Characteristics of Premium Transit Services that Affect Choice of Mode (2014)

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.

C-1 A p p e n d i x C Detailed Survey Results Contents C-1 Overview C-2 Market Research on Transit Service C-21 Demographic Characteristics C-25 Transit Awareness and Use C-26 Traveler Attitudes C-31 Trip Characteristics C-40 Stated Preference Survey Characteristics Overview Market research was conducted in three cities (Salt Lake City, Chicago, and Charlotte), representing a variety of different-sized cities with very different transit systems and different traveler characteristics. The data collected included traditional origin-destination travel data complemented by additional data on premium transit service characteristics, awareness and consideration of modal alternatives, and traveler attitudes. A few new methods of data collection were employed to gather information desired for travel forecasting model estimation: • Maximum Difference Scaling is a method to measure the importance of individual transit service characteristics with respondents choosing the best and worst options from a set of alternatives. There were eight maximum difference experiments in each of the three surveys. • Choice-based Conjoint is a method to measure the stated preference of a combination of transit service characteristics with respondents choosing the best alternative. There were eight stated preference experiments in each of the three surveys. • Attitudinal statements to the travel survey measures the attitudes of travelers on different aspects of travel. There were 15 attitudinal statements for the Salt Lake City survey (6 for transit users and 9 for non-transit users) and 18 attitudinal statements for the Chicago and Charlotte surveys. • Questions about awareness and consideration of transit alternatives were included in the surveys for all three cities. In the initial survey for Salt Lake City, these questions were exploratory. In the second set of surveys for Charlotte and Chicago, these questions were more systematic and comprehensive to allow model estimation for awareness and consideration.

C-2 Characteristics of premium Transit Services that Affect Choice of Mode The transit service attributes collected in this research included two bundles of attributes and a series of other attributes that were not bundled: • Station or stop design features - real-time information about the next transit arrival/departure, security, lighting/safety, shelter, cleanliness of the station, benches, and proximity to services. • On-board features—seating availability, seating comfort, temperature, cleanliness of the transit vehicle, and productivity features (Wi-Fi, power outlets, etc.). • Other features—identification of the transit vehicle, reliability, schedule span, transit frequency, transfer distance, walking distance and parking distance to the station, ease of boarding, and fare machines. These attributes were included in the model estimation work for all three cities and further used in the model implementation and calibration for Salt Lake City. The market research conducted in this study collected data on mode choice behavior (attitudes, awareness and consideration, service attributes, and mode choice) as well as travel times for transit services successfully. These data were collected at a reasonable cost. The authors recommend that future data collection activities for mode choice behavior consider this template for data collection. These survey results are provided to detail the traveler and trip characteristics represented in the survey. The survey was not designed to represent the full population, nor was it designed to be a representative sample because it was used in disaggregate modeling where behavioral differences are represented. These disaggregate models can be used in forecasting as long as the appropriate population proportions or sample weights are used to represent the population. The following survey results should not be interpreted as representative of the full population. Market Research on Transit Service There was no standard method for quantifying non-traditional and/or premium attributes of transit service found in the literature. One example of a challenging measure to quantify is reliability. Questions in other surveys asked about this in different ways that may have led to variation in response between surveys that merely had to do with how the question was phrased (e.g., “1 in 10 trips is 5 minutes late or more” vs. “9 out of 10 trips are on time”). Analytical techniques that are used (e.g., logit vs. other conjoint techniques, vs. simple rating/rankings) have also been found to vary in each study. In this study, the Maximum Difference Scaling (i.e., MaxDiff, also known as “Best-Worst” scaling) approach was used to measure the importance of premium transit service characteristics. In this method, the experiments are designed in such a way that respondents are required to choose their “most preferred” and “least preferred” options from a set of alternatives/items. MaxDiff is a relatively new conjoint technique, pioneered by Jordan Louviere in the early 1990s (Almquist et al. 2009). It is useful for obtaining rankings and relative preferences for a variety of different attributes, such as brand names or service benefits, which can be described in a single statement (e.g., “There are good sidewalks between the BRT stop and your home”). This technique has been used in many settings to test the attributes of transportation services, such as on-board amenities for proposed high-speed rail services in California (Outwater et al. 2010) and station amenities for upgraded New York City subway stations (Spitz et al. 2007).

