Abstract:
|
In a Swedish vehicle speed survey, for a stratified multi-stage sample of road spots, data are collected by use of pneumatic tubes and a traffic analyzer. From registered pulses, vehicles are created and assigned speeds. Typically, some pulses are left over, resulting in an unknown number of missing vehicles. The survey goal is to estimate the average speed, defined as the ratio R of total vehicle mileage and total travel time. An undercount of vehicles is bound to bias the estimators of the totals, whereas the impact on the estimator of R is unclear.
We suggest dividing the traffic passing each spot into weighting classes. The main difficulty is to adjust incompletely observed flow. One proposal is to add vehicles imputed by the analyzer, another to use registration probability weighting. Models for the errors in the number of imputed vehicles, and in the estimated registration probabilities, are used for theoretical evaluations. Utilizing empirical data, the models are assessed and adjusted study variable values compared with the "truth." Our models seem to fit the data quite well, and both adjustment strategies are quite successful in removing bias due to missing vehicles.
|