Factors affecting vehicle occupancy measurement were examined with the aim of improving state and local vehicle occupancy monitoring programs. A comparison was conducted of five data sets looking at the effects on average vehicle occupancy (AVO) of time of day, day of week, road types, HOV lanes, locational differences, and traffic volume. It was found that AVO was higher in the afternoons, on Saturdays, and on HOV lanes. Inconsistent differences were found for the other variables, though there were considerable locational variations. Current vehicle occupancy monitoring programs do not generally take these factors into account and, most likely, produce biased estimates of AVO. Suggestions are made for using these factors in sample designs to represent regional and corridor-level vehicle occupancy levels.