Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 301
Modernizing the U.S. Census APPENDIX G Use of Decennial Census Data in Transportation Planning This appendix reviews the history of the use of census data in transportation planning, the types of data available, and experience with the use of data in actual practice.1 Responses to a survey questionnaire mailed to states and metropolitan planning organizations to determine the past and planned use of census data are summarized. The relationship of census data to current legislative requirements for transportation planning is discussed, as well as the options and costs involved in replacing the data if they are not collected in the 2000 census. The U.S. Department of Transportation (DOT), as well as state and local transportation planning organizations, have relied on the consistent data collection provided by the decennial census for travel-to-work characteristics and vehicle availability information since 1960 when transportation items were first added to the census. Information from the decennial census is used by the DOT as a comprehensive database supporting development of new policies and programs as benchmark data with which to evaluate the impacts and overall effectiveness of previously implemented programs. Decennial census data for small areas such as census tracts and traffic analysis zones are used by state and metropolitan transportation planning agencies to meet the provisions of the Intermodal Surface Transportation Efficiency Act of 1991 (ISTEA), which require a comprehensive transportation planning process at both the state and metropolitan levels. The census provides the baseline origin-destination data on local work trips, household characteristics, and worker characteristics for use in travel forecasting models and for monitoring carpooling, public transit use, and other travel behavior. These data are now provided to all the states and metropolitan planning agencies as a special tabulation in the Census
OCR for page 302
Modernizing the U.S. Census Transportation Planning Package (CTPP) at a cost of only $10 per 1,000 population or 1 cent per capita. Funding for the CTPP comes from funds available through ISTEA. The census data on commuter travel flows and characteristics also are used to evaluate and select projects, develop traffic congestion management systems, and identify transportation corridors needing capacity expansion. Further, travel-to-work and vehicle availability data from the census for small areas are used by state and metropolitan planning organizations to prepare vehicular travel and pollutant emissions profiles, compute regional average rates of vehicle occupancy in the commute to work and evaluate the impact of long range transportation plans on air quality in compliance with the Clean Air Act Amendments of 1990. Finally, census data on the geographic distribution of persons with work disabilities and mobility limitations are used by local transit operators to provide service levels that are fully accessible to all segments of the population under the Americans with Disabilities Act. The federal legislative initiatives cited above—the Intermodal Surface Transportation Efficiency Act of 1991, the Clean Air Act Amendments of 1990, and the Americans with Disabilities Act—delineate the national interest in rebuilding the transportation infrastructure, improving environmental quality, and providing a fully accessible system for accommodating the mobility needs of a diverse population. To respond to these statutes, policy and program development at the federal, state, and local levels needs to be based on decennial census data that provide a context for evaluating past trends and preparing forecasts of the future. The high levels of current use by both state transportation agencies and metropolitan planning organizations, along with the low cost of obtaining and disseminating the data lead to the conclusion that abandoning the tradition of nationwide survey uniformity and geographic consistency of data provided by the decennial census would result in high costs and disruption to program development and evaluation at all levels of government. The continued collection of transportation data in the decennial census is under review. In response to congressional criticism, the Census Bureau is taking a zero-based approach to 2000 census planning. A major thrust of this approach is to question the justification for collecting data on topics such as place of work, mode of transportation to work, travel time, carpooling, time of departure for work, disabled persons with mobility limitations, and number of vehicles available to each household as part of the decennial census. This appendix reviews the history of the use of census data in transportation planning, the types of data available, and experience with the use of data in actual practice. Since the 1990 census results have only recently become available for transportation planning applications, much of this appendix deals with experiences with the 1980 census data. Discussion of the 1990 census is included wherever possible. The historical review of transportation data takes the perspective of building on the experience of previous censuses as has been its tradition in the past.
