Researchers

Professor Gillian Raab, Napier University
Susan Purdon, National Centre for Social Research
Kathy Buckner, Napier University

Aims and objectives

The project has two broad aims:

• to develop a web-based resource which will take the survey analyst through exemplars of the use of different methods that are currently available for the analysis of complex surveys

• to hold a series of workshops aimed at researchers who analyse survey data. The participants will use the draft training material to carry out analyses of selected data sets. The workshops will be integral to the development of the web resource

The overall goal will be to spread good practice in survey design and analysis both within and beyond the academic sector.

How will the project help survey researchers?

Recent developments in survey methodology, such as the calculation of complex standard errors and multiple imputation methods for missing data, are now implemented in a variety of software packages.

The survey researcher needs to know:

• what methods are available
• what are their advantages and limitations
• how easy is it to implement them
• how do the results compare from different approaches and different software when used to analyse the same data

The applicants and the workshop participants will address these questions and the results will be presented as exemplars in the web resource.

The survey data on which the exemplars will be based will be chosen from recent UK survey data that is publicly available and will include surveys of the Scottish population.

Web training resources for survey analysis

The project is due to start in February 2004.

We have identified three themes:

• inferences and standard errors for complex designs
• dealing with missing data
• analysing longitudinal data

A two-day workshop will be held at Napier University (Edinburgh) on each of these themes.

At least two software approaches applied to each of two survey datasets will be used in each workshop.

Participants will be active survey researchers preferably with some experience of the selected datasets.

A maximum of ten participants will take part in each workshop. There will be no fee and expenses will be paid.