Forthcoming and Past Meetings

Below are details of our forthcoming and past meetings. Slides from the presentations, where available, are at the bottom of this page.

Forthcoming Meetings


24th October 2018 -- Predicting volcanic super-eruptions (Postponed from March 2018)

Speaker Prof. Jonathan Rougier (University of Bristol)

Title "Apocalyptic volcanic super eruption that could DESTROY civilisation is much closer than we thought, say experts."

Abstract In a recent paper we re-estimated the global magnitude-frequency relationship of large explosive volcanic eruptions (https://doi.org/10.1016/j.epsl.2017.11.015). I will describe how we addressed the twin problems of mis- and under-recording in the historical record, and how we quantified our uncertainties. I will also discuss our experience of the paper catching the attention of the mainstream media (the title is from the Daily Mail) and, if time permits, some of the wider implications.

Venue George Davies Centre, University of Leicester.

Expected timing 1230 talk.


Past Meetings

18th May 2018 -- A Gentle Introduction to Defined Benefits and Defined Contributions Pension Schemes.

Speaker Prof. Jane Hutton (Warwick University).

Title A gentle introduction to Defined Benefits and Defined Contributions pension schemes.

Abstract Pensions have captured the headlines recently. Large estimated deficits in defined benefit schemes such as Carillion raise questions about the balance of claims of shareholders and future pensioners, and can even contribute to closure of a company, as with BHS. 

This lecture will briefly explain and contrast defined benefit (DB) and defined contribution (DC) schemes. I will then discuss a thorough Canadian review of what groups, such as workers, taxpayers and wider society might gain or lose from different approaches to provision for retirement. If time permits, there will be graphs.

Venue University of Nottingham.


23rd November 2017 -- Sir David Spiegelhalter (joint local group meeting and Uni of Nott'm School of Mathematical Sciences colloquium) & AGM.

Speaker Prof. Sir David Spiegelhalter (RSS president 2017-2018; Winton Professor for the Public Understanding of Risk, University of Cambridge).

Title Does a university education increase the risk of a brain tumour? The ups and (many) downs of trying to be a 'public statistician'.

Abstract Statisticians can have special insights into naughty numbers and shabby statistics in the news, but have not tended to have much of a public role. My job is to try and improve the way that stats and risk are discussed in society, and I will relate both positive and negative experiences from radio, TV, print and online, selecting from  topics such as alcohol, sky-diving, sex, and the risks of burnt toast. You can share the panic at being asked unanswerable questions live on radio, and the joys of doing fairly well in Winter Wipeout, due to careful study of the statistics

Venue University of Nottingham.



12th October 2017 -- An introduction to modelling soccer

Speaker: Dr Rob Mastrodomenico (Global Sports Statistics)

TitleAn introduction to modelling soccer. 

Abstract: The increased data in sports has provided statisticians with the tools to build models of sporting events. This talk will look to show how data from soccer matches can be used to create models with the ability to predict upcoming matches. Starting from a very simple approach we will show how a modified poisson approach is able to characterise the dynamics of the beautiful game. Following from the model definition we fit it on data from England and show how the output can be used in predicting games from the English Premier League.

Venue University of Nottingham; Physics Building, room C27.


23rd May 2017 -- Use of historical data to supplement a future study: opportunities and challenges

Speaker Nicholas Galwey (GlaxoSmithKline, UK).

Title Use of historical data to supplement a future study: opportunities and challenges.

Abstract There are strong arguments, ethical, logistical and financial, for supplementing the evidence from a new clinical trial using data from previous trials with similar control treatments. There is a consensus that historical information should be down-weighted or discounted relative to information from the new trial, but the determination of the appropriate degree of discounting is a major difficulty. The degree of discounting can be represented by a bias parameter with specified variance, but a comparison between the historical and new data gives only a poor estimate of this variance. Hence, if no strong assumption is made concerning its value (i.e. if 'dynamic borrowing' is practiced), there may be little or no gain from using the historical data, in either frequentist terms (type I error rate and power) or Bayesian terms (posterior distribution of the treatment effect). It is therefore best to compare the consequences of a range of assumptions. This presentation will introduce a clear, simple graphical tool for doing so on the basis of the mean square error, and illustrates its use with historical data from clinical trials in amyotrophic lateral sclerosis. This approach makes it clear that different assumptions can lead to very different conclusions. External information can sometimes provide strong additional guidance, but different stakeholders may still make very different judgements concerning the appropriate degree of discounting.

