Accuracy of Trip End Estimates from the Home Interview Survey (167-7)

This report presents the results of a study of the accuracy of home interview survey data in estimating zonal trip ends. The study is based on 100 percent survey data collected by the Texas Highway Department in three apparently homogeneous, adjacent zones in San Antonio. The general data analysis confirmed the homogeneity of the travel characteristics of the zones. A large number of repeated random samples were drawn at various sampling rates and the results used to verify the basic assumptions and general applicability of a set of theoretical relationships between sample size and the expected error of estimation. The analysis of disaggregate zonal data was directed toward the accuracy of home interview data in estimating the population mean (i.e., the mean trips per dwelling unit) and the population variance (i.e., the variance between dwelling units in trip productivity), The results indicate that, at both the 80 and 95 percent probability levels, disturbingly large error ranges (i.e., a large variance of estimates) may be expected when using traditional sampling rates in estimating the population mean and variance for a given zone. The analysis of aggregate zonal data was directed toward the accuracy of home interview data in estimating the zonal trip ends (i.e., the number of trips produced by the zone), The results, likewise, indicate that, at both the 80 and 95 percent probability levels, disturbingly large error ranges (i.e., a large variance of estimates) may be expected when using traditional sampling rates in estimating the zonal trip ends. The results of both the disaggregate and aggregate zonal analyses provide general guidance in zonal delineation and suggest new approaches to trip generation analysis .