Royal Statistical Society Social Statistics Section
21 March 2006, 2.00pm,
at the Royal Statistical Society Society headquarters, 12 Errol Street,
London, EC1Y 8L
Survey Response Rates
A half-day meeting organised by the Social Statistics Section of the RSS.
It is generally accepted that response rates in Great Britain have been
declining over the last couple of decades. This has led to concerns that
non-response error may bias survey estimates. Speakers at this half-day
meeting will present the latest evidence on survey non-response, challenge
the defeatist view that the decline in response rates is irreversible and
explore the complex relationship between response rates and non-response
bias.
Modeling Household Nonresponse using the ONS Census-link study
Gabriele Durrant (University of Southampton)
Household refusal in several government household surveys is explored using
a multilevel modelling approach. The analysis is based on the ONS Census
link study, a rich source of data, including individual and household level
characteristics as well as information about interviewers and geographical
areas. Some initial findings are presented.
How to achieve high response rates: Hard work, balanced strategies and a
little bit of luck
Ineke Stoop (SCP, Netherlands)
Although high response rates in random sample surveys are increasingly
difficult to achieve, recent examples show that they are still possible. The
presentation will focus on the presumed key factors in enhancing response
rates and on the substantial differences in response rates between countries
and survey organisations that are not easily explained.
Non-response and bias
Susan Purdon (National Centre for Social Research)
Based on analysis of a range of NatCen surveys, Susan Purdon will present
empirical evidence on the link between non-response rates and bias. Is bias
constant across surveys of is it survey-specific? And what is the evidence
for additional bias as response rates fall?
The half-day event will be chaired by Gerry Nicolaas (National Centre for
Social Research) and Peter Lynn (ISER, University of Essex) will act as
discussant.
Registration and payment are required. Request a registration form from
[log in to unmask] Details of the meeting can be found at
www.rss.org.uk/diary. The fee is £15 for non RSS fellows, £10 for RSS
fellow, £5 for CStats N.B. The CStat rate applies to MIS, FIS and GradStat.
For a map and directions see http://www.rss.org.uk/about/direction.html.
The event ends with a drinks reception at 5:00 p.m. sponsored by The
National Centre for Social Research.
|