Venue Capital One (very close to Nottingham railway station; google map).



28th March 2017 -- Forensics, Genealogy and Richard III.

Speaker Prof. Kevin Schürer (University of Leicester).

Title Richard III by numbers: How forensic statistics nailed the identity of the last Plantagenet.

Venue Lecture Theatre 2, Centre for Medicine, University of Leicester.

Abstract The story of how the body of Richard III was discovered in a Leicester car park went global in 2013. The discovery and subsequent identification of the skeleton of Richard III beyond any reasonable doubt, was a triumph of multidisciplinary teamwork.

What is less well known is that statistical methods used in forensics and many other disciplines were used by the Leicester Research Team to bring all this evidence together to declare these were the mortal remains of the 'last Plantagenet'.

In this talk Prof. Schürer will describe the fascinating step by step story of how the team used forensic statistical methods. These underpinned the multidisciplinary detective work to link Richard’s body with his present day relatives that led to the reinternment of England’s re-discovered lost monarch.




3rd November 2016 -- A pollster and a politician

Speakers Michael Turner (BMG Research) and Nick Palmer (Maths graduate and MP 1997-2010).

Talk titles
(MT) A painful pursuit of accuracy: challenges faced by an industry that seems determined to self-harm

(NP) Misleading opinion? How do politicians use polls?

Venue Mathematical Sciences Building, University of Nottingham.

Abstracts

(MT)
Dr Michael Turner outlines the obstacles faced by opinion pollsters with a frank examination of the numbers that have helped to shape social and political debate in Britain. Pulling no punches, the talk gives an insider's perspective of the five major issues that pollsters face: money, sampling, adjustments, regulation and unfortunately you, the general public.

(NP)
Leading and following opinion
Accurate predictions
     Age, class, salience, passion
     Was the General Election a sampling error?
     Panels vs face to face
     Large samples vs multiple samples
     Are trends predictive?
The naughty step
     Misleading the public: leading questions
     Misleading the interviewees: push polling
     Misleading yourself: selective interpretation
Why politics needs polls



5th May 2016 -- Statistical modelling in elite sport

Speaker Dr Ioannis Kosmidis (University College London).

Title Statistical modelling of availability-to-train in elite sports.

Venue Room A17, Mathematical Sciences Building, University of Nottingham.

Abstract Recent technological advances have allowed the frequent and simultaneous measurement at small cost of multiple physiological and psychological aspects of individuals and of their fitness and everyday activity. Such data are increasingly being used, for example by elite sport teams to quantify and monitor athlete fatigue and identify the variables that are important for explaining it, with the aim of preventing non-functional overreaching.

This talk will briefly discuss the challenges involved in recording and preparing fatigue-related data for statistical analysis, and present results from ongoing research on modelling availability-to-train in terms of other fatigue-related variables. Special care is being taken to use models that can account for both the heterogeneity between athletes and the natural time dependence between measurements.



3rd March 2016 -- School League Tables

Speaker Dr George Leckie (University of Bristol).

Title School league tables: the good, the bad and the ugly.

Venue Belvoir City Lounge, University of Leicester.

Abstract Since 1992, the Government has published so-called ‘school league tables’ summarizing the average GCSE attainment and progress made by pupils in each state-funded secondary school in England. In this talk, we statistically critique prominent past and current school performance measures. In particular, we question the Government's justifications for scrapping their contextual value-added measure, we show fundamental design flaws in the current expected progress measure, and we highlight various measurement and interpretation difficulties in reporting socioeconomic achievement gaps.

