Improving ADA Complementary Paratransit Demand Estimation (2007)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

The sketch planning model is based on actual paratransit demand and other data for 28 “rep- resentative systems.” This section describes the process for selecting these systems, the specific service area and paratransit system data that were used, and the process for collecting the data. Representative Systems The original project concept called for using data from “transit systems considered exemplary for providing transit services in accordance with ADA requirements and best practices in the transit industry.” As the team investigated candidate systems for use in the research and dis- cussed these systems with riders and advocates, the term “exemplary” came into question. The systems used in the model development are definitely much above average, and all are believed to be in compliance with ADA paratransit requirements regarding capacity constraints. Systems were sought where demand is representative of demand that will occur when a system is in compliance and doing a good job of providing service. However, it is entirely possible that to some members of the eventual audience for this research, the term “exemplary” would imply a degree of excellence that goes beyond actual performance at some of the systems. A system can be meeting the letter and spirit of the law yet be providing a service that falls short of many users’ desires for public transportation. Further, some of the systems may not be using state-of-the-art eligibility methods, and the status of fixed-route accessibility may vary considerably among systems. It was also recognized that passenger perceptions of on-time performance may vary considerably, especially consider- ing the variations in on-time windows in use. Finally, the investigations of the systems done for this research are necessarily limited in scope, so the possibility cannot be ruled out that service or compliance issues exist that have not been identified. Twenty-nine representative systems were chosen for use in model development. These 29 were chosen after beginning with 88 candidate systems identified by respondents to a preliminary survey about factors influencing the demand for ADA paratransit. Candidate systems were con- tacted to determine their interest in participating in the research. Forty-eight systems either declined to participate or never responded to inquiries. The research team investigated whether the remaining 40 candidate systems met the criterion of “no significant capacity constraints.” This investigation emphasized obtaining input from ordinary riders as well as advocates. The transit systems were asked to provide names and con- tact information for riders. In addition, contact information was obtained independently using referral by panel members, local disability organizations, and contacts developed by the research team in previous work. 3 C H A P T E R 2 Data Sources

Exhibit 2-1. List of representative systems. Using these sources, the researchers interviewed at least one and often two or three riders (and in one case, six riders) from each candidate system. The riders were asked about ability to get through on the phone to make reservations, trip denials, trip purpose rules, and on-time performance. In discussing trip denials, riders were specifically asked about trips negotiated over an hour from the requested time and ability to get a trip in response to a request 1 day in advance. In the case of tran- sit systems that operate ADA and other paratransit services (for example, coordinated human ser- vice transportation), the riders were asked to distinguish ADA service from other service. As a result of this process, 11 candidate systems were removed from the list, leaving 29 “representative systems.” One system was eliminated later after no data were obtained to use in modeling. Exhibit 2-1 is a list of the 28 representative systems used in the research. Exhibit 2-2 shows the location of all the representative systems. The Midwest and South are somewhat under-represented. To remedy this, the team attempted to add one more system in Florida, one in Illinois, one more in Iowa, two in Kansas, two more in Michigan, three in North Carolina, one more in Ohio, one in South Carolina, one more in Texas, one more in Virginia, and two in Wisconsin. All these systems either declined, did not respond to multiple calls, or were eliminated after investigation. 4 Improving ADA Complementary Paratransit Demand Estimation AW dnalhciR tisnarT nilknarF neB .1 AV grubskcalB tisnarT grubskcalB - sseccA TB .2 IM gnisnaL ytirohtuA noitatropsnarT aerA latipaC .3 AC drocnoC ytirohtuA tisnarT atsoC artnoC lartneC .4 XT sallaD )TRAD( tisnarT dipaR aerA sallaD .5 AC hcoitnA ytirohtuA tisnarT atsoC artnoC nretsaE .6 7. Fort Worth Transportation Authority (The T) Fort Worth TX AC onserF sserpxE aerA onserF .8 LF apmaT ytirohtuA tisnarT lanoigeR aerA hguorobslliH .9 AV ellivsettolrahC TNUAJ .01 AW elttaeS orteM ytnuoC gniK .11 RO eneguE tcirtsiD tisnarT enaL .21 AW eehctaneW tisnarT kniL .31 AM llihrevaH-ecnerwaL ATR yellaV kcamirreM .41 KO asluT ytirohtuA tisnarT asluT natiloporteM .51 YN kroY weN tisnarT ytiC kroY weN .61 AC miehanA ytirohtuA noitatropsnarT ytnuoC egnarO .71 AI awmuttO metsyS tisnarT awmuttO .81 19. Port Authority of Allegheny County (Access) Pittsburgh PA OC revneD tcirtsiD noitatropsnarT lanoigeR .02 21. Rhode Island Public Transit Authority Providence RI 22. San Mateo County Transit District (SamTrans) San Mateo County CA 23. Southwest Ohio Regional Transit Authority / The Metro Cincinnati OH YN esucaryS ytirohtuA noitatropsnarT lanoigeR kroY weN lartneC .42 RO dnaltroP teM-irT .52 TU ytiC ekaL tlaS ytirohtuA tisnarT hatU .62 AC esoJ naS ytirohtuA noitatropsnarT yellaV .72 AW mahgnilleB ytirohtuA noitatropsnarT moctahW .82

