The Efficiency of Sampling Techniques for NTD Reporting

Abstract

This paper examines the minimum sample size required by each of six sampling techniques for estimating annual passenger miles traveled to meet the Federal Transit Administration's 95% confidence and 10% precision levels for the National Transit Database. It first describes these sampling techniques in non-technical terms and hypothesizes how they are expected to compare in their minimum sample sizes. It then determines the minimum sample size for 83 actual sample datasets that cover 6 modes and 65 transit agencies. Finally, it summarizes the results in minimum sample size to compare the relative efficiency of these sampling techniques. The potential for improved efficiency from using these sampling techniques is great, but the exact degree of improvement depends highly on individual agencies, modes, and services.