National Academies Press: OpenBook

On-Board and Intercept Transit Survey Techniques (2005)

Chapter: Chapter Three - Planning On-Board and Intercept Surveys

« Previous: Chapter Two - Transit Agency Use of On-Board and Intercept Methodologies
Page 9
Suggested Citation:"Chapter Three - Planning On-Board and Intercept Surveys." National Academies of Sciences, Engineering, and Medicine. 2005. On-Board and Intercept Transit Survey Techniques. Washington, DC: The National Academies Press. doi: 10.17226/13866.
×
Page 9
Page 10
Suggested Citation:"Chapter Three - Planning On-Board and Intercept Surveys." National Academies of Sciences, Engineering, and Medicine. 2005. On-Board and Intercept Transit Survey Techniques. Washington, DC: The National Academies Press. doi: 10.17226/13866.
×
Page 10
Page 11
Suggested Citation:"Chapter Three - Planning On-Board and Intercept Surveys." National Academies of Sciences, Engineering, and Medicine. 2005. On-Board and Intercept Transit Survey Techniques. Washington, DC: The National Academies Press. doi: 10.17226/13866.
×
Page 11
Page 12
Suggested Citation:"Chapter Three - Planning On-Board and Intercept Surveys." National Academies of Sciences, Engineering, and Medicine. 2005. On-Board and Intercept Transit Survey Techniques. Washington, DC: The National Academies Press. doi: 10.17226/13866.
×
Page 12
Page 13
Suggested Citation:"Chapter Three - Planning On-Board and Intercept Surveys." National Academies of Sciences, Engineering, and Medicine. 2005. On-Board and Intercept Transit Survey Techniques. Washington, DC: The National Academies Press. doi: 10.17226/13866.
×
Page 13
Page 14
Suggested Citation:"Chapter Three - Planning On-Board and Intercept Surveys." National Academies of Sciences, Engineering, and Medicine. 2005. On-Board and Intercept Transit Survey Techniques. Washington, DC: The National Academies Press. doi: 10.17226/13866.
×
Page 14
Page 15
Suggested Citation:"Chapter Three - Planning On-Board and Intercept Surveys." National Academies of Sciences, Engineering, and Medicine. 2005. On-Board and Intercept Transit Survey Techniques. Washington, DC: The National Academies Press. doi: 10.17226/13866.
×
Page 15
Page 16
Suggested Citation:"Chapter Three - Planning On-Board and Intercept Surveys." National Academies of Sciences, Engineering, and Medicine. 2005. On-Board and Intercept Transit Survey Techniques. Washington, DC: The National Academies Press. doi: 10.17226/13866.
×
Page 16
Page 17
Suggested Citation:"Chapter Three - Planning On-Board and Intercept Surveys." National Academies of Sciences, Engineering, and Medicine. 2005. On-Board and Intercept Transit Survey Techniques. Washington, DC: The National Academies Press. doi: 10.17226/13866.
×
Page 17

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.

