Data from a questionnaire survey that is the principal source of information on individual characteristics are subject to sampling errors and non sampling biases. In aggregate data, sampling errors are usually small because a large sample is more easily obtainable and the data are also free from non sampling biases. This paper develops a data combination and updating method that corrects survey data for nonresponse biases and reduces sampling errors by statistically combining survey data with aggregate data. This method is applied to the estimation of an origin-destination (o-d) table stratified by market segments. The o-d matrix entries and parameters of a nonresponse bias model are estimated by the maximum likelihood estimation method. An application for an intercity rail corridor is presented. This paper appears in Transportation Research Record No. 1203, Demand Forecasting and Trip Generation-Route Choice Dynamics.
Samenvatting