Exhibit 2-2. Location of representative systems. Data Sources 5 Haverhill- Law rence B ellingham S eattle ( K ing County) W enatchee Antioch Concord S anta Clara County Fresno Orange County S alt Lake City Denver Tulsa Fort W orth Dallas Tampa Pittsburgh S yracuse R hode I s land New York City Eugene B enton-Franklin Cincinnati B lacksburg Charlottesville Lansing Ottumw a Portland S an Mateo Co. Merrimack Vall y Measures and Data Sources The principal sources of data for the sketch planning model were a questionnaire sent to the representative systems and the U.S. Census. Because of the small number of systems to be used in this analysis, it was understood that it would be possible to include only the most important and clearly measurable factors in the model. Initially 37 factors were identified that could potentially be included in a system-level model. A shorter list of factors, shown in Exhibit 2-3, was selected for data collection along with specific measures and data sources based on the following: • Ratings from a survey of 160 paratransit professionals, researchers, advocates, and riders about factors that affect demand for ADA paratransit (see Appendix C). • Theoretical considerations and empirical evidence obtained from a comprehensive literature review. • Data availability. • Input from the project panel. Non-ADA paratransit trips (Factor 2) include (1) subsidized taxi trips, even if they are lim- ited to ADA-eligible individuals, that do not comply with ADA service criteria such as limits on number of trips, fare, etc.; and (2) trips provided using eligibility criteria other than ADA eligibility, for example, under senior transportation programs. Human service trips (Factors 3 and 4) include trips sponsored by Medicaid, workshops and training programs serving people with developmental disabilities, adult day care and adult day health care, and senior meal programs. The Census questions about disability mentioned in the last item (Factor 18) are as follows: Question 16. Does this person have any of the following long-lasting conditions: a. Blindness, deafness, or a severe vision or hearing impairment? b. A condition that substantially limits one or more basic physical activities such as walking, climbing stairs, reaching, lifting, or carrying?

6 Improving ADA Complementary Paratransit Demand Estimation Exhibit 2-3. Factors to be used in system-level modeling. Factor Measure Data Sources 1. ADA paratransit demand Annual ADA complementary paratransit trips provided Representative system questionnaire 2. Non-ADA paratransit (including subsidized taxis) provided or administered by the transit agency or a broker Annual trips provided that are not part of ADA complementary paratransit Representative system questionnaire 3. Overall level of human service transportation Proportion of human service transportation needs served by human service agencies Estimates of transit agency and state level staff 4. Human service transportation provided or administered by the transit agency or a broker Annual trips provided for human service agencies, whether or not under a formal arrangement Representative system questionnaire 5. Accessible public transit service Annual vehicle miles of transit service Percentage of revenue vehicles that are ADA accessible Wheelchair boardings, if available Representative system questionnaire 6. ADA paratransit on-time performance Percent of pick-ups after the window Length of on-time window Representative system questionnaire 7. ADA paratransit telephone access Average hold time to make a reservation Representative system questionnaire 8. ADA paratransit eligibility process Percent of applicants interviewed or tested in- person Representative system questionnaire 9. ADA paratransit fare Average fare per passenger Representative system questionnaire 10. Length of time since significant denials were eliminated Months from the last significant denials to the middle of the fiscal year for which data is provided Representative system questionnaire 11. Exact geographic definition of the ADA paratransit service area Maximum extent of service at peak hours Transit system GIS data or maps 12. Total service area population Total population in the service area Geographic definition of service area combined with U.S. Census 13. Age and sex distribution of the service area population Male and female population in the service area age 65+ and age 75+ Geographic definition of service area combined with U.S. Census 14. Ethnic composition of the service area population Non-white or Hispanic population in the service area Geographic definition of service area combined with U.S. Census elbarusaem htiw syad ro llafwons launna naeM etamilC .51 snowfall National Climatic Data Center 16. Household income Total population in the service area with 1999 household income below the poverty level Geographic definition of service area combined U.S. Census 17. Density and/or car ownership Population per square mile within the service area Percent of housing units in the service area with no vehicle available Geographic definition of service area combined U.S. Census ecivres fo noitinifed cihpargoeG a71 ,b dna a61 snoitseuq mrof gnol susneC ytilibasiD .81 area combined U.S. Census

