Start: June 8, 2009
End: June 10, 2009
Place: Southampton (GB)
Courses In Applied Social Surveys
Response rates in many surveys have been falling and the resulting damage to data quality may have serious consequences for data analyses. This course addresses the problem of unit-nonresponse in sample surveys which occurs when individuals fail to participate in a survey, for example because of refusal or non-contact. The course introduces the main ideas and approaches to analyse and to address the problem, either by reducing nonresponse at the data collection stage or by using statistical methods for compensating for the effects of nonresponse at the data analysis stage (in particular using weighting methods). (The course will not focus on item non-response which occurs when sample members take part in the sample but fail to answer some of the survey questions.) . The course will also discuss some more recent advances from an ESRC funded 3-year project on 'Hierarchical Analysis of Unit-Nonresponse' (RES-062-23-0458), which is currently conducted by the course presenters.
* To provide participants with an understanding of the types and sources of non-response in social surveys and their consequences for survey data analysis
* To familiarise participants with research on sources and patterns of non-response and on the impact of alternative strategies to reduce non-response and/or to compensate for non-response
* To instruct participants in methods for reducing non-response at the data collection stage
* To instruct participants in statistical methods for compensating for the effects of non-response at the data analysis stage
* To illustrate the most common methods on nonresponse analysis and adjustment on several case studies, including the illustration of more recent advances in the field.
Course Content: This course will include the following topics:
* Types of non-response
* Factors which affect non-response
* Specification of statistical models to investigate the problem of nonresponse; modelling approaches to non-response (including questions about selection of auxiliary variables etc)
* Methods for reducing non-response (prior to data collection)
* Weighting methods to compensate for unit non-response
* Other nonresponse correction techniques (e.g. Call-Back approach, basic question approach, Heckman approach)
* The course will also discuss two case studies in more detail: 1) Modelling nonresponse using UK Census and interviewer observation data, and 2.) Assessing non-response bias in the PISA school survey in England.
The course will have a strong practical emphasis, with regular workshop sessions to enable participants to work through examples. The course will use SPSS and some basic familiarity with the software would be an advantage. (For course participants new to the package an exercise sheet will be provided prior to the course to enable them to work through examples.)
(Please note: the course will not discuss imputation methods to compensate for item-nonresponse.)
Target Audience: The course is aimed at both researchers who design surveys and wish to reduce non-response and at researchers who analyse survey data and wish to compensate for the effects of non-response. Participants may be researchers in the social sciences or may work in government, survey agencies, official statistics or the private sector. The course is suitable for statisticians or for social scientists who have a basic familiarity with statistical ideas of estimation in surveys.
Pre-requisite: Participants are expected to have a basic knowledge of simple statistical methods (including confidence intervals, testing and estimation; ideally also chi-square test and logistic regression). The course will have a strong practical emphasis, with regular sessions on computers, using real survey data, to enable participants to work through examples. The computer workshops will be using SPSS and some familiarity with the software will be an advantage.
Please bring a calculator for the workshops as well as a USB memory stick in case you would like to save your computer workshop output.
Course Fee: £25 per day for UK-registered students. £50 per day for staff from UK academic institutions (including research centres), ESRC funded researchers and registered charity organizations. £195 per day for all other participants. The course fee includes course materials, lunches and morning and afternoon tea, but not accommodation and travel which is to be arranged by the participant.