For the past four weeks I have been industriously running from work to attend a statistics course (Bivariate Statistics in R). Although my undergraduate degree was in Physics, my M.Ed focused on qualitative data. So, I had not applied any stats in social science, and was a bit uncertain. Hence, my attendance at this course.
Why did I choose R?
R is open source. This is useful for three reasons. Firstly, it means not having to pay for it, which makes students cheer. Secondly, it ensures the reproducibility of the analysis. Finally, open source means that lots of clever people around the world are free to make improvements and add packages to increase functionality.
Although in theory R requires more understanding of programming language (than, for example, SPSS), the syntax is much more consistent. This makes it much quicker to advance your understanding and skills beyond a basic level. Also, using a programming language means that if you come across a bug or something that you want to do that has no source code, you can write the code yourself rather than waiting for someone else.
The graphics in R are more sophisticated (and exportable as PDF’s). Again, you can control every aspect so that your graph looks exactly how you want, rather than how a company believes you should make it look.
Topics Covered
Within this course, we looked at correlation, chi-squared tests, t-tests and ANOVA. It was a really well delivered course and anyone wanting to learn about bivariate statistics, I suggest looks at the course materials (see link above).
Was it useful?
Very. Not just in terms of content (some of which I had encountered before, but not all) or in terms of familiarity with R, but this course really increased my confidence in the importance and application of statistics in the social science and my understanding of the reasoning behind the bewildering array of statistical tests.
What next?
This course was provided as part of a cross-faculty initiative funded by the ESRC. I imagine that many universities provide a similar range of courses for postgraduate students. Although you must weigh the loss of time from reading or writing for your thesis, it is worth checking these out and trying a few that seem relevant to you.
For myself, I plan to try further stats courses. However, I do anticipate that a lot of my data will be qualitative (such as interviews and observations). So, I’m going to have a go at an Atlas.Ti course and education specific courses looking at ethics, interviews etc. The only pity is that, as a part-time student, I struggle to make some of the interesting courses that are scheduled during the day.
Aisling Brown 