OCR for page 303
Modernizing the U.S. Census HISTORY OF TRANSPORTATION QUESTIONS IN THE CENSUS The inclusion of transportation items is a fairly recent occurrence in the history of the decennial census. Although the first census was conducted in 1790, questions pertaining to transportation did not appear until 1960 when three such questions were asked on a 25-percent sample basis. The population items in 1960 included questions on each worker's place of work (city, county, and state) and means of transportation to work; the housing items included a question on the number of automobiles available for use by the members of each household. The principal impetus for adding the question on place of work to the 1960 census was the need for data on commuting interchanges for use as an indicator of economic integration between large cities and their suburbs as part of the criteria for delineating metropolitan statistical areas. The commuting data from the census were certainly of interest to transportation planners, but urban transportation planning in the early 1960s was still being done on the basis of locally conducted origin-destination surveys. By 1970, with the development by the Census Bureau of Address Coding Guides (ACGs) and Dual Independent Map Encoding (DIME) files, interest in the census as a source of transportation planning data increased considerably. The ACGs and DIME files provided the capability of geographically coding place-of-work addresses within the urbanized portion of metropolitan areas down to the level of the census block. The 1970 census again asked questions on place of work, means of transportation to work, and automobile availability. However, the place-of-work question asked for the actual street address of the respondent's workplace, and these addresses were coded to census blocks within the areas covered by ACGs and DIME files. The DOT contracted with the Bureau of the Census to create a series of special tabulations in a transportation planning package. Metropolitan Planning Organizations (MPOs) obtained the data tabulated for their traffic zones on a cost-reimbursable basis. Between 1970 and 1980, several developments resulted in significant expansion in the number of transportation items included in the decennial census. The energy crisis of the early 1970s and the subsequent ongoing concern for the nation's supply of nonrenewable energy sources brought about a sharp increase in the need for statistics for transportation planning and policy formulation. From 1975 to 1977, under the sponsorship of the DOT, the Census Bureau conducted for the first time journey-to-work surveys in 60 metropolitan areas and a national survey in 1975, all as part of the Annual Housing Survey. In recognition of the growing need for analysis of these data, a Journey-to-Work Statistics Branch was created within the bureau's Population Division to carry out the technical planning and developmental work pertaining to the collection, processing, tabulation, and analysis of journey-to-work data from the decennial census and periodic surveys. Also during the decade, the cost of conducting metropolitan origin-destination surveys increased rapidly, and the DOT began to encourage local
OCR for page 304
Modernizing the U.S. Census agencies to look to the decennial census as an alternative source for cost-effective transportation planning data. Thus, due to the significant increase in the need for transportation data at all levels of government, the 1980 census included eight transportation items: six population questions and two housing questions. On the population side, in addition to questions on place of work and means of transportation to work, the 1980 census asked about carpooling arrangements, the number of riders in the carpool, travel time from home to work, and whether persons had a disability that limited or prevented their use of public transportation. On the housing side, the automobile availability question was supplemented with an additional question on the number of light trucks and vans available for use by members of each household. Once again for the 1980 census, the DOT contracted with the Census Bureau to create a series of special tabulations in a Transportation Planning Package. MPOs obtained the data tabulated for their traffic analysis zones on a cost-reimbursable basis. To increase the utility of the census data for local transportation planning, the Census Bureau developed an innovative procedure to assign incomplete place-of-work responses to census blocks so that they too could be tabulated at the traffic analysis zone level. THE 1990 CENSUS The 1990 census transportation statistics program marked the continued refinement of transportation data available from the census, technical improvement in the geographic coding of place-of-work responses to small areas within metropolitan regions, and the creation and dissemination of innovative transportation data products. The 1990 census again included questions on place of work, means of transportation to work, carpooling, carpool occupancy, and travel time to work. An important new question on time of departure from home to work was added to the census questionnaire to allow tabulation of commuting patterns and characteristics by peak hours of travel. The questions on the number of automobiles available to each household and the number of trucks or vans available to each household were combined into one question on the total number of vehicles (cars, trucks, and vans) available. The question on public transportation disability was replaced with a more general questions that identified persons whose disabilities limited their ability to get around outside the home. Two innovative technical advancements in place-of-work coding were made for 1990. The first innovation was the joint development by the Census Bureau and the DOT of the Census/Metropolitan Planning Organization Cooperative Assistance Program. This program gave local MPOs the opportunity to assist the Census Bureau in improving the accuracy of place-of-work data for their region. Planning organizations took part in three activities: providing files of employers and their locations to the Census Bureau, working with major employers
OCR for page 305
Modernizing the U.S. Census to ensure that their employees reported accurate workplace addresses, and assisting the Census Bureau in identifying the locations of workplaces that census clerks could not code. Over 300 MPOs took part in these cooperative activities. The Federal Highway Administration (FHWA) made the costs incurred by the MPOs for this work an eligible activity for use of federal aid highway planning funds. The second advancement in place-of-work coding was the implementation by the Census Bureau of an automated place-of-work coding system. Place-of-work addresses were keyed to create machine-readable files that were then matched to address coding and major employer files to assign geographic codes to the place-of-work responses. Cases that could not be coded on the computer were sorted and clustered and referred to clerks for research and computer-assisted clerical coding. The automation of place-of-work coding allowed the Census Bureau to accomplish the coding operation efficiently and cost effectively. THE URBAN TRANSPORTATION PLANNING PACKAGE (UTTP) The availability of block-level data on commuting origins and destinations from the 1970 census made possible the development for the first time of the Urban Transportation Planning Package (UTPP). The 1970 UTPP was a special tabulation of census data for individual metropolitan areas tailored to the geographic areas that are used in transportation planning. Local transportation