Background

Delayed speech or language development is the most common developmental problem. It affects 5-10% of preschool children. Children's language development may be delayed for a number of reasons such as hearing problems and general learning disabilities. Some children however, have delayed or unusual language development for no known reason. If a delay occurs before the age of four, this can resolve spontaneously for many children. For others however, problems do not ameliorate and continue into the school years and even into adulthood.

Children with 'pragmatic language impairment' have difficulty with the use of language in social situations. For example, they may be overly talkative, fail to understand conversational turn-taking, have difficulty with conventions such when to stay on a particular topic or change topic, interpret jokes and sarcasm over-literally and find it difficult to gauge a listener's needs for information. Such children may sometimes however, have a good vocabulary and understanding of language in non-social situations. Children may develop this kind of language impairment because of genetic factors or because of missed opportunities for social interaction early in development.

Our aim was to select the children from the participating schools who appeared to have the most difficulties with pragmatic language skills and to test a new method of providing help and support for these children using a computer task designed specifically for this purpose.

The 'Maze Game'

We know from our previous research (Murphy and Faulkner 2000, 2006, 2011) that when children collaborate on problem-solving tasks together, it is the children with good communication skills who work most effectively. In particular, we found that children who were successful offered more explanation, gave explicit directions and guidance and addressed questions clearly to their work partners. Children with pragmatic language impairments are likely to have difficulties with all of these skills. We therefore designed a computer task that obliged children to think about and use language and communication skills. From our previous work with children in pairs, we had observed that when two children work together on a computer task side-by-side, using one screen only, the more confident, able or dominant child usually takes the lead, leaving the other child (who may actually need more practice) to take on a more subordinate, passive role. The task that we devised for this project necessitated the use of two computer laptops. Children worked in pairs and each child sat at one screen, with the laptops facing each other. In this way, children could see their own screen, but not that of their work partner. (See illustration below.)

Children working as a pair on computerised 'Maze Game'.

The game involved driving a 'car' around a maze with features such as houses, castles, trees, ponds and animals. The arrow keys on the computer keyboard were used (some children found the mouse too difficult to use) to drive the car around the maze to collect rewards, such as magic wands, stars and treasure chests from various locations. There were also obstacles to be avoided, such as snakes, which could cause the children to 'lose' the game. Each child had, however, a slightly different view of the game.

Child 1(the driver's) view showed the maze and maze features, but not the location of the rewards or the obstacles, which were hidden from view. The other child (the navigator) could see these items clearly on their view of the game. This child (the navigator) however, could not drive, but instead had to provide directions for the other child to drive to collect the rewards and avoid the obstacles. An illustration of the driver's and navigator's views of the game are shown below:

Driver's view: Obstacle (snake) is unseen; reward (treasure chest) is also unseen.

---

Navigator's view: Obstacle (snake) is visible; reward (treasure chest) is also visible.

In this way, the driver and navigator are dependent on each other to complete the task. This obliges both of them to think about how to use skills such as asking questions, giving clear instructions and asking for clarification. These are the very skills that children with pragmatic language impairments are likely to find problematic.

Participant Selection Process

There were six participating lower schools in the Borough of Bedford. The parents of all Year One children in these schools were asked if they would give consent for their children to participate in the study and 61% of parents agreed to do so. This gave us a total of 215 children. All these children were assessed using the Test of Pragmatic Skills devised by Shulman (1986). The test asks children to take part in conversations, respond to requests and ask for clarification from an adult tester. The expected score for children aged between 5 years 0 months and 5 years 11 months is 28, the minimum score is 0 and the maximum score is 41. We selected all children scoring below 22; this gave us 32 children who were targeted for extra training using the computer task. These children's scores ranged from as low as 10 up to 22.

For further information about these children's other skills with language, we also tested their knowledge of vocabulary and word use using the British Picture Vocabulary Scale, however, this was not used for selecting the children. It is important to note that although the Test of Pragmatic Skills and the British Picture Vocabulary Scale can give us a measure of a child's language abilities for research purposes, these are not diagnostic tests and children of concern should be properly assessed by speech and language therapists.

Testing of Children

The 32 children that we had selected were paired with another child of the same gender who had scored highly on the Test of Pragmatic Skills and who became their working partner on the computer task. All 32 children played the Maze Game with their partners and were audio-recorded as they did so. We measured how many rewards they managed to collect, and, from the audio-recordings, measured the number and kinds of questions that they asked, the directions and guidance they gave to partners, clarification asked for and responses to questions.

Training of Children

After testing at baseline, the 32 children were randomly divided into Group 1 (16 children) and Group 2 (16 children). The children in Group 1 then received three training sessions on the computer task with one of the researchers whilst the children in Group 2 received no intervention. All 32 children were then tested again with the Test of Pragmatic Skills and repeated the Maze Game with their working partners thus enabling us to compare children who had received training with those who had not. After this, the children in Group 2 then received the three training sessions, whilst the children in Group1 received none. All children were then tested again one final time which enabled us to see whether any improvement in the Group 1 children was maintained, and if the Group 2 children had improved.

