Introduction | Programme | Themes | Keynote Speakers | Location
We are constantly driven to increase efficiency, lower costs, reduce respondent burden, make better use of data already collected, handle more complex systems, and reconcile apparently inconsistent results from different sources. What are survey professionals doing to answer these challenges and maximise data value?
The conference aimed to explore survey research methods in the area of data integration: making the most of existing data and metadata by using them as a platform to aid further research, using them for deeper secondary analysis, and combining multiple sources of data.
ASC two-day conferences aim to have more depth and be more focussed than our one-day meetings or the larger ASC international conferences. They are residential, giving much greater opportunity for interaction and discussion between participants and presenters. This conference followed the pattern of the successful conferences held at Chilworth in 1998 and Latimer in 2001. As on those occasions we joined forces with The Office for National Statistics, The Market Research Society and The Royal Statistical Society in building the programme.
The conference consisted of four half-day sessions, each with a specific theme, an invited keynote speaker and three contributed papers (no parallel sessions). The proceedings are prepared in advance and distributed to participants at the conference. The session topics (with the keynote papers listed below) are:
Process Integration
Methodology and software for complex models
Models for data, metadata and knowledge
Multi-mode and multi-source surveys
Click here to view the full programme of Keynote and Contributed papers, with abstracts and presentations.
The conference started on the evening of Wednesday 14th September, with registration from 4pm and the conference Reception and Dinner later in the evening. The scientific part of the programme was from 9:30 am to 5pm on the following two days.
We have four broad themes under the integration banner which the selected papers address, though some papers may include more than one of these themes.
Efficiency and cost-effectiveness: Surveys can be expensive, and respondent co-operation is declining. Panels are becoming predominant in Internet research, in order to secure a pool of willing respondents. Are we making the most of existing data to respond to these pressures? Can we use previous information about the same population to improve the design of our sampling procedures? Can we use previous information about respondents (whether regular returns from businesses, or panels of respondents, or information from other (administrative?) sources) to reduce the respondent effort for the next round?
Consistency: How do we manage and/or reconcile differing answers to the same questions from different data sources? Should we try to establish which might be the correct (or best) answer, or is it better to recognise the differences and combine the information from all sources through a statistical model? And how do we communicate results in models to stakeholders who only believe in ‘real data’?
Modelling complex systems: How do we obtain good information about large, complex and evolving systems, too complex to be surveyed on a single occasion, such as a city transport system? How do we combine observations on parts of the system (e.g. underground, bus, train, car, bicycle) on different occasions (hours of the day, days of the week, seasons) and by different modes (vehicle counts, passenger counts, on-mode surveys, passenger diaries) so as to obtain a coherent view of the whole? And how (again) do we convince the consumers of this view (engineers, policy makers) of the validity of our answers?
Making existing data re-usable: In order to use existing data, we need to be able to find it. How can existing data be made accessible to new users, and how can they tell whether the data is really appropriate to their needs? What are the metadata needs for reuse of data (structural, functional and semantic) – how do we know whether definitions and coding are consistent, and how do we cope when they are not? Are there special analytical tools needed to work with single or multiple sets of pre-existing data?
Session |
Speaker and Topic |
Office for National Statistics |
Process IntegrationKeynote: Karen Dunnell, National Statistician, ONS Towards a single continuous population survey for the UK The paper discusses ONS plans to redesign its existing continuous household surveys (GHS, EFS, LFS, Omnibus) into a single module-based survey. It covers: rationale, methodology, efficiency and statistical benefits. |
Royal Statistical Society |
Methodology and software for complex modelsKeynote: Nicky Best, Imperial College Modelling
complexity in health and social sciences: Researchers in substantive fields such as social, behavioural and health sciences face some common problems when attempting to construct and estimate realistic models for phenomena of interest. The available data tend to be observational rather than collected via carefully controlled experimentation, and are typically fraught with missing values, unmeasured confounders, selection biases and so on. These features often render the use of standard analyses misleading; instead a comprehensive set of inter-dependent sub-models are needed to model the data complexities and core processes that researchers want to understand. It is also invariably the case that a single dataset fails to provide all the necessary information, and many complex research questions require the combination of datasets from multiple sources. Bayesian graphical models provide a natural framework for combining a series of local sub-models, informed by different data sources, into a coherent global analysis. This talk introduces the key ideas behind Bayesian inference and graphical models in this context and shows how they can be used to easily construct models of almost arbitrary complexity. The ideas are illustrated by applications involving the integration of survey data, census data and routinely collected health data. The use of the WinBUGS software for Bayesian modelling is illustrated. |
Association for Survey Computing |
Models for data, metadata and knowledgeKeynote: Andrew Westlake, Survey & Statistical Computing Combining Data and Knowledge in Models: We collect data in order to increase our knowledge, but we always have some knowledge before we start. Our existing knowledge raises the questions for which we need more information, and it also guides us in deciding what further data to collect and how to collect it. Models allow us to generalise from specific observed data to a wider situation. When we analyse data we (usually) update our knowledge. If we can find a formal representation for our knowledge, then a standard statistical technique provides a way to formalise the process of updating our knowledge. This can be the basis for the integration of multiple data sets that relate to different aspects of the same system. While of general importance, this approach is the only way of developing an integrated understanding of complex systems which are too extensive to observe with a single data set. But complex methodology is difficult to understand, so we must also address the issues of convincing users from the application domain that our models are appropriate and valid, and of making the results obtained from the methodology accessible. The talk addresses these issues and illustrates them with experiences from the Opus project (www.opus-project.org) which, amongst other things, is looking at the problems of simultaneously modelling all forms of passenger movement in London. |
Market Research Society |
Multi-mode and multi-source surveysKeynote: George Terhanian, President, HI Europe
|
LocationSet in landscaped gardens and woodland, the Newland Park Estate is an ideal place for a relaxed conference, with a residents’ bar, parking and satellite TV. Close to London, Windsor and the Chiltern Hills, it is the home of the Chalfont Campus of Buckinghamshire Chilterns University College. For more details of the site (and more pictures), click on the images. Click here for a map. |
  |
| Market Research Society | Richard Cornelius |
| Office for National Statistics | Tony Manners |
| Royal Statistical Society | Suzanne Evans, Antony Fielding, Paul Hewson |
| Association for Survey Computing | Randy Banks, Raz Khan (Chair), Tim Macer, Andrew Westlake |
| Social Research Association | Wendy Sykes |
| Back to: Top |
Page last updated on 26 January, 2006 |