planning organizations prepared specifications for the blocks that made up their traffic analysis zones, and the Census Bureau then produced a standard set of tabulations for those zones on a cost-reimbursable basis. Specifications for the content of the UTPP were submitted to the bureau by FHWA. About 120 UTPPs were prepared after the 1970 census. The Census Bureau again produced the UTPP after the 1980 census. This time specifications were developed and submitted to the Bureau by an ad hoc committee of transportation planners under the auspices of the Transportation Research Board, an organizational component of the National Research Council. Funding for development of the necessary computer programs and administration of the 1980 project was provided by the DOT. The number of packages produced increased to over 150. Significant innovations in the dissemination of the journey-to-work data were achieved for the 1990 census. Two transportation planning packages were produced: statewide packages for each state and the District of Columbia and urban packages for the transportation study area defined by each MPO. Production of the transportation planning packages by the Census Bureau was sponsored by the state departments of transportation under a pooled funding arrangement with the American Association of State Highway and Transportation Officials (AASHTO). This pooled funding arrangement supported the production
OCR for page 306
Modernizing the U.S. Census of data for the entire country instead of only those areas that decided to purchase the data as in previous censuses. Funding to develop the 1990 CTPP program was provided by the FHWA and the Federal Transit Administration (FTA). The total cost was $2.6 million or about $10 per 1,000 persons compared to $12 per 1,000 persons in 1980 (1980 dollars). The number of packages produced (both state and metropolitan) increased to 415 for the 1990 census. The 1990 CTPP continued the program established in 1970 and conducted in 1980 in the same general format. A working group was established to develop the specifications for the 1990 CTPP for both the metropolitan data set and the statewide data set. This ad hoc group consisted of members from AASHTO, the National Association of Regional Councils, FHWA, FTA, the Census Bureau, plus experts in the field from states and MPOs. A similar group developed the 1980 UTPP. The statewide tabulations were the first product produced from the 1990 CTPP. A file was produced for each state, containing data for the households, persons, and workers who live in each city and county in the state, data on all workers working in each city and county in the state, and data on commuter flows between counties and cities. The urban tabulations were similar to the data provided in the statewide package, except that the data were tabulated on the basis of census tracts or traffic analysis zones. The urban data were delivered after the delivery of the statewide packages. An urban package was produced for each metropolitan area for which the Census Bureau had an address coding capability. To make the CTPP data easily accessible and widely available, the DOT's Bureau of Transportation Statistics released the CTPPs on CD-ROM with software to display and retrieve the data from a personal computer. Use of the 1980 UTPP Data—A Case Study Since the early 1970s, the Delaware Valley Regional Planning Commission (DVRPC), the metropolitan planning organization for the Philadelphia region, has relied on census data for transportation planning and travel forecasting because of rising costs of large-scale data collection such as regional home interviews, employment, and land-use surveys. DVRPC used the 1970 census data to check and validate traffic simulation models for producing traffic analyses based on up-to-date information. Census work trips, housing statistics, car ownership, employed persons, and employment data were used to estimate trip generation and distribution patterns between transportation simulation zones. In addition, information about the journey to work and other characteristics of workers has been used by DVRPC, local and state governments, transit operating agencies, and private corporations to make a variety of decisions on transportation and locational matters. Uses of the 1980 UTPP in the Delaware Valley region were similar to those applications outlined in the Transportation Planners' Guide to Using the 1980
OCR for page 307
Modernizing the U.S. Census Census (Sosslau, 1983). These include the study of bus circulation patterns, location of park-and-ride lots and express bus service, study of accessibility and special population segments, analysis of highway and transit trips, planning of highway and public transportation systems, planning and analysis of projects, update of traffic simulation models, analysis of work-trip trends, location of shopping centers and service industries, analysis of parking requirements, and studies of employment. DVRPC used census data in various studies and continues to use such data in transportation planning and nontransportation planning activities because it is the only comprehensive information at the regional and local levels. The six major uses of the 1980 UTPP in the Delaware Valley region are: Establishing a Database for Transportation Planning DVRPC prepared a data bank for transportation planning at the block-group and tract levels. This information includes population, employment, work trips, car ownership, and other socioeconomic variables required for traffic simulation and transportation analysis and planning. Such data were extracted from the 1980 UTPP. All data items were edited for reasonableness based on other census data and DVRPC surveys, traffic counts, and employment files. These data were used in most transportation system and project planning studies. Preparation of Data Summaries and Evaluation of Trends DVRPC completed a report on the journey-to-work trends in the Delaware Valley region (Delaware Valley Regional Planning Commission, 1994). This report compared the 1970 and 1980 journey-to-work information, means of transportation for commuting to work, employed persons, and employment at the county and regional levels. It also analyzed the commuting flow between the counties of the Delaware Valley region and surrounding counties and cities. The report was well received by planners and decision makers because it provided factual information about trends in development and travel patterns in the region. Other tables showed the trends in employment and mode of travel for all DVRPC counties. Update of DVRPC Traffic Simulation Models The DVRPC travel forecasting models were updated using the 1980 UTPP. The 1970 UTPP was used to check and update the DVRPC traffic simulation models. These models were updated again using 1980 census data. The DVRPC travel simulation models follow the traditional steps of trip generation, trip distribution, modal split, and travel assignment and utilize the computer programs included in the federally sponsored Urban Transportation Planning System (UTPS). Use in Highway and Transit Corridor Studies The 1980 UTPP data, especially the journey-to-work information, were used in three transit corridor studies by DVRPC to check the travel demand or ridership for each transit submode, including high-speed rail line, express bus and park-and-ride service, and local bus service. The use of these data minimized any large-scale data
OCR for page 308
