Citation

Oppenlander JC, Bunte WF, Kadakia PL. Highw. Res. Board bull. 1961; 281: 68-86.

Copyright

(Copyright © 1961, National Research Council (U.S.A.), Highway Research Board)

DOI

unavailable

PMID

unavailable

Abstract

The minimum number of observations required for the estimation of various vehicular-speed percentiles is determined. Based on the assumption that spot speed data has an approximately normal distribution, an equation has been derived to estimate the required sample size in terms of the percentile, tolerance limit, desired confidence level, and standard deviation. The standard deviation of the speeds is the only variable of the sample size determination that is directly influenced by vehicular speeds. The other three factors are arbitrarily selected. To aid in the selection of these minimum sample size requirements, it was possible to determine relationships between the estimated population standard deviation of vehicular speeds and various factors peculiar to the study sites under consideration. For two- lane highways in rural areas, linear regression analysis established a significant correlation of standard deviation of the speeds with the average annual daily traffic (adt), the 30th highest hourly traffic volume (30th hv), and the possible capacity of the spot speed site for the combination of daytime and nighttime vehicular speed observations. If the standard deviation of the vehicular speeds at a given spot speed site on a two-lane highway is not known from previous speed studies, then the adt value of a given location is recommended for estimating the standard deviation of vehicular speeds at this site. This correlation was the most significant of those site factors studied, and the adt is usually known for most highway sections. Although the small number of speed sites did not permit an accurate evaluation of standard deviation and adt for four-and six- lane rural facilities, the averages of these values are given to permit a first approximation of the necessary sample size requirements. Under traffic conditions in rural areas, the standard deviation approaches its maximum value because most motorists are free to choose their desired speed. On this basis, the standard deviations are representative of low traffic volumes and are acceptable for either rural, intermediate-area, or urban highways. By using charts prepared for various percentiles, tolerance limits, and desired confidence levels, this standard deviation indicates the required minimum sample size for the desired statistical accuracy of the spot speed study.