February Meeting
*********************
Wednesday 22nd February at 5pm (with tea from 4.30pm) the Epidemiology Seminar Room, Centre for Health Sciences Research (CHSR), 4th Floor, Neuadd Meirionydd, University Hospital of Wales, Heath Park, Cardiff. Directions: Please email me for directions to the hospital ([log in to unmask]). The Centre for Health Sciences Research can be found as number 121 on the map. Parking is available in the long stay multi-storey car park. Map: http://www.cardiff.ac.uk/locations/maps/heathpark/index.html
Speaker: Nicky Best (Dept of Epidemiology & Public Health, Imperial College London)
The inquiring mind of the statistician - the Bristol and Shipman cases
Abstract
In this talk, I will discuss some of the statistical work I was involved in for the Bristol Royal Infirmary and Shipman Inquiries. In both cases, we used routine data sources to provide statistical evidence of unusual performance. This poses a number of statistical issues such as how to account for multiple sources of random and systematic variation in the data, and how to formally define 'unusual performance'. For the Shipman Inquiry, we also investigated the feasibility of establishing a prospective system for monitoring mortality rates in General Practice using routine data. This talk aims to present these issues and our findings in a (reasonably!) non-technical manner.
March Meeting
*******************
A joint meeting will be held between the Official Statistics Section and South Wales group of the RSS on Wednesday 29th March at 5pm in the Cinema, National Assembly for Wales (with tea from 4.30 in the foyer).
Speakers: Hasan Al-Madfai (Glamorgan University) and Ray Thomas (Open University)
Denominators and ILO Unemployment
Abstract
Unemployment, usually measured using the Labour Force /Annual Population Survey (LFS/APS) or claimant unemployment records, can be seen as a lagged proxy of economic output. But unemployment rates are used in formulating economic and social policies, for inter-country comparisons of economic activity and for societal planning. Therefore, it is vital that unemployment is measured consistently and robustly in order for these applications to be meaningful.
The conventional unemployment rate uses the economically active population as a denominator, while long-term unemployment rates use total unemployment as the denominator. Neither of these rates take into account inheritance of variability and so both misrepresent movements in unemployment. Conventional rates have fostered government policies based on denominator-induced dynamics that are contradicted both by labour market theory and labour market behaviour.
Conventional rates present long-term unemployment as a lagged variable. But time series and cross-sectional analysis of claimant unemployment provide evidence that long-term unemployment, often associated with structural and demand deficiency situations, is more responsive to the labour market conditions than short-term unemployment. Such analysis also shows that variability in unemployment levels largely reflects variation in entry into unemployment. These failures to acknowledge the influence of inheritance can be attributed to the failure of ILO criteria to recognize the concept of entry to unemployment. The maintenance of integrity in National Statistics requires that the LFS/APS questionnaire is modified so that it records entry to unemployment.
Rebecca Cannings-John
Statistician
Department of General Practice
Cardiff University
Centre for Health Sciences Research
School of Medicine
3rd Floor, Neuadd Meirionnydd
Heath Park
Cardiff
CF14 4XN
Tel: 029 2068 7150
Fax: 029 2068 7219
www.uwcm.ac.uk/study/medicine/general_practice
|