9On-board and intercept survey projects typically begin with an expressed need. Transit agency senior management asks the basic question, “How are we doing?” A survey is then commissioned to measure customer satisfaction. The mar- keting department wants to measure customer awareness of a recent advertising campaign or determine how often riders are visiting the agency website. The service planning depart- ment needs O&D data to develop a computer model of ridership patterns for use in route planning or wants to know how customers feel about a recent routing change. The bus department would like to know how customers reacted to recently introduced low-floor buses or wants to test a new bus lighting system under consideration. Planning an on-board/intercept survey requires defining project goals, choosing where and how to conduct the sur- vey, identifying the study population and sampling frame, and deciding what degree of precision is needed in the results. This chapter addresses these issues. PROJECT GOALS Planning an on-board/intercept survey involves first and foremost consideration of the survey goals. What informa- tion does the survey need to collect? Overall, the goals or purposes of on-board and intercept surveys provided by transit agencies address five major areas: • Travel modeling, • Long-range and areawide planning, • Route planning and scheduling, • Marketing, and • Customer communications. Surveys typically ask questions in several rather than just one of these areas, although rarely in all five. O&D surveys are generally used in route planning, long-range planning and, to a lesser extent, scheduling and modeling (Table 6). Surveys covering customer attitudes and demographics address areas of areawide and route planning, marketing, and customer communications. Project purposes can be specifically targeted to upcoming decisions or integrated into a planning process. In these cases, survey methodology and selection and wording of questions can be developed in a strict decision framework. What methodology is best suited to this particular purpose? What questions will be used in the decision or planning process? Why is the information generated by each question useful? By asking these questions, unnecessary or extrane- ous questions can be dropped and the survey instrument can be streamlined. Examples of targeted project purposes provided by transit agencies responding to the survey conducted for this report are: • Track system transfer rate and distribution of fares [TriMet (Portland, Oregon) Annual Fare Survey]. • Measure the implementation of a new mode of trans- portation in service area (Massachusetts Bay Transporta- tion Authority Silver Line Project). • Determine customer perceptions of route changes and their travel patterns (Metropolitan Transportation Authority New York City Transit B15/Q3 Study). • Evaluate summer loop service that uses trolley replica vehicles (Milwaukee County Transit System Trolley Evaluation). • Measure ride free area ridership (to use in conjunction with automatic passenger counter data) (King County Transit Ride Free Area Survey). • Analyze travel habits of bus customers traveling on the West Side of Chicago and in the near west suburbs (CTA West Side Customer Travel Survey). • Determine if a proposed route change would result in more or fewer transfers [Greater Portland (ME) Transit District Route 1 O&D Study]. Surveys may also have the more generalized purpose of profiling the agency’s riders and their experience of transit service. Agencies reported that the information is useful in describing who benefits from transit service; comparing with the overall population and with peer markets and showing historical trends. Examples of relatively generalized survey purposes are: • To determine overall customer evaluation of Dallas Area Rapid Transit (DART) service quality, mar- keting, and scheduling (DART Customer Satisfaction Survey). CHAPTER THREE PLANNING ON-BOARD AND INTERCEPT SURVEYS

• To assess ridership behaviors, attitudes, and usage [Sub- urban Mobility Authority for Regional Transportation (SMART) Transit User Survey]. Finally, it is not unusual for surveys to have both specific and relatively general purposes. Examples are: • To collect O&D information, including travel purpose information for the regional travel-demand model; to obtain a rider demographic profile; and to learn more about our ridership needs and travel habits and patterns to better serve those needs through transit planning [Citizens Area Transit (Las Vegas, Nevada) Origin and Destination Survey]. • To assess who our customers are and to generate input for the passenger estimation model [Metropoli- tan Atlantic Rapid Transit Authority (MARTA) Sys- temwide Survey]. WHERE AND HOW TO CONDUCT SURVEYS Central decisions in survey planning involve whether to have riders complete the surveys themselves or to conduct personal interviews, where to conduct the survey (on board or in sta- tions), how to distribute and collect questionnaires in the case of self-administered surveys, and whether to offer any type of incentives. Each of these issues is discussed in this section. Self-Administered or Personal Interviews? Transit agencies use self-administered surveys more often than personal interviews. More than two-thirds of the surveys reported by transit agencies for this study were entirely or primarily self-administered, whereas only 20% were entirely conducted through personal interviews. (The remaining 11% were partially self-administered and partially conducted as personal interviews.) By definition, self-administered surveys involve respon- dents completing survey forms themselves and then returning the forms to the agency. For most of the surveys reported by transit agencies, questionnaires were distributed by survey workers dedicated to the task. Other options are to use other employees, such as bus operators, or to make questionnaires available in a convenient location, such as on board the bus, for riders to pick up. 10 Table 7 summarizes the advantages and disadvantages of self-administered surveys and personal interviews, and the types of situations in which transit agencies tend to use one or the other method. A primary strength of self-administered surveys is that a number of respondents can complete the survey simulta- neously. Survey workers’ time can be used efficiently because survey staff can distribute questionnaires to a number of riders as the riders board the bus. Therefore, fewer surveyors are needed to obtain a given number of surveys than with personal interviews. Furthermore, a questionnaire can be offered to every rider, avoiding the need to sample those boarding a bus or train or passing through a station. Coverage is therefore maximized and sampling error is minimized. Self-administered surveys must be carefully designed so that respondents can easily understand the questions, follow the flow of the questionnaire, and respond accurately. Without a trained interviewer to guide respondents through the survey the importance of good questionnaire design is paramount. (Chapter five discusses design issues.) Self-administered surveys also need to avoid or minimize complexities such as skips between questions. Interviewer- administered surveys can more readily skip selected questions based on answers provided to previous questions. Self- administered surveys are thus most appropriate when the research need is to ask all respondents a uniform set of ques- tions. Good examples are simple O&D questions, customer satisfaction, and customer opinion surveys. The use of self-administered methods can exact a price in data quality. Respondents may misunderstand questions, leading to measurement error. Respondents may also com- plete only a portion of the questionnaire. Survey workers have very limited ability to ensure that respondents complete the questionnaire. Unless asked a question, survey workers cannot usually ensure the accuracy of responses or check the logic or consistency of answers. Personal interviews reduce or eliminate problems with respondent understanding of questions and with item non- response. Interviewers can also skip questions based on previous responses, particularly when using handheld computers or similar devices. Personal interviews can generate a very high response rate. Response rates, computed as interviews completed as a per- centage of persons approached, exceeded 80% for surveys con- ducted by MARTA, King County Metro, and Gulf Regional Planning Commission (GRPC, Gulfport, Mississippi). Because they are more time-intensive than self- administered surveys, personal interviews tend to be used Use Route planning 86 Long-range planning 76 Schedule planning 43 Modeling Other Total number responding, 21. Percentage 38 24 TABLE 6 ANALYTIC USES FOR ORIGIN AND DESTINATION DATA

