Abstract
Presented is a statistical approach for examining multiple years of travel time data to determine if any trends can be quantified reliably by using only 4 years of observed data. Empirical data for 80 routes covering more than 4025 km were collected in the summers of 1994, 1995, 1996, and 1997 and compiled into a database for public release in the fall of 1997. From this database, seven routes were selected then sampled over all 4 years. These seven routes are primary arterials of regional significance in northeastern Illinois. A multiple regression analysis was performed on the seven arterials to determine if there was a statistically significant difference in the observed mean route travel times over the 4-year period. The regression analysis indicated that only one of seven routes experienced a statistically significant change in directional route travel time between 1994 and 1997. Because only one of seven routes was significantly different, the analyst may merge multiple years of data for an individual route into a single data set to build a more robust database. Also, quantifiable change in travel time may be difficult to perceive year by year. This has significant implications for the design of a sampling strategy that needs to measure performance on approximately 28 980 directional route km of roadways. It would be better to sample fewer routes more intensely on a regular interval than to sample many routes lightly every year.
Get full access to this article
View all access options for this article.