Courses
Courses on Survey Methodology in Europe
MASTER´S PROGRAMS
Barcelona
Design, evaluation and analysis of questionnaires for survey research
Brussels/Leuven
Sampling design
Modeling measurement error in social survey
Unidimensional and multidimensional scaling
Duisburg
Grundkurs Survey Methodology (Fundamentals of Survey Methodology)
Fragebogenkonstruktion (Questionnaire Design)
Erhebungsmodi (Data Collection Methods)
Sampling
Essex
Survey Methods I: Design and Measurement
Survey Methods II: Sampling, Inference and Management
Florence
EUI workshop on Survey design and research
London
MSc Policy Analysis and Evaluation
MSc Social Research Methods
MPhil/PhD Social Research Methods
Madrid
Master in Social Research Methodology. Innovation and applications
Manchester
Survey Research
Advanced Survey Methods
Southampton
Introduction to Survey Research
Analysis of Complex Survey Data
Basic Survey Sampling
Further Survey Estimation Methods
Survey Data Collection
Further Sampling Methods
Utrecht
MA Methodology and Statistics of Behavioural and Social Sciences
SUMMER SCHOOLS
Lisboa
Ljubljana
Essex
London
SHORT COURSES
Barcelona
United Kingdom
MASTER´S PROGRAMS
Barcelona: Design, evaluation and analysis of questionnaires for survey research
Design, evaluation and analysis of questionnaires for survey research (I)
In this course we combine three topics: the evaluation and testing of causal hypotheses, the design and evaluation of measurement instruments, and the combination of the two using Structural Equation models. Survey research is the most commonly used data collection method in the social sciences. Like all data collection methods this procedure contains errors which will affect the results of the data analysis. Given this situation, in this course we will discuss the analysis of survey data taking into account measurement error.
Content:
First we will discuss the designs for testing causal hypotheses (experimental and non-experimental) and the connected statistical analyses. After that the course will concentrate on non-experimental research. In that part the transformation of verbal theories into testable propositions will be given a lot of attention as an essential part of non-experimental research.
After this general introduction of the SEM approach we will move on to the problem of
measurement errors. Most of the time observed variables contain measurement errors and systematic errors because they are not only affected by the variables they are supposed to measure but also by other variables. Structural equation modeling can also be used to design and evaluate measurement instruments. Therefore this will be the second part of the course.
The third part will concentrate on the evaluation of causal hypotheses taking into account measurement error. In order to do so, we introduce a general model for structural equations which is a combination of a simultaneous equations model and a measurement model. This general model allows the specification, estimation and testing of all known linear structural equation models.
However, because the models get rather complex with a structural part and a measurement part, we will discuss an alternative for the analysis. It is a two steps procedure where in the first step composite scores are estimated for all concepts including their quality and in the second step the analysis is done with a simple model but correcting for the biasing effect of the limited quality of the composite scores.
Lecturer: Willem Saris
Period: 1st trimester
Design, evaluation and analysis of questionnaires for survey research (II)
Some people may think that formulating survey questions is very simple because we all can make questions. Others may think that it is an Art which one can not learn. Our point of view is that there is now enough knowledge about the consequences of the choices which are made when questions are formulated that survey design can be a scientific task. Therefore an overview of these choices and the consequences of these choices will be given in this course. Because errors will always remain in the questionnaires, we will also discuss how one can cope with these errors in the analysis of survey data.
Content:
In the first part of the course we will suggest an approach that can help to formulate proper question for the concepts one would like to measure.
In the second part an inventory will be given of the characteristics of survey items which might play a role in determining the quality of a survey question. In this part the different choices which a survey researcher will make while designing a survey item will be made explicit in order to reflect even on habitual choices.
The third part starts with a discussion of criteria for data quality and then gives a summary of the existing knowledge about the effect of the different choices on the quality of the questions as far as this knowledge is used in the SQP program. Subsequently the program SQP is introduced which predicts the data quality on the basis of the characteristics of the components of survey items which have been found to be of influence on the data quality of survey items.
In the fourth part of the course we will demonstrate how this knowledge can enhance the analysis of survey data by developing complex measurement instruments, improving comparative research and correcting for measurement error in multivariate analysis.
