Models for Longitudinal and Hierarchical Data: An Introduction - 12 October 2004

This course was taught at the Office for National Statistics, London, by Jouni Kuha (LSE)

This course introduces regression models for continuous response variables applicable to problems where the data have a hierarchical (grouped) structure. Two common examples are longitudinal data where the response is measured several times for each respondent, and multilevel data where the basic units of analysis, which are often individual people, are grouped in higher-level units such as neighbourhoods, constituencies or schools. Standard regression models are not fully appropriate for hierarchical data because the grouping induces correlations between observations in the same group. This course describes models which account for these correlations and thus allow valid estimation of regression coefficients as well as of further substantively interesting quantities which can only be estimated from hierarchical data. The topics covered will include the definition, interpretation, estimation, inference and diagnostics for such models.

Because the methods build on ordinary least squares regression models, a good understanding of these is expected. The course will consist of 50% lectures and 50% practical exercises. All class demonstrations and practical exercises will be carried out using the statistical software programme Stata with the gllamm routines. The classes will be structured in such a way that prior familiarity with Stata will not be essential, although it will be useful. Participants who wish to learn more about Stata beforehand can find the software and links to several introductory guides at www.stata.com.

[top]

Instructors

Dr Jouni Kuha is Lecturer in Statistics and Research Methodology in the Methodology Institute at the London School of Economics. His research interests include model selection, measurement error, missclassification and missing data and social statistics.

[top]

Cost and Accommodation

The cost of attending the course will be:

  • £25 (UK Institution)
  • £37.50 (Non-UK ECPR Institution)
  • £50 (Non-UK, Non-ECPR Institution)

[top]

Registration

Registration for this course is closed. Please keep an eye on our website for further courses that may be of interest to you.

[top]