3.2 Designing group work
This subsection is intended to provide a checklist and assessment of some of the most important elements in the design of group work. Lecturers may find it helpful as a prompt or reference when considering their own group design.
At the start of the module, students bring or quickly develop a set of expectations. It can be helpful to consider what elements can be manipulated to increase the probability of more effective group work. Expectations are influenced by word of mouth from previous students, so to some degree the expectations are a function of the previous years' operations. As an 'intangible service', perceptions about the level of the course and by association the level of their required contribution are additionally influenced by the introductory materials that students encounter. Care should be taken over presentation details to give impressions of professionalism, clarity and academic rigour.
Helping students to understand the arguments for using group work
It can be helpful to explain to students the benefits of using group work. Although this takes time out from discussing economic theory, it can help to motivate students to make much better use of the opportunities that are provided through group work. For example, the lecturer might emphasise the 'deep learning' and employability benefits described in section 2.1.
Role of lecturer
The role of the economics lecturer or tutor for group work activity will typically move from a formal provider of knowledge to a more supportive role. With proactive interference in the group's direction and learning processes, and because of his or her greater subject and group work experience, the economics tutor is well placed to facilitate the group's progress. However, tutors need to consider whether the benefits gained from interference to ensure positive outcomes outweigh the disadvantages of reduced group autonomy. For identification of dangers that can occur with tutor interference and practical suggestions for limiting the damage, see the chapter on 'Group facilitator behaviours that can damage group work' (Race, 2000, pp. 99-109).
Ruel and Bastiaans (2003, p. 28) provide a more fine-grained classification of different roles that the tutor may take on. They distinguish between:
- close involvement in the processes of the group;
- showing links and explaining concepts;
- giving initiative space to the group;
- providing the stimulation for development.
We can generalise that one form of good practice is a comprehensive set of published operational procedures. If they lay out the boundaries and roles for group activity, it may prompt students towards greater participation and engagement, in contrast to their original expectations of a tutor-led process. This point about the advantages of comprehensive content can be added to the earlier point about clarity and professionalism of presentation in order to influence initial expectations favourably.
Types of task
When assessing whether a group task is fit for the purpose intended, it can be helpful to consider categories such a: disjunctive, conjunctive, additive and discretionary (Steiner, 1972).
- Disjunctive tasks are those that only need one group member to be capable of achieving the task. This can be a useful method of spreading the correct answer around the student body, but it does not oblige group members to think, as they can rely on the most knowledgeable member of the group.
- Conjunctive tasks are those where each group member has to contribute. If these are successfully completed, we could expect more group learning to take place, but the requirement for full participation increases the chances of group failure.
- Additive tasks refer to group work where each additional group member can add something to the output.
- Discretionary tasks are so called because group members have discretion in how they combine their efforts.
To illustrate these concepts by group tasks, we can reference some of the cooperative learning exercises in Robin Bartlett's chapter in Becker and Watts (1998, pp. 18-29):
- 'Counting Triangles' in a diagram. This could be a disjunctive task for the group, as only one member needs to calculate it correctly. However, by adding the requirement that any member of the team may be selected at random to explain their answer, it acquires more of the characteristics of a conjunctive task.
- A 'Cooperative Writing' exercise. This describes what might be categorised as an additive task. Bartlett outlines this in practical detail. A typical example could be:
- A group is given a topic that is being covered on the course (e.g. 'Should the UK adopt the euro?')
- The group is asked to divide this topic into sub-components.
- If there are five group members, there will be five sub-components (for example, the Chancellor's five economic tests: business cycles compatible and sustainable; flexibility to deal with problems; effect on long-term investment; effect on financial services industry; effect on growth, stability and employment).
- Each member is given one of these subjects and asked to write on it for 10 minutes.
- They then pass their work to a colleague for comments, alterations and suggestions (say, 5 minutes).
- Then each of the five scripts is passed to another colleague for further comments and alterations (5 minutes) This process continues until all students have seen and commented on all five scripts.
- One group member is then allocated (randomly or voluntarily) to take all five scripts and write up the final draft, possibly to be submitted the next day.