Modernizing the U.S. Census collection and decreased the rising costs of surveys required for transportation planning. Application in Strategic Planning and Economic Development DVRPC used the 1980 UTPP information on employment to evaluate the significant changes in the type and location of industries and commercial establishments. This evaluation resulted in recommendations and strategies aimed at attracting new industries and high-technology firms to the Delaware Valley. Also, employment information was useful in the redevelopment of declining areas of old urban centers and provision of the required physical improvements for their rehabilitation. Provision of 1980 UTPP Data to Public Agencies and Private Corporations Finally, DVRPC provided the 1980 UTPP data to any public or private agency involved in planning or urban studies. This included studies for housing, finance, real estate, health facilities, social services, economic base, and economic development. Planning agencies and private companies in the Delaware Valley region were interested in obtaining the UTPP information for their various studies. Summary Generally, the 1980 UTPP for the Delaware Valley region contained data of good quality for transportation planning, economic base and employment location studies, urban development analysis, and planning and evaluation of public services. However, the evaluation of UTPP data indicates a few programming, statistical, and bias problems. Most of these problems were resolved before DVRPC used the UTPP for trend analyses, information purposes, traffic simulation, highway and transit project studies, and strategic planning. The errors in the 1980 data were generally smaller than those found in the 1970 UTPP. Trip and employment information should be adjusted before it is used in transportation planning studies because it does not include all workers or jobs. DIFFERENCES BETWEEN CENSUS AND TRAVEL SURVEY DATA Most of the tabulations of the CTPP and its predecessor, the UTPP, focus on workers and their travel. The balance is about households, vehicles, and persons. Vehicles include automobiles, trucks, and vans available to a household. Mode of transportation is synonymous with the census term ''means of transportation." Journey-to-work questions asked in the census differ in some respects from those usually asked by planners in travel surveys (McDonnell, 1984). Several points should be kept in mind when census data about work trips are used: The address where the individual works most often is recorded in the
OCR for page 309
Modernizing the U.S. Census census questionnaire. When a worker holds two jobs, the second job location is not entered. Some workers go to different work locations on a given day. If such workers report to a central location, this location is to be entered as the workplace. If there is no central location and the worker went to various work locations, the smallest geographic area common to the starting places (for example, Westchester County, New York) is to be entered. The data imply direct trips from residence to workplace and do not request information about indirect work trips (stops on the way to work). The census asks about work "at any time last week." Thus, typical (usual) workday information is received rather than average workday information. The difference between an average day and a typical day is significant in transportation planning because on an average day some 10-20 percent of all workers may not commute from home to work for one reason or another. The census asks how the person "usually" got to work the previous week. This approach does not measure periodic use of transit among commuters who drive or carpool, but also use transit on some days of the week. Similarly, questioning about "usual" carpool size in the census probably results in overestimation of carpool size. Carpools are usually formed on a given day, a carpool member might not go to work, might be out of town, and so on, resulting in a number of passengers lower than that reported for the usual case. The census asks where the respondent was employed "last week." It does not ask, as travel surveys do, whether a trip to work was made "yesterday." Journey-to-work questions are asked of both full-time and part-time workers, and only the combined responses were reported by the Census Bureau. For the reasons listed above, transportation planners need to adjust census data for use in planning models, as described below. Journey-to-Work Adjustments The Washington Metropolitan Council of Governments compared census journey-to-work data with those collected by the MPO (Wickstrom, 1984). The census source in this case was a supplementary journey-to-work survey conducted by the Census Bureau for the U.S. Department of Housing and Urban Development as part of the 1977 Annual Housing Survey. The journey-to-work supplement was similar in form to the 1980 census questionnaire. The census questionnaire asked where the respondent was employed "last week." It did not ask, as travel surveys usually do, whether a trip to work was made "yesterday." In Washington, D.C., it was found that a factor of 0.85 was required to adjust the census ''usual-day" data to travel demands on a specific day as sought by transportation planners. Public transit trips tended to be underreported in the census data because
OCR for page 310
Modernizing the U.S. Census only the usual model was requested. A Washington, D.C. survey of transit riders showed that only 89 percent of bus riders and 76 percent of rail riders used public transit 4 or more days per week. For both forms of transit combined, 85 percent were regular users. Comparisons also were made of person work trips and transit work trips. For the Washington, D.C. region, census data were a little more than 6 percent low for total trips and a little more than 5 percent low for transit trips. Overall employment data were also compared. The census data did not count second jobs and, except in areas where commutershed information was available, the failure to count work trips into the region from counties outside the SMSA resulted in underreporting the volume of travel demands. Problems Inherent in Census Data The 1980 census UTPP journey-to-work data were reviewed extensively and utilized in the transportation planning process in the Washington, DC area. There were more than 75,000 individual records representing 1,650,000 workers in 1980. It should be recognized that the census data brought some inherent problems as well as provided a new data source for transportation analysts and planners. The following problems were inherent in the data: There were certain basic definitional differences between the way the census viewed the journey to work and an actual trip. Transportation planners would prefer information on the most recent journey-to-work trip, including data on places that were travelled on the way to work. Actual trips often involve some other intermediate trip purpose and, hence, differ from census information on the "usual" journey to work. Comparisons of census employment location coding with an independent data source indicated that more effort was needed to code accurately to the traffic zone or district level. Although trips to downtown and to urban areas were compatible, trips to outlying suburban centers were underrepresented in most cases. Not all trips were coded to the zone level of geography. Certain key data items useful for transportation planning were not collected in the census. This included information on the cost of parking, departure and arrival times for the work trip (peak hour, peak period, and nonpeak), and a listing of all modes of travel used. The peak hour data were subsequently collected in the 1990 census. Considerable additional staff effort was needed to produce a file considered suitable for use in recalibrating models or for use at the individual planning project level.
