On understanding variability in data: a study of graph interpretation in an advanced experimental biology laboratory

Data analysis is constitutive of the discovery sciences. Few studies in mathematics education, however, investigate how people deal with (statistical) variability and statistical variance in the data to be interpreted. And even fewer, if any, focus on the uncertainties with which scientists wrestle before they are confident in the data they produce. The purpose of this study is to exhibit the work of coping with variability in one advanced research laboratory, as exemplified in a typical data analysis session. The study shows that when the scientists are confronted with novel data, their understanding of the variability does not arise in straightforward fashion, and a lot of normally invisible (interactional) work is required to constitute understanding. Tentative conclusions are provided for the implication to mathematics education.