- 'Jigsaw'. This is an exercise where each group has a different piece of economic data and to analyse the issue they need to understand the information held by the other groups. For example, a group may be asked to evaluate macroeconomic forecasts or policy decisions. Groups can be given separate economic figures for elements such as inflation, unemployment and money supply, but they will need analysis of the other groups' data before they can make confident predictions or policy suggestions. The exercise involves the group analysing their own set of data to reach conclusions; discussing and presenting their conclusions with other groups; analysing or critiquing the other groups' conclusions. How the individual works and combines with others in his or her group appears to be a matter of discretion and 'Jigsaw' might accordingly be classified as a discretionary task.
Most experiments and simulation exercises are natural vehicles for group work. The activities are frequently designed to help students to understand the dynamics of economic interactions within and between groups. For example, Lage et al. (2000, p. 33) describe a production exercise in which additions to groups are used to illustrate marginal concepts in production.
Group size is a widely considered variable in group design, as it affects outcomes in terms of performance and practicability. The most appropriate size of group may be a function of the task. In a disjunctive task, if one member of the group can do the task, the whole group can get the right answer. This would be the case if the task were to identify the right answer in a multiple-choice test. In this case, the probability of the group containing someone capable of doing the disjunctive task will increase with group size.
According to Steiner (1972), the formula 100(1 - Qn) gives the probability that at least one person in the group can perform the task. Q refers to the proportion of the population who cannot do the task, while n refers to the number of students in the group. For example, as shown in Table 1, Steiner calculates that if we estimate that only 40 per cent of a cohort of students could successfully complete a task and we have groups of five, there is a 92 per cent probability of each group containing someone who can complete the task. Of course, this assumes that students are randomly allocated across groups and this condition will be violated if students select the group composition themselves.
Table 1. Steiner's calculation of the probabilities of a group containing an individual who can complete a task
|Size of group||Percentage who can do the task|
Conjunctive tasks are more demanding, in that each student in the group must be able to complete a particular activity. The group will move at the pace of the slowest member and the probability of the conjunctive task being successfully completed will accordingly decrease with larger groups.
Table 2 presents a calculation of the probability of success in a conjunctive group task. In this table, P is the percentage of the class population that has the ability to do the task and n is the size of the group. The probability that the group can perform the task - that is, that all members of the group can perform the task - is 100(Pn).
A comparison between the two tables illustrates how much more likely a group of any size is to be successful, in terms of task completion, when it pursues a disjunctive rather than conjunctive task.
For example, let us assume we have a group of five students, that we ask the group to explain the Lucas price-surprise model and that we calculate that 60% of our class can answer the question.
- As it stands, this is a disjunctive task. Only one person in the group needs to know how to explain the answer. We accordingly calculate the probability of getting the right answer from the group as 100(1 - 0.45): that is, 99 per cent.
- If we ask each student in the group to be able to explain the answer, we have a conjunctive task and the probability of getting this answer from all the group is 100(0.65): that is, 8 per cent.
See Bligh (2000, pp. 129-30) for more comprehensive tables on the probabilities according to Steiner's theory.
Additive tasks refer to projects where each member's contribution 'adds' to the submission. The larger the group, the more contributions, and so we would expect larger groups to produce more. However, the group's output will be subject to the law of diminishing returns and at some point, depending on the task and process employed, additional students in the group will result in successively smaller additions to the group's total output.
We see that the output or probability of success of task (conjunctive, disjunctive or additive) is influenced by group size. A study of 50 economic groups ranging from 8 to 18 members, which considered such variables as size and their link with group marks, is provided by Watkins and Daly (2003).
In relation to the process, we can generalise that in larger groups the majority of individuals will learn less. This is supported by the concept that learning is assisted by participation and there is less of this participation in the larger group. The findings of Bales and Borgatta (1955) with regard to the effects of group size on the distribution of participation in a task-oriented group are seminal in this respect. They found that as group size increased, the variations in participation also increased.
We can generalise that smaller groups have process advantages over larger ones in terms of greater cohesion, less tension and increased motivation to co-operate. There should be less behaviour geared solely to satisfying personal requirements and the group members may be more constrained by group norms in the smaller group. The larger the group, the easier it is to hide and possibly the less likely students are to be equitably evaluated. Generally, smaller groups could be expected to mitigate motivation loss in the form of social loafing and free riding.
