Fourteen roadside locations were chosen to give a full range of truck operating conditions to include: low steady speed; low-speed acceleration; high steady speed; and grades with low speed. Data from similar locations were compared for various registered vehicle weight categories and multiple regression analysis was performed of dBA vs speed, registered weight, and grade. Results showed that for vehicles with more than three axles, registered weight capacity is of little significance, and speed is not important when less than 40 mph. Mean noise levels on grades were found to be significantly higher than those on level read but all grades above 2% gave comparable levels. Alternate methods are suggested for categorizing trucks for noise regulation purposes. Pros and cons of each method are considered. Similarly, methods for choosing regulatory maximum sound levels are put forward with comments on the implications of each.