Study

last updated:November 2006

What impact did the Thinking Together approach have on pupils?    
The research showed that children who had been taught the Thinking Together approach used exploratory talk more often after the programme of lessons than they had before it. (Exploratory talk is the type of talk considered to be effective for thinking and learning). The pupils:

• involved each other
• asked each other questions
• listened carefully to what each other said
• responded constructively, even if their response was a challenge, and
• gave reasons for their opinions.

They also used more exploratory talk than children from matched ‘control’ schools who had not been taught the Thinking Together approach. The researchers found similar improvements in the quality of talk around computers when the children were prepared for thinking together in small groups beforehand. Teachers of the target classes reported on the positive impact the approach had on inclusion. For example one teacher said:

‘Most of the social groups have really knitted well and they have a sense of togetherness, and a sense of helping each other’.  

The researchers’ detailed analysis of the number of times pupils used words such as ‘because’, agree’, ‘I think’, ‘could’, ‘why’ (which they had identified as key features of exploratory or quality talk) showed precisely how the children’s ability to reason together had changed. The target groups’ use of the words increased significantly after the programme. No increase in the use of these keywords was apparent in the talk of the control groups.

The researchers also looked at pupils’ individual and group scores on tests of non-verbal reasoning. Overall, the scores of the target groups increased more than the scores of the control groups. In two schools where the researchers’ observations revealed that the programme had been most carefully and comprehensively carried out, group scores increased by over 10 per cent. The gains made by individual target class children were also significantly greater than those made by children in control classes.

The findings showed that children learned to reason better as individuals when they put into practice the strategies that they first tried in dialogue with others.  This fits well with Vygotsky’s theory of social interaction as a means of constructing learning. Practitioners can find out details about his theory in our earlier RfT summary about Vygotsky’s work.

How can dialogue guide pupils’ thinking?
The kind of talk that is most effective for building children’s knowledge and understanding together is 'exploratory talk', a term that was first used by Douglas Barnes in 1975 (see further reading). Barnes' work was developed by Neil Mercer in the 1990s. Mercer observed three kinds of talk in British primary schools:

• disputational talk, characterised by assertions, disagreement, short exchanges and individualised decision-making. There are few attempts to pool resources, or to offer constructive criticism of suggestions.
• cumulative talk, in which speakers build positively, but uncritically on what others say.  Partners use talk to construct a ‘common knowledge’ by accumulation. Cumulative talk is characterised by repetitions, confirmations and elaborations.
• exploratory talk (the most educationally effective kind of talk), which occurs when partners engage critically, but constructively with each other's ideas. Statements and suggestions are offered for joint consideration. These may be challenged and counter-challenged, but challenges are justified and alternative hypotheses are offered. With exploratory talk, reasoning is more visible than with the other two types of talk. Progress emerges from the joint agreement that is eventually reached.

Mercer argued that thinking and communicating were inseparable in effective classroom talk.  His view was derived from a modern interpretation of Vygotsky’s view of learning in which there is not necessarily a ‘more expert other’ involved in the discussion. Vygotsky suggested that learning occurs when a learner interacts with a more knowledgeable person (such as a child with an adult) who can guide the learners’ thinking. While this model of learning is useful, it does not explain the learning that occurs between children.  Mercer argued that learning in these situations is based on children having to explain and justify their decisions to each other – the characteristics of exploratory talk.

Why and how did the researchers believe exploratory talk should be promoted?
Video recordings of children’s discussions made by the researchers showed how children seldom hold good discussions – they rarely pick up on each other’s ideas and consider them critically, resolving any differences through further discussion (exploratory talk). Usually, children tend not to listen to each other, one child dominates the discussion, the children argue unproductively (disputational talk) or the children happily go along with whatever another child says without any reflection or debate (cumulative talk). They suggested the reason this happened was because teachers often assume that pupils know how to talk and work together and so rarely give them explicit guidance or training in how to make a good discussion happen.

The researchers consulted teachers and the research literature on collaborative learning for ways of promoting good discussion. They arrived at a number of pragmatic social ground rules that they suggested pupils needed to act upon during group discussion:

• all relevant information is shared
• the group seeks to reach agreement
• the group takes responsibility for decisions
• reasons are expected
• challenges are acceptable
• alternatives are discussed before a decision is taken
• everyone in the group is encouraged to speak by other group members.

