|
|
Multilevel and latent variable modelling of discrete choices and rankings
|
When
|
2.00pm on Friday 2nd December
|
|
Where
|
B617, Leverhulme Library, Columbia House
|
|
Presentations
|
|
|
Speaker
|
Anders Skrondal
|
|
From
|
LSE
|
|
Abstract
|
Two important examples of nominal data are discrete choices, such as choice of political party in elections, and rankings, such as the preference order of different parties. It is natural to formulate models for such responses in terms of the latent 'utility' or 'attractiveness' of the alternatives, giving rise to the well-known multinomial logit model. It is sometimes not recognised that this model can include both subject-specific covariates such as age and alternative-specific covariates such as the positions of the parties on the left-right political continuum. When the data have a multilevel structure, dependence among the observed responses from the same cluster (given the covariates) can be thought of as arising from residual correlations among the underlying utilities. It is useful to structure these correlations using latent variables varying at different levels of the hierarchical dataset. Three types of latent variables are used: random coefficients of either subject-specific of alternative-specific covariates and common factors. The methodology is applied to party choice and rankings from the 1987-1992 panel of the British Election Study. Three levels will be considered: elections, voters, and constituencies.
|
|
For further information
|
Ulla Jakobsen (Administrative Assistant) Ext. 6879
Department of Statistics, Columbia House
|
|
|