Data from the 1971 national population census are used to develop a mathematical relationship between the average number of cars per household for an area and the proportions of households with different numbers of cars within the area. The relationship may be interpreted to mean that a household is twice as likely to acquire a first car than to acquire a second car. Similar data from the 1962 and 1971 reading area travel surveys give a modified relationship which suggests some under-recording in the surveys of those cars supplied by employers which householders can use. The relationships are sufficiently invariant under changes of scale, from the old local authority areas to geographical county areas and from traffic zone to survey area, to permit the amalgamation of contiguous areas. (Author/publisher).
Abstract