as a naturally occurring role. In terms of team role scores, Belbin sets this at a score of 70 or above.


However, a person may have more than one role scoring 70 or above. The question then arises as to whether the role scoring, 90 is more natural than the one scoring (say) 70. Given that Belbin maintains teams of fewer than nine members can be balanced, the assumption is that, in the case above, both roles count. What is crucial to the concept of team balance is its relationship to team performance. Belbin's theory says a balanced team will be higher performing than an unbalanced team. What is not clear, however, is whether a team which is balanced in terms of roles scoring 80 or above is likely to be higher performing than a team which, though still balanced in Belbin's terms, scores only 70 or above.


Conversely, will a team that has only five roles naturally represented perform worse than a team which has, seven roles naturally represented? This point has been made by Fisher, Macrosson and Walker who observe that whereas Belbin claims that top teams (high performing teams) have a full complement of personality types, "there is no information in either Belbin's account or in the open literature to tell those firms which cannot muster the full complement of team personality types how poorly or how well their teams are likely to perform" (15).



SUMMARY AND RESEARCH AIMS


The conclusions for the above are twofold. First, there is some confusion about which measure of team role identification should be used to determine how balanced a team might be. Secondly, there is some confusion about how balanced a team should be to be high performing. Is it simply a dichotomy between balanced and unbalanced, or are there degrees of balance which are associated with degrees of performance?


Research with a number of different teams in a range of public and private sector organisations offered the opportunity to test the two different measures of team role identification and to investigate their relationship to different measures of team role balance. The aims of the research can be stated as:


To investigate the relationship between team role balance and team performance using a range of different measures of team roles and team role balance.



METHODOLOGY


Sample

Ten teams were identified as the subjects of this study. These teams contrasted with the teams used for Belbin's original research which were 'artificially' formed from people attending management development courses. The teams in the study were actual management or departmental teams currently operating in their respective organisations. Numbers of team members per team ranged from four to nine, there being 59 members of teams in total. Two teams were in the private sector (one in financial services and the other in the brewing industry) and eight in the public sector (local and county councils and a hospital). All teams had been operating in their current form for at least two years with approximately 75% stability of membership. The functions of the teams varied. For instance, two teams managed two different centres for adults with learning difficulties, another team managed a large hospital's non-clinical services. One of the two private sector teams provided the accounts function for the organisation, whilst the other private sector team managed the human resource development function. All the teams could be defined, loosely, as having a management function, although, as can be seen from the examples given, this differed significantly over the range of organisational levels.



Identifying team members' team roles

Team members' team roles were identified separately in two ways: first, through using only the results of the SPI, which every member completed and which the Interplace programme shows as a set of team role profiles (for example, see Figure 1): and secondly, through using the results of the combined SPI and observer checklists which the Interplace programme shows as a graph (16). Figure 2 gives an example of this.



Note: the bars represent the individuals with the highest combined

score in that team role.





It can be seen from Figures 1 and 2 that scores for the SPI results can be identified only from the sten scores (multiples of ten), while the combined SPI and observer scores can be identified at smaller intervals. Five out of the total of fifty-nine team members, across three teams, had only three observer results each, which was less than the recommended four. This was not considered to affect the results in any significant way.

Measuring team balance

All measures of team balance rely on the presence or absence of team roles above a certain score. As discussed above, Belbin gives 70 and above as the score to be attained if a role is to be considered a person's 'natural' role. Thus, in any team, the greater number of roles represented at score 70 or above, the more balanced the team and vice versa. However, in keeping with the aims of the research to use a range of measures of team balance, scores were used as follows:


Self-perception (SPI) scores only at score 70 and above.

Self-perception (SPI) scores only at score 80 and above.

Self-perception (SPI) scores only at score 90 and above.

Combined (SPI + observer) scores at score 70 and above.

Combined (SPI + observer) scores at score 75 and above.

Combined (SPI + observer) scores at score 80 and above.

Combines (SPI + observer) scores at score 90 and above.


Measuring team performance

Given the type of work done by the teams, there were no obvious objective measures, such as sales figures, number of complaints or components made per hour, available for assessing their performance. This contrasts again with Belbin's research teams judged on by their performance in the management games they were required to play. Therefore, some form of subjective measure was required. Two issues arise: which measure to use, and who should do the measuring?