OCR for page 311
Modernizing the U.S. Census Positive Aspects of Census Data The following positive conclusions may be made based on experience with using the 1980 UTPP for the Washington, D.C. Area: The census data were used extensively. It is unlikely that any other comprehensive data source could have been developed and used within the time and cost associated with the census data. The parts of the UTPP that provided zone of residence, zone of work, and zone-to-zone commuter-flow data were used the most. The county-to-county totals were also extremely useful. Because the Baltimore and Washington, D.C. areas were tabulated as one region in the UTPP, there was an opportunity to obtain data that were not handled adequately by two separate data collection and study processes in the two regions. Although not a file of individual workers, the data were capable of being used for model verification and development. The data verified the need for revisions to travel forecasting procedures. Acceptance of the census data by participants in the planning process, including elected officials, was high. USE OF DECENNIAL CENSUS DATA BY TRANSPORTATION PLANNING AGENCIES To gain insight into the extent to which 1980 decennial census data were used by transportation planning agencies, we mailed a questionnaire to 28 such agencies throughout the United States. Both state departments of transportation and MPOs were contacted. Similar questions also were asked of those agencies about their planned use of the 1990 census data. Questions asked included: (1) type of census data used, (2) use of the data, (3) usefulness of the data in addressing federal requirements, (4) degree of difficulty expected in order to replace the data, (5) method of replacing the data, and (6) whether or not funds and staff would be available to replace the data. Twenty-two agencies (almost 80 percent) responded to the questionnaire. Of these, 10 were state departments of transportation and 12 were MPOs. The ten states included California, Texas, Wisconsin, Washington, Florida, Minnesota, Maryland, Ohio, North Carolina, and New York. The 12 MPOs included Pittsburgh, Philadelphia, Denver, Chicago, San Francisco, Washington, D.C., Boston, Los Angeles, New York, San Diego, Baltimore, and Charlotte. Tables G.1 and G.2 show the overall response from all 22 agencies. Separate summaries are made for the past use of 1980 data and the planned use of the 1990 data. Work-trip data by mode of travel and origin-destination data by zone along
OCR for page 312
Modernizing the U.S. Census with socioeconomic data were used in 1980 by 18 of the 22 respondents. The lowest use reported (15 agencies) was associated with the employment by place-of-work information. All categories, including household and population data, as well as use of the special cross-tabulations provided by the CTPP showed an increase in planned use by agencies for the 1990 data when compared to 1980. Heavy use was made of the data for providing inputs to travel forecasting models (and as a check of model results) and for corridor studies. Long-range planning, transit planning, and alternatives analysis also ranked high on the list. All categories of planned use for 1990 data showed marked increases, with 20 of the 22 agencies planning use of the data in transit planning and corridor studies. Fifteen of the agencies also responded positively to the three categories that reflect new federal requirements through the Clean Air Act Amendments and the Intermodal Surface Transportation Efficiency Act—congestion management, transportation control measures, and conformity determination. In regard to agency responses as to the usefulness of the census data in addressing federal requirements, 20 of the 22 responding agencies reported that the data were either essential or very useful in meeting the requirements for the development of statewide or metropolitan area plans. Seventeen reported the same degree of usefulness in meeting alternatives analysis requirements. Transit planning and the new requirement of developing congestion management plans were similarly noted by 15 agencies. Results were similar for both states and MPOs (not shown separately here), with three-quarters of the MPOs adding air-quality conformity requirements to the list. Eighteen of the 22 agencies said that they would experience great difficulty in replacing the work-trip data by mode (by traffic analysis zone) or the origin-destination data (zone-zone). Fifteen reported that replacing the special tabulations from the CTPP would prove very difficult. Only the employment data provided were not cited as being very difficult to replace by a majority of the respondents. MPOs reported a greater degree of difficulty in all categories listed and by a considerable margin. Since the 1980 census did not contain a statewide element, and since the states had not received the 1990 data at the time of the survey, the lower response rate by states is understandable. When asked how the census data would be replaced (if not available), 19 of the 22 agencies cited a home interview as the most likely method. Some cited workplace surveys, roadside or transit surveys, or secondary sources as alternatives. MPOs were more inclined toward home interviews, with states citing secondary sources more frequently. When asked if funds and staff could be made available to replace the census data with some alternative collection activity, only 1 of the 22 agencies responding gave a "yes" answer to this question. This agency, a state department of transportation, reported that it does not rely on census data because its planning needs require data in a more timely manner than census can provide (2 to 3 year lag). Eleven agencies reported that staff and funds could probably be made
OCR for page 313
Modernizing the U.S. Census available, while 10 said that replacing the data was unlikely. Of these, 5 were states and 5 MPOs. RELATIONSHIP OF CENSUS DATA TO LEGISLATIVE REQUIREMENTS Data from the decennial census are the backbone of the statistical system that supports the transportation planning process of our nation. The DOT, as well as state and local transportation planning agencies, have relied on the consistent data collection provided by the decennial census since 1960 when transportation questions were first added to the census questionnaire. Today, these organizations are increasingly reliant on census data to implement the requirements of the Intermodal Surface Transportation Efficiency Act of 1991 and the Clean Air Act Amendments of 1990. In its October 1993 report, Transportation Infrastructure: Better Tools Needed for Making Decisions on Using ISTEA Funds Flexibly, the U.S. General Accounting Office (GAO) concluded that current-travel demand models must be improved to provide better information for analyzing the impact of transportation projects on air quality. The GAO urged the Secretary of Transportation to ensure that the next generation of travel-demand models and analytical procedures