Effects of Real-Time Passenger Information Systems on Perceptions of Transit

Effects of Real-time Passenger Information Systems on Perceptions of Transit

Services: Investigations of The Ohio State University Community

THESIS

Presented in Partial Fulfillment of the Requirements for the Degree Master of Science in the Graduate School of The Ohio State University

Mahsa Ettefagh, B.S.

Graduate Program in Civil Engineering

The Ohio State University

2013

Thesis Committee:

Dr. Rabi G. Mishalani, Co-advisor

Dr. Mark R. McCord, Co-advisor

Dr. Gulsah Akar

Mahsa Ettefagh

2013

Abstract

Information systems are increasingly adopted in public transportation to improve service. When well utilized, transit also plays a major role in reducing roadway congestion and mitigating transportation’s negative impacts on the environment, benefits that accrue to transit users and nonusers alike. Nonuser’s recognition of such benefits can be important when agencies seek community support of transit initiatives, and it is possible that state-of-the-art technologies enhance the progressive image of transit.

A two-wave survey of The Ohio State University (OSU) community was conducted to study the influence of real-time passenger information systems on travelers’ perceptions of important dimensions associated with transit service. The first wave was conducted before OSU’s Campus Area Bus Service (CABS) implemented a state-of-the- art passenger information system. The second wave was conducted approximately a year after this implementation. Results show statistically significant increases in the value

CABS users ascribe to the service provided and their perceptions of personal safety while accessing and using transit services after the implementation of the information system.

The results also show statistically significant increases in both users’ and nonusers’ perceptions of the positive contribution CABS makes towards mitigating negative environmental impacts and decreasing congestion. Results further reveal statistically significant association of positive responses with a stated awareness of the

ii implemented information system. Socioeconomic variables -- such as gender, whether an individual is a current or past user of other metropolitan transit systems, and university affiliation (student, faculty, and staff) -- are included in the models to control for their possible influences on the community’s perceptions. Interesting statistically significant associations between these socioeconomic variables and positive perceptions are found as well.

In summary, the findings indicate the positive effects passenger information systems have on both transit users and nonusers among emerging college-educated professionals, a demographic important for future transit demand and support.

iii

Dedication

This document is dedicated to my family.

Acknowledgments

I would like to express my gratitude to my advisors, Dr. Mark R. McCord and

Dr. Rabi G. Mishalani for inspiring me to join the transportation field. Their weekly support and suggestions has always guided me through the right direction. I appreciate all the time they spent helping me through this research and always encouraging me to move forward. I have learned life lasting lessons from both of them which will be extremely valuable in my personal and professional life.

I would like to thank The Ohio State University, the Department of Civil,

Environmental, and Geodetic Engineering, along with Dr. Mark R. McCord and Dr. Rabi

G. Mishalani for the opportunity to join the Transit Lab as a Graduate Research

Associate. I would also like to extent a special thanks to Allan Johnson for his generous support of the Campus Transit Lab and the Allan Johnson fellowship, which I received during the academic year 2012-2013.

I am grateful to the U.S. Department of Transportation (DOT), Research and

Innovative Technologies Administration through the Region V University Transportation

Center (NEXTRANS), the Ohio State University (OSU) Transportation and Parking

(T&P) Services (now Department of Transportation and Traffic Management), the OSU

Transportation Research Endowment Program, and the OSU Department of Civil,

Environmental and Geodetic Engineering for their financial support of components of this research and of my graduate studies. I also acknowledge the assistance of Prem Goel

(OSU, Department of Statistics), Yoram Shiftan (Technion Israel Institude of

Technology, Department of Civil and Environmental engineering ), Sarah Blouch and

Chris Kovitya (formerly, OSU T&P), and Doug Moore, Ginny Barry, and Julie Wilder

(COTA). The views, opinions, findings, and conclusions reflected in this thesis are the responsibility of the author and co-advisors and do not represent the official policy or position of any of these entities or individuals.

Thanks to all my fellow research members who contributed to my research and helped me throughout my graduate studies. Special thanks to Ted Reinhold, Greg Hertler,

Greg Burch, Cheng Chen, Andrew Landgraf, Chenbo Shangguan, and Xiao Wei.

Finally thanks to all my family and friends for their continuous support and encouragement throughout this experience.

Vita

2011…………………………………...B.S. Civil Engineering, The Ohio State University

2011 to present ……………………….Graduate Research Associate, Department of

Civil, Environmental and Geodetic

Engineering, The Ohio State University

Fields of Study

Major Field: Civil Engineering

vii

Table of Contents

Abstract ...... ii

Dedication ...... iv

Acknowledgments...... v

Vita ...... vii

Table of Contents ...... viii

List of Tables ...... xi

List of Figures ...... xiv

Chapter 1: Introduction to Real-time Passenger Information System Use on OSU Campus

...... 1

1.1 Background ...... 1

1.2 Scope of Research ...... 3

1.3 Organization of Thesis ...... 5

Chapter 2: Data Preparation and Analysis Methodology ...... 6

2.1 Overview ...... 6

2.2 Survey Questionnaire ...... 6

viii

2.3 Response Rates of the Two Surveys ...... 8

2.5 Model Estimation Methodology ...... 13

Chapter 3: Effect of Provision of TRIP on Travelers’ Safety Perceptions towards CABS

...... 16

3.1 Overview ...... 16

3.2 Safety Walking to CABS Stops ...... 16

3.3 Safety Waiting for CABS Buses ...... 23

3.4 Safety Riding the Buses ...... 29

Chapter 4: Effect of TRIP on Travelers’ Perception of CABS’s Contribution to Reduction in Environmental Impacts and Congestion ...... 34

4.1 Overview ...... 34

4.2 “Green” Campus ...... 35

4.3 Traffic Reduction on Campus ...... 45

Chapter 5: Effect of TRIP on Travelers’ Perception of Valuableness of CABS’s Services

...... 56

5.1 Overview ...... 56

5.2 Valuableness of CABS Services ...... 57

Chapter 6: Conclusion...... 65

6.1 Summary and Conclusions ...... 65

6.2 Future Research ...... 70

References ...... 72

Appendix A: Wave 2 Questionnaire ...... 73

Appendix B: Survey Data Variable Names and Descriptions ...... 80

List of Tables

Table 2.1 Responses of wave 1 and wave 2 of online survey...... 9

Table 2.2 Names and description of variables used in this thesis ...... 11

Table 3.1 Results from model investigating the difference between responses to feeling safe walking to CABS stops before and after TRIP (data: wave 1 and 2) ...... 18

Table 3.2 Results from model investigating the potential socioeconomic factors on travelers’ responses to feeling safe walking to CABS stops (data: wave 1) ...... 20

Table 3.3 Results from model investigating effect of TRIP on travelers responses to feeling safe walking to CABS stops (data: wave 2) ...... 22

Table 3.4 Results from model investigating the difference between responses to feeling safe waiting for buses before and after TRIP (data: wave 1 and wave2) ...... 24

Table 3.5 Results from model investigating the potential socioeconomic factors on travelers’ responses to feeling safe waiting at stops (data: wave 1) ...... 26

Table 3.6 Results from model investigating effect of TRIP on travelers responses to feeling safe waiting at stops (data: wave 2) ...... 28

Table 3.7 Results from model investigating the difference between responses to feeling safe riding the buses before and after TRIP (data: wave 1 and wave2) ...... 30

Table 3.8 Results from model investigating the potential socioeconomic factors on travelers’ responses to feeling safe riding buses (data: wave 1) ...... 32 xi

Table 3.9 Results from model investigating effect of TRIP on travelers responses to feeling safe while riding the buses (data: wave 2) ...... 33

Table 4.1 Results from model investigating the difference between responses to CABS promoting green campus before and after TRIP (data: wave 1 and wave2) ...... 37

Table 4.2 Results from model investigating the potential socioeconomic factors on perception of CABS promoting green campus (data: wave 1) ...... 38

Table 4.3 Results from model investigating effect of TRIP on peoples responses to CABS promoting green campus as an interaction with CABS usage(data: wave 2) ...... 41

Table 4.4 Results from model investigating effect of TRIP on peoples responses to CABS promoting green campus as an interaction with MPT familiarity(data: wave 2) ...... 43

Table 4.5 Results from model investigating the difference between responses to CABS role in traffic reduction before and after TRIP (data: wave 1 and wave2) ...... 46

Table 4.6 Results from model investigating the potential socioeconomic factors on perception of CABS traffic reduction role (data: wave 1) ...... 48

Table 4.7 Results from model investigating effect of TRIP on peoples responses to CABS role in traffic reduction as an interaction with CABS usage (data: wave 2)...... 50

Table 4.8 Results from model investigating effect of TRIP on peoples responses to CABS traffic reduction role as an interaction with MPT familiarity (data: wave 2) ...... 53

xii

Table 4.9 Results from model investigating effect of TRIP on peoples responses to CABS role in traffic reduction as an interaction with CABS usage (data: wave 2)...... 55

Table 5.1 Results from model investigating the difference between responses to CABS services being valuable to travel needs before and after TRIP (data: wave 1 and wave2) 58

Table 5.2 Results from model investigating the difference between responses ease of access to CABS information before and after TRIP (data: wave 1 and wave2) ...... 59

Table 5.3 Results from model investigating the potential socioeconomic factors on CABS valuableness to travel needs (data: wave 1) ...... 61

Table 5.4 Results from model investigating effect of TRIP on the value of CABS to travelers (data: wave 1) ...... 63

Table B.1 Survey Data Variable Names and Description, (Continued on the next page) 80

xiii

List of Figures

Figure 2.1 Survey data processing flowchart ...... 12

Figure 2.2 Modeling flowchart ...... 15

xiv

Chapter 1: Introduction to Real-time Passenger Information System Use on OSU

Campus

1.1 Background

Bus transit systems provide valuable service in many urban areas around the world. Public transportation offers people options and flexibility for travel to meet a variety of purposes. Public transportation also serves the travel needs of a large segment of the population that is transit dependent. However, the benefits of public transportation systems are not limited to only providing transportation options for users. Public transportation also plays a major role in reducing congestion on the roads, when well utilized, and therefore enhancing environmental sustainability. Increased use of transit systems leads to a decreased use of cars resulting in a reduction of greenhouse gas emissions and the negative environmental impacts caused by surface transportation.

People will use transit if they find its services valuable and convenient. Public

Transportation systems that are reliable, adhere to a published schedule, and serve more origins and destinations are usually more valuable and satisfactory to travelers. However, safety concerns can be an impediment to public transportation use. People may not feel safe when walking to the bus stops, especially in areas with high crime rates. Anxiety and safety concerns might also arise while individuals are waiting at bus stops. Similarly,

1 riding a public transportation vehicle could also cause safety concerns for travelers compared to driving personal vehicles.

Public transportation is usually not profitable based on operating revenues. To provide good transportation service, transit agencies require subsidies. Many transit agencies rely on levies from voters in the form of sales taxes, for example, to go towards funding public transportation. Since public transit usage is relatively low in some areas, for these levies to pass, transportation agencies need the support of nonusers of the service, as well as users. Nonusers may be more enticed to vote for taxes to go towards public transportation sector if they realize its positive role in reducing congestion and promoting a “green” environment, in addition to providing mobility to the transit dependent and general public.

Many transit agencies have been implementing automatic services and information technologies in order to improve system performance and level of service for their passengers. Agencies can use the automatically collected data for their scheduling and planning purposes, and travelers can use the real-time information provided to plan their trips, choose bus routes, and plan arrival times at the desired bus stops. In addition, providing real-time information could decrease the anxiety of travelers when waiting for buses and potentially decrease users’ safety concerns. It is also possible that even nonusers would perceive the provision of high technology systems as an effort to improve transit services and decrease congestion and environmental impacts. This improved perception could translate into increased public support for funding levies.