11 when the questionnaire is short and when a relatively small sample size is needed. The King County Transit Ride Free Area survey consisted of a few basic questions about trans- fers, fare payment, and boarding and alighting. The Port Authority Trans Hudson (PATH) (Jersey City, New Jer- sey) O&D survey was seven questions regarding O&D, access mode, transfers, trip purpose, and trip frequency. Personal interviewing is important when some respon- dents do not have the literacy or English language skills necessary to complete a written questionnaire. GRPC offered riders a choice of being interviewed or completing the sur- vey; most chose to be interviewed, in part owing to language skill considerations. The Charlotte Area Transit System (CATS) also offered the option of either self-administered or interviews; approximately 15% of riders completing the survey asked to be interviewed, primarily persons with disabilities that made self-completion difficult. A Fort Worth Transportation Authority survey was primarily self- administered; however, a bilingual surveyor, assigned to bus routes with significant Hispanic ridership, conducted some interviews in Spanish. PATH interviewers used handheld personal digital assis- tants to record respondents’ gender and their answers to the seven survey questions. The survey was programmed into the handheld devices for this purpose. This was the only survey among those reported by transit agencies that used electronic devices for recording respon- dents’ answers. The use of electronic devices appears not to be widespread for two main reasons. First, they are not feasible for most self-administered surveys. Second, trials of electronic devices for on-board interviews found that inter- viewers had difficulty entering information on a moving bus. However, according to one company that specializes in this area, palm devices have been used successfully for intercept interviews at special events and of recreational visitors. Thus, as the PATH experience illustrates, they may be prac- tical and useful for in-station environments. Self-Administered Personal Interview Strengths • Need fewer surveyors to obtain a given number of completed surveys because multiple respondents can complete the survey simultaneously • Can potentially survey all riders boarding a bus or train; no need to select a sample from among those boarding • Higher level of respondent understanding of questions • Ensures that all questions are answered • Obtains responses from persons with limited literacy skills Weaknesses • Respondents may misinterpret questions (measurement error) • Respondents may not complete the entire questionnaire (item nonresponse) • May result in lower response rate than personal interviews • Depends on ability of respondent to read questionnaire • Difficult to use branching and skip patterns • Can be time-intensive for surveyors; may need larger number of surveyors • Possible bias from nonrandom selection of riders interviewed • Cost Situations likely to be used • Projects needing large number of respondents • Same questions asked of all respondents • Relatively long questionnaires • Short questionnaires • Need relatively smaller number of completed surveys • Respondents unable to complete survey due to language, literacy, and/or disability • Use as adjunct to self-administration at choice of respondent Implications for survey planning • Survey instrument must be well designed, with clearly worded questions and clear navigation • Length of survey may need to be minimized • Need to interview riders where they will take the time needed to complete the interview TABLE 7 CHARACTERISTICS OF SELF-ADMINISTERED SURVEYS AND PERSONAL INTERVIEWS

