The magic of multiple regression
It is a commonly held belief that sociologists ‘don’t do stats’. Hence why the ESRC and the Nuffield Foundation have launched Q-Step which is a programme to promote better quantitative social science training in the UK. I deliberately chose to do advanced statistics as part of my PhD. I felt that there is this whole way of discovering things about the world that I only have a tiny bit of knowledge about and as a researcher I want to know about every type of methodology there is so that I can best be prepared to answer a specific research question in the most suitable way.
To supplement my class based learning, I have been doing an online course all about multiple regression. It is a very nicely written and quite fun to do. Module one covers the foundations, so frequencies, barcharts and what not (I finally understand Standard Deviation!). The second module is multiple regression. Although I haven’t quite finished module one yet as I am really trying to learn thoroughly rather than quickly, what I have learnt about multiple regression from my statistics class is that it is pure magic – you can actually predict the future from the multiple regression model. I could not believe my eyes!
The biggest statistical magician of them all, in my view, is Prof Hans Rosling, a Swedish medical doctor and statistician. Last night on BBC2 there was a lecture by him on population growth. What makes his lectures so interesting is of course the incredible graphics and humor, but I would say more importantly it is the way he brings the statistics to life with real examples. Perhaps because he was a medical doctor he is always aware of the people behind the numbers. He banished a lot of myths we have about population growth, poor people in the global south, and reassured us that the world isn’t as bad as we think it is by looking at how much healthier the world has become. Of course there are still one billion people in extreme poverty but that poverty can actually be abolished in a very short time period – if we want to.
Statistics can be very powerful in tackling inequalities and dispelling stereotypes and prejudices and I think are essential part of a sociologist’s toolkit.
(c) 2014 Annika Coughlin