There are various ways of judging the performance of teams in the absence of objective measures. One way is to observe and rate the team's behaviour on some set of agreed criteria. Another is to interview all who may have a view about the team and its performance. A third is to administer a pre-prepared questionnaire to team members and their managers. Some researchers have used senior management as judges of a team's performance as well as, and sometimes instead of, team members' own judgements(17).


All these methods have their advantages and drawbacks. Observing team members' behaviour is very time-consuming and requires a degree of participant observation unavailable to most researchers. Interview data is, itself, qualitative and, unless obtained in a very structured way, does not lend itself to comparison. Many questionnaires purport to measure team performance; however, these assume equality of importance of the items measuring team performance, regardless of the purpose or activities of the team in question. Therefore, a team performance measure was sought which (a) took account of team members' own perceptions of what team performance criteria were important, (b) allowed the team leader's view to be incorporated with those of the other team members and (c) enabled a team's performance to be compared with that of others. Team members and team leaders were chosen as the vehicle both for defining the criteria on which their team should be judged and for using these criteria to measure their own team's performance.


This method produced a form of self-rating, but one which operated on a group (including the team leader) basis. This combines the self-rating SPI measure with other people's ratings so as to avoid the bias of self-measurement. This measure also combines the views of both the team members and the team leader. There is considerable support for such team performance measures: ... 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"(18). They also help overcome what Furnham et al speak of as "... the extreme difficulty in measuring salient, ecologically valid and reliable, team-dependent outcome variables in order to establish some criterion of team success"(19).


Given this decision on the method to be used, the technique for collecting the data on team performance was as follows. In the context of an interview with each team member and the team leader, repertory grid technique was used to elicit constructs (criteria) relating to team performance on which team ratings could be made (20). Several elements were used to elicit the constructs. These included the team of interest as well as each respondent's identification of a 'good' team (the best team known to the respondent); a 'bad' team (the worst team known to the respondent) and an 'okay' team (one which was somewhere in between the other two). Each respondent's perceptions of a well-acted play and a badly-acted play, in terms of their performances, were also used as elements to broaden the 'compare and contrast' activity, which is an essential part of the repertory grid technique.


The Manchester Computing Centre's GAP programme was used to produce, for each team, a principle components analysis of the combined results of all the team members and the team leader. This enabled the production of a cognitive map of the positions of the elements in relation to the constructs and allowed a calculation of, the distance of one element from another. Figure 3 is an example of a cognitive map for one of the teams.



Figure 3 shows the criteria used to measure team performance (the words and phrases around the circle) and the position of the team in question (in this case T4) and all the other elements with respect to these criteria. The performance of a team can be determined from its position in relation to the positive and negative constructs (that is the team performance criteria - each of which occupy approximately half the map area), and its position in relation to the other elements, including the good, bad and okay teams. Therefore, a cognitive map for a team represents a composite view of the team members and team leaders with respect to (a) the performance criteria important to that team, and (b) their ratings of their team on these criteria as well as in relation to other significant elements such as the good, bad and okay teams.



Results


Table 1 shows the ranked performance of the ten teams. Team performance rankings were determined as follows:



Team

Distance from

Good Team

Distance from Bad Team

Distance from Okay Team

Position in positive/ negative zone of map

F

no different than by chance

25% nearer than by chance

55% nearer than by chance

negative

1

G

no different than by chance

no different than by chance

35% nearer than by chance

negative

2

H

no different than by chance

no different than by chance

25% nearer than by chance

borderline

3

J

no different than by chance

25% further than by chance

no different than by chance

borderline

4

B

25% nearer than by chance

no different than by chance

25% nearer than by chance

positive

5

A

25% nearer than by chance

no different than by chance

35% nearer than by chance

positive

6.5

C

25% nearer than by chance

no different than by chance

35% nearer than by chance

positive

6.5

E

25% nearer than by chance

25% further than by chance

25% nearer than by chance

positive

8

D

35% nearer than by chance

25% further than by chance

35% nearer than by chance

positive

9

I

55% nearer than by chance

25% further than by chance

75% nearer than by chance

positive

10


Table 1 Evaluation of team performance

Note: Statistical nearness of the research teams to team members' perceptions of 'Good', 'Bad' and 'Okay' teams and the positive or negative positions of the teams on the cognitive maps are shown. Rank 10 indicates the highest performing team.



1 Order teams in terms of positive, borderline and negative positions on cognitive maps which created three categories - top, middle and bottom rankings.