On-Board or In-Station Surveying? Closely related to the choice between self-administered sur- veys and personal interviews is the choice between adminis- tering the survey on board or in a stationary environment (e.g., a rail station or transit center). Seven in 10 on-board/intercept transit surveys are con- ducted on board buses or trains, whereas 20% are conducted in stations and the remainder used both venues. Transit agen- cies choose the on-board approach for a variety of reasons. • A steady flow of riders pass survey workers. On buses in particular and trains to some degree all riders pass through a single point to pay their fare or show a pass. This situation provides a good opportunity for survey workers to offer a questionnaire to each and every person boarding the bus or train. • There is adequate time for respondents to complete the survey. The “cost” of completing the survey tends to be relatively low, because respondents are spending time on the bus or train anyway. • Questionnaires can be returned on the spot. Most respondents can complete the questionnaire before leaving the bus or train and return it in person. Surveys completed and returned in this manner will not be mis- placed or forgotten, as may happen with surveys taken off the vehicle to be completed later. • It is a relatively safe environment for surveyors. The presence of bus operators and train crews provides a measure of safety for survey workers, whereas subway stations or bus stops in untrafficked areas may raise issues of personal safety for survey workers. • It facilitates obtaining a representative sample of tran- sit riders. The advantage of a moving bus or train is that it picks up passengers along an entire route. Therefore, a survey may be more likely to obtain a representative sample of riders by surveying on board rather than at bus stops or in subway, light rail, or commuter rail stations. This is an important consideration for any O&D study and for any study in which rider character- istics vary significantly between neighborhoods. The choice of conducting surveys in stations or transit centers tends to be driven by circumstances that are particu- lar to the survey or transit property. When nearly all riders pass through a downtown transit center or a few transfer points, it may be most time-efficient for survey staff to work at these central hubs rather than be spread out across a sub- stantial number of buses. Some surveys are conducted in rail stations. TriMet con- ducted its Ticket Vending Machine (TVM) Fascia Redesign survey immediately after respondents had used the TVM. Furthermore, TriMet chose the airport station to interview riders who were not experienced TVM users. 12 Station interviewing may also be attractive when the num- ber of stations is relatively small, as with the PATH O&D survey. Notably, these surveys were conducted as personal interviews and were relatively short, so as to minimize the length of the interruption in respondents’ trips. Distribution and Collection of Self-Administered Surveys Self-administered surveys tend to be distributed and col- lected by survey staff assigned for this purpose. One or two survey workers are typically assigned to each bus or train car; rail surveys are more likely than are bus surveys to use two surveyors per car. Surveyors offer a questionnaire and pen or pencil to passengers as they board. Passengers return questionnaires to survey workers or place completed questionnaires in envelopes provided at the front and rear doors. Questionnaires often provide the option of return by mail. The mail-back option is particularly important if respon- dents do not have sufficient time to complete the survey on board, if they may need help from a family member or friend, or if there is some extenuating circumstance such as they simply need their glasses to be able to complete the survey. Mail-back questionnaires include a postage-paid mailing address on one panel; riders can fold and staple or tape the questionnaire and drop it in the mail. Use of the mail option varies. For some surveys, the Washington Metropolitan Area Transit Agency (WMATA) and CTTran- sit (Hartford, Connecticut) reported receiving more than 30% of returned surveys by mail, whereas fewer than 10% of total surveys were returned by mail in two CTA surveys and in a TriMet survey. Some surveys are distributed by bus operators or are left on seats or in timetable holders. One-quarter of the on-board surveys reported by transit agencies involved bus operators distributing the questionnaires (including surveys distributed by a combination of bus operators and survey workers). Agencies that use bus operators exclusively to distribute surveys tend to be smaller properties such as Pace Suburban Bus, SMART, Central Florida Regional Transportation Authority (known as LYNX), city of Lodi (California), and the Capital District Transportation Authority. Apparently, use of bus operators has been more successful at these smaller agencies; two larger agencies commented that bus operators have been too busy to distribute questionnaires to all riders. There may also be union issues of operators work- ing “out of title” if asked to distribute surveys. Prior Notice to Riders For 46% of the surveys, the agencies notified passengers about the survey in advance through media, on-board

