Indian Journal of Science Communication (Volume 2/ Number 1/ January – June 2003)

 << Back  


Effective Science Communication : Practical Work, Conversational Texts and Prompt-posters for Triggering Discussion

Paul Webb, Paddy Lynch, Derek Potgieter, Notozi Mgobozi, Des Cross, Raj Kurup, Pam Austin
Viv England and Scott Linneman

Department of Science, Mathematics and Technology Education, University of Port Elizabeth
P O Box 1600, Port Elizabeth 6000 (South Africa); Email : cenppw@upe.ac.za

Abstract

In this pilot study the underpinning hypothesis is that discussions in science classrooms can be triggered in a number of ways and that these triggers provide different potential as to both the nature and quality of the discourses that ensue. These, in turn, have different effects on the cognitive and affective development of learners. This study, which researches a series of matching triggers focusing on magnetism at grade seven level in disadvantaged South African schools in urban, peri-urban and rural milleus focuses on practical work, conversational texts and prompt-posters as triggers. The effect of these triggers (and resultant discussion) on children's reasoning, as measured by using Raven's Standard Progressive Matrices, is examined. The data generated by this small scale exercise indicate that the use of practical work and conversational texts generate statistically significant cognitive gains as measured by changes in scores in pre- and post-tests using Raven's Standard Progressive Matrices.

Key Words

Discussion in science classroom, Prompt-posters, Triggers, Science education, Whole-class discussion

Introduction

Constructivist research (particularly notions of social constructivism) has sensitised science educators to the importance of classroom discussion by learners' in science classrooms for effective learning to take place (Sprod 1995). Solomon (1994) champions a change in metaphor in science education to that of a child on the edge of a circle of initiates trying both to make sense of the conversation and be accepted into it. This metaphor requires that an increasing emphasis be placed on the use of discussion in classrooms (Russell and Osborne 1993).

Classroom discourse pattern have been studied fairly extensively (Flanders 1970) and the dominance of Sinclair and Coulthard's (1975) Initiation-Response-Feedback (IRF) questioning cycle has been confirmed by Lemke (1990). This technique (IRF) has been criticised (Edwards and Mercer 1987, Dillon 1994) as being rooted in the need for teacher control of the discussion, rather than an educational rationale, which leads to shallow thinking and a 'guess what teacher is thinking' approach to learning that promotes an epistemological understanding in learners that scientific knowledge is fixed, revealed and uncontentious (Lemke 1990). By contrast, 'True dialogue' and 'Cross-discussion' are two discourse patterns that Lemke (1990) identifies as best supporting constructivist learning. In turn, they are the discourse patterns least used in classrooms (Lemke 1990).

There are a number of models for promoting effective discussions (Dillon 1994). The 'Philosophy

for Children' programme (Lipman, Sharp and Oscanyan 1980) used 'triggers' (often a purpose-written story with open ended puzzles). The discussion is then based on questions that the children ask about the trigger experience and the role of the teacher is to require good thinking about the questions (usually by modelling and cuing). Gardner (1995) asserts that these types of discussions are neither learner-centred nor teacher-centred; they are 'truth-centred'.

In this study the hypothesis is made that discussions in science classrooms can be triggered in a number of ways (e.g., 'hands on' practical work, conversational texts and 'prompt-posters') and that these triggers provide different potential as to both the nature and quality of the discourse that ensues. As such, this research in progress (and pilot study) is aimed at investigating :

  • l science focused triggers of 'whole-class discussion' and the nature of their effect on learners

  • l the relative effectiveness of the different triggers in each case in terms of promoting social constructivist learning

  • l the nature of the effect of the type of whole-class discussion generated on learners in terms of classroom climate, learning strategies and cognition.

Methodology

A series of 'matching' triggers focusing on magnetism were developed for grade seven classrooms, viz., practical activities, conversational readings, and 'prompt-posters' (allowing 'discussion about 'instances' or situations). Each trigger covered four aspects of magnetism, viz.:

  • l magnetic and non-magnetic materials

  • l magnetic fields and forces

  • l finding direction using magnets and

  • l caring for your magnets.

Twelve grade-seven science educators (four in Port Elizabeth, four in King William's Town and four in Queenstown) were selected to participate in the research. These three different geographical settings were chosen to reflect urban, peri-urban and rural milleus. The educators were matched by their success to date as students on our B.Ed. (Science and Mathematics) programme and were allocated (randomly in each centre) to a trigger or control group. As such, each trigger was applied in a classroom in each centre and there was a control group in each centre. The schools were also broadly matched in terms of the fact that they were chosen as institutions that were neither dysfunctional nor excellent, were from disadvantaged communities and had grade seven class sizes between 30 and 50 learners.

The teachers were introduced to notions of 'whole-class discussion' and trained in the use of the specific trigger they had been randomly allocated. All participating teachers (other than the control group) attended a workshop on 'the place of talk in science classrooms' and introduced to possible strategies and criteria for establishing whole-class discussion. Participants were then tasked to make their learners aware of the structure of good whole-class discussion, e.g. that interactions could be teacher to learner, learner to teacher, or learner to learner and that these interactions could be observations, explanations or questions. Rules of politeness applied, and the 'game strategy' was to enable as many types of interaction to take place during a lesson. The danger of slipping into an Initiation-Response-Feedback (IRF) mode was also emphasised.