being developed by the DOT's Travel Model Improvement Program (TMIP) help states and localities address their transportation priorities and comply with both the air-quality standards set forth in the Clean Air Act and the intent of the ISTEA that a total systems approach be used in decision making. In its official response to the GAO report, the DOT stated that it is imperative that the decennial census continue the transportation data used in these models that are the key to the ultimate success of the TMIP. Significant innovations in the dissemination of the journey-to-work data also were achieved for the 1990 census. Two transportation planning packages were produced: statewide packages for each state and the District of Columbia and urban packages for the transportation study area defined by each metropolitan planning organization. Production of the transportation planning packages by the Census Bureau was sponsored by the state departments of transportation under a pooled funding arrangement with the American Association of State Highway and Transportation Officials. This pooled funding arrangement supported the preparation of data for the entire country instead of only those areas that decided to purchase the data as in previous censuses. Funding to develop the 1990 Census Transportation Planning Package Program (CTPP) was provided by the FHWA and the FTA. CTPP funding totaled $2.6 million for both programming and production of the statewide and urban packages. Transportation data from the decennial census are used by the DOT as a comprehensive database supporting development of new policies and programs,
OCR for page 314
Modernizing the U.S. Census and as benchmark data with which to evaluate the impacts and overall effectiveness of previously implemented programs. The DOT works in partnership with states and local governments to assess project and corridor-level impacts of implemented plans, programs, and specific projects. In supporting the ISTEA and the Clean Air Act, as well as other federal legislation such as the National Environmental Protection Act, Title 6 of the Civil Rights Act of 1964, the Uniform Relocation Assistance Act, and the highway Safety Act, decennial census data facilitate a consistent level of responsible federal oversight and review of state and local plans and programs. For example, census data are an important tool in the environment review process required under the National Environmental Protection Act to assess the potential impacts of yet-to-be implemented projects. In consideration of the Clean Air Act, journey-to-work data from the 2000 census will provide important feedback on the overall effectiveness of today's national air-quality agenda. To respond to the requirements of the Americans with Disabilities Act for fully accessible transportation to all segments of the population, census data on persons with mobility limitations that are traditionally provided by the census provide an opportunity for the DOT to conduct a nationwide assessment of service needs. Intermodal Surface Transportation Efficiency Act Legislative Requirements The 1991 legislative requirements of the Intermodal Surface Transportation Efficiency Act (ISTEA) for census data include four specific categories, described below. Comprehensive Planning Provisions The ISTEA contains specific provisions requiring comprehensive transportation planning processes on a statewide basis, as well as at the metropolitan area level. States, local governments, and regional agencies must analyze the impacts of transportation plans, policies, and programs. The procedures involved are very data-intensive, and small-area data from the decennial census provide much of the required information. Principal among these procedures is travel forecasting. The function of transportation models is to replicate how people travel, to model their travel to and from different locations, by time of day, purpose, and mode. Models are used to forecast how people will travel in the future, with assumptions made about transportation infrastructure development and changes, land-use changes, parking cost and availability, and changes in individual travel behavior. By building these models, planners can evaluate different alternatives. For example, will adding carpool lanes along a particular highway reduce or increase congestion in the future and increase transit service? For most travel
OCR for page 315
Modernizing the U.S. Census models, the forecasting horizon is 20-30 years. Thus, data from the 1990 census are used to test the reliability of current models to predict 1990 travel behavior and to then forecast travel in 2000, 2010, and 2020. The decennial census provides the baseline of household and person characteristics, origins and destinations of work trips, and travel characteristics for small areas such as traffic analysis zones used in regional and local travel demand modeling efforts. These forecasts are used by state, regional, and local agencies to develop, test, and refine methods for projecting future travel needs at the regional, subarea, and corridor levels. Using these models for travel forecasting allows analysis of alternative highway, transit, and multimodal developments with various policy scenarios. In addition to supplying data for travel forecasting, the decennial census provides important information for transportation planners to monitor trends in travel behavior. Census data permit the tracking of travel times and peak hours of travel by mode of travel and by residence and work location. The census also provides estimates and data for trend analyses of rates of carpooling and public transit use in the journey to work. Transportation Improvement Program: Project Selection The ISTEA specifically requires that statewide and metropolitan transportation plans address broad issues such as land development and demographic growth, impacts of transportation facilities on population segments, and regional mobility and congestion levels. These plans must give consideration to the social, economic, and environmental effects, including air-quality effects, of transportation plans and programs. Projects contained in transportation improvement programs must be found to conform to the emissions reduction schedules in a state implementation plan. Census data on commuter travel flows and travel behavior patterns provide important baseline values against which transportation improvement program projects can be evaluated and selected. Traffic Congestion Management The ISTEA requires states, in cooperation with metropolitan planning organizations, to develop traffic congestion management systems. Transportation control measures and travel demand management programs often use census data on the journey to work as baseline values from which to establish goals for increasing average vehicle occupancy and for decreasing single occupant vehicles. Census data can also be used for preparing a comprehensive profile of peak period commuter flows by area and time of day.