Lecturer: Willem Saris
Period: 2nd trimester
Sampling design
Content:
The sub-module starts with a case study, presented in the form of a brainstorming session between instructor and students: The Belgian Health Interview Survey (BHIS, 1997, 2001, 2004, and 2008). Because it is an ongoing complex undertaking, encompassing all main sampling methods, it is extremely motivational. In the second part, the concepts developed in the case study are formalized, yet in a non-technical fashion. In the third part, formal and mathematically underpinned developments are given of the main survey sampling methods: simple random sampling, systematic sampling, ratio and regression estimation, stratified sampling, cluster sampling, multi-stage sampling, sampling with unequal probability and self-weighting. In the fourth part, the students are introduced to SAS and STATA analyses on the BHIS. Further software tools that allow for the handling of survey data are briefly reviewed. In the fifth part, special topics in incomplete data and hierarchical analysis (including mixed-model analyses to deal with clustering) are tackled.
Lecturer: Billiet J., Molenberghs G.
Period: 1st semester
Modeling measurement error in social survey
Content:
Question wording and context and the task performance of the interviewers may have considerable impact on responses to survey questions. The introductory part deals with the most common rules for formulating different kinds of questions in Surveys. The first part of the course discusses the most debated field experiments dealing with the wording and context of survey questions. Understanding social-psychological and cognitive processes can provide practical rules that can reduce measurement errors due to response effects. The idea that one may conceive some artefacts as useful data shall be discussed and illustrated by means of examples derived from substantive surveys. A special focus is on the modelling of measurement error in applications with several of response scales, and with (un)balanced scales (modelling acquiescence).
The second part presents an overview of the most important interviewer effects. Special attention is paid to the role-dependent interviewer characteristics and to the impact of specific training programs. Special techniques for reducing non-responce in both interviews and mail surveys are discussed.
Lecturer: Billiet J.
Period: 1st semester
Unidimensional and multidimensional scaling
Content:
Unidimensional Scaling aims to introduce students into the theory and practice of unidimensional scaling models, in particular, classical test theory (CTT) and item response theory (IRT). In CTT, the sum score of a person on a set of items is the main focus of interest. The concepts of reliability and validity of this score are discussed and both factor analysis and item analysis are presented as ways to examine the properties of the sum score and its constituent item scores. In IRT, not the individual’s sum score, but the individual’s item scores are the main focus of interest. Depending on the type of item responses (dichotomous or polytomous) and on the different types of predictors included (characteristics of persons, items, or a combination of persons and items) different IRT models emerge. Several of these models are discussed in more detail. Moreover, the proposed framework for IRT-models is related to the field of generalized (non)linear mixed models.
Multidimensional Scaling: Multidimensional scaling is an extensive family of models for the spatial or geometrical representation of data and for interpreting their underlying structure and relationships. Its variants allow for a wide variety of data-types at different levels of measurement and uses a range of composition models. Both least squares and maximum likelihood estimation methods have been proposed in the literature, with different assumptions and properties. In the course, the following topics will be covered:
- The basic (non-metric and metric) MDS model; monotonic regression for ordinal transformations; assessment of fit and diagnostics; interpretation of configurations (including external property fitting and clustering of data and in solutions).
- Maximum likelihood algorithms, including dimensionality tests and tests of the appropriateness of the model.
- Three-way (“individual Differences”) scaling.
Lecturer: Storms G.
Period: 2nd semester
Language: German
Grundkurs Survey Methodology (Fundamentals of Survey Methodology)
Content:
The course will give an introduction to the principles of survey methodology including the history of survey methodology, Sampling, inference, survey errors, methods of data collection and non-response.
Lecturer: Rainer Schnell
Period: 1st Semester (October-February)
Fragebogenkonstruktion (Questionnaire Design)
Content:
The course imparts knowledge on the psychology of survey response, developmental interviewing, question writing, question evaluation, pretesting, questionnaire ordering and technical support for the formatting and documentation process of questionnaires.
Lecturer: Rainer Schnell
Period: 2nd Semester (March-June)
Erhebungsmodi (Data Collection Methods)
Content:
This course covers alternative data collection techniques used in surveys, especially face-to-face, mail and telephone surveys and the consequences of the data collection method for sampling, questionnaire design and response effects. Recent developments in the area of websurveys and mixed-mode-surveys will also be treated.