13 announcements, or other means. The most frequent means were on-board announcements, “take-one” brochures, car cards, and signs. Other means included press releases, print and electronic media, newsletters, and the Internet. Several agencies commented that the advance notice helped boost response rates. Incentives Incentives are often offered to riders to encourage participa- tion in the survey and at times to encourage full and accurate completion of key survey questions. One-quarter of the transit agency surveys used some type of incentive, ranging from a free pass to a free pen to being entered in a lottery or drawing. Transit agency staff interviewed believes that incentives do increase response rates and that incentives provided to all respondents at the time of the survey are more effective than lotteries or drawings. However, there is little systematic evidence to document this belief or quantify the magnitude of the effect for transit surveys. Extensive research on the use of incentives in mail surveys provides useful experience in a similar realm. Dillman (2000) reported that including $1 or $2 with mail surveys has been shown to increase response rates by 12 to 31 percentage points. In mail surveys, incentives are more effective when sent with questionnaires rather than as a later reward; “token financial incentives of a few dollars” sent with mail questionnaires “have been shown to be significantly more effective than larger payments sent to respondents who have returned the questionnaires” (Dillman 2000). Incentives provided with the questionnaire constitute a “goodwill ges- ture that puts the sponsor and questionnaire in a positive light and sets the stage for the respondent to reciprocate with an appropriate gesture of completing the questionnaire.” By contrast, prizes have “relatively small, if any, effect on response” (Dillman 2000). IDENTIFYING STUDY POPULATION AND DRAWING THE SAMPLE Survey methodologists distinguish between the theoretical population, study population, sampling frame, and sample (Trochim 2004). These are important distinctions and, although not complicated for most on-board and intercept surveys, need to be carefully considered. Theoretical Population The population that the researchers wish to generalize to is the theoretical population. In a survey of voter preference, for example, the theoretical population would be persons who will vote in the next election. The theoretical population in on-board and intercept transit surveys is typically composed of bus, rail, and other transit users. The theoretical popula- tion may be all riders or a subset such as riders on a particu- lar route or traveling in a particular area or at a particular time of day. Study Population The population that the researcher can gain access to is the study population. In a survey of voter preference, the study population might be voters with telephones. In that case, the study population differs somewhat from the theoretical population because not all voters have telephones. The study population in on-board and intercept transit surveys is often congruent with the theoretical population, because all riders can, at least in theory, be reached on board transit vehicles or at transit stops or stations. This is one of the major advan- tages of on-board and intercept surveys—the survey can reach all riders whether or not they live in the service area, or have telephones, or are literate. Three-quarters of the surveys reported by transit agencies defined the study population as “all riders” in some fashion: all bus riders, all rail riders, all fixed-route riders, or all transit riders. The remaining surveys defined the study population in a variety of ways based on geography or time period; for example: • Riders within a geographic study area, • Riders on one route or a group of routes, • Riders traveling within a time period (most often week- days), or • Riders traveling through a selected station. In addition to these distinctions, on-board and intercept sur- veys often qualify respondents by age so that children and sometimes young adults are not included in the study population. A fundamental research question is whether the study population is defined as riders (people) or trips. Although at first glance the differences may seem trivial, they do have significant implications for how the survey is conducted and interpreted. Focusing on riders is most appropriate for customer satis- faction, attitudinal, and demographic questions where the objective is to obtain information on a cross section of customers. The objective of the survey is to collect informa- tion from individuals who use transit. In the survey process, customers who are encountered by surveyors more than once are surveyed only the first time. Focusing on trips is most appropriate when the informa- tion will be used to profile characteristics of trips such as O&D patterns and trip purposes. The objective is to obtain a

