The objective of this study was to develop a procedure to forecast highway traffic volumes through the use of growth factors representative of Kentucky highways and reflecting effects of important socioeconomic and demographic variables. A two-stage modeling process was used. In the first, linear regression models were used to relate average daily travel on Kentucky roadways to personal income, price of fuel, and total miles of streets and highways. In the second, cross-tabulation models were used to relate growth in volume at a specific site--expressed relative to the statewide ADT--to highway functional classification, rural/urban location, county population growth, SMSA/non-SMSA designation, and volume level. The growth model yields estimates not only of the most likely rate of growth at a particular site but also of the range experienced at similar sites statewide. In collecting historical volume data for model calibration, a file was developed representing approximately 15,000 sites on the state highway system at which counts had been taken during the 1963-86 base period. On average, there were about 5.8 counts for each of these sites. Data in the file is instantly accessible by microcomputer users through dBASE software. A piecewise linear regression routine is used to provide estimates for past years during which counts were not taken. The historical volume database has been designed for continuous updating as new counts are made. Similarly, the forecasting models have been designed for convenient recalibration at annual intervals.