The multi-disciplinary nature of statistics is evident in the way it's being utilised in research, examples of which range from predicting election results and modelling the spread of infections to creating a football player index, creating 3D models of heritage sites and improving weather forecasting. And these are just some of the areas of research that statistics is playing a significant part in, according to an RSS analysis of the most recent Research Excellence Framework (REF).
Last year, the RSS conducted analysis into hundreds of ‘impact case studies’ submitted to the 2014 REF; which were introduced to illustrate the impact that academic research has had across the UK and internationally, beyond academia. The RSS analysis sought to find out how many of these case studies included statistics.
Olivia Varley-Winter, RSS policy and research manager, explains: 'The REF’s Mathematical Sciences panel had said that a substantial proportion of the outputs in statistics, probability and operational research were "world leading". The aim of our report is to discuss the part statistics played in these case studies. Using the online impact case studies database, we identified more than 300 cases where statistics was either a key research subject area or where the findings had been published in statistical journals.'
Many of the case studies identified by the analysis were produced from a range of university departments, not necessarily the mathematics and statistics departments. Additionally, nearly two-thirds of the 300+ case studies identified were submitted to REF panels other than Mathematical Sciences – eg Business and Management, Computer Science, Economics, Geography and Engineering.
The RSS analysis also includes summaries of eight separate case studies that between them, illustrate the breadth of research in which statistics plays a key role. Among these is research conducted by the University of Warwick which successfully predicted the outcome of the 2010 general election using data modelling and calibration of probability forecasts across constituencies. Another research project, at the Maxwell Institute, used Bayesian computational methods to fit epidemic models, which have since been applied to control diseases in Scotland and also in America to reduce the impact of Sudden Oak Death.
One of the more surprising applications of statistics was in a project by the Glasgow School of Art, which used data acquisition, processing and modelling techniques to create 3D images of world heritage sites such as the Sydney Opera House and Mount Rushmore.
Other case studies include research into ‘nowcasting’ in economic statistics, improved weather forecasting and the development of a football ‘player performance’ index to inform squad selection in the Premier League.
This project’s report and data tables are available on our website. Both complement a 2013 report on the future of academic statistics in the UK, produced by our then-VP for Academic Affairs, Kevin McConway.