completed survey for each customer trip in the sample frame. Thus, a rider who is encountered by survey workers twice is asked to complete two surveys. A further distinction is between trips and trip segments. Following the standard practice in O&D surveys, a “customer trip” for this discussion is defined as a journey between two activities; for example, home to shopping, whether one or more than one bus or train is used for the jour- ney. A “trip segment” is each time the traveler boards a bus or passes through a turnstile. The distinction is important because a person making a transfer has twice the chance of being surveyed than one who does not. The transferring rider can easily be overrepresented in a sample if the intent is to measure customer trips. With an appropriate weighting of transferring riders, this issue can be corrected. In theory, then, customer satisfaction and attitudinal surveys would appropriately define the study population as “riders.” O&D surveys would define the study population as customer trips or trip segments. Many surveys, however, have both O&D and attitudinal sections. In this case, the decision would be based on the importance of the O&D data and the ability to avoid double counting riders in the attitu- dinal section. One option is to ask riders to complete the O&D section for each trip or trip segment, but to complete the attitudinal section only once. This is a workable approach provided the study population can understand this instruction. As a practical matter, transit agencies have experienced difficulty in enlisting riders to complete more than one survey. Completing multiple surveys may seem redundant to customers and overly burdensome. Therefore, although O&D surveys may request that customers complete multiple surveys, many do not actually do so. This may not be a sub- stantial problem in a large agency where customers are unlikely to be encountered by survey workers more than once. However, it can be a substantial problem with large survey efforts in mid-size or smaller agencies. In either case, the issue of riders not wanting to complete more than one sur- vey is one reason that agencies emphasize weighting of O&D surveys by route, direction, and/or time of day, as shall be discussed later, to attempt to offset the possible biasing effects of self-selection in the survey process. A final comment on this issue concerns the importance of being clear in the presentation of survey results. Although it may be tempting to speak generally about the survey showing how riders feel or how they travel, care should be taken to identify the study population. A survey in which the study population is riders, for example, can refer to riders who used transit on the specified survey dates. A survey that uses trips as the study population can be clear that results are specific to trips, with some riders represented multiple times. 14 Sample Frame The sample frame is the listing of the study population from which the sample will be drawn. In a telephone survey of voter behavior, the sampling frame might be all residential phone numbers. On-board and intercept transit surveys can define the sampling frame in a number of ways. For on-board surveys, the sampling frame is typically based on bus or train routes or vehicle trips (defined as service from one end of the route to the other). The sample frame is therefore customers on these listed routes or vehicle trips. On-board bus and rail surveys reported by transit agencies generally adopted this approach. One variation to this approach is to survey only high- ridership routes. Broward County Transit’s 2003 on-board bus survey was conducted on the 10 routes with the most ridership. This approach focuses survey resources on routes that are likely to generate the largest number of completed surveys. At the same time, there is the possibil- ity that lower-ridership routes not surveyed would produce different responses. The results should thus be viewed as representative only of riders on the most used routes. Surveys conducted at bus stops or rail stations typically define the sampling frame as stops or stations. For the light rail portion of the Denver-area Regional Transportation District (RTD) Customer Satisfaction Survey, the sampling frame was station platforms. Likewise, PATH’s O&D survey defined the sampling frame as all PATH stations. Another approach is to survey at centralized nodes, such as transit centers or transfer facilities. The advantage of this method is that riders from a cross section of routes can be surveyed without having to disperse surveyors over every route. GRPC conducted interviews at a transit center and four transfer facilities for this reason. From a data accuracy standpoint, this approach is most appropriate if the large majority of customers pass through one of these central nodes. The sampling frame may be further defined by time of day, choice of weekday or weekend, and direction of travel. These further refinements are designed to ensure that a rep- resentative sample is drawn for each segment; for example, each line, direction, or time-of-day combination. The use of this sampling frame in a stratified sample is discussed later in this chapter. Sample The sample is the group of people selected to be surveyed. As Trochim (2004) points out, the sample is not the group of people that actually complete the survey. Rather, the sample is the group of people that the researcher attempts to contact or recruit.

