CCSR Seminars October - December 2003

Throughout the term, CCSR has a regular seminar series. Seminars are held from 4pm-5.30pm in room NG29 of the Dover Street building, All are welcome - no is booking necessary. Queries to 0161-275-4736

Copies of the poster for printing are available as PDF and Word files.

• Tuesday October 14th
  Quality issues in the 2001 Census
  Ludi Simpson, CCSR
  Slides (powerpoint)
  Did the strategy to reduce differential undercount succeed? What is the nature of error in the Census for small areas, including the new adjustment of small numbers such that no 1s or 2s appear in any output? What evidence can independently confirm the quality of the population estimates at national and sub-national level? A review of the evidence from fieldwork and from a study for the Local Government Association.
  
  
• Tuesday October 21st
  Geographic destiny vs. regional policy in the European Union
  Anca Carrington, University of Bradford
  How does location affect economic growth? Analyses using spatial regression analysis and distribution dynamics. How should one identify and understand outliers? The focus of the seminar is on the empirical investigation of the role of geographical location as determinant of economic growth. The main argument is centred on the relevance of locational fundamentals in explaining growth and convergence in the European Union. I will place the novelty of the proposed approach in the context of quantitative techniques used in the literature.
  
  
• Tuesday October 28th
  Triangulation in social research: pluralism in employment research
  Wendy Olsen, CCSR
  Slides (powerpoint), References (word),Paper on Additive Separability (word)
  In sociology the mixing of quantitative and qualitative methods presents an ongoing problem. Recent developments in the philosophy of science have emly argued that the two traditions should not have a separate-but-equal status, and that they should instead interact. I use detailed examples from the study of unemployment and of wage-rates to illustrate the pluralist position and its strengths and weaknesses. In this debate, the stakes are high because the ‘factual’ nature of social science itself is at risk.
  
  
• Tuesday November 4th
  The new Economic and Social Data Service
  Vanessa Higgins & Jo Wathan, CCSR
  The seminar will give an overview of ESDS - a new national data service available to academics and non-academics. We will focus on ESDS Government which is led by CCSR and promotes an increased and more effective use of the large-scale government survey datasets.
  
  
• Tuesday November 11th
  Ten things we didn't know a year ago about Scotland and religion
  David Voas, CCSR
  Religion is a new topic on the British census, and for Scotland in particular we now have unprecedented detail on the evolution and distribution of religious affiliation. A close analysis of the results gives us important evidence not only about society north of the border, but also about what influences religious change in any modern society.
  
  
• Tuesday November 18th
  Methods for investigating the causes of zoning effects
  Mark Tranmer, CCSR
  Slides (powerpoint)
  It is well known that the average population size of areas (the scale), and changes to area boundaries for a given population size (the zoning) will affect the results of statistical analyses that involve area level data. Some ideas for investigating the causes of zoning effects are presented.
  
  
• Tuesday November 25th
  The extent of cohort or generational change in employment for women from minority ethnic groups
  Joanne Lindley, CCSR
  Ethnic employment differentials are partly explained by demographic differences such as changing age profiles and family composition, as well as differing qualification levels. Partner’s characteristics, particularly country of birth and educational level, are also significant influences. Data from the Quarterly Labour Force Survey for 1992-1995 and 2000-2002 are used.
  
  
• Tuesday December 2nd
  Some simple options for the Bayesian data intruder
  Duncan Smith, CCSR
  I demonstrate that commonly used steps to perturb sensitive information in census tables do not always disguise the original table counts. I then demonstrate some simple Bayesian techniques that a data intruder could employ in order to try to discover those attributes that apply to individuals with ‘high probability’.
  
  
• Tuesday December 9th
  The impact of disclosure control on data quality
  Kingsley Purdam & Mark Elliot, CCSR
  What is the impact on analyses, after perturbing the Samples of Anonymised Records to simulate measures to reduce the risk of disclosing personal information? The statistical disclosure control software Argus will be explained. There are implications for other large scale survey data.
  
  
• Tuesday December 16th
  Does it matter what estimation method I use to provide small area populations at risk in SMR calculations?
  Paul Norman
  Slides (powerpoint)
  The calculation of Standardised Mortality Ratios (SMRs) requires numerators of mortality or morbidity events and denominator populations at risk for the geographical areas of interest. I will assess various strategies to estimate the populations at risk in small areas, and their impact on SMRs.