This study investigates individuals’ perceptions of transit in relation to several of

2 the aspects discussed above, with an emphasis on how these perceptions are impacted by the provision of real-time information to passengers. In order to study the influences of providing real-time information on travelers’ satisfaction and opinions about the transit system, The Ohio State University’s Campus Area Bus Service (CABS) is used as a test site. Specifically, the data from a two-wave online survey, conducted by the OSU

Campus Transit Lab (CTL, 2013) team, is used in this study. The first wave of the survey was conducted in autumn of 2008, before the provision of real-time passenger information. The second wave was conducted a year after the implementation of the passenger information system in spring of 2010. The results of the first wave are used as benchmarks for investigating the perceptions of the OSU community of various aspects of the impacts of CABS. The results of the second wave are used to determine changes in perceptions resulting from the implementation of the real-time information system. This study investigates responders’ perception of CABS and the possible effects real-time information could have on these perceptions in relation to safety, environmental impacts and traffic congestion, and the overall value of transit service to travel needs.

1.2 Scope of Research

The Ohio State University (OSU) campus is one of the largest in the U.S., serving more than 55,000 students, 3,000 faculty, and 5,000 staff members. The OSU’s CABS operates 7 different routes with a fleet size of approximately 40 buses. CABS serves over

4 million passengers per year, and the majority of the OSU population commutes to campus. More than70% and 96%, respectively, of OSU undergraduate and graduate

3 students live off-campus. In addition, many faculty and staff park their personal vehicles in parking lots located in the outskirts of the main campus area. CABS provides frequent services between West Campus parking lots and the central campus academic core. The buses also connect multiple residential areas and OSU’s Medical center to the central

Campus core. High volumes and multiple service areas make CABS a system that can be considered representative of other transit systems operating in many urban areas.

To provide real-time information to travelers, CABS implemented a passenger information system, called Transportation Route Information Program (TRIP), in autumn

2009. Users can access real-time and general information from variable message boards at stops, text messaging options, mobile apps, a website, and static and dynamic bus location maps. In the following sections of this thesis, the real-time passenger information system will be referred to as “TRIP”.

In this study, the effects of TRIP on the attitudes and perceptions of the OSU community towards safety issues, environmental and congestion impacts, and the value of CABS services are investigated. For each specific dimension of attitude and perception, a set of socioeconomic factors that may affect people’s perceptions on different dimensions are considered. The socioeconomic factors considered in this study are gender, prior or concurrent familiarity with other metropolitan public transportation systems, and responder’s affiliation categories (whether the respondent was undergraduate student, graduate student, staff, or faculty). The potential effects of TRIP on the stated issue are investigated in the presence of these socioeconomic factors.

1.3 Organization of Thesis

The rest of this thesis is organized as follows. The general description of the survey, data processing and classification of the responses, and the analysis methodology are presented in Chapter 2. In the subsequent chapters, the empirical results of investigating the relation between the provision of the new real-time information system and the various dimensions of perception and attitudes are discussed. Specifically, in

Chapter 3, the effect of providing real-time information on individuals’ feelings towards safety when walking to stops, when waiting for buses, and when riding the buses is investigated. Chapter 4 focuses on the potential influences of the information system on individuals’ perceptions towards CABS’s positive role in traffic congestion reduction and environmental impacts. Chapter 5 investigates the perceptions of the value of CABS services and how providing information may improve these perceptions. In addition, possible enhancements to accessing information with regards to the value of CABS services after the provision of this information system are investigated in this chapter.

The last chapter of this thesis summarizes the general findings and potential future research directions are discussed.

Chapter 2: Data Preparation and Analysis Methodology

2.1 Overview

This chapter explains the process of collecting two waves of a survey of the OSU community, the transformation of data into binary indicator variables, the determination of explanatory variables, and the methodology used to estimate different model specifications.

2.2 Survey Questionnaire

As part of a prior effort, two waves of a survey of the OSU community were designed and implemented by Campus Transit Lab (CTL), aimed at studying the perceptions and attitudes of the community towards public transportation. The second wave questionnaire is in Appendix A. It should be noted that the first and second wave questions are identical except for the last eight questions (questions 61 through 68) shown in Appendix A which were only asked in the second wave. After several iterations to complete the design of the questionnaires, the Ohio State University Statistical

Consulting Service (SCS) was contracted to implement the designed surveys on-line.

SCS obtained a random sample of email addresses of four OSU community categories

(undergrad students, grad students, faculty, and staff) and the forms were sent to the sample of subjects to participate in the survey. The first wave of the online survey was 6 sent out in autumn of 2008 (asking about individuals experiences during spring quarter of

2008) and the second wave was sent out in spring of 2010.

The survey data have information about respondents’ demographic characteristics, mode of transportation to and on campus, their perceptions and evaluation of CABS services, and whether they noticed CABS use of modern technology (TRIP) or not (only asked in the second wave). What follows is a summary of these questions:

 Nine demographic and socioeconomic related questions (e.g. gender, affiliation,

etc.) (questions 1 through 9),

 Ten to thirteen questions related to the mode of transportation to and on campus

(the number depends on a subject’s response on certain questions) (e.g. mode of

transportation, parking location, etc.),

 fourteen questions related to perceptions and evaluation of CABS services, safety,

and contribution to reducing traffic congestion on campus and enhancing the

environmental sustainability of the campus (i.e. making the campus “green”), and

 In addition, the second wave included a question about noticing the new real-time

information (TRIP).

The third set of questions above, which are designed to elicit respondents’ perceptions of and attitudes towards CABS, are presented using a Likert scale. The possible responses are labeled as 1: “Strongly disagree”, 2: “Disagree”, 3: “Neutral”, 4:

“Agree”, 5: “Strongly agree” and the option of responding “NA” was also available.

Depending on whether responders are CABS users or not, the fourth set of questions inquire whether they noticed the use of new real-time passenger information system by

CABS, any route changes, or any changes in the number of buses serving the campus area.

2.3 Response Rates of the Two Surveys

The response rates for each wave are shown in Table 2.1 (McCord et al., 2009 and

Mishalani et at. 2011) To understand the adequacy of these rates, they are compared with response rates to other OSU surveys. In 2009, a survey was conducted regarding attitudes and perceptions of OSU undergraduate students on global warming (McCord et al.,

2009.) A random sample of 24,900 undergraduate students was selected and 3570 responded to this survey. The 14.3% response rate is similar to the response rates of undergraduate students in the wave 1 and wave 2 surveys used in this study (13.95% and

13.72%, respectively). In aggregate, however, larger rates of 24.5 % and 23.46% are achieved for waves 1 and 2 respectively compared. It should be noted that wave 1 and 2 had different proportions of responders from each category. The substantially higher response rates from faculty, staff and graduate students are consistent with those achieved in a 2008 survey devoted to Transportation and Parking issues at OSU (McCord et al.,

2009.)

Table 2.1 Responses of wave 1 and wave 2 of online survey

Wave 1 Response Rates Wave 2 Response Rates Category Surveyed Responses Rate Surveyed Responses Rate Faculty 4,480 1,233 27.52% 4,500 1,165 26.90% Staff 4,479 1,758 39.25% 4,500 1,525 35.22% Grad Students 2,994 571 19.07% 4,500 1,061 24.50% UG Students 5,999 837 13.95% 7,500 990 13.72% Overall 17,952 4,399 24.50% 21,000 4,741 23.46%

2.4 Survey Data

This study focuses on the effect of the provision of passenger information system

(TRIP) has on individuals’ perceptions of and attitudes towards safety, environmental sustainability and congestion, and value of CABS services. This section explains how the survey data from two waves are processed. A flowchart of the processing followed is shown in Figure 2.1. First the responses were organized and formatted based on the designed questionnaire code book. Formatted responses from each wave were then transferred into Excel files with each row indicating all the responses from one specific individual. The fields of responses that are not of interest for the purpose of this study were filtered out of the Excel files. The remaining fields were further formatted into binary indicator variables with 1s and 0s. In the case of Likert scale responses, “strongly agree” and “agree” responses were aggregated to a categorical variable “agree” and assigned a value of 1 to indicate agreement. The “neutral”, “disagree”, and “strongly disagree” responses were similarly aggregated to a categorical variable “not agree” and assigned a value of 0 to indicate lack of agreement. There were some cases in which responders had not answered certain questions. The entries in the corresponding fields 9 were replaced with “NA” values. In this case no distinction is made between these “Not

Available” values and those provided by respondents as “Not Applicable.”

The fields used in this study and their contents following this processing are shown in Table 2.2. These fields from either the dependent or explanatory variables used in the models developed in Chapters 3 through 5. For example, agreeing to feel safe waiting at CABS stops (Field number 11 in Table 2.2) is the dependent variable used in studying the effect of TRIP on feeling safe waiting at stops. And affiliation (Field numbers 1, 2, 3, and 4 in Table 2.2), gender (field number 5 in Table 2.2), MPT (field number 6 in Table 2.2), and noticing TRIP (field number 15, 16, 17, and 18) are used as possible explanatory variables.

The newly formatted data were saved as a .CSV file and read in the statistical computing program R. Responses from waves 1 and 2 were pooled in R with a wave indicator specification assigned to each individual.

Table 2.2 Names and description of variables used in this thesis

Field number Field Description Field Content 1 Faculty 1 if faculty, 0 otherwise 2 Staff 1 if staff, 0 otherwise 3 Graduate students 1 if graduate student, 0 otherwise 1 if undergrad student, 4 Undergraduate student 0 otherwise 5 Gender 1 if female, 0 if male 1 if familiar with MPTs, 6 Using other Metropolitan public transportation 0 otherwise 7 CABS role in traffic reduction on campus 1 if agree, 0 if not agree 8 CABS role in promoting green campus 1 if agree, 0 if not agree 9 CABS services are valuable to travel needs 1 if agree, 0 if not agree 10 Feeling safe walking to CABS stops 1 if agree, 0 if not agree 11 Feeling safe waiting for CABS buses 1 if agree, 0 if not agree 12 Feeling safe riding CABS buses 1 if agree, 0 if not agree 13 Ease of access to CABS information 1 if agree, 0 if not agree 14 CABS usage 1 if users, 0 if nonusers Users who noticed changes in information 1 if users noticed TRIP, 15 CABS provides 0 otherwise Users who noticed changes in areas served by 1 if users noticed route changes, 16 CABS 0 otherwise Nonusers who noticed CABS use of modern 1 if nonusers noticed TRIP, 17 technology 0 otherwise Nonusers who noticed changes in number of 1 if nonusers noticed changes in 18 CABS buses number of buses, 0 otherwise

Figure 2.1 Survey data processing flowchart

2.5 Model Estimation Methodology

The main motivation for this research is the interest in the effects of advanced real-time information systems (TRIP in this case) on the perception of and attitudes toward transit (CABS in this case) in the presence of additional socioeconomic independent (explanatory) variables. To quantify the effects of information systems, the variables as defined in section 2.4 are considered. Binary choice models are used when individual n has a choice of agreeing (alternative i) or not agreeing (alternative j) to a specific question. To model this binary choice, the Logit model is used. Following a typical formulation (Ben-Akiva and Lerman, 1985), the choice probability of individual n choosing alternative i is given by:

Pn (i) = (2.1) ( ) where Vn is the difference in the systematic utilities associated with individual n choosing alternative i and alternative j. The specification of Vn is linear in the parameters, as follows:

Vn = β1 Xn1 + β2 Xn2 + ... + βk Xnk + ... + βK XnK (2.2)

Where,

Xnk = independent variable (0 or 1),

βk = the coefficient associated with explanatory variable Xk, and

K = total number of explanatory variables.

In all of the models used in this study, Xn1 is set equal to 1 resulting in β1 representing the coefficient of an alternative specific constant. The other explanatory

13 variables are developed in chapters three through five, specified explicitly for each of the models.

