This report provides guidance for transit agencies to estimate transit usage for reporting to the National Transit Database (NTD) when their counting procedure that is designed to perform full counts misses some trips. Transit usage refers to unlinked passenger trips (UPT), passenger miles traveled (PMT), and average passenger trip length (APTL) in terms of annual totals and annual average daily by schedule type for annual reporting and monthly total UPT for monthly reporting. The guidance is provided in two self-contained guidebooks for bus and rail, respectively. Bus service includes all four fixed-route bus modes defined in the NTD: motor bus (MB), commuter bus (CB), bus rapid transit (RB), and trolleybus (TB). Rail includes light rail (LR), streetcar rail (SR), and hybrid rail (YR). For both mode types, the guidance focuses on data from automatic passenger counters (APC).
The guidance details a methodology for determining transit usage for each mode type through stratified extrapolation of incomplete counts rather than intentional sampling with APCs. The guidance views the total transit usage determined from the methodologies as estimates rather than 100% counts. It also views each methodology as an alternative sampling technique. The guidance identifies the conditions under which transit agencies may estimate annual total transit usage with this methodology as a pre-certified alternative sampling technique. For example, agencies must pass an equivalence test by demonstrating that their APC data are statistically equivalent to paired manual data within ±7.5% at the 95% confidence level. The guidance provides detailed steps for agencies to conduct the equivalence test in an Excel environment. When agencies meet the identified conditions and follow the guidance, they may use Appendix A in each guidebook as the document of certification by a qualified statistician for the alternative sampling technique.
For the NTD program, the guidance fills a gap in current NTD guidance and will result in more accurate UPT and PMT data. For agencies, it prevents under-reporting and saves the need to hire a qualified statistician for certifying the methodology as an alternative sampling technique.