British Irish Region IBS

 View Only

Report from "It's About Time" (incl. 2022 Presidential Address & AGM)

By Kirsty Hassall posted 02-15-2023 03:44


On 29 November 2022, just over twenty statisticians gathered at the Royal Statistical Society from all corners of Britain and Ireland for the first in-person meeting of our regional biometricians in nearly three years. The meeting began with the society AGM where, after receiving reports from the officers, the most notable item of business was an update of our bylaws to include student representation on the committee. This proposal was passed at the meeting, but also requires the agreement of the society at large. Members have hopefully already been contacted to cast their vote about the proposed changes.

The second component of the meeting was called "This 1s 1t!", and invited all those present to share in just "one slide one thing" that they had learned in the past three years. Topics ranged from audio annotation to arranging meetings, from broken sticks to BIR history, from currying to composition, from estimands to ethics, and from hybrid working to half-normals. Many lively discussions ensued over lunch! A subset of the group (pictured) also visited the grave of Rev. Thomas Bayes in the Bunhill Fields burial grounds. (Mark Kelson once quipped that the surrounding area is padlocked to prevent people "frequenting" Bayes' grave, but this $n$ of statisticians was sufficiently large to persuade a passing warden to produce the key.)

Daniel Farewell's presidential address followed lunch, and attempted to convince the participants of the merits of a measure-theoretic approach to statistical methodology. No mass conversion was apparent, but a healthy discussion about the meaning of terms such as "missing data" followed. (Since you ask, data -- meaning given -- by definition cannot be missing, and can be helpfully thought of as a sub-sigma algebra of a larger such algebra.)

A major highlight of the day was Odd Aalen's lecture on time-dependent mediators in survival analysis. The lucidity of his exposition of treatment splitting as a way of understanding causal mediation was matched only by his humility in acknowledging that some of his previous efforts to estimate mediated effects had been (in his words) "basically nonsense". Odd's talk made a big impression on the audience, and again led to an extended time of Q&A.

We are grateful to all those who made the time to participate in this event, and look forward to the next such occasion!