INFORMATION TECHNOLOGY IN THE TEACHING AND LEARNING OF STATISTICS: EDUCATIONAL ISSUES AND PRACTICES

Statistics education has tended to focus on mathematical and mechanical aspects of knowledge. This does not empower students to apply these aspects of statistical knowledge to solve problems generally arising from a specific context. The teaching and learning of statistics should therefore aim at developing problem-solving heuristics, that is, how to think what to do rather than how to do. Hence, statistics curricula should shift the focus from mathematical calculations to tasks of a practical nature that can be more easily understood by students (Pfannkuch, 2000; Snee, 1993) and stress the ability to apply, evaluate and interpret statistical ideas (Chance, 1997). To achieve this kind of change, teaching should no longer be seen as transfer of knowledge but as a process of engaging students with learning activities in which they develop ownership of knowledge. Information technology (IT) facilitates statistical computing and graphing, and the way educational materials are delivered. Thus, the question of what is an appropriate pedagogy for statistics teaching and learning when integrating IT into classrooms should be addressed from the joint perspectives of statisticians and educators. This thesis did so by developing, implementing and evaluating a socioculturally grounded model of pedagogy in an IT environment and related models of statistical thinking and graphical comprehension in the context of regression modelling. The model of pedagogy in an IT environment was implemented by the teacher-researcher, with students enrolled in the Higher Diploma in Applied Statistics and Computing course in the Hong Kong Institute of Vocational Education. Data were gathered from a questionnaire based survey, an observational study, and an experimental study to investigate patterns of social interaction associated with statistical thinking and graphing. The results of the questionnaire based survey indicated that most students held positive perceptions of learning with IT associated with productive social interactions with their learning partners and the teacher. They found their interaction with their partners collaborative and their communication beneficial to learning as it enabled them to verbalise their thoughts to sustain task-centred discussion. The observation study drew on Mercers (1995) work to categorise student-student talk as exploratory, cumulative, or disputational. Most talk was of the exploratory type, characterised by reasoning and statistical thinking, and only a few instances were classified as cumulative when students were attempting straightforward learning tasks or maintaining harmonious social relations. No instances of disputational talk were observed. A finer grained analysis of students talk, using an adaptation of Kumpulainenss (1994) framework, identified some forms of talk that were used for maintaining social interaction, while other forms were associated with making reasoning explicit and activating higher order thinking for a variety of sophisticated tasks, for instance, deducing practical implications for regression parameters. Nevertheless, students actual progress in regression tasks was linked to the teachers interventions. The teacher adopted a combination of means of assisting performance (Tharp & Gallimore, 1988): modelling, questioning, cognitive structuring and contingency management to guide students thinking towards knowledge construction as well as problem solving. The experimental study compared students performance in a test involving preliminary examination of regression data with and without, and before and after, peer collaboration. It revealed that students who had pre-task discussion demonstrated better understanding of regression data than those who had no discussion. More interestingly, students who had pre-task discussion gave more complex and connected responses (SOLO analysis) than those who had post-task or no discussion. The extent to which peer collaboration affects students performance in statistical thinking and statistical graphing tasks may depend on the quality of group discussion and the nature of students participation and involvement in that discussion. This thesis presents both theoretical proposals and empirical results regarding how students learn statistics in an IT environment. As well as adding the body of work in this field, it provides a practical test case for the use of sociocultural ideas in having students assume greater responsibility for their own learning. The findings from this research demonstrated that collaborative group work with computers can facilitate social interaction and learning in a statistics classroom and help students to appreciate the learning environment fostered by computers