Enhancing Statistical Literacy: Multivariate Thinking in Precollege Social Sciences
James R. Nicholson, Jim E. Ridgway, Sean McCusker
SMART Centre, School of Education, Durham University, Durham, United Kingdom

Social sciences deal with complex phenomena where multiple factors are important. Increasingly there is access to disaggregated data where the effects of these factors and any interactions between them can be explored. However, because dealing with multivariate data has not been a realistic possibility until recently, much of the teaching in pre-college social sciences has been discursive, and based on headline statistics about aggregated data. There is also a tradition of 'qualitative' research, where data from large scale surveys is ignored, in preference to (for example) interview data from small samples. To overstate and oversimplify: there is a culture of theorising about causality without first understanding the phenomena in detail. Consider, for example, inequalities in educational attainment. It is widely reported that girls do better than boys, and that students in higher socio-economic groups do better than those from lower groups. However, if one is constructing theories about educational attainment, it is important to know if the differences between boys' and girls' attainment is: the same across socio-economic groups; in different regions; across all ethnic groups, and to explore changes over time.

Students studying social sciences at school in the UK rarely take advanced courses in mathematics. We suspect that statistical literacy levels among students (and perhaps teachers) are rather low. The use of visualisation tools facilitates data exploration, even at pre-college level. On a project funded by the Nuffield Foundation, we present multivariate data that is central to topics in a pre-college level Sociology course, via interactive displays. The paper will describe work with teachers and students on the development of their statistical literacy, and their understanding of key topics. The work has implications for both the definition of 'statistical literacy' and for approaches to fostering statistical literacy in a non-mathematically oriented student population, and in citizens in general.

Keywords: Statistical literacy; Social sciences; Multivariate thinking; Data visualisation

Biography: James Nicholson has over 25 years experience in teaching mathematics and statistics at secondary level, with extensive engagement in national and international mathematics and statistics education. James is the author of two statistics textbooks and has contributed to a number of mathematics texts. He grew up in Belfast in the North of Ireland before studying mathematics at Cambridge and then teaching in England for 12 years. James returned home to Belfast in 1990 as head of the mathematics department at a large secondary school until 2003. He still lives primarily in Belfast, but a lot of his work is with the SMART Centre at Durham University, working on data visualisation.