detailed Survey Results C-3 Survey Administration The survey was administered via a web-based survey to travelers in the Salt Lake City, Chicago, and Charlotte areas. Web-based surveys are more cost efficient to conduct, can report complex data more easily, and can provide a more interactive experience for the user than other surveys. There were some concerns over demographic bias in the web-based surveys but this was not a concern for our convenience sample. The major difference in administration from Salt Lake City to Chicago and Charlotte was that no in-person intercept field effort was conducted in Chicago and Charlotte and instead all administration was done online. A sample size of 1,500 completed surveys was targeted for each city, with a minimum of 200 completed surveys within four quota cells (auto work or school trip, transit work or school trip, auto other trip, and transit other trip). The sampling plan was designed to include a sufficient range of travelers and trip types to support the statistical estimation of the coefficients of a choice model. By collecting data from a range of traveler and trip types it is possible to identify the ways in which different characteristics affect mode choice behavior. These differences can then be reflected in the structure and coefficients of the resulting choice model. The survey sample that supports choice model estimation does not need to be precisely proportional to the population as long as: • Any behavioral differences are properly represented in the model, and • The model is applied for forecasting using appropriate population proportions and/or sample weights. RSG administered the Salt Lake City Travel Survey from mid-July through early August 2009. The survey was made available to respondents online via email invitation containing a URL to the survey or via onsite intercepts using laptops in various locations throughout the Salt Lake City area. For the online recruiting, Utah Transit Authority’s help was invaluable, as they provided 40,000 email addresses from their database which contained both UTA riders and non- riders. Additionally, the survey was administered online via business recruiting through major organizations, employers, and universities in the area, again using the email invitation method. For the onsite recruiting, the researchers surveyed over a 5-day period in Salt Lake City using laptop computers. The survey administration effort yielded an overall dataset with just over 2,000 respondents (total 2,017). This sample (TABLE C-1) provides a strong data set to conduct the analysis. TABLE C-1. Salt Lake City trip purpose and mode of survey reported trips. Auto 987 466 1,453 Transit 480 84 564 Total 1,467 550 2,017 Salt Lake City Survey

C-4 Characteristics of premium Transit Services that Affect Choice of Mode The Charlotte survey fielded from May 16, 2011–June 15, 2011, with a total of 1,527 respondents completing the survey. The survey was fielded first to the Charlotte Area Transit System (CATS) rider email list, from which 222 completes were obtained. Once the email list had been exhausted, the survey was fielded through a reputable online survey panel provider, which provided the remaining 1,305 completed surveys. Respondents taking the survey were screened to ensure they qualified. Any respondents living outside the Charlotte area (Mecklenburg or Carabus counties) or who did not make a trip using auto or transit in the past week were terminated from the survey. Additionally, any respondents completing the survey in less than 7 minutes were not considered “completed” surveys, as this time was deemed too fast to have paid thorough attention to the survey. The median survey completion time was 21 minutes. TABLE C-2 presents the types of trips that were reported during the survey. TABLE C-2. Charlotte trip purpose and mode of survey reported trips. 902 259 1,161 150 215 365 1,052 474 1,526 Transit trips were over-sampled to ensure that this mode had enough samples for mode choice and choice set model estimation. Work trips were defined as commute trips to/from work; school trips were defined as trips to/from school; and other trips were trips that were made from home to a place other than work or school. The age and income characteristics of the sample are shown in the next section. The Chicago survey fielded from June 23, 2011–July 5, 2011, with a total of 1,515 respondents completing the survey. Respondents taking the survey were screened to ensure they qualified. Any respondents living outside the Chicago area (Cook, DeKalb, Kane, McHenry, or Will counties) or who did not make a trip using auto or transit in the past week were terminated from the survey. As with the Charlotte data, any respondents completing the survey in less than 7 minutes were removed from the dataset. The link on Metra’s website was up for 1 week and 19 completes were obtained from this source; other local transit agencies were unable to provide email lists or web links due to their own conflicting survey efforts at the time of administration. The remaining 1,496 completed surveys were obtained through a reputable online survey panel provider. The median survey completion time for Chicago was 20 minutes. TABLE C-3 presents the types of trips that were reported on during the survey. Charlotte Survey Chicago Survey A mix of respondents in terms of age, income, and other demographics was surveyed.

detailed Survey Results C-5 TABLE C-3. Chicago trip purpose and mode of survey reported trips. 450 340 790 364 361 725 814 701 1515 Transit Service Attribute Data Collection The study’s analytical approach involved a three-part survey that was conducted in Salt Lake City, UT in Phase 1 and in Chicago and Charlotte in Phase 2: • The first part of the survey was designed to gather data on awareness of transit options. • The second part of the survey presented choice based conjoint (CBC) stated preference mode choice experiments to travelers where they were asked to choose a mode based on different levels of attributes, including some attributes that were constructed as “bundles.” • The third part of the survey used MaxDiff conjoint techniques, which was used to evaluate the individual attributes that make up the bundles. The second and third parts of the survey instrument were designed for the study to use conjoint analysis to measure preferences among transit features. The survey was intended to gather information to research awareness as well as non-traditional attributes. More detailed information on transit awareness and frequency of transit use was obtained in the Phase 2 surveys based on Phase 1 experience. The survey instrument and sampling designs were largely similar for all three cities to allow for comparisons in the data. The goal of the third part of the survey was to estimate the relative utility of a variety of transit service attributes. Confidence intervals around utilities for each attribute can be calculated to allow statistical differences between attributes to be demonstrated. MaxDiff was thought to be well suited to this exercise for the following reasons: • MaxDiff experiments are simple for the respondent to understand and evaluate. • MaxDiff can be used to evaluate a relatively large number of attributes (RSG has successfully tested in excess of 50 attributes, although a smaller set is reasonable for this study given the other parts of the survey and the complexity arising from having so many attributes). • The setup produces results that show the relative importance of the items being rated, thus avoiding the problem where respondents rate most items as “important,” making it more difficult to distinguish among individual items.

