No abstract

Voting data typically comprise the marginal distributions of votes cast at two successive elections. The fact that these obtain separately for many voting precincts or areas enables one to estimate the actual transition probabilities for movements between the options available to the voter at each election. An aggregated compound multinomial model is proposed. This allows log-linear dependence on various covariates and specifies a simple and illuminating structure of random effects. The information in such aggregated data is compared with that which would obtain had all of the transitions been observed. The model is applicable to many other types of aggregated data and meets many of the difficulties inherent in "ecological regression." It is illustrated with an analysis of the British European election of 1984.

SUMMARY In ecological inference one uses data which are aggregated by areal units to investigate the behaviour of the individuals comprising those units. Aggregated data are readily available in many fields and within a wide variety of data structures. In the structures considered, the aggregate data are characterized by the absence of available data in the internal cells of a cross‐classification. The aim of the ecological methods is to estimate the expected frequencies of such internal cells, which may be conditional on chosen covariates. Four methods of ecological inference are reviewed and their properties and appropriateness considered. These methods are then applied to data for which the internal cells are known and their performances compared.

The context and problems of election night forecasting are described. The linear models proposed attempt to reduce parameterization by splitting constituencies into three types-special, "three-party" and "two-party". The analysis is based on a modified ridge regression approach. The election night February 1974 performanceof the method as implemented for the BBC is described. Some results concerning univariate and multivariate ridge regression are given.

No abstract