Estimates of Total Household Trips for Areas with Zero Sample: a Maryland Statewide Analysis of Household Trip Production

EasyChair Preprint 7836

Abstract

This study applies Small Area Estimation (SAE) methods to a Regional Travel Survey (RTS) in order to derive statewide household person trips at the PUMA level when no sample is available. Several methods have been tested; those include: area and unit synthetic model with both Ordinary Linear Square and Poisson regression, and the Fay-Herriot model. Empirical results have been obtained for the State of Maryland, using the 2018 MWCOG Regional Travel Survey and the American Community Survey (ACS). RTS provides both the direct estimates of household person trips and the auxiliary variables for the synthetic model estimation, but only for PUMAs included in the survey. ACS provides the auxiliary variables for the entire state, including PUMAs with no RTS sample. Based on out-of-sample tests, it can be concluded that the area level linear model with RTS auxiliary variables performs better when compared to the other specifications proposed. This model was then applied to estimate household person trips for the area with no sample. We finally applied the Fay-Herriot method to the PUMAs in RTS and found that the combination of direct and synthetic estimation reduces the Coefficient of Variation. This application demonstrates that SAE methods can produce reliable transportation statistics by linking information from several datasets and could potentially reduce survey data collection costs.

Keyphrases: American Community Survey, Big Data, Regional Household Travel Survey, hierarchical Bayesian modeling, small area estimation, survey data, vehicle probe data

@booklet{EasyChair:7836,
  author    = {Mohammad B. Al-Khasawneh and Cinzia Cirillo and Partha Lahiri},
  title     = {Estimates of Total Household Trips for Areas with Zero Sample: a Maryland Statewide Analysis of Household Trip Production},
  howpublished = {EasyChair Preprint 7836},
  year      = {EasyChair, 2022}}