C-6 Characteristics of premium Transit Services that Affect Choice of Mode • In the context of studies that are similar to the current study, the values of the attributes can be expressed in terms of minutes of travel time or dollars of transit fare by including attributes in the MaxDiff experiments that represent travel time and travel cost savings. • Further, the MaxDiff model specification can easily accommodate “bridging attributes” that may be used to link the model results with the mode choice model, allowing the analyst to evaluate the relative importance of non-traditional transit attributes (such as cleanliness of station, on-board temperature, ease of boarding, etc.) within the context of trip mode choices. This allows the results of the MaxDiff experiments to support recommendations about adjustments to standard mode choice model parameters. As a general rule, all MaxDiff attributes, either statements or images, must be parallel in construction. For example, every attribute in these surveys is defined by both premium and standard attributes. Likewise, it is important for all attributes to be positive and as clear as possible. MaxDiff attributes may have levels (for instance, “BRT bus arrives every 10 minutes” and “BRT bus arrives every 20 minutes”); however using levels does allow for the possibility of some “no brainer” comparisons if both the statements appear in the same experiment. But this possibility does not preclude meaningful comparisons between items. RSG keeps the number of attributes with levels to a minimum to ensure that respondents are challenged as they trade off which transit service benefit is most important and which is least important to them. FIGURE C-1 shows a screenshot of the MaxDiff experiment. Respondents were asked to indicate their “most likely” and “least likely” choices among three alternatives shown. Transit Service Attributes for Market Research A comprehensive set of attributes that affect the level of service offered by transit facilities and differentiate premium transit services from standard transit services was analyzed. For this project, using the important transit attributes identified in the literature review as a basis, RSG constructed a list of attribute statements to include in the MaxDiff and CBC sections of the survey. TABLE C-4 presents a list of the transit attributes that are analyzed in this study. There were several considerations that went into determining the attributes list which are described in this section: • Length of survey and associated respondent fatigue—The complexity of the sur- vey and the various analysis techniques being employed necessitated breaking the survey into three parts and could make the survey quite long. In order to reduce respondent fatigue and still be able to prove the concept of this project, variables to test were carefully selected so that they would best allow measuring the most important factors that differentiate premium transit services from ordinary bus services. Certain traditional attributes that are referred to as anchor attributes were selected to be shown in both the MaxDiff and CBC sections in order to allow linking results from the two techniques, such as travel time and travel cost. Beyond these anchoring attributes, the overall list of MaxDiff statements was reduced from 100 to 37, focusing on non- traditional variables that were felt to be most appropriate. It is generally recommended that there be a maximum of approximately 60 statements when using simple attributes (such as images, or short phrases) or a maximum of approximately 40 statements when using more complex attributes that require a long, detailed statement.

Detailed Survey Results C-7 • Difficulty of describing certain non-traditional attributes—Non-traditional attributes can be difficult to describe and quantify because respondents’ opinions of them can be emotional and varied. For example, the researchers chose to include two safety-related statements for the MaxDiff section that have been found to be important in previous studies reviewed in the Literature Review (Chapter 2) because some respondents may feel the need for safety more acutely than other respondents, and what makes one respondent feel safe may not be the same for another. In the CBC section the researchers tested the affect of premium versus standard attributes in both on-board and station/stop amenities. Here, the researchers grouped into “bundles” various commonly accepted aspects of premium amenities and standard amenities in order to be able to reduce the number of variables to be tested. The researchers defined these groupings for the respondent to make it easy for them to understand what is meant by premium versus standard. The “bundle” attributes and definitions are shown in the following section. • Selection of appropriate attributes that are mode neutral and are applicable across transit services in varying geographies—Third, the researchers selected variables not limited by transit mode or by relatively limited markets; rather, the researchers included as many variables as possible that would be applicable across many transit systems in many markets. Examples of variables that the researchers felt were not applicable across modes are queue jumping, which would pertain only to buses, or the stopping position at a platform, which would pertain to trains. • Consideration of whether the attribute can actually be represented with traditional modeling methods—Finally, a few of the level-of-service features, such as queue jumping, signal priority and dedicated right-of-way can be implemented in models with specified level-of-service improvements relative to “typical” bus services. While the authors wonder whether there is intrinsic value beyond time savings for such features, the choice was made not to include these in the list of attributes in part because it would be very difficult to parse these effects in a MaxDiff experiment. FIGURE C-1. Screenshot of MaxDiff experiment.

C-8 Characteristics of premium Transit Services that Affect Choice of Mode TABLE C-4. Transit service attributes.

Detailed Survey Results C-9 To ensure respondents would understand exactly what differences exist between premium transit features vs. standard transit features, a clear definition of the features was included. Respondents were first shown the definition page and could return to it at any time during the eight experiments by simply rolling over the information button (blue circle with “i”) with their mouse. Examples of the definition pages for premium vs. standard on-board features, stations/stop design, and arrival/departure information are shown in FIGURE C-2. FIGURE C-2. Premium vs. standard transit feature definitions. Traveler Attitudes Respondents for all three surveys were asked to describe a specific, recent, home-based trip, including travel mode, access and egress modes, transfers, trip duration, cost, stops made along the way, and destination. To help understand different types of respondent segments, each respondent was shown a list of attitudinal statements, as shown in TABLE C-5. In the first survey in Salt Lake City, these differed somewhat depending on whether they were categorized as a transit user or non-user. In the second and third surveys, in Chicago and Charlotte, respectively, these statements were asked for all respondents. In addition, some changes to the attitudinal statements for Chicago and Charlotte were made to reflect the importance of the various statements in the Salt Lake City factor analysis models (i.e., some statements were dropped, other statements were added or revised).

