Talking Statistics: A reflection on some of the problems with statistical language

Abstract

For most of my life I managed to swerve statistics. I learned a little at school and studied some (very theoretical) statistics as part of my mathematics degree. As a teacher, in schools and then university, I did not teach anything beyond GCSE statistics. It was not until I got a post in Mathematics and Statistics support a decade ago that I had to begin to learn the subject properly.I had excellent support from the sigma network, attending a memorable SPSS Bootcamp and other events which helped me enormously. But I was conscious that there were various aspects of statistics that presented problems for me. One was the way in which statistics differed from mathematics in being much less cut and dried. If a student had made a mistake in a calculation or argument it was fairly easy to spot and correct. However, when a student said ‘my supervisor said I should do a t-test’ and this did not seem the most appropriate way forward it was much harder to advise. I also realised that I was finding the language of statistics harder to master than I felt it ought to have been. It is this aspect that I will focus on in this article as I tentatively suggest that students might also have such problems.