Choosing and using data and contexts for learning

Abstract

The value for student statistical learning and engagement of real data and real contexts has been stated and established for so long and so thoroughly by statisticians and statistics educators that it is almost superfluous to restate it. Over the last twenty-five years, there has also been increasing emphasis on the importance of using complex real data. Initially the emphasis was in terms of data with a number of variables, whether the data are to be collected by students or accessed, but alongside rapidly growing technological advances and their incorporation in teaching statistics, advocacy for complexity has broadened to encompass large datasets, many variables, datasets already collected that need wrangling or “treatment” before teaching, or non-traditional data. In parallel with emphasis on real data and contexts, there has also been advocacy on critiquing reports on data and data-based commentary and analysis in the media or in accounts in other disciplines. Real, complex data and real contexts provide opportunities for rich and authentic statistical learning, but are not without a range of challenges and the need for careful thought and sound expertise in statistics and its learning for those teaching statistics, especially at the introductory level. Often discussion of teacher planning and pedagogy focuses on the data and its nature, but challenges can also feature significantly in context, required curricula and externally-imposed constraints on student time and assessments. Contexts in particular require careful selection and thought, especially at foundation and introductory level. Contexts must be readily accessible to the relevant student cohort so that they provide a suitable vehicle for statistical learning. If a context requires more than basic understanding from students or if a context is too dominant, authentic statistical learning is inhibited by context learning or by non-transferability of learning. In designing learning experiences, learning purpose embeds content, pedagogical structure and external constraints, the last of which can be considerably restrictive at school, tertiary or workplace levels. Good context and data choice must therefore take account of the student cohort in regard to both prior and current learning, and discipline situation. All those who have been involved in choosing contexts and datasets know how much work is involved in preparation of them for student use, even for extra-curricular open-ended investigations without curricula and assessment restrictions. Preparing good classroom-ready learning resources within a given curriculum requires significant statistical and teaching expertise. Those involved in teaching statistics into other disciplines at tertiary level know the diplomacy and combined knowledge of students and statistics required to balance the desires and demands of other disciplines, as well as students' and institutional restrictions on time and assessments. Contexts with datasets suggested by serviced disciplines are too often unsuitable for their students due to overly-complicated or advanced discipline-based contexts with limited data learning potential. In addition, the approach desired by researchers in any discipline can be a top-down, casestudy type of approach rather than a student-driven investigative type of approach. However, despite all the challenges, contexts and data relevant to students' lives or their program of study are invaluable in their engagement and learning potential, provided the criteria above are met. Hence reports on resources, pedagogies, strategies and research therein that meet such criteria, provided the student cohort, teaching situation, curricula and learning experience circumstances are well-described, are of value for all those teaching statistics. In this issue, we have five papers reporting on such in different student and teaching situations, and a paper discussing an in-depth research study comparing effects of static and interactive visualisation. In [2], the contexts are a number of media reports on COVID-19, two of which made waves within the students' own country, and a third which attracted international attention. The contexts are carefully chosen to enable students to apply worry questions such as in [3] to develop statistical literacy in context. Anyone who has given students in introductory statistics a free hand in choosing a context to investigate, knows how sporting data can be as attractive to students as it can be difficult for them to understand the statistical challenges and pitfalls in the way of researching what they want, which is often associated with individual stars or successful teams. In [6], a scraped and organised dataset using the US National Basketball Association (NBA) Legends dataset, with R code provided for those who DOI: 10.1111/test.12338