2 Within these categories, order by nearness to good team.

3 Where tied rankings occur, order by distance from bad team.

4 Where tied rankings still occur, order by nearness to okay team.


This produced a team performance ranking with only two teams tied.




Team

Perf-ormance

Ranking

CO

SH

PL

RI

IMP

CF

TW

ME

SP

Total number of roles

Roles ranking

F

1

+

+

-

-

+

+

+

+

+

7

6

G

2

-

+

-

+

+

-

+

-

+

5

1

H

3

+

-

+

+

+

+

+

+

+

8

9.5

J

4

+

+

-

+

+

+

+

-

+

7

6

B

5

+

+

+

+

+

+

-

+

-

7

6

A

6.5

+

-

+

+

+

+

+

+

+

8

9.5

C

6.5

+

+

-

+

-

+

+

-

+

6

2.5

E

8

+

-

-

-

+

+

+

+

+

6

2.5

D

9

+

+

+

-

+

+

+

-

+

7

6

I

10

-

+

+

+

+

+

+

-

+

7

6


Table 2 Distribution, number and ranking of team role types (self-perception ratings only) in teams ranked according to level of team performance.

Note: Team roles scoring at sten 70 or above.

Key: + = team role present; - = team role absent.

CO=Coordinator; SH=Shaper; PL=Plant; RI=Resource Investigator; IMP=Implementor; CF=Completer Finisher; TW=Teamworker; ME=Monitor-Evaluator; SP=Specialist



Team

Perf-ormance

Ranking

CO

SH

PL

RI

IMP

CF

TW

ME

SP

Total number of roles

Roles ranking

F

1

+

+

-

-

+

+

+

+

+

7

7.5

G

2

-

+

-

-

+

-

+

-

+

4

1

H

3

+

-

+

+

+

+

+

+

+

8

9.5

J

4

+

+

-

+

-

+

+

-

+

6

5.5

B

5

+

+

+

+

+

+

-

+

-

7

7.5

A

6.5

+

-

+

+

+

+

+

+

+

8

9.5

C

6.5

+

+

-

-

-

+

+

-

+

5

3

E

8

+

-

-

-

-

+

+

+

+

5

3

D

9

-

+

-

-

+

+

+

-

+

5

3

I

10

-

+

+

+

+

+

+

-

-

6

5.5


Table 3 Distribution, number and ranking of team role types (self-perception ratings only) in teams ranked according to level of team performance.

Note: Team roles scoring at sten 80 or above.

Key: + = team role present; - = team role absent.

CO=Coordinator; SH=Shaper; PL=Plant; RI=Resource Investigator; IMP=Implementor; CF=Completer Finisher; TW=Teamworker; ME=Monitor-Evaluator; SP=Specialist



Team

Perf-ormance

Ranking

CO

SH

PL

RI

IMP

CF

TW

ME

SP

Total number of roles

Roles ranking

F

1

-

-

-

-

+

+

+

+

-

4

4.5

G

2

-

+

-

-

+

-

-

-

+

3

2

H

3

+

-

+

+

+

+

+

-

-

6

9

J

4

-

+

-

+

-

+

+

-

+

5

7

B

5

+

-

+

+

-

+

-

+

-

5

7

A

6.5

-

-

+

+

+

+

+

+

+

7

10

C

6.5

+

-

-

-

-

+

-

-

+

3

2

E

8

+

-

-

-

-

+

-

+

-

3

2

D

9

-

-

-

-

+

+

+

-

+

4

4.5

I

10

-

+

+

-

+

+

+

-

-

5

7


Table 4 Distribution, number and ranking of team role types (self-perception ratings only) in teams ranked according to level of team performance.

Note: Team roles scoring at sten 90 or above.

Key: + = team role present; - = team role absent.

CO=Coordinator; SH=Shaper; PL=Plant; RI=Resource Investigator; IMP=Implementor; CF=Completer Finisher; TW=Teamworker; ME=Monitor-Evaluator; SP=Specialist



Tables 2, 3 and 4 indicate the distribution of team roles according to the presence (indicated as +) or absence (indicated by -) of individuals whose self-perception (SPI) team role scores were, respectively, at 70, 80 and 90 and above. The teams are shown in order of team performance, where rank 10 indicates the highest performing team. The roles ranking (final column) is in terms of the number of team roles represented at the respective levels identified above, as shown in penultimate column. Rank 10 indicates the highest number of team roles in any team, and therefore the most balanced team. Rank 1 represents the lowest number of team roles, and therefore the least balanced team. Where rankings are tied, the mid point between the rankings has been taken. For illustration in Table 2, team J is ranked 4 in terms of performance. It has 7 out of a possible 9 team roles represented, which gives it a ranking of 6 in terms of team role balance.





