Urban Studies (1979),16, 165-177 A Longitudinal Study of Urban Populations

Summary. Data drawn mainly from census publications are used to compute trends in population, household, employ-ment and other characteristics for towns of 50,000 and over in England and Wales over the period 1951 to 1971. Com-ponent analysis reveals that the predominating factors illustrating the differentiation between towns are socio-economic status and household characteristics, especially overcrowding. Cluster analysis is used to define twelve classes of towns whose characteristics and trends are compared. This classification is shown to relate closely with a socio-economic group index. Both the component interpretations and the classification possess temporal continuity. More than 50 per cent of the population of England and Wales lives in relatively large urban authority areas, and there is available a considerable amount of information relating to them. From this material itis possible to construct a partial picture of measures, such as employment figures, birth rate ratios, and household composition, which can be said to characterise these urban populations. Simple arithmetic manipulation is sufficient to highlight the principal trends in these characteristics over time, but more sophisticated mathematical techniques will yield more complex relationships. Moser and Scott (1961) described the application of component analysis to data, drawn principally from the 1951 census, referring to towns in Great Britain. The components which were extracted enabled the most influential characteristics of the populations of the towns to be determined and made possible the grouping of populations with similar combinations of characteristics. That was a once-only study and there is considerable information to be derived from the repetition of a similar exercise at appropriate time intervals in order to investigate possible changes in the characteristics themselves and also the relative stability of the classification or grouping of the towns over the period. For instance, the effect of the New Towns on the 1951 classification would be of interest to planners Such an exercise has been completed and a full report, including a considerable amount of both the [First received May 1978; in final form January 1979] original and resulting data, has been publishe