DETERMINATION OF SAMPLE SIZES FOR TRAVEL TIME STUDIES

Abstract

travel time studies have been carried out following guidelines included in the Institute of Transportation Engineers’ (ITE) manuals of traffic and transportation engineering studies. 1,2 Those guidelines define minimum sample sizes as a function of permitted errors in the estimate of the mean speed, average ranges in speed and specified levels of confidence. This methodology was developed originally using concepts adapted from quality-control theory where the lower and upper control limits for the mean are used to determine the need for additional sampling. 3 Two basic assumptions used in quality-control theory are that estimates of the population’s mean µ and standard deviation σ are reliable and that a simple direct relationship between σ and the average sample range can be established. For travel time studies, these assumptions imply that all travel time runs are made under the same traffic and environmental conditions and that a comprehensive set of initial runs is made to obtain reliable estimates for µ and σ. Traditional practice in industrial quality control suggests around 20 samples of four to five runs each for this purpose. 4 However, for most travel time studies, obtaining such a high number of initial samples is not practical. For this reason, it is usually accepted that one initial sample composed of four to five runs, 3 or even as low as two runs, 2 R is sufficient to characterize the route under consideration. Based upon the observed range, a decision can be made about the need for additional sampling. Unfortunately, as it is shown in this feature, using the ITE guidelines may result in lower-than-required sample sizes (or in lower-than-expected confidence levels). Because the ITE guidelines are so widely used, it would be of interest to determine the reason for the discrepancy and to provide a method to correct the situation. SAMPLE SIZE FORMULATIONS There are basically two types of formulations that are commonly used for estimating minimum sample sizes: (a) based on sample ranges; and (b) based on sample standard deviations. This section describes and compares these formulations.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 50 citations exist upstream.