Team

Perf-ormance

Ranking

CO

SH

PL

RI

IMP

CF

TW

ME

SP

Total number of roles

Roles ranking

F

1

+

+

-

+

+

+

+

+

-

7

6

G

2

-

+

-

-

+

+

+

-

+

5

1.5

H

3

+

+

+

+

-

-

+

-

+

6

4

J

4

-

+

+

+

-

+

+

-

+

6

4

B

5

+

+

+

+

+

+

-

+

+

8

8

A

6.5

+

+

+

+

+

+

+

+

+

9

10

C

6.5

+

+

-

-

+

+

+

-

-

5

1.5

E

8

+

+

-

-

+

+

+

-

+

6

4

D

9

+

+

-

+

+

+

+

+

+

8

8

I

10

+

+

-

+

+

+

+

+

+

8

8


Table 5 Distribution, number and ranking of team role types (combined results of self-perception and observer ratings) in teams ranked according to level of team performance.

Note: Team roles scoring at 70 sten or above.

Key: + = team role present; - = team role absent;


Team

Perf-ormance

Ranking

CO

SH

PL

RI

IMP

CF

TW

ME

SP

Total number of roles

Roles ranking

F

1

+

+

-

-

-

+

+

+

-

5

4

G

2

-

+

-

-

-

+

+

-

+

4

1

H

3

+

+

+

+

-

-

+

-

-

5

8.5

J

4

-

+

-

+

-

+

+

-

+

5

6

B

5

-

+

+

+

+

+

-

+

-

6

7

A

6.5

+

-

+

+

+

+

+

+

-

7

8.5

C

6.5

+

+

-

-

+

+

+

-

-

5

4

E

8

-

+

-

-

+

+

+

-

+

5

4

D

9

+

+

-

+

+

+

+

+

+

8

10

I

10

-

+

-

+

+

+

+

+

+

7

4


Table 6 Distribution, number and ranking of team role types (combined results of self-perception and observer ratings) in teams ranked according to level of team performance.

Note: Team roles scoring at 75 sten or above.

Key: + = team role present; - = team role absent;


Tables 5, 6, 7 and 8 have been prepared on the same basis as Tables 2 to 4, except that the team role identifications are based on the combined self-perception (SPI) and observer scores. The Tables represent, respectively, individual team members with scores of 70 and above, 75 and above, 80 and above and 90 and above.



Team

Perf-ormance

Ranking

CO

SH

PL

RI

IMP

CF

TW

ME

SP

Total number of roles

Roles ranking

F

1

+

+

-

-

-

+

+

-

-

4

3.5

G

2

-

+

-

-

-

-

+

-

+

3

1

H

3

-

+

+

+

-

-

+

-

-

4

3.5

J

4

-

+

-

+

-

+

+

-

+

5

6

B

5

-

+

+

+

+

+

-

+

-

6

8

A

6.5

+

-

+

+

+

+

+

+

-

7

10

C

6.5

-

+

-

-

+

+

+

-

-

4

3.5

E

8

-

+

-

-

-

+

+

-

+

4

3.5

D

9

+

+

-

+

+

+

+

-

-

6

8

I

10

-

+

-

-

+

+

+

+

+

6

8


Table 7 Distribution, number and ranking of team role types (combined results of self-perception and observer ratings) in teams ranked according to level of team performance.

Note: Team roles scoring at 80 sten or above.

Key: + = team role present; - = team role absent;



Team

Perf-ormance

Ranking

CO

SH

PL

RI

IMP

CF

TW

ME

SP

Total number of roles

Roles ranking

F

1

+

-

-

-

-

-

-

-

-

1

2.5

G

2

-

-

-

-

-

-

-

-

+

1

2.5

H

3

-

-

+

+

-

-

-

-

-

2

5.5

J

4

-

+

-

+

-

+

+

-

+

5

10

B

5

-

-

+

+

-

+

-

-

-

3

7

A

6.5

-

-

-

+

+

-

+

+

-

4

8.5

C

6.5

-

-

-

-

-

+

-

-

-

1

2.5

E

8

-

-

-

-

-

-

-

-

+

1

2.5

D

9

-

+

-

-

-

-

+

-

-

2

5.5

I

10

-

+

-

-

+

+

-

-

+

4

8.5


Table 8 Distribution, number and ranking of team role types (combined results of self-perception and observer ratings) in teams ranked according to level of team performance.