C-10 Characteristics of premium Transit Services that Affect Choice of Mode Traveler attitudes are obtained for 18 attitudinal questions from the surveys in Charlotte and Chicago and 15 attitudinal questions from the survey in Salt Lake City. There are five ranges of responses to these attitudinal questions (strongly disagree, somewhat disagree, neutral, somewhat agree, strongly agree). FIGURE C-3, FIGURE C-4, and FIGURE C-5 show the range of responses, sorted from the least agreement with the statement to the most agreement with the statement. In Charlotte (FIGURE C-3), travelers rate statements that are pro-transit as the ones they least agree with and several statements that are anti-transit are ones they most agree with. The one exception is the statement “I am not afraid to ride transit,” which has strong agreement among Charlotte travelers. Protecting the environment has very strong agreement as well as willingness to carpool or ride transit to reduce emissions. Charlotte travelers are not willing to pay higher tolls to reduce congestion and to some degree are willing to tolerate delays if comfortable. They value saving time over choosing a particular mode. In Chicago (FIGURE C-4), travelers are in more agreement with pro-transit statements than in Charlotte and are also not afraid to ride transit, which has very strong agreement. Protecting the environment also has very strong agreement, as well as willingness to carpool or ride transit to reduce emissions. Chicago travelers are also not willing to pay higher tolls to reduce congestion and to some degree are willing to tolerate delays if comfortable. They also value saving time over choosing a particular mode. The Salt Lake City attitudinal questions ( FIGURE C-5) were completed in Phase 1 and do not match exactly those used in Charlotte and Chicago, but do have many similarities. TABLE C-5. Traveler attitude statements from three surveys. √ √ √ √ √ √ √ √ √ √ √ √

detailed Survey Results C-11 √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ Note: A check in this table indicates that the statement was included in that survey. TABLE C-5. (Continued).

FIGURE C-3. Charlotte traveler attitudes—least to most agreement.

FIGURE C-4. Chicago traveler attitudes—least to most agreement.

FIGURE C-5. Salt Lake City traveler attitudes—least to most agreement.

detailed Survey Results C-15 In addition to the attitudinal questions, two other questions were asked that are considered latent variables (i.e., variables that cannot be directly measured): • Willingness to walk (how far is the respondent willing to walk for a specific trip) • Informed about transit (how informed is the respondent about transit services) As shown in FIGURE C-6, “20 minutes” seems to be the threshold for walk to transit. There are around 90% of respondents who are not willing to walk for 20 minutes to reach a transit stop/station. On average, people using transit mode are more willing to walk than people who drive. About 14% of respondents who drive in their reference trips are not willing to walk at all, compared to only 1% of transit riders who are not willing to walk at all. The willingness to walk question was added to the surveys for Chicago and Charlotte and was not asked in the Salt Lake City survey. FIGURE C-6. Willingness to walk to transit for Charlotte and Chicago. The survey respondents were asked to select how informed they are about the survey area’s public transit services regarding types of service available, routes, schedules, fare options etc. (FIGURE C-7). There is no significant difference about respondents’ awareness of area public transit for the three cities, except that Charlotte has slightly fewer respondents (8% percentage points) being very informed.

C-16 Characteristics of premium Transit Services that Affect Choice of Mode FIGURE C-7. Awareness of transit for Charlotte, Chicago, and Salt Lake City. Stated Preference Stated preference questions were included to evaluate tradeoffs travelers make when choosing a mode. Survey respondents were asked to complete eight stated preference questions with varying travel time, costs, and transit service features to force choices among three options shown in each question. An experimental design was created that would allow the calculation of value in terms of time (minutes) or in cost (dollars) for all the traditional variables used in forecast demand modeling. Additionally, because the goal of this research is to determine the effect premium vs. standard transit features have on people’s choices and to improve estimations of mode choice models and transit path builders, bundles of non-traditional variables (premium transit vs. standard transit features) were included in the stated preference design to allow calculation of a specific value for each category. The values for these non-traditional bundles were to be linked to the values of their specific components generated through the MaxDiff analysis, a process that is described in Appendix E. Respondents were shown three trip options on each page: • Option 1a: Trip made by car (if car was available) • Option 1b: Trip made by randomly selected transit type (bus or train, if car was not available) • Option 2: Trip made by bus • Option 3: Trip made by train These options were shown in random order to prevent bias. An example of a stated preference experiment page is shown in FIGURE C-8.