The first three ground rules in the list were shared with cumulative talk – rules that served to bind the group, share information together and construct knowledge together through seeking agreement. The next two rules focused on the explicit reasoning that characterises exploratory talk as opposed to the other types of talk. The sixth ground rule, that alternatives are discussed, reflected the findings of research on collaborative problem solving, which has found that the groups that do best are those that consider alternatives before making decisions. The researchers added rule seven in the light of their experience of working with groups of children. They found that simply offering the right to participate wasn’t enough – in practice, children needed their peers to actively encourage them to speak and put forward their views.

How did the teachers prepare the children for working together?   
Working closely with primary teachers, the research team produced a series of nine 'Talk Lessons' to teach the social ground rules and to apply them within normal curriculum teaching and learning. Each lesson lasted for about one hour and focused on one or more of the ground rules of exploratory talk. The first few lessons dealt with skills such as listening, sharing information and co-operating.  Later lessons encouraged children to make critical arguments for and against different cases. The children were given opportunities to practice discussing alternative ideas, giving and asking for reasons and ensuring that all members of the group were invited to contribute.

Teachers taught the children how to use the ground rules through explicit modelling and coaching in whole group and small group discussions. They also gave the children opportunities to use the rules when working in small groups without the teacher. The explicit modelling involved the teacher at the front of the class illustrating the ways in which s/he wanted the children to talk together.  These included:

• asking ‘why?’
• using ‘because’ to give reasons for statements
• asking other children what they thought, and
• reaching agreement before making a final decision.

A key lesson in the programme involved eliciting the ground rules in the children’s own words. This was the third lesson after the children had all had some practice in collaborative activities. In a guided discussion the teacher drew from the class the kind of rules that they thought should be used in group work. The resulting list was displayed on the wall. During the rest of the programme and in other lessons where collaborative learning was used, the teacher or children could then refer to the rules on the wall and say ‘remember our rules’. Although each class produced a different set of ground rules they were all similar to the ground rules for exploratory talk identified by the researchers. For example, one class composed the following ground rules for talk:

Class 5D’s Ground rules for talk

• Discuss things together. That means:
• ask everyone for their opinion
• ask for reasons why
• listen to people.
• Be prepared to change your mind.
• Think before you speak.
• Respect other people's ideas – don't just use your own.
• Share all the ideas and information you have.
• Make sure the group agrees after talking.

The children worked in mixed ability groups of three for the talk lessons. This provided groups with a range of opinions and ideas and ensured that each group had a fluent reader/writer. To encourage a perception that all contributions to the group work were equal, there were no set ‘roles’ within the groups, other than that of occasional scribe or reader. Practitioners may like to read a case study of how a teacher decided on the composition of one particular group of children and what they learned from working with each other.

How did talking together help the children learn to think better?
The researchers’ analysis of discussions by groups of children when working together revealed how the programme of talk lessons helped to improve the children’s thinking skills. For example, the researchers examined the talk of George, Trisha and Susan as they worked together to solve a reasoning problem. The group had scored eight more correct answers on the reasoning tests after the programme than they had before it. The researchers compared the talk that resulted in an incorrect answer on the test with the successful talk after the programme that led to the group achieving the correct answer. The problem involved deciding which of six possible shapes (a black diamond, a circle, a square, a circle within a square, a black diamond within a square or a diamond within a circle within a square) completed the sequence.

Unsuccessful talk before the programme
In the pre-programme talk, the discussion started with a series of exchanges typical of disputational talk, in which the participants simply assert their opposing views without giving their reasoning – ‘It’s 2’.  ‘No it’s not’.  ‘It is 2’.  ‘No it’s not’. Susan then suggested, ‘It’s that one, 6’.  This was taken up by Trisha and both she and Susan offered reasons:

Trisha: It’s that, it has to be that … it has to be 6 because look, they’ve only got that (pointing to the pictures).
Susan: Look, first they are starting with one of them things over there (pointing) and then it has to be black.

The (wrong) answer ‘6’ was apparently agreed upon and George wrote it down.  However, Susan then appeared to change her mind, although she didn’t give a reason for her new opinion.  A dispute then followed about who should write the answers on the answer sheet.

