Edinburgh RSS local group
Meetings 2015 - Archive

Event in November:

"Statistics as a force for good - a data manifesto for a better society"
Tuesday, 10 November 2015 , 5:30 - 7:15pm
Speaker: Hetan Shah, Executive Director, Royal Statistical Society

Abstract: Hetan will give an overview of the work of the Royal Statistical Society, with a particular focus on its work to improve public policy and statistical literacy. This will include the Society's 'Data Manifesto', their research on the public's 'data trust deficit' and also on public misperception in statistics. He will talk about the campaign the RSS ran in the run up to the UK general election to have political candidates pledge to take statistical training from the RSS. More generally he will talk about the role of volunteers in the RSS, and the opportunities to get more involved. He will also come with open ears to have feedback from you about what more the RSS could be doing and how it could work better with you.

Event in October:

"A World Statistics Day - Sustainable Development Goals: the challenges facing the statistics community"
Tuesday, 20 October 2015 , 5:30 - 8pm


Glenn Everett, Director of the Measuring National Well-being Programme at the ONS
Glenn Everett will provide an overview of the new sustainable development goals. He will describe the work that the UN Statistical Commission has initiated, to develop a set of global indicators that will be used to report on progress.

Neil Jackson, Chief Statistician at the Department for International Development (DFID)
Neil Jackson will consider the challenges that national statistical systems will face in reporting on the new global indicators. He will discuss the actions that are being taken by the international community to build statistical capacity, and the particular role that DFID has been playing.

Jenny Church, Chair of the RSS' International Development Working Group
Jenny Church will launch a policy statement by the RSS, calling for a data revolution that recognises the importance of data for policy making and for accountability in all countries of the world. And puts useful and useable information into the hands of everyone who needs it, especially the poor and the marginalised.

Event in September:

"Statistical Issues in the Calculation of Greenhouse Gas Emissions"
Tuesday, 29 September 2015 , 5:45 - 7:15pm
A joint meeting with the RSS Environmental Statistics Section.


Martin McFie (Senior Assistant Statistician, Scottish Government)
TITLE: An Overview of Greenhouse Gas Emissions in Scotland
This talk will provide an overview of results from the Scottish Greenhouse Gas Inventory for 2013. The inventory is the key tool for understanding the origins and magnitudes of Scotland's anthropogenic greenhouse gas emissions and is published as part of the Official Statistics release "Scottish Greenhouse Gas Emissions". Martin will outline the scope of the inventory and methods used in its compilation. He will provide a brief description of improvements, revisions and uncertainties.

Prof. Ute Skiba (Centre for Ecology and Hydrology)
TITLE: Improving the Agricultural Greenhouse Gas Emission Inventory for the UK, Focussing on Soil Nitrous Oxide Emissions
Agriculture is responsible for 81% and 44% of the total UK anthropogenic annual nitrous oxide and methane emissions, with nitrogen fertiliser/manure application and livestock rearing being the dominant sources. Emissions are currently calculated using very general, non-specific emission factors, but will soon be replaced with UK specific emission factors (funded through the DEFRA GHG Platform projects). The underlying research to derive these new emission factors will be presented.

Vera Eory (SRUC: Scotland's Rural College)
TITLE: That is the assumption we've made.
Information on the uncertainty of scientific evidence used in public decision making is essential for designing robust policies. However, this information is often not available in relation to the economics of greenhouse gas (GHG) mitigation in agriculture. This talk presents a work which looked at this uncertainty in relation to agricultural soils in Scotland. The results show that the uncertainty of the GHG mitigation potential can be high, however, the ranking of the measures is relatively robust.

Event in June:

"Looking to the Future of the Past: a New Era in the Analysis of Historical Census Data?"
Monday 1 June 2015 , 6 - 7pm

Speaker Dr. Eilidh Garrett:
Affiliation: Department of History, University of Essex

In Britain researchers using historical census data have, until very recently, tended to be straightjacketed by the data published in the Census Reports or constrained by the time and resources necessary for delving into the enumerators books. New digital resources are now opening up exciting opportunities to develop research and analysis of nineteenth and early twentieth century censuses in ways which could not have been contemplated previously. This presentation will consider how the censuses have contributed to past, present and possible future approaches to the study of the Fertility Transition using examples drawn from Scotland, England and Wales and Northern Ireland. The discussion will then be widened to consider other avenues of research which may benefit from the 'new' resource provided by the individual level data for complete populations spanning the decades between 1851 and 1911.

Event in May:

"What is driving changes in crime in Scotland?"
Tuesday 12 May 2015 , 6 - 7pm

Speaker: Professor Susan McVie
Affiliation: School of Law, University of Edinburgh

This presentation will draw on the early results of a large programme of research on crime and victimisation in Scotland being developed by the Applied Quantitative Methods Network (AQMeN). Using a range of existing data sources, the study has examined change in crime over time at the national, the regional and the local levels. Study findings reveal some interesting disparities in patterns of crime at each of these three levels and cast doubt on the notion of a global crime drop. Potential causal factors which correlate with periods of declining crime show time-varying tendencies, and there are divergent spatial patterns in the data that indicate different experiences across Scottish communities, which may explain why public perceptions of crime do not always match reality. This presentation will address the thorny question of what is driving changes in crime in Scotland and look to provide some answers against a backdrop of major methodological challenges.

Event in January:

Thursday 29 January 2015 , 6 - 7:45pm

A joint event with the University of Edinburgh Business School, and PRMIA (the Professional Risk Managers International Association) on "Quantitative Approaches to Financial Risks":

Venue: The University of Edinburgh Business School

Speaker one: Juthamon Sithipolvanichgul
Affiliation: University of Edinburgh Business School
Title: Scoring Method of Enterprise Risk Management

The main objective of this paper is to propose Enterprise Risk Management (ERM) scoring method from integrating well-implemented ERM components where contribution scoring can be standardized by constructing a sample of firms from all public listed firms in Thailand and classified level of ERM adoption. Stepwise regression model is used to find relationship of ERM implementation level and Tobin's Q. There are positive significant different between ERM and firm value. Finally, not only financial companies will advantage implementation of ERM but also non-financial companies as well.

Speaker two: Dr Douglas McLean
Affiliation: Insurance-ERS, Moody's Analytics, Edinburgh
Title: Challenges in Asset and Liability Proxy Modelling in Insurance

The Least Squares Monte-Carlo technique is used in the Life Assurance sector as a means to avoid the computationally intensive nested stochastic simulations necessary in one and multi-year value at risk modelling. In the one year setting, an insurer's liabilities are forecast forward one year and then, typically, a general linear model is fitted to the outturned data which is returned as part of a noisy market consistent simulation. Often this data can be nonlinear and heteroscedastic violating general linear modelling assumptions. We outline methods to detect such problems using a linear and nonlinear regression techniques.