Click for next page ( 61


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



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 60
60 CHAPTER EIGHT CONCLUSIONS AND FURTHER RESEARCH On-board and intercept surveys are used by virtually all Personal interviewing tends to achieve a higher response transit agencies to collect data on customer trip characteris- rate, fewer item nonresponses, and possibly better under- tics, customer ridership and demographic characteristics, standing of questions than self-administered surveys. and customer attitudes about service. Survey results are used for a wide variety of purposes including travel modeling, On-board and intercept surveys are highly flexible and areawide and route planning, scheduling, marketing, and adaptable to the project purposes and characteristics of the customer communications. Transit agencies reported that agency's service configuration and of the study population. survey results can be highly useful, accurate, and timely. Given the need to tailor survey methodology to the particu- larities of the situation, there is no uniform way to carry On-board and intercept survey methodologies are essen- out on-board and intercept surveys. In designing a survey to tial to transit agencies in a variety of situations. These fulfill project objectives, maximize data quality, and control methodologies may be the only cost-effective way to gather costs, a number of factors need to be considered. Important information from riders where the incidence of transit users considerations relating to project purposes include: in the general population is low. In major cities with a high incidence of transit users among the citizenry, on-board and What is the study population?--Whether the study pop- intercept methodologies are highly useful for surveys on ulation is all riders, riders on a particular route, or those specific routes or among specific customer segments. using a particular station will affect the choice of on- board versus in-station locales. Other major strengths of on-board and intercept surveys What amount and level of detail of information is include the ability to obtain a representative sample of the needed?--Longer surveys need to be conducted in targeted population; the wherewithal to obtain accurate, situations where respondents have sufficient time to reliable, and detailed information from riders; and the complete the survey for immediate return or are likely means to survey during the immediate experience of the to complete the survey at a later time for return service. On-board and intercept surveys often provide through the mail. The ability of respondents to provide higher response rates than alternative methodologies such accurate detailed information and the level of accu- as telephone, mail, and on-line surveys, and at lower cost. racy needed will affect the choice of personal inter- views or self-completed surveys. Although offering many advantages, on-board and intercept What aspects of the customer experience does the surveys are not the optimal methodology in a variety of situa- survey ask about?--For example, surveys may need tions. Telephone or other methodologies are necessary when to take place proximate to customers' experience of a the objective is to survey non-users. On-board and intercept ticket vending machine. surveys also cannot be used when the survey questionnaire is Considerations affecting data quality are: extensive or complex. Response rates--Response rates affect the amount of Surveys are typically returned to survey workers, but may nonresponse error, which arises from the possibility also be deposited in envelopes or boxes on board or through that those not completing a survey would have the mail. answered differently than those who did complete a survey. Response rates are affected by a broad range of Self-administered surveys are sometimes distributed at factors including the enthusiasm and diligence of sur- transit centers and transfer points instead of on board. This vey workers who distribute questionnaires or conduct approach is particularly suitable when most or all of the study interviews, the level of rider interest, whether self- population passes through the survey locations. administered surveys or personal interviews are used, length and complexity of the questionnaire, use of On occasion agencies use personal interviews instead of incentives, and the frequency of surveys being con- self-administered surveys. Personal interviews can be con- ducted. Other factors, less subject to transit agency ducted either on-board or in rail stations, transit centers, or at influence, are rider income, education level, and other transfer points. The interviews are generally relatively brief. demographic characteristics and rider literacy levels.

OCR for page 60
61 Clarity and organization of the questionnaire--Word- sample; determining the sample size needed for the desired ing of questions, question ordering, and layout can level of precision in the results; and stratification and weight- affect the ease of completing the questionnaire and the ing of subgroups being surveyed based on route, time of day, accuracy and completeness of responses. direction, and other factors. Additional important tasks are How and where the survey is conducted--Data quality is recruitment, training, and supervision of survey workers; affected by whether self-completed questionnaires or per- ensuring worker safety; and data cleaning, data processing, sonal interviews are used, where the surveying is carried data tabulation, and analysis. out, and the means of returning self-administered surveys. In on-board and intercept surveys, the devil is in the Costs for on-board and intercept surveys vary widely, details. Although transit agencies have developed many even for surveys conducted using nearly identical method- effective practices through experience, there is insufficient ologies. Factors that affect costs are: methodologically sound research to guide decisions in two key areas: (1) impact of design and layout of questionnaires Whether the survey is conducted by personal interview and (2) impact of the use of incentives. Additional research or uses self-administered questionnaires. is needed to explore these two issues. The research should Length and complexity of the survey. formulate and test different questionnaire designs and Whether the survey collects detailed origin and desti- different incentive levels to measure the affects of these nation surveys, which tend to incur higher costs for factors on response rates, item nonresponse, and, for origin geocoding and data processing. and destination information, the quality of address and Total number of completed surveys (larger surveys may trip information. Alternative designs could also be tested benefit from economies of scale). to determine the impact of questionnaire length, use of Density of riders (e.g., number of riders per hour on matrixes, and use of horizontal versus vertical lists of surveyed bus or train routes or in stations). answer choices. Response rates. Whether self-administered questionnaires are distrib- Ideally, alternative questionnaire designs and incentive uted by dedicated survey staff or bus operators. levels should be tested on routes (or stations) where charac- teristics of the sampled population are held constant, to Planning and implementing surveys also involves deci- derive reliable conclusions about the impacts, if any, of alter- sions about identifying the study population and selecting the native layouts and incentive levels.