The overall approach to modeling is shown in Figure 2.2. First the generalized

Logit estimations on the variables are run in R. The estimated coefficients and statistical summaries for each specification are then produced in R and reported. After studying each of the estimated specifications, if the model is supporting the initial hypothesis and the signs of the estimated coefficients are meaningful, the results are reported and the estimations are interpreted. Otherwise, insignificant explanatory variables are taken out of the model or other potential explanatory variables are added and new specifications are estimated in R.

Figure 2.2 Modeling flowchart

Chapter 3: Effect of Provision of TRIP on Travelers’ Safety Perceptions towards

CABS

3.1 Overview

Concerns about personal safety can be a deterrent to the use of public transportation. This portion of the study aims to explore how perceptions of safety when using transit are affected by providing a real-time passenger information. Specifically, relationships between perceptions of personal safety on CABS and noticing TRIP are modeled using data from the two waves of the online survey. Field numbers 10, 11, and

12 shown in Table 2.2 from both waves are used to study the perceptions of safety when walking to bus stops, waiting at the bus stops, and riding the buses. Different models are estimated for these three safety components to investigate if providing real-time passenger information system on campus (via TRIP) on campus has an effect on travelers’ safety concerns when using the buses.

3.2 Safety Walking to CABS Stops

To investigate the effects of TRIP on safety walking to bus stops, responses by

CABS users to the “I feel safe walking to CABS stops” statement were used (field number 10 in Table 2.2). This statement was asked in both waves of the survey. Since the emphasis is on the perception of safety for people walking to stops to use transit, only responses from individuals who used CABS were considered in this analysis. In both 16 waves of the survey, the question of “During this quarter, how often do you use CABS

(whether to travel to campus or on campus)?” (field number 12 in Table 2.2) was asked.

The responders had three options to choose from: “I never use CABS”, “I use CABS occasionally”, or “I use CABS regularly.” The individuals who responded “I use CABS occasionally” or “I use CABS regularly” were aggregated into the “users” category and those who selected “I never use CABS” were aggregated into the “nonusers” category.

This specification will remain the same for the remainder of this thesis.

If TRIP has an effect on how travelers feel about safety walking to CABS stops, the second wave would be expected to produce a higher probability of positive responses

(“agree” or “strongly agree”) than the first survey. To investigate this hypothesis, the simple specification shown in Table 3.1 was estimated. In this model, the data from both surveys were pooled together. The dependent variable is set to 1 if the individual chose

“agree” or “strongly agree” in response to the statement “I feel safe walking to CABS stops.” The explanatory variable X2 indicates if the individual response belonged to the second wave (X2n = 1) or the first survey (X2n = 0). As explained earlier, the first variable

X1 is set to 1 so that the first estimated coefficient represents an alternative specific constant in the logit specification. The maximum likelihood estimates of the estimated coefficients (β1 = 1.9792 and β2 = 0.389) are found in the third column of Table 3.1. The standard errors, t-statistics, and P-values for each variable are presented in the subsequent columns of the table. Other statistical results of this model can also be found under the

“summary statistics” row in Table 3.1.

Both coefficients are statistically significant as indicated by the t-statistics. The positive signs of β1 and β2 imply that whether the individual responded to the first wave

(X2n = 0) or second wave (X2n = 1), he or she has a probability greater than 0.5 of agreeing to this statement. That is, in both waves, people generally felt safe walking to

CABS stops. The positive coefficient for X2n (β2 = 0.389) indicates that CABS users felt safer in the second wave (when TRIP was in effect) compared to the first wave.

Table 3.1 Results from model investigating the difference between responses to feeling safe walking to CABS stops before and after TRIP (data: wave 1 and 2)

Estimated Variable Description Std. Error t-stat P-value Coefficient (βs) X1 Constant 1.9792 0.077 25.57 < 2e-16 1 if wave 2 X2 0.389 0.117 3.33 0.000865 0 if wave 1 Summary Statistics Number of observations 3235 Number of cases = 3235 LL(0) = -2242.33 LL(c) = -1072.36 LL(β) = -1066.75 ρ² = 0.5242 (-2)[LL(0)-LL(β)] = 2351.14 (-2)[LL(c)-LL(β)] = 11.21

It would also be useful to control for possible socioeconomic factors that might influence an individual’s perception of safety walking to CABS stops. In this study, gender, familiarity with other types of metropolitan public transportation systems

(MPTs), and affiliation (i.e. faculty, staff, graduate students, and undergraduate students) are considered. There were different proportions of the four categories of affiliation in first and second wave. This could potentially have an effect on the estimated specifications. Females may be expected to be more concerned about safety issues 18 compared to males, all other things being equal. In addition, individuals who are familiar with other types of metropolitan public transportation systems (MPTs) may feel safer using transit in general, than those who have no prior or concurrent experience with public transportations.

To investigate the possible effects of these factors on an individual’s perception of safety, the specification presented in Table 3.2 was estimated using only wave 1 data when TRIP was not available. In the specification shown in Table 3.2, the dependent variable is set to 1 if the individual chose “agree” or “strongly agree” to the feeling safe walking to bus stops statement and 0 otherwise. An “undergraduate student” affiliation was used as the base, and the three other categories were explicitly included as independent dummy indicator variables. Gender was chosen as another one of the explanatory indicator variables. This variable was set to 1 (X2n = 1) if the respondent was female and 0 (X2n = 0) if male. Individuals who are familiar with other MPTs were assigned a value of 1 (X3n = 1), and those with no prior or concurrent experience were assigned a value of 0 (X3n = 0). Whether the individual was considered to be familiar with

MPTs was based on his or her response to the question “Other than OSU’s Campus Area

Bus Service (CABS), do you currently use metropolitan public transportation or have you regularly used it in the past?” (field number 4 in Table 2.1.)This question was asked in both waves of the survey.

Table 3.2 Results from model investigating the potential socioeconomic factors on travelers’ responses to feeling safe walking to CABS stops (data: wave 1)

Estimated Variable Description Std. Error t-stat P-value Coefficient (βs) X1 Constant 2.5483 0.197 12.929 < 2e-16 1 if Female X2 -0.8888 0.174 -5.104 3.32E-07 0 if Male 1 if MPT X3 0.1803 0.174 1.040 0.2986 0 otherwise 1 if Faculty X4 0.3408 0.268 1.271 0.2038 0 otherwise 1 if Staff X5 -0.4276 0.187 -2.283 0.0225 0 otherwise 1 if GRAD X6 0.1174 0.248 0.473 0.636 0 otherwise Summary Statistics Number of observations 1551 Number of cases = 1551 LL(0) = -1075.07 LL(c) = -574.82 LL(β) = -551.53 ρ² = 0.49 (-2)[LL(0)-LL(β)] = 1047.08 (-2)[LL(c)-LL(β)] = 46.57

As indicated by the t-statistics shown in Table 3.2, the estimated coefficients for the constant, female indicator, and staff are statistically significant. The estimated coefficient for the constant term is positive and significant. The magnitude of this coefficient is such that the combination of any other set of explanatory variables will lead to a positive β1X1 +β2X2+… +βk Xk. Therefore no matter what socioeconomic category a traveler belong to, he or she has a probability greater than 0.5 of feeling safe walking to

CABS stops. The negative sign of the female indicator variable (β2 = -0.8888) agrees with the initial hypothesis that females in general feel less safe compared to males. The positive sign of the MPT indicator indicates that individuals who were using other transit

20 systems at the time of the surveyor had previously used them, in general felt safer walking to CABS stops compared to those who had no prior or concurrent experience with other forms of MPTs. However the effect is not significant. Out of the three categories of OSU affiliation, only staff seems to have felt less safe walking to bus stops compared to the rest of the travelers. This could potentially be due to their work hours and familiarity with the campus area. For example, about 50% of staff work during night shifts, especially in the OSU’s medical and health sciences area. They may feel less safe walking during the night hours to bus stops.

In order to capture the effects of TRIP on perception of safety walking to stops, only wave 2 data were used in the following analysis. As seen in the wave 1 results

(shown in Table 3.2), certain socioeconomic factors affect travelers’ safety perceptions with regards to walking to stops. The socioeconomic factors found to have significant coefficients were included in the model using the wave 2 data. A general specification with all the possible interactions between the significant variables from the Table 3.2 specification (i.e., gender and staff) and noticing TRIP was estimated. For this model only the data from wave 2 were used since the question of noticing changes in the information CABS provides was asked in the second wave after the provision of TRIP.

However, this specification led to the use of many variables, thereby reducing the quantity of data available per variable. When estimating models under such specifications, some of the resultant estimates were insignificant, while others had signs inconsistent with expectations. Therefore, a simplified specification was estimated, also using the second wave data considering CABS users.

In this simplified specification, gender and staff were kept as indicator variables to control for their effects on travelers perception of safety while the role of TRIP was investigated. The dependent variable is again set to 1 for individuals responding “strongly agree” or “agree” to feel safe walking to the CABS stops statement. The explanatory variables consist of staff (X2n = 1 if staff), females (X3n = 1 if female), and users who had noticed the new information system (X4n = 1 if TRIP). The estimation results are shown in Table 3.3.

Table 3.3 Results from model investigating effect of TRIP on travelers responses to feeling safe walking to CABS stops (data: wave 2)

Estimated Variable Description Std. Error t-stat P-value Coefficient (βs) X1 Constant 2.2852 0.184 12.447 < 2e-16 1 if Female X2 -0.2897 0.192 -1.510 0.131 0 if Male 1 if Staff X3 -0.4037 0.208 -1.939 0.052 0 otherwise 1 if noticed TRIP X4 0.604 0.189 3.204 0.001 0 otherwise Summary Statistics Number of observations 1459 Number of cases = 1459 LL(0) = -1011.30 LL(c) = -426.64 LL(β) = -418.55 ρ² = 0.59 (-2)[LL(0)-LL(β)] = 1185.51 (-2)[LL(c)-LL(β)] = 16.18

In Table 3.3, all the variables’ coefficients are significant. The estimated coefficient for the constant term is positive and has an appreciably higher magnitude compared to all other coefficients. The magnitude of this coefficient is such that the

22 combination of any other set of explanatory variables will lead to a positive

β1X1 +β2X2+… +βk Xk. Therefore no matter what socioeconomic category a traveler belong to, he or she has a probability greater than 0.5 of feeling safe walking to CABS stops. The estimated coefficient for the X2 variable is again negative, indicating that females felt less safe compared to males. Staff travelers, which are represented by the X3 variable, again have a negative estimated coefficient (β3 = -0.4037). Again this may be due to the wide range of work hours that they may have and familiarity with campus area.

The X4 variable indicates whether CABS users noticed the new information system. The coefficient for this variable is positive and significant (β4 = 0.604, t-stat = 3.204). The positive sign of the noticing TRIP variable agrees with the original hypothesis of the positive role of TRIP in increasing traveler’s perception of safety walking to CABS stops.

Knowing when the next bus will arrive at a specific stop, allows travelers to plan their arrival time more precisely. This planned arrival time may make the users feel more confident while walking to stops.

3.3 Safety Waiting for CABS Buses

Similar to section 3.2 regarding safety walking to CABS stops, a question with regards to feeling safe waiting at the bus stops was asked in both waves. To investigate the effects of TRIP on safety waiting at CABS stops, responses by CABS users to the “I feel safe waiting for CABS buses” (field number 11 in Table 2.2) statement were used.

Again, only the responses of CABS users were used in this analysis since the emphasize is on the perception of safety for travelers waiting at the bus stops. If TRIP had an effect

23 on how safe people felt while waiting for CABS buses, the second wave would be expected to produce a higher probability of positive responses (“agree” or “strongly agree”) than the first wave. To investigate this hypothesis, the simple specification shown in Table 3.4 was estimated. The data from both waves were pooled together and used in this model. The dependent variable is set to 1 if the individual chose “agree” or “strongly agree” in response to the “feeling safe waiting for buses”. The explanatory variable X2 indicates if the individual response belonged to the second wave (X2n = 1) or the first wave (X2n = 0).