Further workshops provided material, trigger strategies and topic sequencing for magnetism. For each trigger the magnetism unit was broken into four phases, viz. magnetic and non-magnetic materials; magnetic fields and forces; finding direction using magnets and caring for your magnets. Thus a situation had been developed in which teachers had both been alerted to the characteristics of good whole class discussion and enabled to implement it in their classrooms using either practical work, prompt-posters or conversational readings in their classroom.

Before the teachers applied these triggers in their classroom, measures of classroom climate (WIHIC), cognitive discourse strategies (DS) and reasoning (Raven's Standard Progressive Matrices) were applied to their learners. As the learners participating in this study are all Xhosa home-language speakers, the tests were administered by a researcher who is also a Xhosa home-language speaker.

Also, as our baseline study of the natural situation in the classrooms of participating teachers indicated

that a substantial amount of Xhosa is used by both teacher and learners, recognition of code switching was made explicit to all participants and encouraged. There are strong theoretical justifications for the side-by-side necessity for Xhosa and English discourses in order to promote understanding (Halliday 1985, Chick et al. 1985). The degree of code switching that occurs varies according to the type of discussion interaction that occurs, e.g. teacher-learner, learner-teacher, or learner-learner (Milroy and Muysken 1995).

The educators then applied the four topics using their allocated trigger over a period of a school term. It was intended that at the end of the intervention period (one school term) the measures of classroom climate, learning strategies and cognition were again to be applied to all participating learners (as well as to the control groups), but it ended up that only classes in Port Elizabeth were post-tested.

In this paper, the possible effects of whole class discussion on measures reasoning (i.e. scores on the Raven's Standard Progressive Matrices) are treated statistically and analysed. Raven's Standard Progressive Matrices (CPM) consists of graphical puzzles and is widely used in education and psychology as a test of 'non-verbal' reasoning. Raven's tests appear to be particularly appropriate for exploring the links between language practices as they correlate well with similar tests of reasoning and with measures of academic achievement (Raven, Court and Raven 1995, Richardson 1991). To account for the fact that the four different groups (practical activities, conversational readings, 'prompt posters' and control group) did not start at exactly the same baseline when pre-tests were conducted, it was necessary to apply analyses of co-variance (ANCOVA) with pre-test scores being the co-variates to compare the adjusted post-scores. Cromwell's coefficient of ‘x’ was used as an indicator of reliability. The data obtained will later be triangulated with the data generated by the attitudinal measure (WIHIC) and the cognitive discourse strategies that deal with how learners cope with language difficulties in bilingual classrooms.

Results

Classroom observation and video recordings have been made both pre- and during the course of the intervention, but have yet to be analysed in full. Classroom climate and discourse strategies pre- and post-tests have also been administered and await analysis. The Raven's Standard Progressive Matrices pre- and post-tests of reasoning (60 items) have been analysed and yielded the following data.

The frequency distribution of the pre-test of the Port Elizabeth sample (n = 146) indicated that the spread of the results was wide, there was a substantial fraction of very low scores and that the average score was low. The 50 percentile score fell at a value of 21, while the United Kingdom (UK) 50 percentile norm for 12 year olds is at a value of 38. However, the post-test 50 percentile of the Port Elizabeth sample (including the control group) fell at 35, approximating that of the UK 50 percentile norm and which represents a considerable improvement.

The mean pre-test scores of the four groups have been given in Table 1. Analysis of covariance rejects the nul hypothesis that there is no difference between the four groups at the 99% level of confidence (p < 0.01).

There is a statistically significant difference between the change in Raven's Standard Progressive Matrices scores of both the practical activities group and the conversational readings group versus the control group (p<0.05 in both cases). There is no statistically significant difference between the change in scores of the practical activity group and the conversational reading group (p = 0.7608), but they both differ significantly from the change in scores of the prompt poster group (p = 0.0285 and 0.0103, respectively). There is no statistically significant difference between the prompt poster and the control group data (p = 0.6886).

Table : Number of learners, pre-, post-, adjusted post-test and change in mean score in each of the trigger groups and control group
Group  n Pre-Test  Post-Test Adjusted Post Test Change
Practical work 30 24.2 30.8 32.3 6.67
Prompt posters 33 26.8 27.8 28.5 2.88
Conversational readings 35 27.4 32.9 32.9 6.26
Control 48 28.3 30.6 29.1 2.21

Practical activities generated a great deal of small-group and whole-class discussion. Observation of the teacher using the 'Big Book' conversational reading revealed that the discussions generated often led to the teacher bringing out pertinent apparatus and allowing the children to repeat what was illustrated in the book. These activities led to considerable small-group discussion, almost totally in the vernacular (learners had been told that Xhosa and English were equally acceptable in true discussion). The teacher using the prompt-posters was not able to present either the content of the lesson, or the methodology, as intended. She was nervous and could not generate the desired whole class discussion at any of the stages of the lessons observed, despite support during the debriefing sessions. Also, the use of 'flash-cards' to focus on the meaning of words used in the posters appeared to restrict the discussion to a much more closed discourse than was desired.