Successful talk after the programme
The discussion after the programme contained more exploratory talk and longer, more elaborate explanations than the earlier discussion had done.  For example, Trisha opened the discussion with an elaborated hypothesis and a question that encouraged debate:

‘That has got to be a diamond, a square with a diamond with a circle, number 6, do you agree?’

George asked for more explanation, challenging Trisha not into conflict, but to encourage her to be more explicit: ‘No, what do you mean?’ In her effort to be more explicit, Trisha appeared to see that she was wrong and changed her claim. The group then came to agree on an answer:

Trisha: Look. That’s got a square with a diamond with a circle in, that’s got a square with a diamond in and that’s got a square with a circle in, so that’s got to be a square.
Susan: It’s got to be a blank square.  Yeh it is.
George; Do you agree on number 5, do you agree on 5?
(George writes ‘5’, which is the correct answer).

How did the Thinking Together approach make classrooms more inclusive?
The researchers noticed that when there was no explicit focus on talk, children for whom English was an additional language (EAL) often sat in groups without speaking or became disruptive out of boredom. They found that having ground rules for talk encouraged the children to draw others in. The researchers gave an example of the interactions before and after the programme of a group of three children that included a normally quiet girl with EAL (Nuresha). The children were videoed as they worked together on a reasoning problem. The problem involved selecting one picture of a jigsaw piece out of six possible pieces, which they thought would fit into a blank space on a geometric pattern.

Before the programme
In the first video, Nuresha did not speak at all and was disengaged from the task – she sat back from the table, sometimes looking around the room or playing with her ruler.  Shortly after starting on the task, Vijay and Kyle started to disagree over the answer and who should record it, without attempting to provide reasons for their opinions or seek each other's views (disputational talk):

Vijay: It’s this one, isn’t it?
Kyle: No.
Vijay: It’s this one, isn’t it?
Kyle: No.
Vijay: Yes.
Kyle: No.

After the programme
The children worked on exactly the same task in the follow-up video, but in a different way. This time, Nuresha became more involved in the group’s shared reasoning – encouraged by the other two children, who now listened to one another and accepted each other’s viewpoints:

Vijay: (to Nuresha) You have to say ‘what do you think, Vijay or Kyle?’
Nuresha: I think that (number 2).
Kyle: I think that (number 4).
Vijay: Nuresha, look.
Nuresha: I think, that, that, that.
Kyle: No, because, look, because that goes round. It goes out. It goes out.
Vijay: Or that one.
Kyle: No, because it hasn’t got squiggly lines.
Vijay: It has to be that.

The researchers also gave an example of how the Thinking Together approach helped two pupils in a special school for children with emotional behavioural difficulties – pupils who often find it difficult to articulate their thoughts and feelings or appreciate the thoughts and feelings of others. The pupils worked together at a computer to fill in speech and think bubbles for two characters who were squaring up for a fight over a skateboard. Their dialogue showed how they realised that even though the characters felt angry, apologising or asking nicely would diffuse the situation. We explore other ways of combining the Thinking Together approach with computers on the following pages.

Practitioners may like to read a case study that explored other benefits of collaborative group work, as perceived by the students themselves.

How were computers used to support the Thinking Together approach?
The researchers suggested that the computer offers an important way of supporting children’s learning when working in small groups. They argued that children often treat computers as a person that thinks and knows things.  But children are also aware that computers are machines and not people. The researchers suggested that this ambivalent nature – being both an object (a machine) and yet subject-like (interactive) gives the computer a unique and distinctive role in mediating learning conversations. The fact that computers are really objects, infinitely patient and without judgement means they can provide a safe context for children to try out ideas, while their ability to behave as interactive partners in games and conversations enables them to focus and direct children’s learning dialogues.

The researchers’ studies showed that if computers were to support learning dialogues effectively, teachers needed to remind the children of the ground rules for exploratory talk that their class had agreed upon before asking them to work together in small groups around computers.  The researchers’ video evidence showed that when children were not prepared for working together before working with a computer, they did not work effectively together. For example:

• one child appointed him/herself as leader, sat centrally to the keyboard, and read out and entered instantaneous responses to questions
• quiet and less confident children watched, agreed or withdrew from the ‘discussion’
• friends tended to agree readily with each other, while other children always disagreed
• the most heated debates were about who pressed the next key or used the mouse etc
• the children chatted about other things rather than complete the task
• the children competed against each other.