Note: Team roles scoring at sten 90 or above.

Key: + = team role present; - = team role absent.

CO=Coordinator; SH=Shaper; PL=Plant; RI=Resource Investigator; IMP=Implementor; CF=Completer Finisher; TW=Teamworker; ME=Monitor-Evaluator; SP=Specialist



Type of balance measure

rs

significance

level (two-tailed)

Self-perception (SPI) only score 70 and above

.0000

1.000

Self-perception (SPI) only score 80 and above

-.2876

.420

Self-perception (SPI) only score 90 and above

.0907

.803

SPI with observers score 70 and above

.4158

.232

SPI with observers score 75 and above

.2328

.517

SPI with observers score 80 and above

.5751

.082*

SPI with observers score 90 and above

.2273

.528


Table 9 Correlation of team performance with team balance according to different measures of balance.

* Significant at the p < .10, two-tailed



Correlations

Spearman's rank correlation coefficients, rs, were calculated for team performance and the number of team roles present (as representing team role balance) for each of the measures shown in Tables 2 to 8. Table 9 summarises these results. Only when both self-perception (SPI) and observer scores are used, in combination, at the score of 80 and above does team performance correlate with team balance at p <.10, two-tailed; that is, at a level of 10%.



DISCUSSION


It can be deduced from the results shown in Table 9 that there is no statistically significant relationship between team role balance and team performance when balance is measured using the SPI scores only to indicate presence or absence of natural team roles. This is the case whether the cut off score is 70 and above (as Belbin suggests), or whether the more stringent criteria of 80 and above, or 90 and above, are used. The only significant relationship is that between team performance and team role balance measured using the combined self-perception and observer scores of 80 and above. Using combined cut-off scores of 70 and above, 75 and above, and 90 and above to indicate team role balance yields no relationship with team performance.


This result can be compared with a study of six product development teams in a UK software company (21). If found that team role scores at 90 and above were ? before a significant relationship between team role balance and team performance could be found - in this case, measured by revenues generated and cost performance identified by administering a questionnaire to senior management. However, there are several of differences from the present study. First, Mottram's measures of team roles (based on 16PF scores) were used to compute the self-perception scores (22). Secondly, Belbin's earlier framework of only eight team roles was used. Thirdly, the Implementor role was excluded on the grounds that management members of the teams would, of necessity, ensure completion of the work on time. Finally, only self-perception measures were employed.


In spite of these anomalies, this study supports the present one in finding that, in measuring team role balance, a criteria more stringent than that suggested by Belbin is required in order to find a positive relationship between team role balance and team performance. In addition, the present study finds no case for using the SPI on its own. It appears that a combined SPI and observer measure is likely to prove more reliable. This, to some extent, follows Parkinson's reasoning that there will be a difference between a person's team role measure using the SPI and that obtained when observers' views are taken into account (23).


Implications for future research

This study is limited in a number of ways. It is limited in the number of teams surveyed, although, given the research time required to survey a single team, additional data will accumulate slowly. Clearly, more work needs to be done to confirm or refute the results to date. Secondly, other related factors influence team performance in addition to the degree of balance found in any team. Belbin discusses the relationship between a team's stage of project development and the need for particular team roles relevant to each stage. He also maintains that the team roles of team leaders should be compatible with the culture of the team and suggests what these might be. Therefore, the notion of 'balance' could change if these other variables are taken into account.


A third issue is that of how to measure team performance. Stewart and Stewart have written comprehensively on the business applications of repertory grid, including its use in questionnaire design and to investigate organisational climate and managerial effectiveness -uses which are very similar to that involved in this study (24). Team members judging the performance of their own team in their own terms allows them to 'buy-in' to the measurement and give it validity Galpin maintains that self-rating of team performance is more critical than the ratings given by managers of teams (25). In this study, however, each team leader's views were included in the composite results. Even so, how to measure team performance for teams which, in Katzenbach and Smith's terms, "recommend or run things" rather than "making or doing things" remains problematic (26). One way to strengthen this measure would be to gain the views of "customers" and higher management. Initially this would lengthen what is already a lengthy process.