Table 3.4 Results from model investigating the difference between responses to feeling safe waiting for buses before and after TRIP (data: wave 1 and wave2)

Estimated Variable Description Std. Error t-stat P-value Coefficient (βs) X1 Constant 1.81805 0.073 24.950 < 2e-16 1 if wave 2 X2 0.38051 0.109 3.480 0.0005 0 if wave 1 Summary Statistics Number of observations 3240 Number of cases = 3240 LL(0) = -2245.80 LL(c) = -1183.25 LL(β) = -1177.14 ρ² = 0.48 (-2)[LL(0)-LL(β)] = 2137.32 (-2)[LL(c)-LL(β)] = 12.22

As indicated by the t-statistics shown in Table 3.4, both estimated coefficients are significant. The positive signs of β1 and β2 imply that travelers have a probability greater than 0.5, whether they responded to the first wave (X2n = 0) or second wave (X2n = 1).

That is, in both waves, travelers generally felt safe waiting for CABS buses. The positive

24 coefficient for X2n (β2 = 0.10936) indicates that travelers felt safer in the second wave compared to first wave.

Similar to the case of safety walking to CABS stops, possible socioeconomic factors that may influence an individual’s perception of safety waiting for CABS buses are investigated. Again, affiliation, gender, and familiarity with other MPTs are used as possible explanatory variables here. To investigate the possible effects of these socioeconomic factors on an individual’s perception on safety waiting at stops, the specification presented in Table 3.5 was estimated using only wave 1 data. In this specification, the dependent variable is set to 1 if the individual chose “agree” or

“strongly agree” to the feeling safe waiting for CABS buses statement and 0 otherwise.

Again “undergraduate students” were used as the base, and the three other categories were explicitly included as independent indicator variables. The gender explanatory variable was set to 1 (X2n = 1) for female, and the MPT variable was set to 1 (X3n = 1) for travelers who were familiar with other MPTs.

Table 3.5 Results from model investigating the potential socioeconomic factors on travelers’ responses to feeling safe waiting at stops (data: wave 1)

Estimated Variable Description Std. Error t-stat P-value Coefficient (βs) X1 Constant 2.35742 0.184 12.791 < 2e-16 1 if Female X2 -1.00057 0.165 -6.046 1.49E-09 0 if Male 1 if MPT X3 0.20653 0.162 1.275 0.202 0 otherwise 1 if Faculty X4 0.45042 0.251 1.795 0.073 0 otherwise 1 if Staff X5 -0.12827 0.180 -0.713 0.476 0 otherwise 1 if GRAD X6 -0.06729 0.216 -0.311 0.756 0 otherwise Summary Statistics Number of observations 1554 Number of cases = 1554 LL(0) = -1077.15 LL(c) = -630.11 LL(β) = -604.20 ρ² = 0.44 (-2)[LL(0)-LL(β)] = 945.91 (-2)[LL(c)-LL(β)] = 51.84

The estimated coefficients for the constant, female indicator, and faculty variables are significant in Table 3.5. The estimated coefficient for the constant term is positive and has an appreciably higher magnitude compared to all other coefficients. The magnitude of this coefficient is such that the combination of any other set of explanatory variables will lead to a positive β1X1 +β2X2+… +βk Xk. Therefore no matter what socioeconomic category a traveler belong to, he or she has a probability greater than 0.5 of feeling safe waiting for CABS buses. The negative sign of the female indicator variable (β2 = -

1.00057) agrees with the initial hypothesis that females in general feel less safe waiting compared to males. Out of the three categories of OSU community, only faculty seems to

26 have felt safer waiting at stops compared to rest of travelers. This could be due to the work hours of faculty members which are usually during the day. That is, they most probably use CABS during the hours of day when the campus is busy. The safety issues are usually a concern to individuals who travel during the less busy hours and night time hours. Further studies can be conducted to investigate this hypothesis.

In order to capture the effects of TRIP on perception of safety waiting at stops, only wave 2 data were used in the following analysis. As seen in the wave 1 results

(shown in Table 3.5), gender and whether or not the individual was a faculty member affect travelers’ perceptions of safety while waiting at stops. Therefore, these two significant socioeconomic variables were included in the model using the wave 2 data. A general specification with all the possible interactions between the significant variables from Table 3.5 and noticing TRIP was estimated. However, this specification again led to the use of many variables thereby reducing the quantity of data available per variable. In the resulting models, some of the estimates again turned out to be insignificant, while others had unexpected signs. Therefore, the simpler specification shown in Table 3.6 was estimated without any interaction terms. In this specification, gender and faculty were kept as indicator variables to control for their effects on travelers perception of safety waiting at bus stops while the role of TRIP was investigated. The dependent variable is set to 1 for individuals responding “strongly agree” or “agree” to the feeling safe waiting for CABS buses statement. The explanatory variables consist of gender (X2n = 1, if female), affiliation (X3n = 1, if faculty), and noticing TRIP or not (X4n = 1, if noticed

TRIP).

Table 3.6 Results from model investigating effect of TRIP on travelers responses to feeling safe waiting at stops (data: wave 2)

Estimated Variable Description Std. Error t-stat P-value Coefficient (βs) X1 Constant 2.1283 0.178 11.989 < 2e-16 1 if Female X2 -0.6635 0.187 -3.555 0.00038 0 if Male 1 if Faculty X3 1.0719 0.398 2.693 0.00708 0 otherwise 1 if noticed TRIP X4 0.6183 0.176 3.522 0.00043 0 otherwise Summary Statistics Number of observations 1463 Number of cases = 1463 LL(0) = -1014.07 LL(c) = -481.53 LL(β) = -463.17 ρ² = 0.54 (-2)[LL(0)-LL(β)] = 1101.81 (-2)[LL(c)-LL(β)] = 36.73

In Table 3.6, all the variables’ coefficients are significant. The estimated coefficient for the constant term is positive and has an appreciably higher magnitude compared to all other coefficients. The magnitude of this coefficient is such that the combination of any other set of explanatory variables will lead to a positive β1X1

+β2X2+… +βk Xk. Therefore no matter what socioeconomic category a traveler belong to, he or she has a probability greater than 0.5 of feeling safe waiting for buses. The estimated coefficient for the X2 variable representing gender is negative (β2 = -0.6635) in

Table 3.6. The negative sign of X2’s coefficient indicates again that females felt less safe waiting for buses compared to males. Similar to the results of the wave 1 estimation shown in Table 3.5, the estimated coefficient for the faculty indicator variable also turns out to be positive and significant. Again, this could be due to their normal working hours

28 in which they use CABS. Knowing when the next bus will arrive decreases the uncertainly of travelers and may consequently decrease their anxiety while waiting for the buses. Indeed, the coefficient for the X4 variable indicating whether the individual noticed the new information system, is positive and significant (β4 = 0.6183). The positive sign of the noticing TRIP variable agrees with the original hypothesis of the positive role of TRIP in increasing travelers’ perceptions of safety while waiting for

CABS buses.

3.4 Safety Riding the Buses

A question regarding feeling safe while riding the CABS buses was also asked in both waves. To investigate the effects of TRIP on safety riding CABS buses, responses by

CABS users to the “I feel safe riding CABS buses” (field number 12 in Table 2.2) statement were used. Again, only the responses to CABS users were relevant to this statement. In addition to providing real-time bus arrival times to passengers, the CABS buses were equipped with variable signs inside buses that indicate the next bus stop and available transfers at each approaching stop. The system also announces this information to bus riders. Such a system reassures passengers that they are on the correct bus route and their location along the route. This reassurance would reduce anxiety and may lead to an increase in the sense of safety.

If TRIP had an effect on how safe people felt while riding CABS buses, the second wave would be expected to produce a higher probability of positive responses (“agree” or

“strongly agree”) than the first wave. To investigate this hypothesis, the specification

29 shown in Table 3.7 was estimated using only CABS users’ responses. Again, the data from both waves were pooled together and used in this model. The dependent variable is set to 1 if the individual chose “agree” or “strongly agree” in response to the feeling safe riding CABS buses statement. The explanatory variable X2 indicates whether the individual response belonged to the second wave (X2n = 1) or the first wave (X2n = 0).

Table 3.7 Results from model investigating the difference between responses to feeling safe riding the buses before and after TRIP (data: wave 1 and wave2)

Estimated Variable Description Std. Error t-stat P-value Coefficient (βs) X1 Constant 2.6885 0.103 26.016 <2e-16 1 if wave 2 X2 0.395 0.158 2.498 0.0125 0 if wave 1 Summary Statistics Number of observations 3238 Number of cases = 3238 LL(0) = -2244.41 LL(c) = -675.08 LL(β) = -671.92 ρ² = 0.70 (-2)[LL(0)-LL(β)] = 3144.98 (-2)[LL(c)-LL(β)] = 6.32

As indicated by the t-statistic shown in Table 3.7, both estimated coefficients are significant. The positive signs of β1 and β2 imply that whether the individual responded to the first wave (X2n = 0) or second wave (X2n = 1), he or she has a probability greater than

0.5 of feeling safe while riding the buses. The positive coefficient for X2n (β2 = 0.395) indicates that travelers felt safer in the second wave after the provision of real-time passenger information compared to first wave.

Similar to previous investigations, different effects of socioeconomic groups on how travelers felt about riding the buses were explored using the wave 1 data. The 30 specification shown in Table 3.8 was estimated again using gender, MPT, and affiliation as explanatory indicator variables. The estimated coefficient for the constant term is positive and has an appreciably higher magnitude compared to all other coefficients. The magnitude of this coefficient is such that the combination of any other set of explanatory variables will lead to a positive β1X1 +β2X2+… +βk Xk. Therefore no matter what socioeconomic category a traveler belong to, he or she has a probability greater than 0.5 of feeling safe riding CABS. The t-statistic values shown in Table 3.8 indicate that all of these explanatory variables are insignificant in how travelers felt towards safety riding the buses. There is no specific reason as to why any of these socioeconomic groups should feel different about safety while riding CABS buses.

Table 3.8 Results from model investigating the potential socioeconomic factors on travelers’ responses to feeling safe riding buses (data: wave 1)

Estimated Variable Description Std. Error t-stat P-value Coefficient (βs) X1 Constant 2.8561 0.243 11.753 <2e-16 X2 1 if Female -0.256 0.213 -1.200 0.23 0 if Male X3 1 if MPT 0.2421 0.231 1.049 0.294 0 otherwise X4 1 if Faculty 0.3812 0.362 1.053 0.292 0 otherwise X5 1 if Staff -0.3786 0.252 -1.503 0.133 0 otherwise X6 1 if GRAD -0.223 0.305 -0.730 0.465 0 otherwise Summary Statistics Number of observations 1557 Number of cases = 1557 LL(0) = -1079.23 LL(c) = -371.26 LL(β) = -366.59 ρ² = 0.66 (-2)[LL(0)-LL(β)] = 1425.29 (-2)[LL(c)-LL(β)] = 9.34

Again to investigate the effects of TRIP on perceptions of safety while riding the buses, only wave 2 data were used in the following analysis. As seen in wave 1 results, none of the socioeconomic factors seemed to have a significant effect on how travelers feel about riding CABS. Therefore, they are not included in the specification shown in

Table 3.9. In this specification, only the noticing TRIP explanatory variable is used. Both of the estimated coefficients are significant in Table 3.9. The positive sign of the constant and noticed TRIP variable indicate that travelers in general feel safe riding the buses. The estimated coefficient for noticing TRIP variable (X2 = 1, if TRIP) is also positive. The

32 positive sign of the X2 variable illustrates the positive effect of providing information to travelers inside the buses on their responses to feeling safe while riding the buses.