The researchers found that when children had been taught to use the ground rules, computers effectively supported learning conversations between children. Their findings showed that different types of computer activity supported learning in different ways.  For example:

• group strategy games played against the computer motivated and supported the development of joint reasoning, and
• computer software learning activities prompted Initiation Discussion Response Feedback (IDRF) exchanges, where the computer was used to stimulate discussion and then direct it through giving feedback.

We explore these ways of supporting learning conversations with computers on the following pages.

How did playing strategy games against the computer motivate and support joint reasoning?   
The researchers found that working together to try to beat the machine was a motivating context for shared reasoning and problem solving, provided the children knew the ground rules for exploratory talk. The researchers gave examples of discussions by two groups of pupils that showed the difference that preparing the children for working collaboratively at the computer made. The children were expected to locate an elephant in a grid-map, by keying in co-ordinates.  After each guess, the programme provided information about how near the guess was to the actual position of the elephant.

In the first example, the two boys took turns to make random guesses, not really based on the information the computer offered.  Their talk was disputational – they laughed or made derisory comments when their partner made an incorrect guess. They were sufficiently motivated to keep trying until one of them guessed the right answer and could say with satisfaction, ‘I won!’ Their exchanges did not help them learn about co-ordinates or to develop problem-solving strategies together:

Lester: I know where it is. (Sean takes his turn and fails to find the elephant)
Lester: I told you it weren’t over there. (He then takes his turn also without success)
Sean:  Eh, heh heh heh (laughing gleefully).
Lester: Which one just went on? I don’t know. (Says something unintelligible).
Sean:  1, 2, 3, 4, 5, 6. (Counting squares on the screen).
Lester: I know where it is.
Sean:  I got the nearest.

In the second example, the children were expected to find a rhino on a grid with negative as well as positive co-ordinate squares. This time, the children’s class had worked through the Talk lessons. The children’s interaction at the computer was quite different. The children collaborated to beat the computer. They discussed and agreed pairs of co-ordinates before one of them typed them into the computer.

Working together to find the rhino
The children had just been told by the computer that the rhino was 12 squares away.

Andy: 1,2,3,4,5, … 12  … (pointing to screen).
Baz: What, oh - 2, -5.
Andy: Maybe then, cos look …. Cos when you went 12 you went that way but if you go that way it’s 2 away and it makes 12, look see it goes 1,2,3,4,… So I think it’s that one, do you?
Baz: Yeh, OK, let’s try it: -1, -3.
Andy: No, it can’t be actually, no ...
Baz: -2, -4 it might be.
Andy: Yeh, it’s got to be that, if it’s not I’ll be surprised.

How did computers prompt and direct discussion between pupils?
The researchers argued that the activity around much educational software often fits the kind of exchange known as IRF which involves Initiation by the teacher, Response by a pupil and Follow up by a teacher. There is often an initiation by the computer, a response by the user and feedback or follow-up from the computer. They gave examples of how this kind of exchange was transformed when the computer prompted the children to discuss their response together – the IDRF exchange (Initiation, Discussion, Response, Feedback).  During their discussion, pupils came up with ideas and supported them with reasons before testing them out on the computer.

An example of an IDRF exchange prompted by a piece of tutorial software

(On the computer screen)

Rough surfaces cause:

a) as much friction as a smooth surface?
b) more friction than a smooth surface?
c) less friction than a smooth surface?


Rachel: Which one do you think it is?
Cindy:  'c'
Rachel: I think 'b' (Laughs).
Cindy: I don't. Look 'changes more surfaces than a smooth surface' (Misreading the screen).
Rachel: Yeh I know but - wait, wait - listen, if you rub two smooth surfaces together right, will it be slippery or stable? (Rubs hands together).
Cindy:  Stable - depends how tight you've got it.
Rachel: Cindy listen! If you've got oil on your hands and you rub them together will they be slippery or not? (Rubs hands together).
Cindy: Well you see (She rubs her hands in a parody of Rachel but in a way that makes them miss each other) 'cos they don't rub together they go ...
Rachel: Cindy! (in mock exasperated tone) If you've got ...
Cindy:  Yeh, they will be slippery! (laughs).
Rachel: Yeh, exactly. So if you've got two rough surfaces and you rub them together it will not be as slippery will it?
Cindy: No.
Rachel: So that proves my point doesn't it?
Cindy: mmm.
Rachel: Yes, do you agree? Good. (She clicks on answer 'b').