Lecturer: Rainer Schnell
Period: 1st Semester (October-February, starting 2011)
Sampling
Content:
Based on the knowledge of the class “Fundamentals of Survey Methodology”, this course will cover the general theory and application of simple random samples, stratification, cluster- and multistage-sampling, design-effects, non-response adjustment and weighting.
Lecturer: Rainer Schnell
Period: 1st Semester (October-February, starting 2011)
Survey Methods I: Design and Measurement
Overview:
This module introduces students to the principles and practice of modern survey design. The module exposes students to the considerable literature on survey methodology that informs best practice in contemporary survey research. Survey methodology has, over the past two decades or so, developed into a more or less unified field of research and practice. It brings together insights from, inter alia, cognitive and social psychology and statistics to explain how human behaviour and survey design decisions interact to produce data of varying quality. Key to this perspective is the concept of 'total survey error’. This framework is used throughout the module to discuss the multiple sources of error that modern survey design methods aim to mitigate. The initial focus of this module is on introducing social science graduates to the fundamentals of survey design and to the concept of survey error. A variety of different types of design are introduced with their relative costs, benefits and indications for particular types of study purpose. The focus then moves to survey measurement. Considerable attention is paid to the theory and practice of designing effective questions for surveys along with emerging methods for testing them. The final topic introduces students to a variety of modes of data collection and the significance of survey mode on data quality. Throughout the module, concepts and methods will be illustrated with real examples and case studies – many of them drawn from the survey work that takes place at ISER.
Teaching is carried out with a combination of lectures and computer lab and seminar sessions. In these lab sessions and seminars, students will carry out a series of practical exercises as well as having the opportunity to discuss focused readings relevant to the lecture topics.
Lecturer: Nick Allum
Period: Autumn
Survey Methods II: Sampling, Inference and Management
Overview:
This module is intended to follow on directly from Survey Methodology I. Grounding discussion in the total survey error framework, the first part of the module deals with sampling error. It introduces students to the theory and practice of probability-based methods for survey sampling and the derivation of standard errors for complex sample designs. Problems arising from non-response are approached with a discussion of theoretical considerations and practical solutions for mitigating non-response errors. Strategies for reducing non-response errors that are covered include aspects of the survey process (incentives and fieldwork procedures) and in post-estimation (weighting and imputation). In the final part of the module, some of the general features of survey management are discussed, including data management, documentation and meta-data, ethical obligations, field-force and client relations.
Teaching is carried out with a combination of lectures and computer lab and seminar sessions. In these lab sessions and seminars, students will carry out a series of practical exercises as well as having the opportunity to discuss focused readings relevant to the lecture topics.
Lecturer: Nick Allum
Period: Spring
Introduction:
Research based on data obtained by ‘asking questions’, Modes of surveying and interviewing, The (social psychology of the) interview process, The necessity of anticipation in the research process
What to ask and how?:
Questions and questionnaires, Examples from actual studies
Selection and Fieldwork:
Selecting Respondents, Special populations: elites, Interviewer training, Fieldwork
Analyzing answers:
Coding and categorizing, Reliability and Validity, Missing data, Calibration,Weighting
Miscellaneous:
Longitudinal analyses using surveys, Planning and Organisation, Documentation and Archiving
London:
MSc Policy Analysis and Evaluation
MSc Social Research Methods
MPhil/PhD Social Research Methods
Language: Spanish
Survey Research
Teaching:
The course is taught over 10 weekly sessions and comprises lectures, practicals and workshops.
The course includes an actual survey thus giving students practical, hands on
experience of research in practice.
The course will be structured around the following headings: Introduction to social surveys; Sampling; Questionnaire design; Piloting; Fieldwork.
The course aims to:
Introduce students to the role of surveys in social research; Provide an introduction and practical experience of the key elements of conducting a survey – development of a research question, questionnaire design, sampling, fieldwork and data entry; Provide a practical learning forum for students to consolidate and further develop their academic knowledge about research methods.