Table 3.9 Results from model investigating effect of TRIP on travelers responses to feeling safe while riding the buses (data: wave 2)

Estimated Variable Description Std. Error t-stat P-value Coefficient (βs) X1 Constant 2.6271 0.173 15.222 < 2e-16 1 if noticed TRIP X2 0.8486 0.258 3.288 0.00101 0 otherwise Summary Statistics Number of observations 1467 Number of cases = 1467 LL(0) = -1016.85 LL(c) = -263.04 LL(β) = -257.60 ρ² = 0.75 (-2)[LL(0)-LL(β)] = 1518.50 (-2)[LL(c)-LL(β)] = 10.88

Chapter 4: Effect of TRIP on Travelers’ Perception of CABS’s Contribution to

Reduction in Environmental Impacts and Congestion

4.1 Overview

Surface transportation systems are a large contributor to pollution and greenhouse gas emissions. Using public transportation in the United States has been estimated to save approximately 1.4 billion gallons of gasoline and about 1.5 million tons of carbon dioxide annually (APTA, 2010). Therefore, public transportation services, when effectively utilized, play a vital role in creating environmental sustainability. Congestion and the resulting increase in travel times, especially during rush hour periods, are also major problems around the world. Increased use of public transportation systems, leads to decreased use of personal vehicles, which can positively improve the quality of life in urban areas.

The benefits of public transportation systems and their positive environmental effects impact not only users, but also the general public. Even those who may never use the services can benefit from these effects. Individuals who recognize public transportation’s benefits might be more willing to vote for governmental funding and taxes that support public transportation services.

In this portion of the study, the effects of providing real-time passenger information on individuals’ perception of CABS’s roles in promoting a “green” campus and in reducing car traffic around campus are investigated. The hypothesis is that the use of state-of-the-art technologies may increase the individuals’ perceptions of transit as a progressive mode of transportation and that these perceptions of progressive public transportation would enforce its role in promoting green environment and in reducing congestion. In the next sections, a series of models are estimated to investigate the effects of TRIP and different socioeconomic factors on respondents’ perception of CABS positive role in promoting sustainable environment and a “green” campus. In section 4.3 models are estimated to investigate the hypothesis that individuals who notice TRIP have a higher positive attitude towards CABS role in congestion reduction around campus than those who do not notice TRIP.

4.2 “Green” Campus

To investigate the effects of TRIP on individuals’ attitudes towards CABS role in promoting a “green” campus, responses to the statement, “Providing bus service around campus should be part of OSU’s efforts to promote a green campus” (field number 8 in

Table 2.2) statement was used. This statement was asked in both waves of the survey. If the new information system has an effect on individuals’ perception of CABS role in promoting green campus, the second wave would have a higher proportion of positive responses (“agree” or “strongly agree”) than the first wave. Since the attitudes of the general public on the perceptions of environmental impacts are of interest, the responses

35 from both CABS users and nonusers were used in the following estimations. To investigate this hypothesis, the simple specification shown in Table 4.1 was estimated. In this model, the data from both waves were pooled together. The dependent variable is set to 1 for individuals who chose “agree” or “strongly agree” in response to the statement

“Providing bus service around campus should be part of OSU’s efforts to promote a green campus.” The explanatory variable X2 indicates if the responder belongs to wave 2

(X2n = 1) or wave 1 (X2n = 0).

As indicated by the t-statics values in Table 4.1, both coefficients are significant.

The positive coefficient of the constant is appreciably larger than the β2 coefficient. This indicates that whether the individual responded to the first wave (X2n = 0) or the second wave (X2n = 1), he or she has a probability greater than 0.5 of agreeing to this statement.

That is, in general, the OSU community in both waves had a positive attitude towards

CABS environmental contributions. The coefficient for the X2 variable, which is 1 when individuals answered in the second survey, is negative. This result does not agree with the original hypothesis of TRIP having a positive effect towards promoting a green campus.

However, this could be due to different mixes of responders and other socioeconomic factors that may affect the responses to this statement.

Table 4.1 Results from model investigating the difference between responses to CABS promoting green campus before and after TRIP (data: wave 1 and wave2)

Estimated Variable Description Std. Error t-stat P-value Coefficient (βs) X1 Constant 2.21932 0.054 41.253 <2e-16 1 if wave 2 X2 -0.13883 0.073 -1.915 0.0555 0 if wave 1 Summary Statistics Number of observations 8195 Number of cases = 8195 LL(0) = -5680.34 LL(c) = -2749.80 LL(β) = -2747.96 ρ² = 0.52 (-2)[LL(0)-LL(β)] = 5864.76 (-2)[LL(c)-LL(β)] = 3.68

To investigate the socioeconomic factors that may influence individuals’ responses to CABS role in promoting a “green” campus, the specification presented in

Table 4.2 was estimated using only wave 1 data. In this specification, the dependent variable is again set to 1 for individuals agreeing to CABS’s role in promoting a green campus and 0 otherwise. Similar to before, the X2n variable is set to 1 for CABS users and 0 for nonusers. Gender variable (X3n = 1 if female, 0 if male) and familiarity with other metropolitan public transportation MPT (X4n = 1 if MPT, 0 otherwise) are the next two explanatory variables shown in Table 4.2. Again, an “undergraduate student” affiliation was used as the base, and the three other categories were explicitly included as independent variables (X5 = 1 if faculty, X6 = 1 if staff, and X7 = 1 if grad students).

Table 4.2 Results from model investigating the potential socioeconomic factors on perception of CABS promoting green campus (data: wave 1)

Estimated Variable Description Std. Error t-stat P-value Coefficient (βs) X1 Constant 1.3985 0.155 9.007 < 2e-16 1 if User X2 0.5806 0.130 4.474 7.69E-06 0 if Nonuser 1 if Female X3 0.1371 0.114 1.206 0.22783 0 if Male 1 if MPT X4 0.4664 0.162 2.888 0.00388 0 otherwise 1 if Faculty X5 0.9436 0.177 5.325 1.01E-07 0 otherwise 1 if Staff X6 0.3599 0.152 2.369 0.01785 0 otherwise 1 if GRAD X7 0.6652 0.210 3.173 0.00151 0 otherwise Summary Statistics Number of observations 3823 Number of cases = 3823 LL(0) = -2649.90 LL(c) = -1217.80 LL(β) = -1188.44 ρ² = 0.55 (-2)[LL(0)-LL(β)] = 2922.93 (-2)[LL(c)-LL(β)] = 58.72

All of the estimated coefficients are significant except the coefficient for gender

(X3). The positive sign of all the estimated coefficients indicates that no matter what socioeconomic category travelers belong to, they have a probability greater than 0.5 of agreeing to CABS’s role in promoting a green campus. The positive sign of the X2 variable indicates that responders who use CABS services have a higher realization of

CABS’s role in promoting a green campus compared to nonusers. The positive sign for

MPT explanatory variable (β4 = 0.4664) indicates that those who are using or have used 38 other MPTs in the past, gave a higher recognition of CABS’s contribution to promoting a green campus than those who do not. Explanatory variables for faculty, staff, and graduate students all have a positive sign. This indicates that undergraduate students feel slightly less positive about CABS’s environmental contribution than the other groups.

The undergraduate students may have a little less consciousness of CABS’s positive environmental contribution compared to other categories.

In order to capture the effects of TRIP on OSU community’s perception of CABS positive role in promoting a green campus, only the data from the wave 2 were used in the following analysis. As seen in the wave 1 results (shown in Table 4.2), different affiliations, CABS usage, and familiarity with other MPTs significantly affect individuals’ responses to this statement. A general specification with all the possible interactions between these significant socioeconomic factors and noticing TRIP was estimated. However, this specification led to use of many variables, thereby reducing the quantity of data available per variable. When estimating models under such specifications, some of the resultant estimates were insignificant, while others had the wrong signs. Therefore, the interaction between users with TRIP and the interaction between MPT with TRIP were estimated separately. The two specifications are presented in Table 4.3 and Table 4.4 respectively.

Table 4.3 shows the interaction between CABS users and noticing TRIP variables with regards to individuals’ attitude towards CABS’s role in promoting a green campus.

The dependent variable here is again set to 1 for individuals responding “strongly agree” or “agree” to the green campus statement. A variable representing nonusers who did not

39 notice TRIP is used as the base and the following three other variables are explicitly included as independent variables. These three interaction explanatory variables consist of CABS users who notice TRIP (X3n = 1, 0 otherwise), CABS users who did not noticed

TRIP (X4n = 1, 0 otherwise), and CABS nonusers who noticed TRIP (X5n = 1, 0 otherwise). MPT, faculty, staff, and graduate student variables are also included in this specification to control for their effects on the responses.

In Table 4.3, all the variables are significant. The positive sign of the constant and all of the other estimated coefficients indicates that no matter what socioeconomic category travelers belong to and whether they noticed TRIP or not, both users and nonusers have a probability greater than 0.5 of agreeing to CABS’s role in promoting a green campus. The estimated coefficient for the X2 variable, which represents MPT users, is positive. The positive sign of the MPT variable indicates again that those familiar with other MPTs recognized CABS positive environmental impacts more than others. Similar to the results of the wave 1 estimation shown in Table 4.2, the estimated coefficients for the faculty, staff, and grad students indicator variables using wave 2 data also turns out to be positive. Again these positive signs on the affiliation coefficients suggest that undergraduate students are slightly less conscious about the positive effect of

CABS towards promoting a sustainable environment.

Table 4.3 Results from model investigating effect of TRIP on peoples responses to CABS promoting green campus as an interaction with CABS usage(data: wave 2)

Estimated Variable Description Std. Error t-stat P-value Coefficient (βs) X1 Constant 0.9565 0.157 6.086 1.16E-09 1 if MPT X2 0.4084 0.161 2.530 0.011397 0 otherwise 1 if user and noticed X3 TRIP 1.4985 0.185 8.101 5.46E-16 0 otherwise 1 if user and didn’t X4 notice TRIP 0.6091 0.174 3.504 0.000458 0 otherwise 1 if Nonuser and X5 notice TRIP 0.9006 0.220 4.089 4.34E-05 0 otherwise 1 if Faculty X6 0.772 0.213 3.630 0.000283 0 otherwise 1 if Staff X7 0.6794 0.176 3.857 0.000115 0 otherwise 1 if GRAD X8 0.3695 0.180 2.051 0.040275 0 otherwise Summary Statistics

Number of observations 2583 Number of cases = 2583 LL(0) = -1790.40 LL(c) = -878.15 LL(β) = -830.27 ρ² = 0.54 (-2)[LL(0)-LL(β)] = 1920.25 (-2)[LL(c)-LL(β)] = 95.75

The X3 variable in Table 4.3 represents CABS users who noticed the new information system in the second wave. This variable has a positive estimated coefficient

(β3 = 1.4985). The X4 variable represents CABS users who did not notice TRIP and also has a positive coefficient (β4 = 0.6091). The X5 variable represents the individuals who

41 do not use CABS services but have noticed the new information system on campus. The estimated coefficient for X5 is positive (β5 = 0.9006). Comparing these three interaction terms, it is observed that the X3 variable has the highest positive coefficient compared to the other two interaction terms (X4 and X5). This means that users who noticed TRIP recognize CABS positive environmental contribution more than the other groups. From comparing the X3 to the X4 coefficient and the X5 coefficient to zero, it can be concluded that not only CABS users but also nonusers who had noticed TRIP have a higher positive attitude towards CABS role in promoting green campus compared to those who did not notice the new information system. In other words, the positive effect of providing information by CABS on OSU community’s realization of CABS role in promoting green campus is confirmed. Comparing the X4 coefficient with the “base” variable of zero, the initial hypothesis is again confirmed that users, in general, have a higher positive attitude towards CABS environmental contribution compared to nonusers. All of these results are found in the presence of MPT and affiliation categories variables in order to control for their socioeconomic factors.