This study has accumulated approximately six hundred individually-generated team performance constructs (an average of ten per team member), although, as expected, the meanings of some of these overlap. What is possible now is to carry out a content analysis of these to identify the most commonly used across all teams and those which are unique to a particular type of team. From this, a series of questionnaires could be devised for use in future research. These could be tailored to the specific circumstances of the team surveyed and for use by people outside the team. The use of questionnaires would certainly shorten the time taken to collect data and, hopefully, allow additional data collection to take place at an accelerated rate.


CONCLUSIONS


The data presented here have shown that the most commonly used measure of team role balance (use of only the SPI with natural roles occurring at the recommended score of 70 and above) has been found to be flawed in its potential for predicting team performance. Even if combined SPI and observer scores are used, team role scores of 80 and above are required to determine team role balance as an indicator of team performance. For the present, therefore, it must be concluded that self-perception is no basis on which to build a team.


NOTES AND REFERENCES

1 M. West, Effective Teamwork, BPS Books, Leicester, 1994.

2 M. West, and J.A. Slater, 'Teamwork: myths, realities and research', The Occupational Psychologist, 24, 1995, pp 24-9.

3 J.R. Katzenbach and K. Smith, 'The discipline of teams', Harvard Business Review, March-April 1993, pp 111-20.

4 K.D. Benne and P. Sheats, 'Functional roles of group members', Journal of Social Issues, 4, 1948, pp 41-9.

5 R.F Bales, 'A set of categories for the analysis of small group interaction', American Sociological Review, 15, 1950, pp 257-63.

6 M. Belbin, Management Teams, Why they Succeed or Fail, Heinneman, London 1981.

7 R.M. Belbin, Team Roles at Work, Butterworth Heinemann, Oxford, 1993.

8 J. Davis, P. Millburn, T. Murphy and M. Woodhouse, Successful Team Building, How to Create Teams that Really Work. Kogan Page, London 1992.

9 C. Margerison and D. McCann, Team Management, W H Allen, London, 1990

10 J. Spencer and A. Pruss, Managing Your Team, Piatkus, 1992.

11 M. Woodcock, Team Development Manual, Gower, Aldershot, 1989

12 A. Furnham, H. Steele and D. Pendleton 'A psychometric assessment of the Belbin Team-Role Self-Perception Inventory', Journal of Occupational and Organizational Psychology, 66, 1993, pp 245-57.

13 R. Parkinson, 'Cutting edge', Organisations and People, 2, 1, 1995, pp 22-5.

14 Belbin, Team Roles at Work, op cit, p 89.

15 S.G. Fisher, W.D.K Macrossan, and C.A. Walker, 'The structure of new product teams', Selection and Development Review, 10, 5, 1994, pp 1-3 (p 2).

16 The Belbin Associates Interplace IV computer programme was used for all the Belbin team roles analyses

17 See for instance, Fisher etal, op cit.; N. Brewer, C. Wilson and K. Beck 'Supervisory behaviour and team performance amongst police patrol sergeants', Journal of Occupational and Organizational Psychology, 67, 1994, pp 69-78; G. Borelli, J. Cable and M. Higgs, 'What makes teams work better?', Team Performance Management, 1995, p 3; P. Dainty, and A. Kakabadse, 'Brittle, blocked, blended and blind', Journal of Managerial Psychology, 7, 2, 1992, pp 4-17.

18 T. Galpin, 'How to manage human performance', Employment Relations Today, Summer 1994, pp 207-25 (p 245).

19 Furnham etal, op cit, p 245.

20 The particular repertory grid technique used in this study is explained in detail in B. Senior, 'Team performance: using repertory grid technique to gain a view from the inside', Journal of Managerial Psychology, 11, 3, 1996, pp 26-32.

21 Fisher etal, op cit.

22 R.D. Mottram, 'Building effective management teams using the in The Analysis of Personality in Research and Assessment, Independent Assessment and Research Centre, London, 1988.

23 Parkinson, op cit.

24 V. Stewart and A. Stewart, Business Applications if Repertory Grid, McGraw Hill, London, 1981.

25 Galpin, op cit.

26 Katzenbach and Smith, op cit.