Course Content:
The social survey is a research tool of fundamental importance to
government and social researchers. The course addresses a need for training in the
understanding of survey data and in aspects of survey design and data collection. It covers key generic and subject specific training needs specified in the ESRC’s postgraduate training guidelines. The course often includes presentations from external speakers from research organisations. In the past sessions have included speakers from local authorities and commercial companies such as MORI.
Assessment:
The assessment for this module is an essay (of not more than 3,000 words)
which should outline and discuss how you would set about conducting a survey to answer a specific research question of interest. You should include a short example question module designed to collect the appropriate information with which to address the specified research question which would form part of a larger questionnaire.
Lecturer:
Dr. Kingsley Purdam
Period:
Term 1
Timetable:
Thursdays 2.00pm – 4.00pm, Venue Mansfield Cooper 2.1
Advanced Survey Methods
Description:
This module will extend the students’ skills of conducting survey research by focussing on more advanced methodological aspects of surveys. It covers the most important features of design and analysis in complex surveys. We will discuss different sampling strategies and see how these impact on the analysis. Students will get an insight in further aspects of survey methodology such as interviewer effects and non-response. We will spend some time discussing methodological issues in a longitudinal context, such as clustering and attrition. A major focus of the course relates to how these methodological aspects affect the analysis. We will look into two different statistical approaches of dealing with all these features of complex surveys: i.e. the design and model-based approach. A substantial part of the course will consist of computer sessions whereby the techniques of handling complex surveys are practised. In this way, the students will gain experience of applying methods for handling clustering, stratification and non-response in survey data.
Aim:
This course provides an insight into the design and methodological issues of longitudinal and other complex surveys. It also introduces software and methods for handling complex survey data.
Objectives:
At the end of this module, students should be able to:
Know several methodological aspects of conducting a survey.
Assess the strengths and weaknesses of the design of secondary survey data.
Assess how aspects of survey design will impact on the analysis.
Use the stata software to analyse complex survey data.
Understand the difference between the model-based and design-based approach to handling complex survey designs.
Lecturer:
Leen Vandecasteele
Timetable:
Thursday 2 – 4, Venue: HBS Hanson room, HBS 2.2
Introduction to Survey Research
Lecturer: Dr Gareth James, ONS
Date: 11 Oct – 15 Oct, 2010
Analysis of Complex Survey Data
Lecturer: Prof Chris Skinner, Univ. of Southampton
Date: 15 Nov – 19 Nov, 2010
Basic Survey Sampling
Lecturer: Prof. Danny Pfeffermann, Univ. of Southampton
Date: 6 Dec – 10 Dec, 2010
Further Survey Estimation Methods
Lecturer: Dr James Brown, Univ. of Southampton
Date: 10 Jan – 14 Jan, 2011
Survey Data Collection
Lecturer: Ms Sue Daltan, ONS
Date: 31 Jan – 4 Feb, 2011
Further Sampling Methods
Lecturer: Prof Chris Skinner, Univ. of Southampton
Date: 14 Mar – 18 Mar, 2011
Advanced Survey Methodology
Modern society relies on the availability of reliable and valid data. One of the prime sources of data is the survey. This course offers an overview of state-of-the-art survey design, including web surveys.
Lecturer: Dr Edith de Leeuw, Prof Dr Joop Hox
Period: 1st semester
SUMMER SCHOOLS:
Lisboa: Information will be available after May 2011 at : www.atitudessociais.org or www.ics.ul.pt
Ljubljana: ECPR Summer School in Methods and Techniques: http://www.ecprnet.eu/summerschools/Ljubljana/default.asp
Essex: Essex Summer School on Social Science Data Analysis: http://www.essex.ac.uk/summerschool/
London: LSE Summer School in Survey Methods and Analysis: http://www2.lse.ac.uk/study/summerSchools/summerSchool/courses/Survey%20Methods.aspx
SHORT COURSES:
Barcelona: Short courses and seminars in survey methodology, Research and Expertise Centre for Survey Methodology (RECSM), Universitat Pompeu Fabra: New Developments in Survey Methodology
United Kingdom:
CASS short course programme in survey methods and data analysis: CASS short courses
The Survey Skills Programme runs workshops for career-young social researchers followed by hands-on practical placements within the survey: Survey Skills Programme
| 