Detailed Survey Results C-17 FIGURE C-8. Example of stated preference experiment. To ensure survey respondents would understand exactly what differences exist between premium transit features vs. standard transit features, a clear definition of the features was included. Respondents were first shown the definition page and could return to it at any time during the eight experiments by simply rolling over the information button (blue circle with “i”) with their mouse. The bundle attributes’ definitions of premium, modernized, and informative versus standard and are shown in TABLE C-6. The attribute levels for the stated preference experiments are shown in TABLE C-7 and TABLE C-8 for each mode option (auto and transit).

C-18 Characteristics of premium Transit Services that Affect Choice of Mode TABLE C-6. Definitions of premium and standard transit variables by category.

detailed Survey Results C-19 TABLE C-7. Stated preference experiment attribute levels.

C-20 Characteristics of premium Transit Services that Affect Choice of Mode TABLE C-8. Stated preference experiment attribute levels, Continued. *Reliability was modified for the Chicago and Charlotte surveys to better reflect the way in which transit agencies use and communicate this information. The new definition was x% of trips delayed by y minutes or more. **Real time information was included in the station/stop design bundle in the Chicago and Charlotte surveys because this resulted in inconsistencies in the mode choice modeling for Salt Lake City. In addition, span of service was added to the Chicago and Charlotte stated preference experiments (and tested in the models) because the attribute was deemed important to making the decision to use transit.

detailed Survey Results C-21 Demographic Characteristics Gender composition of the Salt Lake City sample is quite different from Chicago and Charlotte (FIGURE C-9). There are about 56% of male respondents in Salt Lake City while the other two cities only have around 35%. Among the Charlotte and Chicago female respondents, 70% are workers and the remaining 30% are either homemakers, retired, or not employed during the time of the survey. A similar percentage (72%) of Salt Lake City female respondents are workers, who are either full-time (54%) or part-time (19%). Among the male respondents, about 79% are workers and the rest (21%) are either homemakers, retired, or not employed during the time of the survey. About 85% of male respondents from Salt Lake City are employed, with 74% full-time and 11% part-time. FIGURE C-9. Gender statistics for Charlotte, Chicago, and Salt Lake City. On average, Chicago respondents are older than those the of other two cities (FIGURE C-10). The median age of Chicago respondents falls into the 45-54 yrs group, while the median age of the other cities is within the 35-44 yrs range. FIGURE C-10. Age statistics for Charlotte, Chicago, and Salt Lake City.

C-22 Characteristics of premium Transit Services that Affect Choice of Mode Chicago stands out as slightly different from the other two cities in terms of employment status (FIGURE C-11). Among Chicago respondents, retired persons are almost twice as many as in the other two cities, and full-time workers are more than 10% less than in other two cities. It also has higher percentage of homemaker and not-employed workers (Salt Lake City does not have the category “self employed”). FIGURE C-11. Employment status statistics for Charlotte, Chicago, and Salt Lake City. The characteristics of the student population are presented in FIGURE C-12. The Chicago data represents about 14% student respondents. Among the student respondents, almost 62% are full-time students and the other 38% are part-time students. Salt Lake City data has 21% student respondents and a similar full-time/part-time split as Chicago. The Charlotte data has only 11% student respondents (half that of Salt Lake City), and full-time/part-time students are half split. FIGURE C-12. Student status statistics for Charlotte, Chicago, and Salt Lake City.

detailed Survey Results C-23 The household size and composition characteristics are presented in FIGURE C-13. Salt Lake City respondents tend to have bigger families, with 24% of them living in households with five or more members, while the other two cities only have about 10% of respondents in households with five or more members. FIGURE C-13. Household size statistics for Charlotte, Chicago, and Salt Lake City. Of the three cities, Charlotte has the highest percentage (11%) of respondents earning more than $150,000 and the smallest percentage (17%) of respondents earning less than $35,000 (FIGURE C-14). FIGURE C-14. Income statistics for Charlotte, Chicago, and Salt Lake City.

C-24 Characteristics of premium Transit Services that Affect Choice of Mode (FIGURE C-15). This reflects the household size composition feature in Salt Lake City. The mean motorized vehicle ownership rate per household is around 1.96 (Charlotte), 1.71 (Chicago), and 2.26 (Salt Lake City). FIGURE C-15. Auto ownership statistics for Charlotte, Chicago, and Salt Lake City. There are no dramatic differences for housing mobility among the three cities except that Chicago shows a slightly higher household mobility. The duration of respondents living in the metropolitan region of interest is shown in FIGURE C-16. FIGURE C-16. Duration of living in the area statistics for Charlotte, Chicago, and Salt Lake City. Thirty-five percent of households in the Salt Lake City sample own three, four, or five- plus cars, while the Charlotte and Chicago samples only have 22% and 19%, respectively

detailed Survey Results C-25 Transit Awareness and Use The survey respondents were asked to select answers to indicate how informed they are about the survey area’s public transit services regarding types of service available, routes, schedules, fare options, etc. There is no significant difference about respondents’ awareness of area public transit for the three cities, except that Charlotte has slightly fewer respondents (8% less) indicating that they are very informed. There is a positive relationship between usage of public transit and awareness of the transit system. People are less likely to use public transit if they are not aware of it (see FIGURE C-17). However, the causality is hard to identify because it is also reasonable to argue that the people using transit are more likely to be aware of it than the people who did not use it. FIGURE C-17. Relationship between awareness and transit usage for Salt Lake City. Chicago had the highest percentage of respondents who indicated they have used transit in the past 12 months, followed by Salt Lake City. Charlotte had the least (FIGURE C-18). FIGURE C-18. Transit usage for Charlotte, Chicago, and Salt Lake City.