15 The sample may consist of the entire sample frame. For example, Pace Suburban Bus distributed its Customer Satisfaction Index/User Survey on all Pace buses over a 3-day period. More often, however, a subset of the sample frame is selected, typically through simple random sample procedures. For example, Los Angeles County Metropolitan Transportation Authority (Metro) created a digital file of all bus trips and used computer software to randomly select bus trips to assign to survey workers for an on-board customer satisfaction survey. Similarly, for a systemwide survey, MARTA randomly selects 1,000 bus and rail trips to be surveyed each year. Simple random samples, as in these examples, are method- ologically attractive because they maintain the basic principal of probability sampling: that each unit has an equal chance of being selected. Simple random samples can be difficult to field, however. Random selection of routes (or bus stops or stations) can result in survey workers spending an inordinate amount of time moving between assignments. MARTA circumvents this problem by filling in surveyor time with other assignments through the course of the year. Another approach is to group bus trips, for example, to form coherent itineraries for survey workers. Each “package” of bus trips is then run through the random selection procedure. Santa Clara Valley Transportation Authority (VTA) used this approach for its 2000 On-Board Survey. Although the sampling frame may consist of bus trips, subway stations, or light rail platforms, the sample itself inevitably consists of riders. The transition from vehicle trips or stations to riders is generally straightforward. In self- administered surveys, the sample typically consists of all persons riding the buses or trains that are in the sampling frame. However surveys are distributed, the transit agency attempts to enlist participation from every rider. At times, however, it is necessary to select from within the group of people boarding the bus or otherwise falling into the sample frame. If survey workers are conducting personal interviews, for example, it is probably not possible for them to interview every person who boards the bus or passes through a station. Without a carefully implemented method of selecting which persons are approached, there is an acute chance that surveyors will gravitate toward persons who appear friendly and/or are similar to the surveyor. Transit agencies typically attempt to maintain the randomness of the sampling procedure by selecting every nth person. Thus, the King County Metro Ride Free Area survey selected for inter- viewing every third rider that boarded during non-peak times and every fifth rider that boarded during peak times. MARTA surveyors interviewed every fifth person to board the bus or train for a 5-min interview on travel patterns. Not all agencies adopt a strict sampling procedure. Some agencies reported encouraging bus operators who are charged with distributing surveys to attempt to distrib- ute the surveys to a cross section of riders. The integrity of the sample can be affected by this approach, although clearly there may be logistical and budgetary reasons necessitating it. Stratified Sampling Although most on-board and intercept surveys reported by transit agencies used a simple probability sample, approxi- mately one-quarter of these agencies reported using a stratified sample. The objective of stratification is to ensure that key subgroups of the population are represented in the overall population. Among on-board and intercept surveys, stratifica- tion is most often applied to time of day and route. The objec- tive is to ensure that peak and non-peak hours, or each route, are represented in the sample and to control for variations in response rate within strata. In a stratified sample, the sample frame is divided into mutually exclusive and exhaustive subsets. A simple random sample of elements is then chosen independently from each group or subset. A San Diego Association of Governments (SANDAG) On-Board Transit Passenger Survey, which covered 164 bus and rail routes in a large geographic area, was stratified based on route, direction, and time of day. Surveyors were assigned trips within each of four time periods for each route and direction combination. In deciding how to stratify, a main goal is to divide the sample frame into relatively homogeneous groups. In the case of bus routes, agencies believe or assume that bus riders on a given route and rail riders at a given station and time period are relatively similar. Some agencies stratify at a more generalized level. For example, the Regional Transportation Commission of Southern Nevada stratified based on quadrants of the agency’s service area. The Ann Arbor Transportation Authority (AATA) strati- fied for five groupings of bus routes. The same sampling fraction can be used within each strata; for example, sampling one in every 10 bus trips for each route, time of day, and direction combination. Almost universally, transit agencies reported using the same sampling fraction for each strata, which is called a propor- tionate stratified random sample. An alternative is to use different fractions to produce a disproportionate random sample. This approach provides statistically significant results for relatively small subgroups; for example, people using transit during late night hours. A proportionate stratified sample would likely produce too few surveys from this group to be statistically meaningful. TriMet oversampled light rail riders to obtain a meaningful number of responses from light rail riders.

