MARGINAL WEIGHTING PROCEDURES FOR EXPANDING SMALL-SAMPLE SURVEYS
This research investigates two methods of expanding a small-sample (n less than 300) transit-oriented survey to best replicate general population characteristics. In addition, two related questions of adequate sample size and methods of expanding onboard transit surveys are also investigated. Results show that the "Marginal Weighting" method of survey expansion is superior to the traditional "Category" expansion method because it jointly considers several characteristics (such as household size and auto ownership), that may have a relationship to transit usage. In addition, the "weighting" technique shows promise for application to many other non-transit related surveys. Minimum sample sizes are determined which will produce a desired number of observations in each population subgroup. A statistically derived minimum sample size closely matched empirical data on the number of samples needed for minimum reliability. Sample size design by the statistical method is further explored for use in future survey designs. The "head count" method for on-board surveying is judged superior to other methods analyzed. Both administrative ease and reliability of ridership estimates support arguments for implementation of this type of survey in the future.
Albany, NY USA 12232
- Kloeber, K W
- Howe, S M
- Publication Date: 1975-11
- Features: Appendices; Figures; References; Tables;
- Pagination: 74 p.
- TRT Terms: Data collection; Population; Ridership; Statistical sampling; Surveys; Transportation; Weight
- Uncontrolled Terms: Sample size
- Subject Areas: Administration and Management; Transportation (General);
- Accession Number: 00142203
- Record Type: Publication
- Report/Paper Numbers: Prelim. Res. Rpt. 97
- Files: TRIS, STATEDOT
- Created Date: Feb 16 1977 12:00AM