 |
 |
Apologies for cross posting. Response rate as an indicator of sample quality: A first look at the results from the Census 2011 ‘household survey link study’ plus two papers quantifying the importance of response maximisation efforts to estimate quality. Location and time: 17 April 2013 17:30-19:00, Royal Statistical Society, 12 Errol Street, London EC1Y 8LX Andy Fallows (ONS): Making the most out of 2011 Census: ONS household survey linked datasets The 2011 Census offers a unique opportunity for ONS to look at the issue of non-response to our household surveys. It allows us to look at the characteristics of non-respondents by a number of demographic variables, as well as allowing us to develop models which help adjust for any non-response bias. With this paper Andy Fallows of ONS will describe some of the analysis that has been done on the linked Census 2011/Crime Survey for England and Wales dataset, how that compares to the analysis from 2001 and finally an overview of the multivariate analysis being done on other ONS household survey linked datasets including the Labour Force Survey. Peter Lynn (ISER, University of Essex): What is the impact of efforts to convert initial refusals and non-contacts into interviews? Survey researchers must strike a balance between the cost of resource-intensive measures to reduce non-response and the benefits in terms of improved estimates. In the context of face-to-face surveys the non-response reduction measures can be particularly expensive. In this presentation Peter Lynn of ISER at the University of Essex will investigate the impact on bias reduction of attempts to convert initial refusals and of attempts to contact hard-to-contact households. He will also explore the extent to which the impact persists in the presence of corrective weighting using data from three large national surveys. Finally, by replicating the methods of an earlier study, Peter will also be able to show whether the effects of extended interviewer efforts have changed over the past decade. Joel Williams (TNS BMRB): What is the risk we run if we obtain low response rates? We often hear that there is an additional risk of non-response bias if a survey response rate drops but that risk is rarely quantified. Consequently, it is not surprising that survey commissioners continue to regard maintenance of response rates as a critical determinant of quality. With this paper, Joel Williams of TNS BMRB will use simple paradata to explore the impact of response rate across a full range of data items in four major UK surveys. He will also relate these findings to the cost of achieving a high response rate so that survey commissioners can gain a better understanding of its value. Dr Patrick Sturgis, Director of the National Centre for Research Methods will act as discussant. Please register by emailing [log in to unmask] You may leave the list at any time by sending the command SIGNOFF allstat to [log in to unmask], leaving the subject line blank.