Table 2. The probability of a group being capable of performing a conjunctive task, according to Steiner's theory
|Size of group||Percentage who can do the task|
So is there an optimal group size? With regard to quality of output we need reference to specific task, process and structure (see section 5 for further reading). With regard to the learning process, it is easier to be intuitive about the benefits of smaller groups. A number of current authors (for example, Cooper et al., 1990; Johnson et al., 1998) support group sizes of four to five students as they believe that larger groups restrict members' participation and so provide less opportunities for them to increase their skills. A summary of the implications of group size is presented in Box 1.
Box 1 A Summary of the effects of group size
- Pairs. Advantages include ease of arranging meetings and less likelihood of free riding.
- Couples. These are likely to put more effort into group work than random pairs.
- Threes. The likelihood of free riding is quite low, but there is a danger that one of the three may feel marginalised at some stage.
- Fours. Passenger behaviour is still less probable than in larger groups and the group can subdivide into twos.
- Fives. The possibility of free riding increases significantly. The role of the leader becomes more important.
- Sixes. The group can subdivide into twos or threes and so reduce the increased risk of passenger behaviour.
- Greater than sixes. The role of leader needs to change in order to get the group working together. There is a greater possibility of free riding. (Race, 2000, chapter 11)
Here are some tips for working with larger groups:
- To maximise leadership effectiveness, consider introducing a split between task-and person-oriented leadership roles.
- With a large number of students it is likely that there will be more conflicts. To assist group cohesion, consider devoting more time to group maintenance processes.
- For reasons of communication, friendship and manageability, large groups are more likely to subdivide into smaller working units. Consider designing the work so that it can be split into subgroups. Workshops with 20-30 students can be very practical in that they can be broken up into subgroups for plenary sessions, so giving students experience in both small and large groups.
- Larger groups may need a more formal structure and allocation of roles in order to ensure greater participation. The economics tutor may need to intervene more directly to introduce and support the formal structures.
- In the context of group performance, designing for larger groups may be more efficient as long as they can exploit the advantages of division of labour.
- If the task is designed as being too simple, a larger group cannot be expected to improve on the performance of a smaller group.
- If we have variable-sized groups, we need to take care over equitable marking for tasks involving random error. A combination of estimates within a larger group is statistically more likely to be accurate.
Feedback is a vital component in developing our own identity, reducing anxiety and driving learning (see section 5.2 for further reading). For information on feedback's relationship with self-image, confidence and esteem, see Bligh (2000, pp. 109-11). When group work is applied in experiments and games, the feedback from their assessment can highlight valuable economic lessons. An international trade game by Sloman (2002) outlines an activity suitable for a number of teams, providing insights into a variety of diverse macro and micro factors. A particularly notable feature of his work, in the context of this section, is the emphasis he places on debriefing students and determining the learning outcomes that this debriefing will facilitate. Group work entails specific considerations in relation to feedback:
- In some circumstances, students may need individual communication due to the social position and the individual receiver's vulnerability when providing feedback within a group.
- When there are a large number of groups presenting reports on their discussion, it can be very boring for students to listen to each group's presentation. In these circumstances, they are likely to pay minimal attention and the tutor needs to design a process to deal with this. For example, the tutor could observe the groups' discussions and give a summary, and then ask the class for any additional points. Alternatively, he or she could ask each group for one point only. This reduces the difficulty for groups that report towards the end of the process. The class is then asked for any additional points at the end.
- Feedback is particularly valuable when received from our peers. Opportunities provided by group practices such as buzz groups should be utilised whenever possible.
Ideally, economics is a way of thinking about the world, providing insights into the way it operates and training students to think analytically. Many of the constituent parts of an economics course are greatly assisted by 'formative feedback', as students discuss positive concepts within the group.