The slides of the talk are available at the bottom of this page.


3rd December 2015 -- Big Data (and AGM)

Speaker Paul Robinson (Head of Advanced Analytics, Bank of England).

Title "Big Data" and central banks: opportunities and challenges.

Venue University of Nottingham.

Abstract  Spectacular increases in the volume of data coupled with IT and theoretical developments have led some to declare the birth of a new golden age. But some remain unconvinced. They argue that the excitement about the vast number of possible relationships has led to a neglect of proper statistical rigour, that we risk drowning in an ocean of spurious correlations. What does this mean for a policy making body such as the Bank of England that has the responsibility to make use of all available relevant information?


8th October 2015 -- Journal rankings

Speaker Prof David Firth (University of Warwick).

Venue Mathematical Sciences, University of Nottingham.

Title Journal rankings: Why (and how) should we care?

Abstract The status of an academic journal has for many years been indicated (to librarians and others) by its "impact factor".  The value of research, and even of individual researchers, is sometimes assessed (by university administrators and others) through the impact factors of journals in which work is published.  The impact-factor measure itself has seen strong criticism in the academic literature, but its use persists.  Statisticians have had relatively little involvement in the "bibliometrics" (or "scientometrics") field so far.  The work reported in this talk aims to highlight the importance of more principled approaches to journal ranking, with emphasis on what is measured and how accurately.

As a case study, and to help catch the minds of research statisticians who might contribute in this area, ranking of the main journals of Statistics is considered in some detail. The methods used are clearly more widely applicable.

This talk is based on the paper "Statistical Modelling of Citation Exchange Between Statistics Journals" (joint with C Varin and M Cattelan) which was discussed at an RSS Ordinary Meeting in May 2015.  The paper, along with data and R-code to facilitate replication or further development of the work, are available at http://warwick.ac.uk/dfirth.



5th May 2015 -- Have I got Statistics for You!

Speaker Andy Sutherland (Former Head of Profession for Statistics, Health and Social Care Information Centre)

Venue Ogden Lewis Seminar Suite, University of Leicester.

Title Have I got Statistics for You!

Abstract Government depends on statistics to take forward its work, and to make arguments and debate issues in public. The public depend on Government Statistics to inform themselves and to hold Government to account. In this presentation Andy reflects on the various methods and mechanisms which can be used for obtaining and presenting figures for use in political debate, how these are used and, sometimes, abused, and how to find the evidence which may underlie contentious or conflicting claims.


12th February 2015 -- Relaunch Event

Speaker Prof Peter Diggle (Medical School, Lancaster University; Institute of Infection and Global Health, University of Liverpool; President of the Royal Statistical Society)

Venue Mathematical Sciences, University of Nottingham.

Title Model-Based Geostatistics for Prevalence Mapping in Low-Resource Settings

Abstract In low-resource settings, prevalence mapping relies on empirical prevalence data from a finite, often spatially sparse, set of surveys of communities within the region of interest, possibly supplemented by remotely sensed images that can act as proxies for environmental risk factors. A standard geostatistical model for data of this kind is a generalized linear mixed model with logistic link, binomial error distribution and a Gaussian spatial process as a stochastic component of the linear predictor.

In this talk, I will first review statistical methods and software associated with this standard model, then consider several methodological extensions whose development has been motivated by the requirements of specific applications. These include: low-rank approximations for use with large data-sets; methods for combining randomised survey data with data from non-randomised, and therefore potentially biased, surveys; spatio-temporal extensions; spatially structured zero-in flation. Finally, I will also describe disease mapping applications that have arisen through collaboration with a range of African public health programmes.

Joint work with Emanuele Giorgi (Medical School, Lancaster University)
Ċ
RSS East Midlands Local Group,
May 12, 2015, 4:10 AM
ć
RSS East Midlands Local Group,
May 12, 2015, 4:10 AM
Ċ
RSS East Midlands Local Group,
Mar 17, 2016, 3:49 AM
Comments