Hosted jointly by the University of Glasgow School of Mathematics and Statistics, the Glasgow local group of the RSS and the Glasgow Science Festival, this event at the University of Glasgow on 15 June was very well attended, attracting more than 100 S5 and S6 students from several local schools. Two speakers, Liberty Vittert (University of Glasgow) and Jennifer Rogers (University of Oxford), each delivered talks that encouraged students to consider the use of statistical thinking across a range of real world situations.
Jennifer Rogers spoke about the broad themes of understanding risk and probability as seen through the lens of clinical trials. She explored the correct interpretation of the medical statistical results that are routinely published in the media and invited students to question how we can assess the efficacy of a treatment when very little is known about an individual’s ability to recover when treated with only a placebo. Sampling variation, sample size and the associated difficulties of identifying different effect sizes in different groups were illustrated using the long run average of a biased dice experiment.
The students were invited to decide whether and how a dice might be biased from frequency plots of observed rolls. Jennifer discussed the difference between relative and absolute risk, and their importance in interpreting the often misleading news headlines about the health consequences of consuming bacon compared to more damaging lifestyle factors such as smoking.
Liberty discussed the importance of critical thinking with regard to the meaning of probability statements in real world settings. The examples discussed ranged from the use of card counting to win poker games, to the use and abuse of probability in criminal trials. Liberty highlighted the importance of thinking about sampling, and developing a statistician’s mindset while discussing the example of Abraham’s fighter planes in WW2, in which students were invited to think about the parts of a returning damaged fighter plane they would repair conditional on the location of combat damage that is observed.
Liberty explained how statistical thinking can add clarity to the apparently counter intuitive abundance of double lottery winners by comparing the likelihood of a specific individual winning the lottery twice (which is astronomically small), with the likelihood of any lottery player winning the lottery twice (small, but not implausibly so).