Conferences
Scientific programme
Prior to the conference itself two short courses will take place on Monday (July 18th 2011) – see below.
The conference officially begins on Tuesday morning (July 19th 2011) at 9:00. The conference will end on Friday (July 22nd 2011) at 12:30.
For a conference overview see the time table (pdf).
Sessions and abstracts are available online here.
The day-by-day program (pdf) gives you an overview when each session is scheduled.
Keynote speakers
 |
Roger Tourangeau
University of Maryland, United States
The Bad and the Beautiful: Measurement Error in
Web Surveys
Tuesday, July 19th 2011, 9:30 – 10:30
|
This talk will examine what is good and what is not so
good about response error in web surveys. On the positive
side, it will discuss the web’s capacity to supply visual
information to respondents, its potential for reducing
cognitive burden relative to telephone or face-to-face
surveys, and its capacity for interacting with respondents
in real time. On the negative side, it will consider the
evidence for satisficing among web panel members and the
tendencies for respondents in web surveys to misinterpret
visual cues, to ignore potentially helpful definitions and
other clarifying information, and to give too much weight
to images. On the whole, web surveys can offer excellent
measurement properties, but, as with more traditional
survey modes, require careful questionnaire design.
|
 |
Chris Skinner
University of Southampton, United Kingdom
Analysis of Complex Survey Data
Thursday, July 21st 2011, 18:00 – 19:00
|
Methods which take account of complex
sampling schemes when analysing survey data have been
around for decades, but were not used widely in practice
before the extensive recent implementation of these
methods in statistical software. This presentation
will survey some of the history of this field, including
some of the more significant debates, e.g. whether or not
to use sampling weights in regression analysis. It will
give an overview of current approaches and challenges,
including methods which seek to combine the treatment of
complex sampling and nonresponse or measurement error.
|
Short courses
Two short courses will be offered to conference participants before the conference. The half-day courses are scheduled for Monday afternoon (July 18th 2011) from 14:00 to 17:30.
Course #1: Multi-level models for Survey Data
Dr Ian Brunton-Smith, University of Surrey, UK
This half day course provides an introduction to the use of multilevel models to explore clustered survey data. This will begin with a discussion of the basic theory of multilevel modelling, and the different types of data structure that it can be used to handle. Then we will introduce the linear random intercept and random slope models, demonstrating the types of information that these can offer researchers. Finally, various extensions to the standard multilevel model will be outlined, including models to deal with longitudinal data, and binary response models. Participants will also be given details of the range of software available to implement these models. Throughout, the emphasis will be on the substantive application of multilevel models, and the interpretation of key parameters.
Dr. Ian Brunton-Smith is a lecturer in quantitative methods and criminology at the University of Surrey. His research interests include survey methodology; criminal statistics; multilevel modelling; structural equation modelling; missing data problems; cross-national data analysis, and public opinion research. He has conducted a range of analyses using multilevel modelling, and teaches a number of short courses that explore the use of multilevel models.
Course #2: Survey nonresponse: reduction, bias and comparability
Prof. Jelke Bethlehem, Satistics Netherlands, and Dr. Ineke Stoop, the Netherlands Institute for Social Research/SCP
Nonresponse is a serious concern because it can affect survey results. This short course describes the problem and presents approaches to reduce nonresponse or to reduce the effects of nonresponse. Topics that will be covered are nonresponse analysis, the R-indicator as a measure of nonresponse bias, the selection and use of auxiliary variables, strategies to enhance response rates, the problem of unequal response rates across groups, collection and use of paradata, and adjustment techniques. The course will end with nonresponse in a cross-national survey, the European Social Survey, where dual aims (high quality and optimal comparability) may conflict. Examples of field strategies and response outcomes in different countries will be given, and of different methods to assess and adjust for nonresponse bias in the ESS.
The course should be interesting for substantive researchers in order to gain some knowledge about the possible impact of nonresponse on their data and for survey methodologists to get a concise overview of current developments in the area of nonresponse research.
Prof. Jelke Bethlehem is senior survey methodologist at Statistics Netherlands. He is also part-time professor in survey methodology at the University of Amsterdam. He has been involved for many years in nonresponse research.
Dr. Ineke Stoop is head of the Department of Data Services and IT, The Netherlands Institute for Social Research/SCP. She is also a member of the Central Coordinating Team of the European Social Survey. Her research interests are comparative social surveys and nonresponse.
| 