OCR for page 316
Modernizing the U.S. Census Corridor Preservation The ISTEA provides a planning framework for early identification of transportation corridors needing some form of capacity expansion. Small-area data from the census provide a basis for defining these corridors and the number and characteristics of residents and jobs affected. Clean Air Act Amendments of 1990 Legislative Requirements Regions cited for being in nonattainment of federal air-quality standards must comply with Environmental Protection Agency and DOT requirements under the Clean Air Act Amendments of 1990 (CAAA). The transportation/air-quality planning requirements of the CAAA require state and local public agencies to prepare comprehensive vehicular travel and pollutant emissions profiles. To prepare these profiles requires analysis of detailed household and worker characteristics, means of travel, commuting patterns, and journey-to-work trip lengths obtained from the decennial census. The CAAA also requires severely polluted areas to compute regional average rates of vehicle occupancy in the commute to work. The census provides these data in a consistent manner nationwide. Under the CAAA, preparation of the state implementation plan and the comprehensive urban transportation planning process must be coordinated. Transportation facilities and projects proposed as part of the long-range transportation plan must be evaluated for their impact on air quality. Thus, forecasted travel volumes along specific routes are translated into forecasted traffic volumes, speeds, and subsequent emissions. The results are used in conformity analyses of state implementation plan. Data from the decennial census are used to calibrate and verify the travel simulation and forecasting models used to develop these parameters. Other Legislative Requirements Understanding regional travel patterns assists transit agencies in developing new services and revising existing services. These services may include vanpools and carpools, in addition to fixed-rail and fixed-route bus services. Small-area census data for traffic analysis zones on journey-to-work characteristics are used for route planning, market analysis, publicity, and advertising. The Americans with Disabilities Act requires states and local transit operators, with oversight and policy review by the DOT, to provide service levels that are fully accessible to all segments of the population. Data from the census that describe the geographic distribution of persons with disabilities that limit their ability to get around outside the home are used to develop and improve transportation services for this specific population.
OCR for page 317
Modernizing the U.S. Census STATE AND LOCAL USES Decennial census data for small areas such as census tracts and traffic analysis zones are used by states and metropolitan planning organizations to meet the provisions of the ISTEA, the Clean Air Act, and the Americans with Disabilities Act. In the survey described earlier, 11 of the 12 MPOs responding stated that they used all of the categories of the census data listed in the questionnaire with the exception of the employment data (75 percent utilization). Uses included application to all planning items required by the federal government, including the new requirements imposed by ISTEA and the Clean Air Act. The data were deemed either essential or very useful by a large majority of respondents for these purposes. Work-trip data by mode and origin-destination data, along with the special tabulations from the CTPP, were cited as the most difficult to replace. Home interview surveys, along with workplace and roadside or transit surveys would be required to replace the data. Five of the 12 MPOs reported that adequate staff and funding would be made available to replace the census data. State transportation agencies reported significantly higher planned use of the 1990 census data than the actual use of the 1980 data. (Special statewide tabulations were made available as part of the CTPP in 1990 for the first time). All items were reported used by a majority of the states responding for 1980 and by 8 of the 10 in 1990. At the time of the survey (in late 1993), state agencies were receiving 1990 data, but had not yet begun actual use. As was the case with the MPOs, state uses for the data included a heavy emphasis on inputs to travel forecasting models, including use as a base year, for the development of model parameters, and especially as a check of model results (9 of 10). These models are used extensively to provide data and forecasts in meeting Clean Air Act Amendment requirements. Heaviest use of the census data was for corridor studies, alternatives analysis, and transit planning application (all 8 of 10). Nine of the 10 states responding reported that the data were most useful for statewide and regional transportation planning. Other items cited were congestion management (a new federal requirement along with state-wide planning), transit planning, and alternatives analysis. Work-trip data by mode were cited as the most difficult to replace by 8 of 10 states. Besides home interview surveys, the use of secondary source data would be the most likely way that states would replace the census data. Five of the 10 states stated that it would be unlikely that adequate funding would be made available to replace the census data.
OCR for page 318
Modernizing the U.S. Census COST AND FEASIBILITY OF REPLACING CENSUS TRANSPORTATION DATA At a sampling rate of 1 in 6 households and a cost of $2.6 million to prepare, program, tabulate, and distribute the CTPP, the unit cost to provide the journey-to-work and associated data to states and MPOs totals less than 25 cents per household interviewed. The majority of state and metropolitan agencies surveyed indicated that they would substitute home interview surveys if census data were not available. Costs of urban home interview surveys generally range between $70 and $100 per household. Because of the high cost, sampling rates for home interview surveys are also much lower than that of the census, typically on the order of 1 in 20 (0.5 percent). The 1990 CTPP also includes statewide summaries as well as urban. Home interview surveys collect data on all trips made by the household, however, so a direct cost comparison is difficult. Most urban areas collect home interviews in addition to census data. Because of the detailed geography (tract and traffic zone) and mode of travel information, other work-related surveys that do not provide such information are not comparable. Even if home interviews are conducted or an alternative survey methodology is found, they would not be national in scope, but only state or local in nature. Many agencies have indicated that they lack money or staff to replace the census. Transportation planning agencies require such detail since the mathematical models using the data are in widespread use to simulate a base year and then to forecast future travel by mode. Approximately 80 percent of the states and MPOs surveyed plan to use the 1990 census journey-to-work data to provide inputs to such models and to establish a base year. Approximately 90 percent of the agencies plan to use the data as a check of such models and to use the data in transit and corridor studies. Perhaps the only real alternative to the provision of census journey-to-work information would be to greatly expand the Nationwide Personal Transportation Survey, now conducted every 5 years by the DOT. Its present sample size of about 25,000 households would have to be expanded by at least a factor of 20 to provide even a minimal sample to obtain adequate geographical data suitable for use by state and metropolitan agencies. That cost is estimated at $50 million, or 20 times the current cost of the Census Transportation Planning Package. The high levels of current use by both state transportation and metropolitan planning agencies, along with the low cost of obtaining and disseminating the data, lead to the conclusion that abandoning the tradition of nationwide survey uniformity and geographic consistency of data provided by the decennial census would result in high costs and disruption to program development and evaluation at all levels of government.
