Median Watch

Eyes on statistics

Working Weekends and Late Nights

Re-posted from the March 2020 issue of Advocate. The smell of pool chlorine reminds us of work. Both of us admit to doing emails while sitting at the local pool waiting for the kids to finish their lessons, as well as at cafés, on airplanes, waiting in queues. At least one of us does the odd email or two from the bathroom. We are academics and therefore work anywhere, anytime (maybe most of the time?

The COVID-19 crisis is amplifying shortcomings in health and medical research.

Re-posted from Campus Morning Mail. Poor quality medical research is nothing new. A major cause is the race to be published first, which means researchers do not adequately check their work. The COVID-19 crisis has put an already pressured system under even more strain, and the cracks are clearly visible. Small studies have been used to justify massive changes in clinical practice, such as the early results on Hydroxychloroquine, which have looked less promising as more work is published.

Playing the scientific record backwards.

I’ve done a lot of research on seasonal models. I’ve always enjoyed playing with the elegantly simple sinusoid function, which can fit a remarkable array of shapes with just a few parameters. Put enough sinusoids together and you can create any sound. When Fourier discovered this, it was considered too simple to be true. A great number of diseases are seasonal, so these models are useful as well as being mathematically pleasing.

It's fun to look at the Y A C M (Yet Another COVID Model)

Yet another COVID model. I did this modelling because I was asked to provide some COVID estimates for a hospital. There have been lots of models in the last few weeks and I don’t want to reduce the signal-to-noise ratio in this vitally important area, but I’m sharing this in case someone finds my approach useful. All the code is here. I have used similar models before to simulate disease numbers over time, for example my PhD student Dimity used microsimulation to examine the long-term effects of climate change (Stephen and Barnett 2017).

Mistakes, I’ve made a few, but then again, not too few to mention

Statistics can be hard Here’s a great quote about working with statistics from the fantastic statistician David Spiegelhalter: “I am often asked why people tend to find probability a difficult and unintuitive idea, and I reply that, after forty years researching and teaching in this area, I have finally concluded that it is because probability really is a difficult and unintuitive idea” (Spiegelhalter 2019). Someone who attended my one-day statistics refresher course expressed disappointment in their evaluation that after a day’s training they did not now feel like an expert.