If the dynamics of the group are poor, there should be flexibility to change various elements of an activity to make it into a more successful learning experience. Where students have been allocated to groups, they might be given the option of change. In the group work component of a first-year economics skills module at Kingston University, groups that are slow to combine effectively are offered the option of being divided into autonomous subgroups. Advantages of such flexibility may include greater commitment from group members because they can be involved in the change and possibly more appropriate groupings. The costs of flexibility can include the extra resources required to implement and monitor, the generation of uncertainty and a lengthening of lead times before the group 'beds down'.
Increased co-operation within a group appears to be generally beneficial. Apparently obvious but still worth pointing out is that members should be motivated by task and process design to co-operate with each other. This would occur when there is a positive interdependence of fate: that is, where one person's success directly helps or is necessary for others to succeed, as happens with teams. Johnson et al. (1981) reviewed 109 studies comparing co-operative with competitive structures for group work and found co-operative to be overwhelmingly superior.
The arguments in favour of increased co-operation include the following:
- It may reduce 'transaction costs' and hence increase efficiency, as we need a less formalised system or reduced levels of monitoring.
- With goodwill and inter-group co-operation, the externalities that are experienced with group work can be better accommodated.
- With greater co-operation we are more likely to get an equitable allocation of rewards for group members.
- Co-operation is a useful method for reducing the risks and uncertainties that individual students will face.
The prevalence of virtual learning environments (VLEs) for economics lecturers has risen from 35 per cent in 2001 to 57.6 per cent in 2003 (Economics LTSN, 2003). They can add structure and a wide variety of group process support functions as well as having a number of beneficial effects on the format of participation (see section 5.6 for further reading). Developments in VLEs and appreciation of their potential appear to be positive drivers towards increased intergroup co-operation.
Studies generally indicate that, the greater the conflict between groups, the more cohesion there is within a group. This appears to be a phenomenon utilised by schools for many years with variations on their house systems. In addition to the process, the results of this inter-group competition may have an effect. When the outcome is in favour of the group, cohesion tends to increase, whilst if the group appears to be losing, there will be a force against group cohesion: that is, success raises and failure lowers morale.
A practical finding by Holt (1987) was that that people will show decreased motivation in groups when the groups are psychologically insignificant. The implication is that if groups can be made to matter to their members, they are likely to commit more to the group effort. Can we make our groups matter more by preparing for seminars as thoroughly as for lectures, by increasing the 'fun' element, by providing activities that increase relevant interaction? Introductory discussions and social or physical exercises can all help increase group cohesion. Simple devices like giving groups names or identities can help in the process of making them more significant, and one of the suggestions by Bartlett (1998) is that colours are used to identify teams. This is on the basis that it helps build rapport and 'building team identity is important for building team cohesiveness'.
Group composition can have major implications for the way and extent to which group work develops (see section 5.5 for further reading). There are two main alternatives to the allocation of students on the basis of practicability (typically, the timetable). Either the tutor selects groups on the basis of some preconceived criteria or the groups self-select.
A logical approach is to divide students between groups, so that their personal characteristics blend to produce a supportive and successful group. An example might be to identify students with high and low characteristics of attributes such as 'confidence', 'ability', 'motivation' and 'knowledge'. They could be allocated between groups with the intention that students with more of a characteristic can help or favourably influence the students with less of that characteristic.
Within the considerable amount of research into the variety of group roles needed to make a group successful, Belbin (1981) is worth highlighting as one of the most influential thinkers on relevant roles and the choice of appropriate individual characteristics. Eight clusters of behaviour are identified as relevant to team roles. See Gibbs (1994, p. 4) for a useful table of team roles, based on Belbin, which you could use in your team work. These roles are innovator, investigator, chairperson, shaper, evaluator, team worker, organiser and finisher.
A large amount of economic group work comprises short bursts of activity within seminars conducted on a weekly basis. In these circumstances, tutors will typically select groups on the basis of student proximity. It is interesting to note the observation by Johnston et al. (2000) when employing this method that, as the semester progressed, students began to sit nearest to the people with whom they felt comfortable working.
There is considerable argument for self-selection in terms of greater cohesion. By making their own group choice, students are more likely to feel greater levels of group identification. The author's experience is that where groups have been formed by tutor selection, the argument for self-selection is one of the most common features identified by students as a way to improve the course. One possible format (Moore, 1998) with self-selection is to adopt a very open approach to group formation:
- In the second week of the course, it is announced that students are free to form their own groups.
