The Reality Of Survey Results; An Urban Goods Movement Case Study

Abstract

Results from a commercial vehicle survey undertaken in Sydney, Australia show relative standard errors varying from 5% to 33%. When looking at similar survey data from around the world, very few explicitly state the accuracy of final results, which can be misleading to users. When errors are not quoted, many users assume the sampling error is so small it is not worthy of concern. However, in practical transport and traffic problems, there are very few situations where this is the case. In order to compare published results it is essential to know the sampling error so that conclusions can be drawn with confidence. This also enables scope for an evolving improvement in accuracy when variables with larger errors can be identified. Accuracy depends a lot on the survey methodology, response rate and rigour of the subsequent statistical analysis. And it also depends on the sample size and variability of the subject. This paper briefly describes the greater Sydney (Australia) commercial vehicle survey, it discusses the statistical analyses and highlights some important issues to consider in future surveys.