The purpose of this study was twofold: (1) to examine the working practices followed in obtaining manual traffic counts on primary highways in Virginia, and (2) to evaluate the accuracy of the counts. The former was achieved by observing and interviewing personnel conducting traffic counts at 26 stations, statewide; for the latter, the results of manual counts were compared with data obtained by using a portable traffic recorder and a motion picture camera at 9 stations. The study concluded that (1) there is a need to better inform both traffic count personnel and their supervisors about why the counting is important and how it should be done; (2) often either too many or too few personnel are assigned to work at traffic count stations; (3) at some intersections on primary highways there are requirements for manually counting the traffic to and from the secondary roads, but the costs incurred for these are not substantially contributing towards accomplishing the objective of the program; (4) a few traffic count personnel were found to be working at sites where they were exposed to unusual traffic hazards; (5) many traffic count personnel have been routinely scheduled to work for long overtime periods at 1.5 times their normal hourly pay rates; (8) altogether, many working practices that were inefficient or contrary to prescribed procedures were observed at the 26 stations; and (7) an analysis of the accuracy of the traffic count data taken at 9 stations showed that the manual counts for 86% of the stations included unacceptable errors. Also, the results of the study indicated that the mathematical factors used to convert 12-hour manual counts into annual average daily traffic figures should be reevaluated and updated. It is recommended that the Traffic and Safety Division (1) develop and distribute an updated version of the pamphlet "Traffic Counting on Interstate, Arterial and Primary Routes"; (2) provide supervisory personnel with adequate training and instructions for spot-checking traffic counting activities; (3) confirm how many people should be assigned to work at each count station; (4) cancel the requirements for manually counting traffic on secondary roads where they intersect high volume primary roads; and (5) prohibit traffic count personnel from using work sites at which they would be exposed to unusual traffic hazards. Another recommendation was that resident engineers should avoid the routine scheduling of overtime work by dividing traffic counting assignments, among more hourly people. The findings also support the recommendation in another report that the Department consider changing the current traffic count program into one using traffic recorder counts supplemented with a limited statistical sample of manual counts.