- Groups must consist of three or four students.
- Students could opt to be placed in groups on the basis of their timetable compatibility.
- Changes in groups are allowed any time after the first test attempt.
- Students are allowed to choose an individual option, but are strongly encouraged to work in groups.
An effective leader can make the difference in achieving greater group cohesion, more inclusion and a higher quality of work. How should the group leader be selected? Democratic selection is frequently favoured, as we would then expect the nominated leader to enjoy support from the group. However, a difficulty is encountered where members of the group do not know each other very well and leadership is needed from the start in order to meet tight deadlines.
Selecting students with the relevant leadership personality characteristics may be problematic, particularly when dealing with very large numbers of students. Brown (1988), in a review of empirical investigations into personality characteristics of group leaders, did not find many reliable correlations because various studies gave different associations. However, they did report some indication that leaders may be slightly more self-confident, dominant, sociable, achievement orientated, experienced, intelligent, and taller and older than the other group members.
Rough-and-ready approaches are suggested by a common finding that an important characteristic of leaders is that they have a tendency to initiate activities in the group. Students with this type of characteristic can be easier to identify: for example, the lecturer can ask interested students to undertake a proactive activity, such as applying by email for the position of group leader, with a rationale as to why it would be useful to them personally.
It has been suggested that there are four critical stages for an effective group process, consisting of the rhyming stages: forming, storming, norming and performing (Tuckman, 1965; Bligh, 2000; Cameron, 2002). Adjourning was subsequently added as a fifth stage in 1977. Forming relates to the initial situation where members get to know each other and are trying to determine appropriate behaviour. Storming is the following stage, where there can be considerable conflict and strong status battles. The group's understanding of the task should be explored in this forming and storming time. At this point the group can move towards more cohesion or complete disintegration. Norming refers to the stage where norms have emerged within the group. The performing stage refers to the position where the group begins to work efficiently as the necessary roles within the group are adopted.
An alternative seven-stage model for phases in group development is provided by Johnson and Johnson (1987). It consists of:
- Setting up the group and providing information (defining and structuring procedures).
- The group getting to know each other while still dependent on the tutor for directions (sticking to procedures and getting acquainted).
- Recognising mutuality and building trust. At this stage, group members start to recognise their independence.
- Some features of rebellion. There may be resistance to some of the responsibilities accepted earlier.
- Ownership of goals and procedures. It becomes the members' group rather than the tutor's. The norms of the group are internalised.
- The group functioning in appropriate ways to achieve its goals: that is, performing maturely.
- Conclusion of the task and winding up of the group.
How to speed up progress
Groups may spend considerable time in inconclusive activity. They can spend long periods trying to arrange mutually convenient times for their next meeting and postponing decisions. Practical options to this problem include:
- Staff members adopting a more 'hands-on' approach, particularly in the early stages. By being present at meetings, tutors might intervene by providing suggestions and prompts designed to get the group moving, although tutors should beware of trying to exert too much influence. A common cause of paralysis is the absence of team members and it can be useful to suggest a practical option such as to record and circulate decisions reached in each meeting to all team members. This should be accompanied by the instruction that any responses or disagreements are sent to the group manager by a particular date. After that date, without any responses, it can be legitimately assumed that there is agreement with the decision.
- Clarifying stages to be completed in the task. A simple example of a five-stage project might be:
- The group selects a manager and the group tutor is notified.
- The group develops a project proposal and presents it to the group tutor.
The project is then carried out in three stages:
- Research stage.
- Production stage.
Competitive pressures can be introduced by giving advantages for groups that complete certain stages ahead of other groups. Using the above example for illustration, the group may be offered a 'first come, first served' choice of a limited selection of project topics that can only be made after selection of a group manager. Furthermore, if the groups are making presentations, these could be allocated in reverse order to the receipt of the groups' project proposals.
An alternative to the use of competitive pressures is the use of a directed process, such as groups being told to reach certain stages by particular times, on the pain of losing marks.