In this example, Rachel appeared to know the answer and persuaded Cindy by giving her reasons.  The researchers pointed out that although this discussion seems to be one-sided, Cindy was genuinely persuaded, and in other interactions Cindy was the one persuading Rachel, so their relationship was more balanced than this episode implies.

Practitioners may like to read a case study that investigated how a computer helped to structure discussions between older students. The students were prompted by the computer to predict the flight path of various projectiles such as a ball rolling off a table, before watching a digital video clip of the experiment. The study examined how the tasks promoted meaningful discussion between the students.

What training and support did the teachers receive?
The researchers outlined the training and support they provided for the teachers taking part in the intervention programme. They first introduced the teachers to the Thinking Together programme in an after-school session that was designed to inform the teachers about the background and aims of the approach.  All head teachers, advisory teachers, participating teachers and learning support staff in the target schools were invited.

The introductory session was followed by a full day of professional development and two further after-school sessions for the teachers in the target schools. All learning support staff in the target schools also participated in a half-day of in-school professional development. During this time, the researchers worked with the teachers to plan lessons that focused on developing the children’s awareness and skills in using spoken language.

The researchers visited all the target schools regularly to support the project teachers and collect feedback on activities as well as observe and gather data. This informal support was strengthened by two more after-school sessions.

The approach to professional development adopted by the researchers in these studies echoes the findings of three systematic reviews of continuing professional development which we described in two earlier RfTs. We reported how the reviewers found a link between the provision of a mix of teacher-to-teacher collaboration plus input and support from specialists and a positive impact on teaching and learning. The use of specialist expertise took a number of forms including:

• instruction by the specialists
• continued support during the intervention
• observation and feedback
• coaching in new techniques and strategies.

Practitioners can find out more about the kind of specialist expertise that made a difference in our RfT summary of the first review of collaborative CPD and our RfT summary of the second and third reviews.

Practitioners may also like to read a case study that details the professional development provided by other university researchers and its impact on collaborative learning. They ran a two-day workshop to introduce the teachers to the basic principles of collaborative learning and provided training in specific communication skills that challenged the children’s thinking. The case study shows how the children then used the communication skills modelled by their teachers in their discussions with each other.

How was the research conducted?
We have focused on three of the researchers’ studies: a study that involved KS1 children, a study that involved KS2 children and a study that explored ways computers provided support for learning conversations.

The researchers gathered a range of complementary data for their studies:

• the children’s scores on two non-verbal reasoning tests (completed before and after the programme had ended)
• video recordings of focal groups from the target classes taken at the start and end of the programme, in which the children carried out a reasoning test activity or worked with a piece of computer software
• observations and written notes of how the children in the target schools worked with each other
• interviews with the teachers, head teachers and support staff (where possible) in the target schools.

The researchers transcribed their videotapes of the focal groups’ talk. Their analysis of the transcripts revealed that the children used words such as ‘because’, ‘if’, ‘I think’, ‘why’, ‘which’, ‘what’, and ‘you’ to put forward reasons and encourage the inclusion of each other’s perspectives. These words were then included in a computerised search for key features in the talk of the focal groups that indicated exploratory talk. They were used to provide a quantitative measure for comparing transcripts, both before and after the programme and between the target and control classes.

Practitioners considering conducting their own classroom research may like to read a case study which describes how a group of teachers investigated ways of encouraging exploratory talk with Year 8 and Year 10 students during collaborative group work in mathematics lessons.

References:
Wegerif, R., Mercer, N., and Dawes, L. (1999) From social interaction to individual reasoning: an empirical investigation of a possible socio-cultural model of cognitive development. Learning and Instruction 9 (5) pp.493-516

This study involved 124 children aged 9-10 years from six middle schools. Sixty pupils from three middle schools took part in the Thinking Together programme. Each of the target classes was matched with a control class of the same age in another middle school (64 children).

Wegerif, R., Littleton, K., Dawes, L., Mercer, N., & Rowe, D. (2004) Widening access to educational opportunities through teaching children how to reason together. Westminster Studies in Education 27 (2) pp. 143-156

This study involved six teachers and their Y2 classes (aged 6-7 years) in three target schools and five teachers and their Y2 classes in three matched control schools.  One of the target schools and one of the control schools had a high proportion of children from low income families who had fairly recently arrived from the Indian subcontinent and for whom English was an additional language.

Wegerif, R. (2004) The Role of ICT as catalyst and support for dialogue. NALDIC Quarterly 1 (4) pp. 4-12

The researchers analysed over 50 hours of video recording of groups of children from a range of schools working with different pieces of computer software.  They identified examples of exploratory talk (such as asking probing questions or giving reasons for opinions) and analysed the impact of the software on supporting or inhibiting this kind of talk.

What are the implications of the research?
Teachers wanting to promote collaborative working and improve their pupils’ thinking skills may like to consider the following implications of the findings of this research:

The researchers stressed the importance of eliciting ground rules in the children’s own words.  Would watching videos or listening to audio tapes of themselves or other pupils working together help with this? How could you develop their comments into rules for discussion for their class?

The teachers involved in the studies organised their classes into mixed gender groups of three. How do you group children for learning? Is this a key area to think about when you are planning for group work? (You may find it helpful to read case study 1 which shows the criteria one teacher used for deciding which children to group together).

Could you make more use of computer activities to create opportunities for learning conversations, whilst supporting your pupils’ in learning how to work collaboratively?
Could you work collaboratively with some of your colleagues to help you investigate ways of promoting exploratory talk with your classes?  Would classifying your children’s talk into disputational, cumulative and exploratory types create a framework for developing your thinking about the way the children interact with each other? (You may find case study 5 a useful starting point because it shows how a group of teachers went about exploring and analysing the talk their classes engaged in).

Would your pupils benefit from reflecting on their experiences of working collaboratively with you and each other? How might you plan and organise such debriefing sessions? (You may find case study 2 a useful starting point because it shows a way of collecting and analysing students’ views of collaborative group work).

Leaders may wish to consider the following implications:

The researchers stressed the importance of teachers modelling how they wanted children to talk together. Would your colleagues find it helpful to have the opportunity to practise and reflect on their skills of questioning, and giving and asking for reasons, through taking part in role-plays with each other etc? (You may find case study 4 a useful starting point because it shows how teachers were trained in how to challenge their pupils’ thinking and understanding during group activities).

Does improving communication feature sufficiently within your proposed development plans as a means of promoting social inclusion? Could your school do more to exploit enhancing talk skills as a social inclusion strategy?

The teachers involved in the projects reported in the RfT benefited from professional development focused on structured group work. Could you do more to support colleagues trying to improve collaborative working in their classrooms, by for example, organising workshops and inviting practitioners from university education departments or other schools?

The researchers found that certain kinds of computer software could support learning dialogues (provided children were prepared for working together beforehand).  Would your colleagues find it helpful to have opportunities to share with each other their experiences of using the computer in this way?

Filling in the gaps   
Gaps that are uncovered in a piece of research have a useful role in making sure that future research builds cumulatively on what is known. But research also needs to inform practice, so practitioners’ interpretation of the gaps and follow-up questions are crucial. We think three kinds of studies would usefully supplement the research we have reported on:

• studies designed to gather evidence of collaborative group work in other areas of the curriculum, such as literacy, history and music
• a review of different approaches to supporting group work, and
• case studies of other external devices for supporting learning conversations, such as interactive whiteboards.

What is your experience?
Do you have any evidence about initiatives aimed at promoting collaborative group work? Do you have action research or enquiry based development programmes running that explore, for example the group work in different areas of the curriculum or making use of technology?  We would be interested to hear about examples of effective activities, which we could perhaps feature in our case study section.

Your feedback
Have you found this study to be useful? Have you used any aspect of this research in your own classroom teaching practice? We would like to hear your feedback on this study. Contact research@gtce.org.uk to share your views with us.