C-26 Characteristics of premium Transit Services that Affect Choice of Mode Transit use for Charlotte and Chicago are presented in TABLE C-9 and TABLE C-10, respectively. (Frequency of transit usage is not available at Salt Lake City data.) Some 35% of respondents in Charlotte never used any transit and 14% of respondents in Chicago never used any transit. TABLE C-9. Charlotte transit usage. Frequency of Using Transit CATS Local Bus CATS Express Bus LYNX Light Rail Never 68.60% 71.30% 46.20% At least once 31.40% 28.70% 53.80% 4 times or less per year* 40.13% 33.45% 47.77% 5-11 times per year 13.38% 9.06% 21.00% 1-3 times per month 14.65% 8.36% 15.99% 1-2 times per week 7.96% 8.01% 5.39% 3-4 times per week 9.87% 11.85% 3.35% 5 or more times per week 14.01% 29.27% 6.51% *Calculated as % of individuals who have taken transit "At least once." TABLE C-10. Chicago transit usage. Frequency of Using Transit CATS Local Bus CATS Express Bus Pace Bus CTA train (the 'L') Metro commuter rail Never 43.80% 62.90% 64.00% 32.30% 30.00% At least once 56.20% 37.10% 36.00% 67.70% 70.00% 4 times or less per year 34.70% 50.94% 51.11% 34.86% 47.86% 5-11 times per year 12.63% 11.05% 15.00% 15.51% 20.29% 1-3 times per month 12.28% 14.56% 11.67% 13.88% 12.00% 1-2 times per week 10.32% 10.24% 7.78% 9.01% 4.57% 3-4 times per week 9.07% 6.20% 6.94% 8.42% 3.29% 5 or more times per week 21.00% 7.01% 7.50% 18.17% 11.86% *Calculated as % of individuals who have taken transit "At least once." Traveler Attitudes Traveler attitudes are obtained for 18 attitudinal questions from the surveys used in Charlotte and Chicago and 15 attitudinal questions from the survey used in Salt Lake City. There is a range of five responses to these attitudinal questions (strongly disagree, somewhat disagree, neutral, somewhat agree, and strongly agree). Respondents from Salt Lake City are most likely to try the transit option (FIGURE C-19), followed by those from Chicago, and respondents from Charlotte are least likely to try transit option.

detailed Survey Results C-27 Statement: I currently make an effort to take public transit whenever I can. FIGURE C-19. Make effort to take transit for Charlotte, Chicago, and Salt Lake City. Respondents from Salt Lake City are more willing to increase the frequency of transit usage (FIGURE C-20). Respondents from Charlotte and Chicago share very similar attitudes toward the possibility of increasing transit usage. Statement: If I wanted to, I could use public transit more frequently. FIGURE C-20. Willingness to increase transit usage for Charlotte, Chicago, and Salt Lake City. Although Salt Lake City respondents tend to be willing to take transit or use transit more frequently, they do have difficulty in planning a trip using transit (FIGURE C-21). About 60% of them disagree that it is easy to plan a trip with transit and only 12% agree. Chicago seems to have better transit coverage than other two cities, given that 46% of Chicago respondents agree that it is easy to plan transit trips.

C-28 Characteristics of premium Transit Services that Affect Choice of Mode Statement: It’s easy to plan a trip using transit. FIGURE C-21. Ease in planning a transit trip for Charlotte, Chicago, and Salt Lake City. The response regarding “I am the kind of person who rides transit” is quite consistent with the question about “It is easy to plan a trip using transit.” Chicago has the highest percentage (39%) of respondents who identify themselves as transit riders, while Salt Lake City has the least (only 11%) as shown in FIGURE C-22. The similarity indicates that ease in planning a trip using transit is positively related to the transit usage of the city. Statement: I’m the kind of person who rides transit. FIGURE C-22. Kind of person riding transit for Charlotte, Chicago, and Salt Lake City. Chicago did a worse job on transit sanitation than did the other two cities, although it has the highest percentage of transit riders (FIGURE C-23). Apparently, sanitation is not a major concern for transit riders (otherwise Chicago would have smaller transit share); also a more challenging sanitation condition is likely to be a consequence of higher transit usage.

detailed Survey Results C-29 Statement: Transit is often dirty. FIGURE C-23. Transit sanitation impression for Charlotte, Chicago, and Salt Lake City. Salt Lake City has the biggest percentage of car users who are reluctant to switch to transit mode (FIGURE C-24). Statement: For me, car is king! Nothing will replace my car as my main mode of transportation. FIGURE C-24. Attitude toward car for Charlotte, Chicago, and Salt Lake City. The three cities’ respondents share similar attitudes regarding the feeling of riding transit (FIGURE C-25).

