Files
Abstract
This paper examines the accuracy of data in annual unlinked passenger trips reported to the
National Transit Database (NTD) at the individual agency level. This examination takes a twostep
approach. The first step compares the ridership reported by member agencies to the
American Public Transportation Association (APTA) and the ridership reported to the NTD as
recipients of transit formula grants. The NTD ridership can be as high as 50 percent more than
the APTA ridership, and such significant positive deviation exists persistently over time across
many agencies. The second step explores potential sources of these positive deviations by
examining their components. Random errors, including both sampling errors and some of the
non-sampling errors, do not help explain these one-sided deviations. Nor do occasional annual
adjustments such as special events ridership to a direct count in the NTD ridership. Much of
these positive deviations appear to be attributable to systematic non-sampling errors that result
from undercounting in direct counts, from unintentional biases in procedures, or perhaps from
intentional manipulation. Limited evidence in the literature, however, suggests that
undercounting in direct counts is small at the systemwide level. The paper then quantitatively
examines how these systematic non-sampling errors affect the allocation of two formula grants
to Florida transit agencies: the Urbanized Area Formula Grant Program at the federal level and
Florida’s Transit Block Grant Program at the state level. The paper also discusses a strategy for
reducing these systematic non-sampling errors.