OCR for page 319
Modernizing the U.S. Census NOTE 1 The panel acknowledges the assistance of Philip Fulton, associate director of the Department of Transportation's Bureau of Transportation Statistics, and George V. Wickstrom, a transportation planning consultant in Kensington, Maryland, for their assistance in preparing this appendix. Dr. Fulton prepared materials on the history of transportation data in the census and the use of census data to meet federal legislative requirements. Mr. Wickstrom coordinated the survey of state transportation and metropolitan planning agencies of their use of census data for their programs. Both Dr. Fulton and Mr. Wickstrom provided drafts of their work that were used in the panel's deliberations and in preparation of this appendix. REFERENCES Delaware Valley Regional Planning Commission 1994 The Journey-to-Work Trends in the Delaware Valley Region 1970-1980. Philadelphia, Pa.: Delaware Valley Regional Planning Commission. McDoneell, J.J. 1984 The urban transportation planning package. Pp. 11-15 in Census Data and Urban Transportation Planning in the 1980s. Transportation Research Record 981. Transportation Research Board, National Research Council. Washington, D.C.: Transportation Research Board. Sosslau, A.B. 1983 Transportation Planner's Guide to Using the 1980 Census. Washington, D.C.: Federal Highway Administration, U.S. Department of Transportation. U.S. General Accounting Office 1993 Transportation Infrastructure: Better Tools Needed for Making Decisions on Using ISTEA Funds Flexibly. GAO/RCED-94-25. Washington, D.C.: U.S. General Accounting Office. Wickstrom, G.V. 1984 Experience with the 1980 census urban transportation planning package in the Washington metropolitan area. Pp. 91-95 in Census Data and Urban Transportation Planning in the 1980s. Transportation Research Record 981. Transportation Research Board, National Research Council. Washington, D.C.: Transportation Research Board. 1994 The Use of Decennial Census Data in Transportation Planning. May. Paper prepared for the Bureau of Transportation Statistics, U.S. Department of Transportation, Washington, D.C.
OCR for page 320
Modernizing the U.S. Census TABLE G.1 Type of Transportation Data Used by State Transportation Agencies and Metropolitan Planning Agencies (22 agencies reporting) Number Percentage Items 1980 1990 1980 1990 Type of Data Used Household, population 17 19 77 86 Other socioeconomic 18 21 82 95 Employment 15 17 68 77 Trip characteristics 18 20 82 91 Origin-destination 18 19 82 86 Cross-tabs (CTPP) 16 18 73 82 Uses for the Data As inputs to models 16 18 73 82 As a base year 12 18 55 82 To develop model parameters 13 18 59 82 As a check of model results 14 19 64 86 Purpose of Use Long-range planning/TIP 15 16 68 73 Corridor studies 16 0 73 91 Alternatives analysis 15 18 68 82 Transit planning 15 0 68 91 Congestion managementa — 16 — 73 TCM analysisa — 15 — 68 Air quality evaluation 13 16 59 73 Conformity determinationa — 15 — 68 Monitoring 12 15 55 68 Other 1 2 5 7 a These purposes are new federal requirements that were added to census data requirements by legislation enacted since the 1980 census. Source: Wickstrom (1994: Table 4).
OCR for page 321
Modernizing the U.S. Census TABLE G.2 Usefulness and Alternative Sources of Transportation Data for State Transportation Agencies and Metropolitan Planning Agencies (22 agencies reporting), 1990 Census Items Number Percentage Usefulness of the Data for (essential or very useful): Statewide and regional transportation plan 20 91 Program development 14 64 Conformity requirements 14 64 Congestion management 15 68 Transit planning 15 68 Traffic control measures 11 50 Alternatives analysis 17 77 Monitoring and evaluation 11 50 Other 4 18 Difficulty to Replace Data (very difficult): Socioeconomic data (by zone) 13 59 Employment data (by zone) 8 36 Work-trip data by mode (by zone) 18 82 Origin-destination data (by zone) 17 77 Special tabulations in CTPP 15 68 Most Likely Way to Replace Data: Conduct or expand home interviews 19 86 Conduct workplace surveys 13 59 Conduct roadside or transit surveys 13 59 Use secondary source data 13 59 Other new data collection 7 32 Could Funds and Staff Be Made Available Yes 1 Probably 11 Unlikely 10 Source: Wickstrom (1994: Table 5).
Representative terms from entire chapter: