CLUSTER SAMPLING TECHNIQUES FOR ESTIMATING TRANSIT PATRONAGE

Sampling of trips is necessary to estimate ridership on most transit systems. Because of UMTA Section 15 and other internal requirements, it is important to know the accuracy of estimates made. Simple random sampling of trips is a technique approved by UMTA, and its accuracy formulas are well kown. However, random sampling of trips is a very unnatural and wasteful method of gathering data because of time lost in traveling from one selected trip to another. A more natural way of sampling is by run-piece, the sequence of trips that a driver follows for a day or half a day. Estimation and accuracy formulas for several techniques of sampling by clusters of trips are presented, with and without conversion factors. Stratification to improve accuracy is also incorporated. A case study in Los Angeles is used to demonstrate the merits of the various techniques. Pittsburgh data are also used to evaluate cluster sampling for route-level estimates. The results confirm the superior value of cluster sampling as compared to simple random sampling in most cases.

Media Info

  • Features: References; Tables;
  • Pagination: p. 105-114
  • Monograph Title: Transit management and replacement capital planning
  • Serial:

Subject/Index Terms

Filing Info

  • Accession Number: 00479333
  • Record Type: Publication
  • ISBN: 0309047064
  • Files: TRIS, TRB, ATRI
  • Created Date: Jan 31 1989 12:00AM