Household classification techniques have been applied to the data from the 1962 and 1971 reading travel surveys to examine three sources of error in forecasting the numbers of trips generated by zones within the survey area. These sources were the calculation of household trip rates from the base-year data sample, estimation of the future proportions of different types of household within each zone and forecasting the survey area population. The empirical results were in close agreement with those obtained when a theoretical model of the errors was applied to reading. It was found that inaccurate population forecasts caused the greatest loss of accuracy followed by inaccurate estimation of the household proportions within each zone. The use of small base-year survey samples for calculating household trip rates had little effect on the accuracy of the zonal trip forecasts. This suggests that considerable savings can be made in the costs of travel surveys and the construction of base-year trip models. (Author/publisher).
Abstract