Similar to the specification shown in Table 4.3, the interactions between having familiarity with other MPTs and noticing TRIP with regards to individuals perception of

CABS role in promoting green campus was investigated. Only the second wave data were used for this specification as well. The results of this estimation are shown in Table 4.4.

Table 4.4 Results from model investigating effect of TRIP on peoples responses to CABS promoting green campus as an interaction with MPT familiarity(data: wave 2)

Estimated Variable Description Std. Error t-stat P-value Coefficient (βs) X1 Constant 1.232 0.125 9.891 < 2e-16 1 if User X2 0.466 0.143 3.255 0.00113 0 otherwise 1 if MPT and noticed X3 TRIP 1.256 0.304 4.133 3.58E-05 0 otherwise 1 if MPT and didn’t X4 notice TRIP 0.337 0.145 2.335 0.01954 0 otherwise 1 if Not MPT and X5 notice TRIP 0.688 0.221 3.112 0.00186 0 otherwise 1 if Faculty X6 0.843 0.157 5.383 7.31E-08 0 otherwise 1 if Staff X7 0.662 0.144 4.609 4.04E-06 0 otherwise 1 if GRAD X8 0.439 0.149 2.941 0.00327 0 otherwise Summary Statistics Number of observations 4212 Number of cases = 4212 LL(0) = -2919.54 LL(c) = -1463.04 LL(β) = -1417.63 ρ² = 0.51 (-2)[LL(0)-LL(β)] = 3003.82 (-2)[LL(c)-LL(β)] = 90.83

The hypothesis here is that those who are somewhat familiar with other MPTs and have noticed the information system on campus have a higher positive attitude towards

CABS environmental contributions. The dependent variable is set to 1 for individuals who agreed to CABS role in promoting a green campus and 0 otherwise. The three

43 interaction explanatory variables consist of MPT users who noticed TRIP (X3n = 1, 0 otherwise), MPT users who did not notice TRIP (X4n = 1, 0 otherwise), and MPT nonusers who noticed TRIP (X5n = 1, 0 otherwise). The base interaction variable is MPT nonusers who did not notice TRIP. Similar to before, CABS users, faculty, staff, and graduate student variables are included in this specification to control for their effects on the responses.

In Table 4.4, all the variables are significant. The positive sign of all the estimated coefficients indicates that no matter what socioeconomic category travelers belong to and whether they noticed TRIP or not, they have a probability greater than 0.5 of agreeing to

CABS’s role in promoting a green campus. The X2 estimated coefficient is positive indicating again that users realized the environmental benefits of CABS more than nonusers. The X3 variable in Table X.4 represents MPT users who noticed TRIP in the second wave. This variable has a positive estimated coefficient (β3=1.256). The X4 variable, which represents MPT users who did not notice TRIP, also has a positive estimated coefficient (β4 =0.3373). The X5 variable represents the individuals who have no prior or concurrent experience with other MPTs but have noticed the new information system on campus. The estimated coefficient for the X5 variable is positive (β5 = 0.688) as well.

The estimated coefficient for the X3 variable has the highest positive value compared to all the other interaction terms (X4 and X5). This indicates that users’ familiar with other MPTs who also noticed TRIP, recognized CABS environmental contribution more than the other groups. From comparing the X3 and the X4 coefficients and the X5

44 coefficient to zero, it can be concluded that not only MPT users but also those who have no prior or concurrent experience using MPT, who had noticed TRIP, have a higher positive attitude towards CABS’s positive role in promoting a green campus.

Comparing the X4 coefficient with the “base” variable of zero, the initial hypothesis is again confirmed that those who have some prior or concurrent experience with other MPT, in general, have a higher positive attitude towards CABS environmental contribution compared to those who do not. All of these results are found in the presence of CABS users and affiliation categories variables in order to control for their socioeconomic factor as well. All of these estimated coefficients are positive which agrees with the results discussed from Table 4.2.

4.3 Traffic Reduction on Campus

In this section, the responses to “Having CABS service reduces the amount of car traffic on campus” (field number 7 in Table 2.2) statement is used in order to investigate the effects of TRIP on individuals’ perception of CABS role in traffic reduction. This statement was asked in both waves of the survey. If TRIP had a positive effect on how people felt towards CABS’s contribution to congestion reduction around campus, the second wave would be expected to produce higher probability of positive responses

(“agree” or “strongly agree”) than the first wave. To investigate this hypothesis, the simple specification was estimated and the results are presented in Table 4.5. In this model, the data from both waves were pooled together and both CABS users and nonusers’ responses were considered. The dependent variable is set to 1 for individuals

45 who chose “agree” or “strongly agree” in response to the traffic reduction statement. The explanatory variable (X2 = 1) indicates if the responder belongs to wave 2 or wave 1

(X2 = 0).

Table 4.5 Results from model investigating the difference between responses to CABS role in traffic reduction before and after TRIP (data: wave 1 and wave2)

Estimated Variable Description Std. Error t-stat P-value Coefficient (βs) X1 Constant 0.957 0.038 25.497 <2e-16 1 if wave 2 X2 0.092 0.052 1.761 0.0783 0 if wave 1 Summary Statistics Number of observations 7515 Number of cases = 7515 LL(0) = -5209.00 LL(c) = -4368.15 LL(β) = -4366.60 ρ² = 0.16 (-2)[LL(0)-LL(β)] = 1684.81 (-2)[LL(c)-LL(β)] = 3.10

As indicated by the t-statistics shown in Tale 4.5, both estimated coefficients are significant. The positive signs on β1 and β2 imply that whether the individual responded in the first wave or second wave, he or she has a probability greater than 0.5 of agreeing to role of CABS in traffic reduction on campus. That is, in both waves, the OSU community recognized CABS’s role in congestion reduction. The coefficient for the X2 variable, which indicates individuals who answered in the second survey, is also positive.

The positive coefficient for X2 variable indicates that responders appreciated CABS role in traffic reduction more in the second wave compared to first.

Similar to the case of CABS’s environmental impacts, possible effects of different socioeconomic categories are investigated using only the wave 1 data. The results of this 46 specification are shown in Table 4.6. Again, affiliation, gender, and familiarity with other

MPTs are used as possible explanatory variables here. The X2 variable is set to 1 for

CABS users and 0 for nonusers. The other socioeconomic variables are defined as before

(gender (X3n = 1 if female, 0 otherwise), MPTs (X4n = 1 if familiar with MPTs, 0 otherwise), faculty (X5n = 1 if faculty, 0 otherwise), staff (X6n = 1 if staff, 0 otherwise), and graduate students (X7n = 1 if graduate student, 0 otherwise)).

Except for the X3 coefficient which represents females and the X7 variable which indicates graduate students, all the other estimated coefficients are statically significant.

The estimated coefficient for the constant has a positive sign. The magnitude of this coefficient is such that the combination of any other set of explanatory variables will lead to a positive β1X1 +β2X2+… +βk Xk. Therefore no matter what socioeconomic category a traveler belong to, he or she has a probability greater than 0.5 of feeling safe walking to

CABS stops. This result agrees with the original hypothesis that not only users but also nonusers agree to CABS positive role in traffic reduction around campus area. This result is encouraging since public transportation needs the support of both service users and nonusers to get governmental funding. The estimated coefficient for the X2 variable which represents users, is positive (β2 = 0.59503). The positive sign of the X2 variable indicates that users have a higher recognition towards CABS role in traffic reduction compared to nonusers. The positive sign for the X4 variable which represents those who have prior or concurrent familiarity with other MPT systems gave a higher recognition to

CABS’s role in congestion reduction than those who do not. Explanatory variables for faculty and staff categories (the X5 and X6 variables) both have a positive sign. This

47 suggests that students are slightly less conscious about CABS traffic reduction impact compared to the other groups.

Table 4.6 Results from model investigating the potential socioeconomic factors on perception of CABS traffic reduction role (data: wave 1)

Estimated Variable Description Std. Error t-stat P-value Coefficient (βs) X1 Constant 0.383 0.113 3.379 0.000728 1 if User X2 0.595 0.087 6.852 7.27E-12 0 if Nonuser 1 if Female X3 -0.104 0.080 -1.310 0.190098 0 if Male 1 if MPT X4 0.224 0.101 2.210 0.027128 0 otherwise 1 if Faculty X5 0.940 0.126 7.459 8.73E-14 0 otherwise 1 if Staff X6 0.252 0.108 2.332 0.019719 0 otherwise 1 if GRAD X7 0.117 0.132 0.886 0.37586 0 otherwise Summary Statistics Number of observations 3471 Number of cases = 3471 LL(0) = -2405.91 LL(c) = -2046.83 LL(β) = -1988.66 ρ² = 0.17 (-2)[LL(0)-LL(β)] = 834.51 (-2)[LL(c)-LL(β)] = 116.33

Next the data from wave 2 were used in order to investigate the possible effects of providing real-time information system on OSU community’s perception of CABS’s role in traffic reduction. Again, a general specification with all the possible interactions between the significant socioeconomic factors and noticing TRIP was estimated. 48

However, this specification again led to the use of many variables, thereby reducing the quantity of data available per variable. In the resulting models, some of the estimates again turned out insignificant, while others had unreasonable signs. Therefore, the interactions between each significant socioeconomic factor (users and MPT) with TRIP were estimated separately. The two specifications are presented in Table 4.7 and Table

4.8 respectively.

Table 4.7 shows the interaction between CABS users and noticing TRIP with regards to individuals’ attitude towards CABS role in reducing car traffic on campus. The dependent variable here is again set to 1 for individuals responding “strongly agree” or

“agree” to CABS role in reducing congestion on campus. A variable representing nonusers who did not notice TRIP is used as the base and the following three other variables are explicitly included as independent variables. These three interaction explanatory variables consist of CABS users who noticed TRIP (X3n = 1, 0 otherwise),

CABS users who did not notice TRIP (X4n = 1, 0 otherwise), and CABS nonusers who noticed TRIP (X5n = 1, 0 otherwise). MPT, faculty, and staff variables are included in this specification to control for their effects on the responses.

Table 4.7 Results from model investigating effect of TRIP on peoples responses to CABS role in traffic reduction as an interaction with CABS usage (data: wave 2)

Estimated Variable Description Std. Error t-stat P-value Coefficient (βs) X1 Constant 0.276 0.101 2.731 0.00631 1 if MPT X2 -0.018 0.106 -0.173 0.86267 0 otherwise 1 if user and X3 noticed TRIP 1.150 0.118 9.713 < 2e-16 0 otherwise 1 if user and didn’t X4 notice TRIP 0.704 0.128 5.484 4.17E-08 0 otherwise 1 if Nonuser and X5 notice TRIP 0.797 0.152 5.239 1.61E-07 0 otherwise 1 if Faculty X6 0.557 0.142 3.920 8.84E-05 0 otherwise 1 if Staff X7 0.132 0.109 1.213 0.22509 0 otherwise Summary Statistics Number of observations 2549 Number of cases = 2549 LL(0) = -1766.83 LL(c) = -1477.77 LL(β) = -1420.43 ρ² = 0.20 (-2)[LL(0)-LL(β)] = 692.80 (-2)[LL(c)-LL(β)] = 114.68

In Table 4.7, all the variables except MPT and staff are significant. The positive sign of all the estimated coefficients except MPT indicator coefficient indicates that no matter what socioeconomic category travelers belong to and whether they noticed TRIP or not, both users and nonusers have a probability greater than 0.5 of agreeing to CABS’s role in congestion reduction. The X3 variable in Table 4.7 represents CABS users who noticed TRIP in the second wave. This variable has a positive estimated coefficient

(β3 = 1.15044). The X4 variable represents CABS users who did not notice TRIP and it also has a positive coefficient (β3 =0.70383). The X5 variable represents the individuals who do not use CABS services but have noticed the new information system on campus and it has a positive estimated coefficient (β5 = 0.79709). The positive sign of the X5 estimated coefficient indicates that individuals who do not use CABS services but have noticed TRIP, agree to the role CABS plays in reducing the amount of car traffic around campus. Comparing these three interaction terms, it is observed that the X3 variable has the highest positive coefficient compared to the other two interaction terms (X4 and X5).