C-30 Characteristics of premium Transit Services that Affect Choice of Mode Statement: I’m the kind of person who rides transit. FIGURE C-25. Feeling of transit riding for Charlotte, Chicago, and Salt Lake City. Charlotte and Salt Lake City respondents share similar feelings regarding stations’ or stops’ pedestrian accessibility, while a higher percentage of Chicago respondents feel that stations or stops are pedestrian friendly (FIGURE C-26). Statement: Getting to and from transit station/stops is not pedestrian friendly and is very unpleasant. FIGURE C-26. Station not pedestrian friendly, unpleasant for Charlotte, Chicago, and Salt Lake City. With respect to park and ride or directly driving, more Salt Lake City respondents are likely to choose driving while more Chicago respondents are likely to choose park and ride (FIGURE C-27).

detailed Survey Results C-31 Statement: I have to drive to get to transit anyway, so I may as well just drive my car the whole way. FIGURE C-27. Park and ride vs. drive for Charlotte, Chicago, and Salt Lake City. Trip Characteristics A specific trip type and mode used is picked randomly from among the trip type and mode used according to what the person said he/she made in the last week. This trip is called the reference trip, about which detailed questions are asked. There are more commute trips made in Charlotte and Salt Lake City, whereas Chicago is more evenly split between commuting and non-mandatory trips (FIGURE C-28). FIGURE C-28. Reference trip split by purpose for Charlotte, Chicago, and Salt Lake City. Chicago had the highest mode share among the three cities (FIGURE C-29). As mentioned previously, these mode shares are not representative of the population, but reflect a desire to achieve an adequate number of transit trips for model estimation purposes.

C-32 Characteristics of premium Transit Services that Affect Choice of Mode FIGURE C-29. Primary mode for Charlotte, Chicago, and Salt Lake City. Among the three cities, Chicago respondents made more non-mandatory/transit trips than did respondents from other two cities. In Charlotte, car is the primary mode for the the commute and for school trips, while transit has a bigger share than auto for non-mandatory purposes (FIGURE C-30). In Chicago, car and transit have a similar share. In Salt Lake City, car is the primary mode for all purposes. FIGURE C-30. Mode share by purpose for Charlotte, Chicago, and Salt Lake City. FIGURE C-31 presents the number of people traveling on the reference trip for Charlotte and Chicago. Charlotte has more car trips with three or more persons than Chicago, and Chicago has a more equal split for party size by mode.

Detailed Survey Results C-33 FIGURE C-31. Party size for Charlotte and Chicago. The respondents were asked how often they make this specific trip using the same mode (FIGURE C-32). TABLE C-11 shows the responses for the people with transit and car reference trips. FIGURE C-32. Trip frequency for Charlotte and Chicago. It is surprising to find that respondents mentioned stop behavior in transit mode for pick- up/drop-off purposes, because the authors expect this to be more prevalent for the car mode (TABLE C-11). Another surprising finding is that people would make a stop during a transit trip to buy coffee or newspapers. In any case, pick up/drop off is more likely to happen during mandatory trips than it is during non-mandatory trips.

C-34 Characteristics of premium Transit Services that Affect Choice of Mode TABLE C-11. Stop-making behavior for Charlotte, Chicago, and Salt Lake City. Mode Car Transit Car Transit City Cha Chi SLC Cha Chi SLC Cha Chi SLC Cha Chi SLC Pick up/drop off household member 13% 18% 18% 10% 16% 18% 5% 7% 0% 7% 6% 0% Pick up/drop off non-household member 2% 0% 0% 3% 8% 0% 4% 10% 0% 11% 10% 0% Buy groceries 27% 11% 11% 37% 22% 12% 29% 29% 36% 16% 21% 100% Coffee, newspapers, etc. 13% 34% 34% 16% 26% 41% 10% 12% 14% 20% 28% 0% Get gas 25% 13% 13% 21% 7% 0% 28% 13% 23% 10% 9% 0% Business/school related stop 5% 4% 4% 0% 4% 6% 3% 4% 5% 4% 5% 0% Pick up/meet other carpool members 1% 3% 3% 4% 4% 0% 0% 1% 5% 4% 4% 0% Other reason 14% 18% 18% 9% 12% 24% 21% 23% 18% 28% 18% 0% City Codes Cha – Charlotte; Chi – Chicago; SLC – Salt Lake City The threshold for walk to transit seems to be “20 minutes.” Around 90% of respondents are not willing to walk for 20 minutes to reach a transit stop/station. On average, people using transit mode are more willing to walk than are people who drive. About 14% of respondents who drive in their reference trips are not willing to walk at all, compared to only 1% of transit riders who are not willing to walk at all. Willingness to walk is discussed in Chapter 2 of the main report. On average, men are more willing to walk than women are, and Chicago respondents are more willing to walk than Charlotte respondents (FIGURE C-33.). FIGURE C-33. Willingness to walk by gender for Charlotte and Chicago. Purpose Mandatory Non-mandatory

detailed Survey Results C-35 In both Charlotte and Chicago, youth are least willing to walk. In Charlotte, people ages 45-64 yrs are the most willing to walk (FIGURE C-34). In Chicago, elderly persons are the most willing to walk (FIGURE C-35). FIGURE C-34. Willingness to walk by age for Charlotte. FIGURE C-35. Willingness to walk by age for Chicago. It is not surprising to see that people using transit mode are much more willing to use various access/egress options than people who drive (FIGURE C-36). These data were not available from the Salt Lake City survey.

