This mathematician hopes to change the way we use statistics
PhD student Alexander Ly (34, mathematics) has spent more than four years researching an alternative statistical method for conducting empirical studies called Bayesian statistics. ‘We have to stop thinking in a binary manner.’
Tell us what your research is about?
‘It’s about looking at using the Bayesian method as an alternative to the commonly used statistical method. Many researchers use a standard method when conducting empirical research. For example: when you use the standard method to test whether a medicine works, the results will be binary. They will be significant or not, so you can state that the drug has an effect, or not.’
What’s wrong with this standard method?
‘It causes problems. Suppose a hundred people test a medicine: it helps a little for person A but doesn't work for person B. Everyone’s different, but the standard method ignores this. The Bayesian method doesn’t give a ‘yes’ or a ‘no’ as an answer, but a number. For example: it’s seven times more likely that the drug works than that it doesn’t work. It’s a more transparent method.’
Why isn’t the Bayesian method used much?
‘It’s complex. I've now designed complex formulae and programmed them into a transparent and open-source computer programme, JASP. It’s also suitable for educational purposes: universities in the United States and France are already using it.’
It sounds like a turning point in science.
‘Many researchers still continue to cling to the old standard, which is less transparent and reliable. As a mathematician, I think this is crazy. Science is about looking for the truth.’