Whether using proportionate or disproportionate sam- pling, weights are developed for each group (strata) in the sample once the surveying is completed. Weights are most often based on ridership. The weight for each stratum is cal- culated based on the ratio of total ridership for the strata and the number of surveys collected from that group. Total ridership may be for a given route, route and time-of-day combination, or station. As an example in a rail survey, WMATA weighted surveys to daily ridership by mezzanine. For weighting by boardings, transit agencies measure total boardings in a variety of ways. SANDAG, for example, used automatic passenger counters to determine the number of passengers boarding for each route and time-of-day com- bination. Ridership can also be based on entries as recorded by bus fare boxes or turnstiles. Another method is for survey workers to count the total number of passengers entering, whether or not they accepted a survey form. (See chapter five for further discussion of the fieldwork protocols for these counts.) Although most responding transit agencies weighted surveys by ridership, more complex methods are sometimes used, particularly for O&D surveys. A good example is PATH, which used an advanced statistical technique called iterative biproportional fitting to weight response by station entry and exit and time of day. MINIMIZING SAMPLING AND NONRESPONSE ERROR IN SURVEY The precision of a survey is determined by the amount of error created in the process of taking a sample and conduct- ing data collection. Sampling error, which arises from surveying a sample of the study population rather than the entire population, is often the focus of discussion of survey error issues. There are other sources of error, however, including nonresponse error, coverage error (discussed earlier), and measurement error (discussed in chapter four). Sampling Error Virtually all on-board and intercept transit surveys involve taking a sample of the study population and are thus subject to sampling error. Because surveys rely on a sample of the population, survey results are likely to be somewhat differ- ent than if the entire population was interviewed. The sampling error is an expression of the difference between the true (but unknown) value and observed values if, hypothetically, the survey were repeated numerous times. To use an example, suppose that an on-board bus survey found that 20% of riders transferred to another bus on the trip. The sampling error might be stated as plus or minus 3 percentage points with a 95% confidence level. The con- fidence interval (plus or minus 3 percentage points) is the 16 degree of precision and reflects the spread of observed val- ues that would be seen if the survey were repeated numerous times. The confidence level (95%) is how often the observed transfer rate would be within 3 percentage points of the true transfer rate if the survey were repeated numerous times. The sample size that is needed for a given survey depends on the population size and level of precision desired. If the researcher wants to be within 3 percentage points, for exam- ple, a sample size of 1,066 is required for a large population, but approximately one-half that number for a population of about 1,000. Table 8 provides sample sizes needed for three levels of precision (10%, 5%, and 3%) at a 95% confidence level for various population sizes. In transit surveys, it is often desired to achieve a given level of precision for each of a number of major routes or for each of several time periods. In this situation, the sample size needs to be computed for each subgroup; for example, each route or day part. A Transit Authority of River City (TARC) survey, for example, developed the sample plan based on achieving a sampling error of 8 percentage points for routes with 1,000 or more average weekday boardings and 12 percentage points for routes with fewer weekday boardings. In addition, the bottom 10 routes in terms of ridership were sampled as one unit with a sampling error of 5 percentage points. For surveys with stratified sampling, as in the TARC survey, calculating the sampling error for the entire survey must take account of the stratified sample design. One cannot simply use the overall number of responses to calculate the sampling error. Stratification may change the efficiency of the sample— in some cases improving efficiency (as when the strata are relatively homogeneous) or reducing efficiency (when the vari- ance of each strata are about the same). The specific situation will affect the sampling error of the total sample. Nonresponse Error Another major source of error is nonresponse error, which results from failure to obtain completed surveys from some Sampling Error for 95% Confidence Level Population ±10% ±5% ±3% 200 65 132 400 78 196 1,000 88 278 6,000 95 361 20,000 96 377 1,000,000 96 384 169 291 517 906 1,013 1,066 Note: Sample size needed for each sampling error; responses with frequency of 50%. TABLE 8 SAMPLE SIZES NEEDED FOR VARIOUS POPULATION SIZES AT VARIOUS LEVELS OF PRECISION

17 portion of the population selected in the sample. It is inevitable that some riders refuse to take a survey, never return a survey that they took, or refuse to be interviewed. These respondents might have responded in the same way as respondents who did complete the survey, or they might not have. In contrast to sampling error that can be computed, there is no standard- ized way to compute the error that arises from nonresponse. However, one can attempt to evaluate and possibly compen- sate for nonresponse. The likely impact from nonresponse can be evaluated by comparing characteristics of respondents with those of the entire population or those within the sampling frame. The comparison is sometimes made for rider charac- teristics such as gender, race, and place of residence, or for trip characteristics such as on and off locations.

Next: Chapter Four - Developing Questionnaires »
On-Board and Intercept Transit Survey Techniques Get This Book
×
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

TRB’s Transit Cooperative Research Program (TCRP) Synthesis 63: On-Board and Intercept Transit Survey Techniques examines transit agencies’ experiences with planning and implementing on-board and intercept surveys. On-board and intercept surveys include self-administered surveys distributed on board buses and railcars, and in stations, as well as interviews conducted in these environments. The report provides an overview of industry practices and covers a broad range of issues addressed in planning a given survey.

  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  6. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  7. ×

    View our suggested citation for this chapter.

    « Back Next »
  8. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!