Random effects modelling – advanced issues (Online)

Random effects models are applied in a range of social science domains (e.g. education, health and economics). Across disciplines, however, they are often used for different purposes, with different specifications, or even with different terminologies.

These differences may well reflect genuine complexities and ambiguities that are associated with their implementation. This two-day course will focus on selected advanced issues in the application of random effects models in social research contexts. It is most suited to empirical social science researchers with some previous experience in using statistical models with random effects.

The participants will be invited to use their own computers in the virtual lab exercises. Participants should have at least one of the software packages Stata, SPSS and R installed on their computers in order to participate in the virtual labs. Example materials will be available in all three packages, although the largest volume of examples are available in Stata format. 
For webinar sessions: 

Required: Some previous knowledge and experience of using statistical models in the social sciences
Desirable: Some previous experience of implementing random effects models in applied research

Software requirements for participation in the virtual lab sessions: 

Required: Access to at least one software from Stata, SPSS or R. Use of a secure computer that will support downloading and storing data files. Previous experience of using ‘command syntax’ code in statistical software to analyse datasets.
Desirable: Access to more than one software from Stata, SPSS and R. Access to MLwiN. 

Course will be delivered online and course times are 10:00 – 16:00 each day 

Course Description

Participants will be supported in order to

• Secure a rigorous understanding the terminology and features associated with random effects models
• Understand different perspectives in important debates concerning the use of random effects
• Enhance practical skills in applying random effects models to statistical datasets

Target Audience: 

The course is best suited to postgraduate students and postgraduate/postdoctoral researchers who have some previous experience of working with quantitative datasets and using statistical models to analyse them. Participants might be drawn from any career stage. 

The course is open to researchers from any sector, although all research examples will be non-commercial in nature (e.g. as typically undertaken in academic, public and third sector research organisations). 

Examples used within the course are often based upon academic sociological research traditions, typically using relatively large-scale social survey data resources. Such scenarios are likely to be most recognisable to participants with backgrounds in quantitative research in sociology, criminology, education, public health, political science, economics, and social geography. The course seeks to make connections between social science disciplines, and might be helpful to participants from any disciplinary background.