Question 17. Because of a physical, mental, or emotional condition lasting 6 months or more, does this person have any difficulty in doing any of the following activities: a. Learning, remembering, or concentrating? As with all the Census disability questions, there is no way to know for sure how answers are related to degrees of disability. These three questions do cover the principal functional abilities for transit use. They also include some issues that generally do not result in ADA paratransit eligibility, such as deafness and difficulty lifting. Still, it is likely that incidence of disability as measured by these questions is highly correlated with ADA paratransit eligibility, at least at the level of large geographic areas. In addition to the questions listed, the Census included a question about “go outside the home disability.” Panel members questioned the relevance of this question, noting that it was intended to measure the need for in-home care. In addition, Census staff have determined that go outside the home disability appears to have been substantially overstated in the 2000 Census, most likely as a result of a confusing skip pattern in the mail-back version of the long form.1 All of the population measures were for an area corresponding as closely as possible to the actual ADA paratransit service area. In a few cases, this area may correspond closely to a divi- sion, such as a city or county, for which published Census tabulations are available. In most cases, the area served, consisting of points within three-quarters of a mile of transit routes, corresponds to some combination of numerous Census tracts or block groups. If maps of the ADA service area are available, ideally as geographic files from a geographic information system (GIS), these can be combined with Census data sets to compute the desired measures. This process can be somewhat time consuming, but it is very important, since the population of the ADA service area may be very different than the population of the urbanized area normally reported by transit systems to the National Transit Database. Data Collection The data collection process began in October 2005. A data collection form for the representa- tive transit systems was created in Microsoft Word. The form was designed so that transit agency staff could complete the form within Word and return it as an e-mail attachment. On October 20, 2005, a draft of the form was provided to the panel for comment. After obtaining comments from one panel member, a draft of the form was provided to two agencies to obtain their comments and their estimate of the time that it would take to complete. Based on their input, further refine- ments were made, after which the Principal Investigator sent an e-mail to all 29 representative systems, reminding them about the project and alerting them to expect a form from Planners Collaborative that would take about 2 hours to complete. Planners Collaborative sent the form on November 2, 2005, requesting responses by November 18. The first response came the following day. However, by November 18 only 13 completed forms had been received, with promises of five more. At this point, the process began of sending reminder e-mails and making phone calls. Follow-up contacts also began at this time, in partic- ular to obtain the best possible information about service area, which is critical to developing usable demographic information for the modeling process. As of the end of December 2005, responses had been received from 28 of the 29 representative systems, including three systems that provided data for multiple years. The data collection process was completed over the next Data Sources 7 1 Sharon Stern and Matthew Brault, “Disability Data from the American Community Survey: A Brief Exami- nation of the Effects of a Question Redesign in 2003,” Feb. 2005. At http://www.census.gov/hhes/www/ disability/ACS_disability.pdf.

Exhibit 2-4. Methods used for creating GIS files. 3 months, resulting in usable data from 28 of the 29 representative systems, including three systems that provided data for multiple years. Activities included the following: • Creating GIS files to describe the ADA service area of each system. As described below, some systems sent GIS files, some sent lists of jurisdictions, and some sent graphics of their service area. Some of these required interpretation and personal follow-up with the respondents. Eventually, GIS files were created for all 28 systems. • Analyzing Census data to calculate values for all the desired demographic variables specific to each system’s ADA paratransit service area. • Working with the respondents to clarify responses, in particular with respect to definitions of on-time performance. Follow up inquiries were made about 33 data items from 16 systems. • Sending a supplementary questionnaire concerning human service transportation and non- ADA paratransit to all 28 systems. This questionnaire was sent to the panel for their review and comment at the end of January 2006. One panel member responded with several sugges- tions that were incorporated into the supplemental questionnaire. Several systems responded that they did not have the information requested. In these cases information was sought from other contacts, typically in state government. Only one of these non-transit agency contacts yielded a response. • Compiling information about snowfall from the National Climatic Data Center. As expected, a particularly time-intensive activity was obtaining a description of the ADA paratransit service area suitable for estimating the size of the population served and the number of people in various subcategories. The methods used are summarized in Exhibit 2-4. In some cases, the GIS layers were specific to paratransit, while in others they were for the fixed-route system, from which three-quarter-mile buffers were created. In one case the layer was approximate and needed to be adjusted based on knowledge of the area. As described in the section about model development, further data collection was necessary after preliminary regression analysis led to identifying data that was incomplete or incorrect. 8 Improving ADA Complementary Paratransit Demand Estimation Methods Used Number of Transit Systems 51 .atad susneC ot dehctam eb dluoc taht sreyal SIG tneS Provided a list of jurisdictions from which it was possible to compute the necessary population measures. 7 Sent maps of their service area in pictorial form from which approximate GIS layers could be created. 3 Described their service areas as consisting of one jurisdiction plus areas around one or two routes, for which it was necessary to create an approximate GIS layer. 2 One system sent screen prints from its scheduling system. These were not usable, but a paper copy of the bus route book was available from which it was possible to create an approximate GIS layer. 1 Total: 28