C-36 Characteristics of premium Transit Services that Affect Choice of Mode FIGURE C-36. Willingness to use various access/egress modes for Charlotte and Chicago. No significant difference of auto type is found between Charlotte and Chicago respon- dents, except that Chicago respondents own a little bit higher percentage of minivans (FIGURE C-37). These data were not available from the Salt Lake City survey. FIGURE C-37. Auto type for Charlotte and Chicago. The characteristics of the transit reference trips are presented in TABLE C-12 and TABLE C-13 for Charlotte and Chicago, respectively. These data were not available from the Salt Lake City survey. In Charlotte, 365 individuals were asked transit reference-trip details. In Chicago, 725 individuals were asked transit reference-trip details.

detailed Survey Results C-37 TABLE C-12. Charlotte reference trips auto availability. Respondent usually has a car available to make reference trip Yes No Respondent had a car available to make specific reference trip Yes 272 13 No 21 59 TABLE C-13. Chicago reference trips auto availability. Respondent usually has a car available to make reference trip Yes No Respondent had a car available to make specific reference trip Yes 345 39 No 51 290 The primary access mode is walk for Salk Lake City (FIGURE C-38) and Chicago (FIGURE C-39), and park and ride for Charlotte (FIGURE C-40). The primary egress mode is walk for both Charlotte and Chicago. There is no significant difference for access and egress mode by purpose. In Salt Lake City and Charlotte, people are more likely to use park and ride for mandatory trips than for non-mandatory trips. Interestingly, people at Chicago use less park and ride for mandatory trips than for non-mandatory trips. The Salt Lake City data only have access mode. FIGURE C-38. Salt Lake City access mode by purpose.

C-38 Characteristics of premium Transit Services that Affect Choice of Mode FIGURE C-39. Chicago transit access and egress mode by purpose. FIGURE C-40. Charlotte transit access and egress mode by purpose. Respondents from Chicago are more likely to make chain transit trips (46% of transit trips) than respondents from the other two cities (25% for Charlotte and 6% for Salt Lake City) as shown in FIGURE C-41. Respondents from Chicago experienced much higher total travel time, which may due to city scale (TABLE C-14). TABLE C-15 presents the auto trip characteristics, including travelers with no car available, where Salt Lake City has the highest percentage, and average stop times, where Charlotte has the highest value (although Salt Lake City data is not available). Salt Lake City has a shorter average travel time by auto than Charlotte, even though Charlotte is a smaller city and has shorter transit travel times.

detailed Survey Results C-39 TABLE C-14. Transit travel time for Charlotte, Chicago, and Salt Lake City. Average time (in minutes) Charlotte Chicago Salt Lake City Wait time 9 10 7 Transit in-vehicle time 26 37 31 Total travel time 65 83 54 Access/egress time Commute Access time 14 16 n/a Egress time 9 16 n/a Non- commute Access time 15 16 n/a Egress time 15 23 n/a TABLE C-15. Auto trips characteristics. Characteristic Charlotte Chicago Salt Lake City Do not usually have a car 2% 2% 6% Average travel time (min.) 26 28.4 22.4 Average stop time 6.5 5.2 n/a Parking cost incurred by (min.) 13.30% 4.80% 7% Tolls paid by 0.50% 12.80% n/a FIGURE C-41. Transit legs for Charlotte, Chicago, and Salt Lake City.

C-40 Characteristics of premium Transit Services that Affect Choice of Mode FIGURE C-42 shows the common reasons for not using transit among the three cities for respondents who considered one or more of the transit modes. The “needed my car for other reasons” is the most-cited reason for Chicago and Charlotte, while “Travel time too long” is the most-cited reason for Salt Lake City. FIGURE C-42. Reasons not taking transit in Charlotte, Chicago, and Salt Lake City. Stated Preference Survey Characteristics Chicago has the highest percentage of scenarios where car is chosen (29%) compared to Charlotte (22%) as shown in FIGURE C-43 and FIGURE C-44 for Charlotte and Chicago, respectively. These data are presented for Charlotte and Chicago only because the stated preference experiments in Salt Lake City were different. In the stated preference (SP) choice model the alternatives are assumed to be always considered if they are displayed to the user. This assumption is supported by the following statistics. Among 1,515 individuals, 550 individuals used bus or considered using bus. Among the remaining 965 individuals who never considered using bus in the revealed preference (RP) response, 428 actually selected bus in at least one of the eight SP scenarios. Similarly, out of 914 individuals who did not consider using train, 505 selected train in at least one of the eight SP scenarios. The usefulness of stated preference responses in logit choice modeling is to identify a wider range of options than are present in real life and therefore capture the trade-off point at which a traveler will change modes in the experiment. …

detailed Survey Results C-41 FIGURE C-43. Number of scenarios car is chosen in Charlotte. FIGURE C-44. Number of scenarios car is chosen in Chicago.