1: Has Covid-19 been good or bad for perceptions of statistics?

Statisticians are increasingly taking centre stage, as data has dominated the headlines over the pandemic. I sense a positive shift in the perception of statistics and data science as interesting and important. More people are becoming data literate. Royal Statistical Society (RSS) fellows have played a vital role in everything from real-time forecasting of infections to raising awareness about diagnostic tests. We have conveyed how important it is that data underpinning government decisions are made public, so people can make informed choices.

2: What is the most often misunderstood aspect of statistics?

It would be good if statistics conjured in people’s minds the art of learning from data and caring about uncertainty.

One common mistake is not defining the reference group. This has implications for judging whether samples are representative of the relevant population and understanding how an association between two variables in a population can disappear, or even be reversed, when the population is divided into subgroups.

3: The RSS has led attempts for greater professional standards for data science. What could the impact be?

Demand for data scientists has rocketed, but there isn’t a professional framework to ensure those working with data are doing so ethically and to a high standard. The standards that the RSS is developing with other learned societies will ensure data scientists have the training and skills they need. This will help organisations make the most of new data-driven approaches, and build the public’s confidence that their data is being used ethically and analysed robustly.

4: Is the relationship between statistics and data science a comfortable one?

Statistical science is a foundational discipline critical for data science. Statistics has always been shaped by advances in science and technology, and is well equipped to draw inference from complex, high-dimensional models needed by application areas.

Computer scientists are well equipped to deal with high volumes of data, and speed and scalability of computations. Interesting challenges are arising from new types of data, such as networks, images and sensors. Data science is the field that brings all these strands together.

5: How important will analysis of big data sets be in the future of statistics?

There is a large appetite from business, industry and government for putting masses of accumulated data to some kind of use. These data sets are not only big, but heterogeneous, subject to unknown selection processes and missingness, multimodal and highly structured, with text, images and shapes. This generates interesting analysis challenges for statisticians, requiring multidisciplinary collaborations. Our role is to actively engage and promote the relevance of statistical thinking in different fields.

6: Do you see statistics playing a greater role in our response to health problems at a societal level in the future?

Statisticians have shown during the pandemic that they can work incredibly responsively – planning studies at pace and producing outputs in real time, while maintaining quality. We need to build upon these achievements, and ensure we have the infrastructure and resources in place to improve our statistical readiness to face any emergent health crisis.

7: How can statistics succeed over misinformation?

Misinformation is a massive issue of our times, which erodes trust and has deep consequences for society. When statistics have been used misleadingly, the Office for Statistics Regulation has held the government to account, but it needs more funding and power to act. The RSS has called for government and public bodies to increase transparency of data, which would allow for better scrutiny. We also need to help journalists improve their statistical literacy, so we can avoid misleading stories hitting the press.

THIS ARTICLE WAS FIRST PUBLISHED IN THE JANUARY 2022 ISSUE OF IMPACT.