1995 INTRODUCTION growing industrial sectors in the United States. In many areas of the country, travel-related services The Intermodal Surface Transportation Efficiency form the primary industries for providing income AMERICAN Act of 1991 (ISTEA) created the Bureau of Transpor- and jobs for many of its residents. In view of the im- tation Statistics (BTS) in the Department of Trans- portance of travel and transportation to the eco- TRAVEL portation (DOT). In December of 1992, the BTS was nomic and social well-being of America, formally established to compile, analyze and pub- comprehensive and timely data relating to this sec- SURVEY lish statistics; to develop a long-term data collection tor are essential to government, private industry, program; to develop guidelines to improve the cred- and others responsible for policy formation and pro- ibility and effectiveness of the DOT’s statistics; to rep- AN OVERVIEW OF THE SURVEY motional activities in the general field of travel and resent transportation interests in the statistical DESIGN AND METHODOLOGY tourism. community; and to make statistics accessible and The main objective of the ATS is to provide policy understandable. At about the same time, the Trans- makers with more accurate and comprehensive in- portation Research Board of the National Academy formation than has previously been available about of Sciences in its report, Data for Decisions: Re- the travel of persons and households in the United quirements for National Transportation Policy States and about the principal characteristics of Making, concluded that the biggest gap in the DOT’s travel and travelers. The survey will provide key data intermodal data program was in “flow data.” The most recent source of data on passenger flows dates back to the 1977 National Travel Survey, conducted by the Bureau of the Census as a component of the quinquennial Census of Transportation. To meet the need for passenger flow data, the BTS conducted the 1995 American Travel Survey (ATS). The ATS field work was performed by the Census Bu- reau, under its authority to conduct a Census of Transportation as stated in Title 13 of the U.S. Code. In the future, BTS plans to conduct the survey every five years. This paper describes the methods used in the to assist in the formulation and evaluation of initia- 1995 ATS. The introduction provides an overview of tives in intermodal transportation. Because policy- the purpose and objectives of the survey followed by makers have a particular interest in understanding a description of the survey and sample designs, sur- the differences in travel patterns by state, the ATS will vey field operations, and processing of survey data. provide detailed information on state-to-state travel The intention of this paper is to provide a reference as well as travel to and from metropolitan areas by to data users that will aid in the analysis of the data mode of transportation. Data will also be available when they become available. for subgroups defined in terms of characteristics re- lated to travel, such as business, vacation or com- A. SURVEY OBJECTIVES bined/business pleasure travel by age, family type, labor force status and income. Subgroups of special The American Travel Survey was designed to ob- interest may include, for example, frequent travel- tain information about long-distance travel of per- ers, trips taken by children traveling without a par- sons living in the United States. The information is ent, and trips taken by retirees. needed to identify characteristics of current use of To achieve survey objectives, both cross-sectional the nation’s transportation system, forecast future and longitudinal estimates were needed. Key cross- demand, analyze alternatives for investment in and sectional estimates include the origins and destina- development of the system, and assess the effects of tions of trips, the proportions of people traveling on Federal legislation and Federal and state regulations various transportation modes, intermodal connec- Susan J. Lapham on the transportation system and its use. tions, reasons for trips, trip duration, trip distance, Bureau of Transportation Economic activity generated from people travel- Statistics, U.S. Department and person and household characteristics that may of Transportation ing represents one of the largest and most rapidly influence aggregate travel deman during a particu- 1995 AMERICAN TRAVEL SURVEY An Overview of the Survey Design and Methodology 1 lar time period. Longitudinal estimates require the ever, information about trips taken up until the collection of information about travel behavior or time a respondent became ineligible was collected. the members of households and persons over the en- Persons who moved into the original sample address tire survey year. For example, it may be desired to after the initial interview were interviewed and a se- measure the cumulative effect of patterns of travel ries of brief questions was asked about all trips taken on the transportation system or the patterns of since January 1, 1995. This retrospective travel in- change in travel by season. formation was used to make weighting adjustments in the annual estimates. Households not interviewed for any reason during B. SURVEY AND SAMPLE DESIGN a particular interview period were contacted and in- 1. Survey Design terviewed at a subsequent time and asked about travel taken during the missed travel quarter. Be- ATS uses a quinquennial design, collecting data cause the recall period was longer for these few re- for calendar years ending in zero and five. A sample spondents, trip information was limited to the of households was selected at the beginning of cal- number of trips, mode of transportation for each endar year 1995. Interviewing of households in the trip, primary purpose of each trip, and the origin 1995 survey, which contained about 80,000 eligible and destination of each trip. Again, the retrospective addresses, began in April 1995 and ended in March 1996. Sample households were interviewed four times during this period, at approximately three- month intervals. The survey was based on a probability sample of households from each of the 50 states and the Dis- trict of Columbia. The survey population for ATS consisted of persons resident in households and per- sons living in group quarters, such as dormitories, rooming houses, religious groups dwellings, and family-type housing on military bases. Persons liv- ing in military barracks and in institutions, such as prisons and nursing homes, were excluded. information on these trips are used only to make adjustments in annual data. 2. Sample Design C. QUESTIONNAIRE CONTENT Sample selection for ATS was based on lists of ad- dresses compiled by the Census Bureau from the de- Basic demographic characteristics and other cennial censuses of population. In addition to the classification variables associated with a household ATS, the Census Bureau uses address-based lists to and its members were recorded during the initial in- provide samples for all of its major household sur- terviews for the survey and were updated or verified veys, including the Current Population Survey, Na- in each subsequent interview. Social and economic tional Crime Survey, and the American Housing characteristics included age, sex, marital status, Survey. race, household type, Hispanic origin, education, la- After the sample of addresses was selected for the bor force status, and income. Detailed information ATS, rules were established to determine which per- about each trip taken by each member of the house- sons should be included in initial interviews at those hold was collected quarterly. The main trip charac- sample addresses. At the initial interview, all persons teristics included the purpose of trip, means of whose usual residence was the sample address were transportation, origin, destination, intermediate included and remained in sample as long as they stops, travel dates, trip duration, number of nights remained at the sample address. Persons who left away, and types of lodging used. Travel distances for the survey population through death, going abroad, each trip were assigned based on transportation net- or changing residence to another address were con- work routing algorithms. sidered ineligible for subsequent interviews. How- 1995 AMERICAN TRAVEL SURVEY An Overview of the Survey Design and Methodology 2 Household-level questions were asked only once the start of the first interview, the interviewer re- for each household and were updated if necessary corded the basic demographic characteristics for throughout the survey year. The detailed trip ques- each person residing at the sample address, plus a tions were asked of the initial respondent. If this re- few characteristics of the household and the hous- spondent was knowledgeable about other household ing unit. The telephone number and address were members’ travels, he or she served as a proxy re- verified and recorded in case callbacks were needed spondent for other household member’s trips. Oth- to obtain additional information. The interview was erwise, the individual household member was conducted on a trip-by-trip and person-by-person contacted and interviewed. Trip characteristics were basis until the questionnaire was completed for all asked about each trip reported by the respondent for trips taken by each household member. each member of the household. Other instruments used in the data collection pro- cess were a series of advance letters and one phone call. These