The training that we devised for the children was based on the training guidelines proposed by Baines et al (2009) in their book 'Promoting effective group work in the classroom'. Baines et al proposed use of the following teaching strategies:

• (a) Modelling: providing the child with a good model of communication to 'copy' on which they can base their own approach to the task.
• (b) Briefing: before commencing the task, describing exactly what needs to be achieved. In this case, there was a different focus in each of the three sessions, how to ask questions, how to give directions, how to ask for clarification.
• (c) Practice: getting the child to practice the above skills for themselves.
• (d) Recap: Providing feedback about the child's practice in a positive and constructive way.

The three sessions each specifically targeted asking questions, giving directions and clarification requests as, previous research on children with pragmatic language impairments has shown that, when faced with a communication task of this nature, these are specifically the areas where they perform most poorly. The children were given three half-hour sessions focussing on each of these skills.

Our manual provides details of each of the training sessions and of the pre- and post-test tasks that the children carried out with their work partners. We piloted the task extensively with many children and from this learned what we felt were the best ways of providing explanations and coaching to the children. We included very specific instructions for the tasks as we wanted make sure that all the children received the same training.

Results

Differences between the children who received training and those who did not

At post-test 1, children who had received the training managed to secure 71% of the rewards available, whereas the children who had not received it managed only to secure 58%. Collecting rewards is only a rough measure of success on the test, but this does seem to indicate that the training may have had a beneficial effect.

The Test of Pragmatic Skills is a more accurate measure of whether or not children benefitted from training and, importantly, whether or not this training has generalised to situations outside of the training. The training was focussed specifically on the computer Maze Game, and therefore, improvements on this task show that children will have learned and retained what was taught. However, in many cases, what is learned on particular tasks is afterwards applied by children only on the same task by children.

The average score of the children in Group 1 (who received the training) was 18.7, this group improved their scores on the Test of Pragmatic Skills on average by +8.33 points after the training. The Group 2 children who did not receive training started at a slightly higher average score, 19.6, but improved by only +5.33. In fact, some of this group had scores that were actually worse. The difference in scores is illustrated in the graph below. The fact that the trained children showed a definite and statistically significant difference in improvement over the non-trained children is a very strong indicator that they are applying this training outside of the computer task situation. The difference is statistically significant at the 1% level, which means that there is a less than 1% probability that we obtained these results by sheer chance.

Test of Pragmatic Skills scores before and after training for Group 1.

Analysis of the children's conversations from the audio-recordings as they played the game showed that the children who received the intervention used significantly more clarification-questions and information-seeking questions than the children who did not. They also gave their partners better directions, although not to a significant extent. Interestingly, they also made significantly many more positive comments about the game and about their partners whilst playing, suggesting that they were enjoying collaboration much more.

Before undertaking the study, we had speculated that possibly the children starting with the highest Test of Pragmatic Skills scores would improve the most as they were more able children, or maybe that those with the lowest scores would improve the most, as there was scope for the most improvement. In fact, neither of these speculations has proved to be true. When we looked at BPVS scores, which give an indication of the child's understanding of language and vocabulary (expected standardised score is 100) and many children who had scores of well over 100 still experienced difficulty with the pragmatic use of language.

Sadly, it appears that the training may not benefit all children. Some children improved little, in fact only about the same as the children who had not had any training. Again, the reasons for this are not obvious – there does not appear to be any one kind of child in this group as far as we know from the measures that we took and all these children were from different schools.

Why did some of the children who did not receive training improve?

The Group 2 children (the control group) who did not receive training improved their score on the Test of Pragmatic Skills, on average, by 5.33 points. One may ask why these children did not remain at the same level. Some improvement is to be expected through maturation as the children are growing older and the brain develops. Also, parents, schools and teachers are not without effect on a child's difficulties, although this effect can be variable. The improvement of +5.33 is an average and there is quite a lot of variation between these children's scores.

Conclusion

In conclusion, at appears that the training that we gave to the children was largely beneficial, although there are reasons that we have not yet managed to pinpoint as to why this did not work for some of the children. One possible reason is that three half-hour sessions is a limited intervention and that these children simply need greater input to improve. Another possibility is that some other factor that we did not measure is impacting on these children's ability to deal with the pragmatic aspects of language and, until and unless this can be addressed, then improvements will not be possible.

It is gratifying to see that for the children who were given the intervention however most (12 out of 16) experienced a substantial improvement. These are early, initial findings for working with children in this kind of way. Further work would be needed to determine exactly what it is that is that aided these children. The task was designed to highlight the importance of communication and to oblige the children to think about this by making the task impossible to complete without good communication between the pairs. Whether the task itself is promoting this thinking, or whether it is the training that stimulates thinking around this aspect of communication could be an worthwhile topic for a future piece of research.

References

• Baines, E., Blatchford, P. and Kutnick, P. (2009) Promoting effective group work in the classroom. Routledge: London.
• Murphy, S. and Faulkner, D. (2011). The relationship between bullying roles and children's everyday dyadic interactions. Social Development, 20 (2), 272-293
• Murphy, S. and Faulkner, D. (2006). Gender differences in verbal communication between popular and unpopular children during an interactive task, Social Development, 15(1), 83-109
• Murphy, S. and Faulkner, D. (2000). Learning to collaborate: Can young children develop better communication strategies through collaboration with a more popular peer? European Journal of Psychology of Education, 15(4), 389-404
• Murphy, S. M., Faulkner, D. M., & Farley, L. R. (2014). The behaviour of young children with social communication disorders during dyadic interaction with peers, Journal of Abnormal Child Psychology, 42(2), 277-289.