Discussion

The use of whole-class discussion as a teaching and learning strategy in South African classrooms is not unproblematic – particularly in schools from the previously disadvantaged sector. However, preliminary data from this study in progress suggests that the whole class discussion generated by means of practical work and conversational readings results in improved reasoning, i.e., raised mean scores in the Raven's test of 28% and 23% respectively. An increase of 11 and 8 percent in the Raven's Standard Progressive Matrices tests were also recorded for the prompt-poster trigger group and the control group, respectively, but these increases in scores were not statistically significantly different from one another and were statistically significantly lower than those achieved by the learners in the practical work and conversational reading groups.

The enhanced Raven's test average for the control group (not statistically significant) can be possibly ascribed to the 'Hawthorne Effect', viz. improvements or changes in behaviour induced simply by the stimulus of the 'attention' focused on the group by the presence of observers who appear to be interested in the group selected as the control. Light may be shed on this matter by a careful study of the videotape data on the control group. Similarly, study of the qualitative videotape data on each of the experimental groups should reveal insights into the verbal and cognitive interactions that took place as a result of the different triggers, as will analyses of the classroom observation records. These sources suggest that the failure of the prompt-poster trigger to result in statistically significant increases in Raven's scores over the control can be ascribed to the fact that the teacher using this trigger was not able to present either the content of the lesson, or the methodology, as intended. She personally struggled with the concepts herself and, as such was nervous and could not generate the desired whole class discussion at any stages of the lessons observed, despite support during the debriefing sessions.

However, the Raven's test data generated by whole-class discussion triggered by practical work and conversational readings were both remarkable and unexpected, considering the short duration of the intervention. Four planned lessons on magnetism during a school quarter could not reasonably be expected to bring about such a marked improvement in reasoning skills. A possible explanation that could be put forward is that, as the teachers teach the whole gamut of subjects to the participant learners, the teacher introduced whole-class discussion to other aspects of science and other learning areas on a regular basis for the duration of the intervention. The possibility that this has been the case, requires investigation. If this is what has happened, the introduction of whole-class discussion as a general teaching and learning strategy via science lessons (or any other learning area specific context) may be a more viable technique than anticipated.

Further interrogation of the pilot data, and full implementation of the study in 2002, should provide pointers to answering a number of questions that this research has raised and give guidelines to generating meaningful whole-class discussion in classrooms, a teaching and learning strategy sorely missing at present in South African schools.

Acknowledgements

The authors are grateful to the National Research Foundation (NRF) of South Africa for financial support for this study.

References

  1. Chick J, Language in classrooms : Towards a framework for intervention, NCCRD Report, Pretoria, 2000.

  2. Dillon J T, Using Discussions in the Classroom, Open University Press, Buckingham, 1994.

  3. Edwards D, and Mercer N, Common Knowledge : The Development of Understanding in the Classroom, Methuen. London, 1987

  4. Flanders N, Analysing Teacher Behaviour, Addison-Wesley. London, 1970.

  5. Gardner S, Inquiry is no mere conversation, Critical and Creative Thinking, 3 (2) : 38-49, 1995.

  6. Halliday M A K, Language, Context and Text : Aspects of Language in a Social-Semiotic Perspective, Deakin, Oxford University Press, 1985.

  7. Lemke J L, Talking Science : Language, Learning and Values, Ablex, New Jersey, 1990.

  8. Lipman M, Sharp A M and Oscanyan F S, Philosophy in the Classroom, Temple University Press. Philadelphia, 1980.

  9. Milroy L and Musken P, Code-Switching and Bilingual Research : One Speaker, Toe Languages. Cross-disciplinary Perspectives on Code-Switching, Cambridge University Press, Cambridge, 1995.

  10. Raven J, Court J and Raven J C, Manual for Raven's Progressive Matrices and Vocabulary Scales, Oxford Psychologists Press, Oxford, 1995.

  11. Richardson K, Reasoning with Raven in and out of Context, British Journal of Educational Psychology, 61 (2) : 129-138, 1991.

  12. Russell T and Osborne J, Constructivist research, curriculum development and practice in primary classrooms : reflections on five years of activity in the Science Processes and Concept Exploration (SPACE) Project, Paper presented at Third International Seminar on Misconceptions and Educational Strategies in Science and Mathematics, Cornell University, 1993.

  13. Sinclair M Mc H and Coulthard R M, Towards an analysis of discourse, The English Used by Pupils and Teachers, Oxford University Press. London, 1975.

  14. Solomon J, The rise and fall of constructivism, Studies in Science Education, 23 : 1-19, 1994.

  15. Sprod T, Cognitive development, philosophy and children's literature, Early Child Development and Care, 107: 23-33, 1995.

<< Back