19 March 2004,Charles Wilson Building, University of Leicester

This one-day workshop was hosted by the School of Education, University of Leicester, and was chaired by Ray Crozier from RCBN. The workshop was led by Harvey Goldstein and Jon Rasbash from the Centre for Multilevel Modelling at the Institute of Education, University of London. The session aimed to present a ‘relatively non-technical’ introduction to multi-level modelling. It was intended to develop a conceptual understanding of the technique, illustrated with examples from educational and psychological research, and did not aim to provide hands-on experience of computer software (MlwiN) for conducting data analysis (participants were altered to a forthcoming workshop on this). The morning session gave an introduction to multilevel (hierarchical) data structures and their analysis and provided illustrations of the multilevel model to analysis of schools data. A copy of the presentation is available on the RCBN website. The afternoon session examined further examples, discussed strengths and limitations of the model, and considered future applications of the model.

There were 23 participants, of whom 15 returned event feedback forms. The workshop was generally very well received, with most returns indicating that the course objectives had been met and that participants’ personal objectives had also been met. A majority of participants (12/23) had stated that they wished to gain an overview of the technique and its logic, and afterwards most respondents reported that their conceptual understanding had been improved and that they could see the benefits of this analytic approach, including application to their current projects. The level of difficulty at which the presentation was pitched seemed appropriate, as most participants indicated they had little prior knowledge of multilevel modelling session, and this was assumed in the presentation. However, some expressed the view that the afternoon session might have been too difficult and introduced too much new material rather than examine further what had been covered in the morning. The workshop presenters received highly favourable ratings. Where reservations were expressed it was to request more opportunity for interaction and for the chance to apply the analysis to own projects (there had been little response to the invitation in course publicity for participants to bring along their own data sets although, of course, this would be difficult to do in advance of knowing something about the technique.

The venue was generally well-received although there were comments about parking problems.