This means that travelers who noticed TRIP recognized CABS role in traffic reduction more than the other groups. Comparing the X3 to the X4 coefficients and X5 coefficient to zero, it can be concluded that not only users but also nonusers who noticed TRIP have a higher positive appreciation towards CABS role in traffic reduction compared to those who did not notice TRIP. Comparing X4 estimated coefficient with zero, the initial hypothesis is again confirmed that CABS users in general recognize CABS’s role in traffic reduction more compared to nonusers. In this model only the faculty category is significant and it has a positive estimated coefficient. This result suggests that faculty category is more conscious about CABS’s role in traffic reduction around campus than the other categories.

Similar to the specification shown in Table 4.7, the interactions between having familiarity with other MPTs and noticing TRIP with regards to individuals’ perceptions of CABS’s role in congestion reduction was investigated. The results of this estimation using only wave 2 data are shown in Table 4.8. The hypothesis here is that those who are

51 familiar with other MPTs and noticed TRIP may have a higher recognition of CABS’s role in traffic congestion reduction. The dependent variable is set to 1 for those who agreed to CABS’s role in car traffic reduction on campus and 0 otherwise. The three interaction explanatory variables consist of MPT users who noticed TRIP (X3n = 1, 0 otherwise), MPT users who did not notice TRIP (X4n = 1, 0 otherwise), and MPT nonusers who noticed TRIP (X5n = 1, 0 otherwise). The base interaction variable is MPT nonusers who did not notice TRIP. Similar to before, CAB users, faculty, staff, and graduate student variables are included in this specification to control for their effects on the responses.

In Table 4.8, all the variables except X4 are significant. The positive signs of all the estimated coefficients indicate that no matter what socioeconomic category travelers belong to and whether they noticed TRIP or not, they have a probability greater than 0.5 of agreeing to CABS’s role in congestion reduction. The X2 estimated coefficient is positive indicating again that CABS users realized the congestion reduction role of CABS more than nonusers. The X3 variable in Table X.4 represents MPT users who noticed

TRIP in the second wave. This variable has a positive estimated coefficient

(β3 = 0.26963). The X4 variable, which represents MPT users who did not notice TRIP, has a positive estimated coefficient but it does not turn out to be statistically significant.

The X5 variable represents the individuals who have never used any other MPTs but have noticed TRIP on campus. The estimated coefficient for the X5 variable is positive

(β5 = 0.40404) as well.

Table 4.8 Results from model investigating effect of TRIP on peoples responses to CABS traffic reduction role as an interaction with MPT familiarity (data: wave 2)

Estimated Variable Description Std. Error t-stat P-value Coefficient (βs) X1 Constant 0.498 0.076 6.569 5.07E-11 1 if User X2 0.526 0.105 5.002 5.67E-07 0 otherwise 1 if MPT and X3 noticed TRIP 0.270 0.164 1.647 0.09964 0 otherwise 1 if MPT and X4 didn’t notice TRIP 0.136 0.107 1.271 0.203787 0 otherwise 1 if Not MPT and X5 notice TRIP 0.404 0.151 2.676 0.00746 0 otherwise 1 if Faculty X6 0.694 0.105 6.618 3.63E-11 0 otherwise 1 if Staff X7 0.302 0.091 3.339 0.000841 0 otherwise Summary Statistics Number of observations 3900 Number of cases = 3900 LL(0) = -2703.27 LL(c) = -2232.83 LL(β) = -2180.29 ρ² = 0.19 (-2)[LL(0)-LL(β)] = 1045.97 (-2)[LL(c)-LL(β)] = 105.07

The X3 and X5 variables indicate individuals who noticed TRIP who either had prior or concurrent familiarity with other MPTs or not respectfully. Both β3 and β5 estimated coefficients are positive. This suggests that whether people were familiar with

MPTs or not, if they noticed TRIP, they recognized CABS role in car traffic reduction more. The specification presented in Table 4.7 showed that once the effect of TRIP is included in the model, having familiarity with other MPTs variable becomes

53 insignificant. In this model, the insignificant estimated coefficient of the X4 variable again confirms this conclusion. In other words, in the presence of noticing TRIP variable in the model, being an MPT user or not does not have a significant effect on how individuals feel about CABS’s role in congestion reduction around campus. Being a

CABS user, faculty, and staff variables are included in this specification to control for their effects on the responses.

After estimating the specification shown in Table 4.8, the earlier specification presented in Table 4.7 (interactions between CABS user and noting TRIP) was modified.

The revised estimation does not include familiarity with MPT explanatory variable since its effect turned out to be insignificant in presence of TRIP variable. The result of this new improved interaction specification is presented in Table 4.9. Staff variable was also taken out of this final specification since its estimated coefficient turned out to be insignificant from Table 4.8.

The estimated coefficients shown in Table 4.9 are all significant and positive.

Again, the positive signs of all the estimated coefficients indicate that no matter what socioeconomic category travelers belong to and whether they noticed TRIP or not, both users and nonusers have a probability greater than 0.5 of agreeing to CABS’s role in congestion reduction. The description of the interaction terms are the same as the ones explained earlier. The relative magnitudes of the estimated coefficients for the three interaction terms are similar to those found in Table 4.7. Again, comparing the three interaction terms, it is observed that the X2 variable has the highest positive coefficient compared to the other two interaction terms (X3 and X4). This means that travelers who

54 noticed TRIP recognized CABS traffic reduction more than the other groups. Again

CABS users who had noticed the real-time information system recognized the important role of CABS in traffic reduction followed by nonusers who noticed TRIP and CABS users who noticed TRIP respectively.

Table 4.9 Results from model investigating effect of TRIP on peoples responses to CABS role in traffic reduction as an interaction with CABS usage (data: wave 2)

Estimated Variable Description Std. Error t-stat P-value Coefficient (βs) X1 Constant 0.344 0.078 4.408 1.05E-05 1 if user and X2 noticed TRIP 1.103 0.114 9.702 < 2e-16 0 otherwise 1 if user and didn’t X3 notice TRIP 0.666 0.125 5.339 9.35E-08 0 otherwise 1 if Nonuser and X4 notice TRIP 0.792 0.152 5.218 1.81E-07 0 otherwise 1 if Faculty X5 0.510 0.133 3.850 0.000118 0 otherwise Summary Statistics Number of observations 2549 Number of cases = 2549 LL(0) = -1766.83 LL(c) = -1477.77 LL(β) = -1420.43 ρ² = 0.20 (-2)[LL(0)-LL(β)] = 692.80 (-2)[LL(c)-LL(β)] = 114.68

Chapter 5: Effect of TRIP on Travelers’ Perception of Valuableness of CABS’s

Services

5.1 Overview

Public transportation provides personal mobility and flexibility to individuals.

Access to public transportation gives people transportation options to get to work, go to school, and many other trip purposes. As mentioned in the introduction chapter, CABS is a free transit service provided by The Ohio State University. CABS has 7 different routes around OSU’s main campus, operates with fleet size of around 40 buses, and serves about

4 million passengers every year. CABS provides frequent and reliable services between

West Campus parking lots and the central campus academic core. The buses also connect multiple residential areas and OSU’s Medical Center to the central campus core. More than 70% of undergraduate students and 96% of graduate students live off-campus. Also many faculty and staff park their personal vehicles on the off-campus parking lots. CABS services are essential to many of these individuals in order to get to their classrooms or work places. Freshmen undergraduate students are not allowed to have a personal vehicle on campus.

By providing real-time information, CABS has tried to improve the level of service for travelers and enhance travelers’ experiences with public transportation.

Providing information on different bus routes, arrival times, and announcing available 56 transfers at the next stop and their location along the route are some of the features CABS has been providing after the provision of TRIP in order to improve their services. In this chapter, potential effects of providing real-time passenger information system on travelers’ perception of valuableness of CABS services are investigated.

5.2 Valuableness of CABS Services

To investigate the effects of TRIP on individuals’ perceptions of and attitudes towards the value of CABS services, responses to the “CABS offers service that is valuable to my travel needs” (field number 9 in Table 2.2) statement was used. This statement was asked in both waves of the survey. If the new real-time information system had an effect on individuals’ opinion on the value of CABS services to their travel needs, the proportion of positive responses to this statement would be higher in the second wave, after the provision of TRIP, compared to first wave. To investigate this hypothesis, a simple specification was estimated and shown in Table 5.1. In this model, the data from both waves were pooled together. Only the responses of CABS users to this question are relevant and used in this portion of the study. The dependent variable is set to 1 for individuals who chose “agree” or “strongly agree” in response to the statement of CABS services being valuable to their travel needs. The explanatory variable (X2 = 1) indicates if the responder belongs to wave 2 or wave 1 (X2 = 0).

As indicated by the t-statics values in Table 5.1, both of the coefficients are significant. The positive signs of β1 and β2 imply that whether the travelers responded in the first wave or second wave, they have a probability greater than 0.5 of believing

CABS services to be valuable to their travel needs. That is, in both waves, the users valued CABS services. The coefficient for the X2 variable, which indicates individuals who answered in the second survey, is also positive. This positive coefficient indicates that travelers valued CABS services more in the second wave compared to the first wave.

Table 5.1 Results from model investigating the difference between responses to CABS services being valuable to travel needs before and after TRIP (data: wave 1 and wave2)

Estimated Variable Description Std. Error t-stat P-value Coefficient (βs) X1 Constant 0.844 0.055 15.310 < 2e-16 1 if wave 2 X2 0.417 0.081 5.169 2.35E-07 0 if wave 1 Summary Statistics Number of observations 3237 Number of cases = 3237 LL(0) = -2243.72 LL(c) = -1853.29 LL(β) = -1839.84 ρ² = 0.18 (-2)[LL(0)-LL(β)] = 807.76 (-2)[LL(c)-LL(β)] = 26.91

A question of “Accessing information about CABS service (e.g., routes, frequency of service, hours of operation) is easy” (field number 13 in Table 2.2) was also asked in both waves. This statement is believed to have an effect on how travelers responded to the value of CABS services statement. A simple model to investigate the responses to the ease of access to CABS information statement was estimated to study the effect of providing TRIP on this variable. If providing TRIP had an effect on the ease of access to CABS information, the proportion of positive responses to the second wave would be higher than the first wave. In the simple specification shown in Table 5.2, the dependent variable is set to 1 for individuals who chose “agree” or “strongly agree” in 58 response to accessing information about CABS services is easy statement. The explanatory variable (X2 = 1) indicates if the responder belongs to wave 2 or wave 1

(X2 = 0).

Table 5.2 Results from model investigating the difference between responses ease of access to CABS information before and after TRIP (data: wave 1 and wave2)

Estimated Variable Description Std. Error t-stat P-value Coefficient (βs) X1 Constant 0.662 0.055 12.120 < 2e-16 1 if wave 2 X2 0.281 0.078 3.610 3.06E-04 0 if wave 1 Summary Statistics Number of observations 3099 Number of cases = 3099 LL(0) = -2148.06 LL(c) = -1916.44 LL(β) = -1909.90 ρ² = 0.11 (-2)[LL(0)-LL(β)] = 476.32 (-2)[LL(c)-LL(β)] = 13.06

The t-statistics for both estimated coefficients in Table 5.2 are significant. The positive signs of β1 and β2 imply that whether the individuals responded in the first wave or the second wave, they believed that accessing CABS information is easy. The positive sign of the coefficient of the X2 variable indicates that travelers had a stronger positive attitude towards the ease of access to CABS information in the second wave (after the provision of TRIP) compared to first wave.

