Small area estimates of daily person-miles of travel: 2001 National Household Transportation Survey |
| |
Authors: | Akhil K. Vaish Shijie Chen Neeraja S. Sathe Ralph E. Folsom Promod Chandhok Kuo Guo |
| |
Affiliation: | (1) RTI International, 3040 Cornwallis Road, PO Pox 14742, Research Triangle Park, NC 27709, USA;(2) Biostatistics & Data Management, Pfizer Consumer Healthcare, 5 Giralda Farms, Madison, NJ 09740, USA;(3) Bureau of Transportation Statistics, Research and Innovative Technology Administration, U.S. Department of Transportation, 1200 New Jersey Avenue SE, Washington, DC 20590, USA;(4) Emmes Corporation, 401 N. Washington St, Rockville, MD 20850, USA |
| |
Abstract: | The National Household Transportation Survey (NHTS) was designed at the national level, and for most states it does not have a large enough sample to produce reliable estimates, especially for subdomains (e.g., age groups) within a state. Using the 2001 NHTS, we produced small area estimates (SAEs) of the percentage of persons among four age groups (17 or younger, 18–39, 40–54, and 55 or older) having high daily person-miles of travel (more than 87.5 miles a day, which is the 90th percentile for daily person-miles traveled) and associated prediction intervals for all 50 states and the District of Columbia. The survey weighted hierarchical Bayes (Folsom et al., Proc of the Sect on Surv Res Methods of the Am Stat Assoc 371–375, 1999) small area estimation (SAE) methodology was used to produce state-level SAEs. This paper describes the methodology and shows that SAE can be an effective technique for producing reliable state-level estimates from large, national surveys like the NHTS. In particular, the prediction interval relative widths for SAEs were, on average, 31–48% narrower than the corresponding design-based confidence interval widths, whereas for small states the reduction was around 47–63%. |
| |
Keywords: | |
本文献已被 SpringerLink 等数据库收录! |
|