Alternative Sampling Techniques for NTD Reporting

Center Identification Number: 527-05

Project Title: Alternative Sampling Techniques for NTD Reporting

Principal Investigator:

Xuehao Chu

Phone: 813-974-9831

E-mail: xchu@cutr.usf.edu

Institution:

Center for Urban Transportation Research

University of South Florida

Fax: (813) 974-5168

Website: www.cutr.usf.edu

External Project Contact:

Ike Ubaka, Transit Office

Phone: 850-414-4532

I. Project Objective

The objective is to develop two alternative sampling techniques for estimating passenger miles for directly-operated fixed-route bus services with one for agencies with total passenger counts and one for agencies without total passenger counts. The ultimate goal is to help transit agencies to meet NTD reporting requirements with minimal expenditure of resources.

II. Project Abstract

The National Transit Database (NTD) is the system through which the Federal Transit Administration (FTA) collects uniform data needed by the Secretary of Transportation to administer department programs. The database consists of selected financial and operating data that describe mass transportation characteristics. Recipients and beneficiaries of Urbanized Area Formula Program grants must submit a NTD report through the life of the grant and/or capital equipment obtained through the grant. Unlike all other data reported in the NTD, passenger miles may be an estimate based on a sampling procedure meeting FTA requirements for confidence (95%) and precision (10%) levels. Transit agencies must sample or collect 100% counts of passenger miles in a mandatory year. The mandatory years are fixed in one-year, three-year or five-year cycles based on urbanized area size, number of vehicles operated in annual maximum service and type of service. In non-mandatory sampling (intermediate) years, agencies are required to report passenger miles in one of three ways: 1) Passenger miles from a 100% count; 2) Passenger miles estimated based on a sample during the current report year (must meet FTA’s sampling requirements); or 3) Passenger miles estimated using the average trip length from the mandatory year multiplied by current year unlinked passenger trips.

To assist transit agencies, FTA has developed acceptable passenger mile sampling procedures for demand response and fixed route services. A transit agency may use any data sampling technique, by mode and type of service, which meets 95% confidence and 10% precision levels. Many transit agencies find the FTA approved techniques too costly and wish to use alternative techniques. However, there is little guidance to transit agencies on alternative techniques and how they compare with the FTA approved techniques in terms of their requirements for labor and costs.

III. Task Descriptions

Task 1: Identify Potential Techniques

This task will identify alternative sampling techniques that have the potential to be widely used. The criteria in selecting the potential techniques include both implementation costs and statistical efficiency. This may be achieved through a review of reports and published journal articles that document such techniques. When needed, additional information may be obtained from FTA or a survey of NTD reporting personnel of transit agencies. The result of this task will be at least one potential technique for agencies with total passenger counts and one potential technique for agencies without total passenger counts.

Task 2: Identify Agencies

This task will identify transit agencies that currently use the above-identified alternative sampling techniques and have sample data available that were collected with such techniques. This may be achieved through a combination of three approaches: 1) a survey of NTD reporting personnel of transit agencies; 2) FTA personnel; or 3) a review of reports and published journal articles that document case studies of such techniques. The result of this task will be a range of agencies for each of the two potential techniques. In addition, this task will acquire the sample data from these agencies

Task 3: Evaluate Techniques

This task would evaluate the potential techniques in a number of ways. The NTD sample data from the agencies will be used to determine the values of key statistical parameters related to these techniques. In addition, this task will develop a spreadsheet that would allow agencies to determine the minimal sample size for their particular conditions. Finally, the transit agencies identified in Task 2 may be surveyed on the labor and cost requirements of these techniques.

Task 4: New Technologies

This task will examine how new technologies being used in transit agencies may influence sampling techniques. Specifically, this task would consider whether new technologies would eliminate the need for sampling, make sampling easier, or make sampling harder. Examples of technologies include automated passenger counters (APCs), electronic registering fareboxes (ERFs), and smart cards. At one extreme, for example, sampling would become obsolete when all buses are equipped with APCs or when smart cards are used systemwide.

Task 5: Final Deliverables

This task would prepare a guide on two alternative sampling techniques for estimating passenger miles for directly-operated fixed-route bus services with one for agencies with total passenger counts and one without counts. Details will be provided on how each should be implemented. Attached to the guide would be a spreadsheet for transit agencies to determine the sample sizes for their particular conditions once they know their statistical parameters. New technologies and their impact on sampling techniques will also be discussed.

CUTR will e-mail draft deliverables with two bonded hard copies to the FDOT Project Manager. After receiving comments from the FDOT Project Manager, CUTR will revise the draft deliverables accordingly. In addition to the 30 copies for the Research Office, CUTR will deliver 45 copies of the guide to the FDOT Project Manager. Each hard copy will include a 3.5-inch disk containing the spreadsheet for determining sample sizes.

IV. Project Schedule

Project Start Date: October 2002

Tasks

Month from Start Date

1: Techniques

2: Agencies

3: Evaluation

4: Technologies

5: Deliverables

Draft

Final

Progress Reports

Note: T = Task efforts

D = Deliverables

V. Project Budget

Alternative Sampling Techniques for NTD Reporting

Budget Categories

Hours X Rate

State Share

    Institute Director Salary

$0

    Faculty Salaries

900 X $42.8

$38,500

    Administrative Staff Salaries

29 X $22

$647

    Other Staff Salaries

$0

    Graduate Student Salaries

267 X $15

$4,000

    Undergraduate Salaries

$0

    Staff Benefits

$12,496

Total Salaries and Benefits

$55,643

    Permanent Equipment

$0

    Expendable Equipment and Supplies

$1,500

    Domestic Travel

$0

    Foreign Travel

$0

    Computer Costs

$0

    Other Costs

$0

Total Direct Costs

$57,143

    Indirect Costs

$2,857

TOTAL COSTS

$60,000

Note: This budget does not reflect any federal participation. The project team will include faculty, students, and secretarial and other support staff who will work directly on the project and whose costs are reflected in the direct costs of the project as listed above.

The results of this analysis will be provided to the FDOT through a series of quarterly progress reports and final deliverables. Copies of the deliverables will be widely distributed to various agencies, offices, and decision-makers in Florida and nationwide.

Recipients and Beneficiaries of Urbanized Area Formula Program grants in Florida and across the nation spend a considerable amount of resources in order to report NTD passenger miles data. Circumstances differ among transit agencies. Certain sampling techniques are better suited to some agencies but not others. A good guidebook on alternative sampling techniques provides an efficient mechanism for transit agencies to select the most appropriate techniques for their particular circumstance. The result could be significant savings in both labor and costs to transit agencies.

By allowing transit agencies to select and use the most appropriate sampling techniques for NTD reporting purposes, the proposed research will allow transit agencies save resources spent on NTD reporting so that more resources are devoted to providing their services in an efficient and productive manner.

The sample data from transit agencies may provide information on the degree of variation in passenger miles per one-way bus trip. If this information shows that the degree of variation is substantially smaller than what have been assumed as the default values in the current FTA procedures, FTA could use this new information to revise the current procedures and reduce the NTD reporting burden on individual transit agencies.