Citation

Shumate RP, Crowther RF. Highw. Res. Board bull. 1961; 281: 87-96.

Copyright

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

DOI

unavailable

PMID

unavailable

Abstract

One of the most vexing problems in the study of speed arises from the lack of background information needed in developing necessary sampling methods. In the most common method, speed measurements are taken at a fixed point. This raises the question of whether or not samples taken at different hours of the day, days of the week, or months of the year may be used to infer changes in speed behavior brought about by an artificially induced variable. A study of vehicle speeds under normal conditions covering a period of six months suggests that a number of observations alone is insufficient as a measure of adequacy of sampling. Samples of vehicle speeds should be collected within fixed time intervals. The data for the study were collected in southern wisconsin on a typical section of the rural state and federal highway system. Specific tentative conclusions reached are: (1) hourly mean speeds show differences greater than chance would account for even after any possible effect produced by differences between days and months is eliminated, (2) the differences between monday, tuesday, thursday, and friday mean speeds are real and material, (3) the differences between monthly mean speeds are larger than can be accounted for by chance, (4) as sample sizes are increased without regard for the time interval involved, differences in the sample mean speeds provide estimates not only of the true changes in speed behavior but also changes in speed arising from differences in the hours, days, and months, (5) the quality of speed estimates can be improved by matching sampling periods by hour of day, day of week, and month of the year.