Based on the results of Table 5.2, the ease of access to CABS information variable was decided to be included as an explanatory variable in addition to the other usual socioeconomic variables (i.e. gender, MPT, and affiliation). In order to investigate the effects of ease of access to CABS information and the other socioeconomic 59 independent variables on travelers’ responses to value of CABS services, the specification shown in Table 5.3 was estimated. The specification presented in Table 5.3 was estimated using only the wave 1 data. The dependent variable is set to 1 for individuals who chose “agree” or “strongly agree” in response to CABS services being valuable to their travel needs statement. The X2 variable is set to 1 for responders who believed accessing CABS information is easy and 0 otherwise. Similar to models discussed in the previous chapters, a gender variable (X3n = 1 if female, 0 if male) and a prior or concurrent familiarity with other MPTs variable (X4n = 1 if MPT, 0 otherwise) are the next two explanatory variables used in this specification. Again, an

“undergraduate student” affiliation was used as the base, and the three other categories were explicitly included as independent variables (X5 = 1 if faculty, X6 = 1 if staff, and

X7 = 1 if grad students).

In Table 5.3, all of the estimated coefficients are significant except the coefficient for gender indicator (X3). The estimated coefficient for the constant, ease of access, females, and MPT are positive. This indicates no matter what socioeconomic group undergraduate travelers belong to, they have a probability greater than 0.5 of believing

CABS offers services which is valuable to their travel needs. The estimated coefficient for the X2 variable is positive (β2 = 1.0335). This sign agrees with the initial hypothesis that travelers who believe accessing information about CABS are easy, value the services more. The positive sign for the X4 variable indicates that individuals who use other MPTs value CABS services more.

Table 5.3 Results from model investigating the potential socioeconomic factors on CABS valuableness to travel needs (data: wave 1)

Estimated Variable Description Std. Error t-stat P-value Coefficient (βs) X1 Constant 0.563 0.157 3.591 0.000329 1 if easy to X2 access info 1.034 0.122 8.439 < 2e-16 0 if otherwise 1 if Female X3 0.088 0.121 0.723 0.46979 0 if Male 1 if MPT X4 0.345 0.133 2.596 9.42E-03 0 otherwise 1 if Faculty X5 -0.598 0.180 -3.313 9.24E-04 0 otherwise 1 if Staff X6 -1.031 0.153 -6.745 1.53E-11 0 otherwise 1 if GRAD X7 -0.402 0.185 -2.172 0.029836 0 otherwise Summary Statistics Number of observations 1461 Number of cases = 1461 LL(0) = -1012.69 LL(c) = -889.69 LL(β) = -820.49 ρ² = 0.19 (-2)[LL(0)-LL(β)] = 384.40 (-2)[LL(c)-LL(β)] = 138.41

The explanatory variables for faculty, staff, and graduate students all have negative estimated coefficients. This result suggests that undergraduate students recognize the value of CABS services to their travel needs more than the other categories.

This result seems reasonable, since undergraduate students are the main users of CABS, and many of them do not have any other transportation options to travel to or on campus.

Usually undergraduate students have classes on different locations around campus and use CABS buses to make short trips necessary to get to their destinations on time.

Meanwhile, graduate students as well as faculty and staff usually stay in one part of the campus for the majority of time and may also have other transportation options to travel within campus. Therefore, it makes sense that undergraduate students feel more strongly towards CABS services being valuable to their travel needs compared to other categories.

In order to investigate the possible effects of providing real-time information system on travelers’ perception of the value of CABS services, only wave 2 data were used in the following analysis. A general specification with all the possible interactions between the significant socioeconomic factors from Table 5.3 and noticing TRIP was estimated. However, similar to the safety models presented in Chapter 3, this specification led to use of many variables thereby reducing the quantity of data available per variable. Therefore, the specification shown in Table 5.4 was estimated using only the

CABS users’ responses from wave 2 data.

In this specification, ease of access to information, MPT, and the three affiliation categories were kept as indicator variables to control for their effects on how travelers value CABS services. The dependent variable here is again set to 1 for travelers responding “strongly agree” or “agree” to CABS services being valuable to their travel needs. The explanatory variables consist of agreeing to ease of access to CABS information (X2n = 1 if agreed, 0 otherwise), familiarity with other MPTs (X4n = 1 if

MPT, 0 otherwise), faculty (X4n = 1 if faculty, 0 otherwise), staff (X5n = 1 if staff, 0 otherwise), and graduate students (X6n = 1 if graduate students, 0 otherwise), and noticing

TRIP (X7n =1 if noticed TRIP, 0 otherwise).

Table 5.4 Results from model investigating effect of TRIP on the value of CABS to travelers (data: wave 1)

Estimated Variable Description Std. Error t-stat P-value Coefficient (βs) X1 Constant 0.922 0.175 5.271 1.36E-07 1 if easy to X2 access info 0.720 0.143 5.057 4.26E-07 0 if otherwise 1 if MPT X3 0.341 0.146 2.336 0.0195 0 otherwise 1 if Faculty X4 -0.515 0.216 -2.386 0.0171 0 otherwise 1 if Staff X5 -0.879 0.174 -5.047 4.48E-07 0 otherwise 1 if GRAD X6 -0.348 0.184 -1.890 0.0587 0 otherwise 1 if noticed X7 0.311 0.139 2.241 0.025 0 otherwise Summary Statistics Number of observations 1407 Number of cases = 1407 LL(0) = -975.26 LL(c) = -719.25 LL(β) = -682.48 ρ² = 0.30 (-2)[LL(0)-LL(β)] = (-2)[LL(c)-LL(β)] = 585.55 73.53

The t-statistics values in Table 5.4 indicate that all of the explanatory variables are significant. The estimated coefficient for the constant, the ease of access, MPT, and noticed TRIP variable are all positive. This indicates that no matter what socioeconomic group undergraduate students belong to, they have a probability greater than 0.5 of believing CABS offers services which is valuable to their travel needs. The X2 variable indicates travelers who believe accessing information about CABS services is easy, and its coefficient is positive. This positive sign indicates that if users agreed to ease of access 63 to CABS information, they valued CABS services more. As discussed earlier, TRIP had a positive effect on whether users believed that it was easy to access the information or not.

TRIP provides more ways to access information about CABS services (via variable message boards, phone apps, a website, and etc.), and users seem to value this convenience. The X3 variable, which indicates those who are familiar with other MPTs, also has a positive estimated coefficient. Again individuals, who are currently using other public transportations or have used them in the past, have a point of reference to compare

CABS services with. The positive sign of this variable indicates that MPT users value

CABS services more than those who have no prior or concurrent experience with MPTs.

The X4, X5, and X6 all have negative estimated coefficients. This again indicates the fact that undergraduate students value CABS services more. The last estimated coefficient presented in Table 5.4 is associated with an indicator for noticing TRIP and has a positive sign. This positive sign indicates that travelers who noticed the use of new real-time information valued CABS services more compared to those who did not notice TRIP.

This confirms the original hypothesis that users associate the new real-time information system as a valuable feature to their travel needs. CABS has increased its level of service by providing real-time passenger information to its users and serves their transportation needs more efficiently.

Chapter 6: Conclusion

In this chapter, the general conclusions of the empirical results and possible future research are discussed. Section 6.1 summarizes the results from investigating the effects of providing real-time information on responders’ perceptions of and attitudes towards safety issues, environmental and congestion issues, and value of CABS services. In section 6.2, potential future research is discussed.

6.1 Summary and Conclusions

Transit agencies have been implementing automatic sensing and information technologies in order to improve system performance and level of service for their passengers. To investigate the potential effects of providing real-time passenger information on travelers’ attitude towards different dimensions of public transportation, data from a two-wave survey of The Ohio State University (OSU) community towards

OSU Campus Area Bus Services (CABS) were analyzed. In autumn of 2009, CABS began to provide real-time passenger information via TRIP. Using the information provided by TRIP, travelers can plan their trips, choose bus routes, and plan arrival times at the desired bus stops. The provision of real-time information could decrease the anxiety of travelers when waiting for buses and potentially decrease users’ safety

65 concerns. The importance of nonusers’ opinions towards public transportation is also important. Transportation agencies need the support of users and the general public to vote for levies to go towards funding public transportation. Therefore, it is interesting to investigate the effects of providing real-time information on nonusers as well. It may be possible that even nonusers would perceive the provision of high technology systems as an effort to improve transit services and decrease congestion and environmental impacts.

This improved perception could translate into increased public support for funding levies.

The data from a two-wave online survey (conducted by the OSU Transit Lab team) are analyzed. The online survey questionnaire was administered to a random sample of email addresses of four OSU community categories (undergrad students, grad students, faculty, and staff). The first wave of the online survey was administered in autumn of 2008 (asking about individuals experiences during spring quarter of 2008) and the second wave was administered in spring of 2010. The two-wave survey data included responders’ demographic characteristics, mode of transportation to and on campus, opinions on CABS role in reducing environmental and congestion impacts, and evaluation of CABS services. In the second wave, data on whether respondents noticed the use of modern technologies (i.e., TRIP) were also collected.

The responses to both waves are transformed into a binary choice set. Responses of “agree” or “strongly agree” are aggregated to a categorical variable “agree”, and responses of “neutral”, “disagree”, and “strongly disagree” are aggregated to a categorical variable of “not agree”. The dependent (choice indicator) variable is set to 1 for the categorical “agree” responses and to 0 for the categorical “not agree” responses. A

66 logit model is used to model the association between the responses and multiple explanatory variables.

The same general approach is used to investigate the effects of providing real- time information on the different dimensions of interest. In all of the models, the data from both waves are first pooled together, and a simple specification with only a wave 2 indicator variable is estimated. The sign and significance of the estimated coefficients of the wave 2 indicator variable are investigated to assess whether the positive responses to the statement increased in the second wave, after the implementation of TRIP.

The associations of different socioeconomic groups (CABS users, gender, familiarity with other metropolitan public transportation (MPT), and four affiliation categories of undergraduate students, graduate students, staff and faculty) with the responses are investigated using only wave 1 data. Only the wave 1 data are used in order to exclude the effects of providing TRIP on individuals’ responses at this stage of the analysis.

Finally, the specific effect of providing TRIP on individuals responses are investigated using the wave 2 data considering the socioeconomic variables found significant based on the wave 1 analysis. In the case of safety issues and value of CABS services, only the responses of CABS users are relevant and used in the specifications.

However, in the case of reduction in environmental impacts and congestion, the responses of the entire OSU community (users and nonusers of the service) are considered.

Regarding the effect of providing real-time information (TRIP) on travelers’ perceptions of safety, the following conclusions are drawn. In general, travelers feel very safe walking to stops, waiting for buses, and riding the buses. Nevertheless, users’ perceptions of safety (walking to stops, waiting for buses, and riding buses) increased after the availability of TRIP. It is hypothesized that knowing when the next bus arrives decreases anxiety levels and waiting time, and hence, respondents feel safer in the presence of real-time information. The positive attitudes towards safety walking to and waiting at bus stops increased significantly more for females than males after the implementation of TRIP compared to this increase for males.

With regards to the effects of TRIP on OSU community’s perceptions towards the value of CABS in reducing congestion and environmental impacts, the following results are found. A large majority of respondents in general appreciated CABS’s positive environmental contributions and congestion reduction. This effect is noticeable for both

CABS users and nonusers and could translate into additional support from transit users and the general public for funding levies to public transportation. The positive attitude of respondents towards CABS’s contribution to promoting a “green” campus and to reducing congestion increased after the implementation of TRIP. Respondents who noticed TRIP recognized CABS’s role in promoting a green campus and reducing traffic significantly more than those who did not notice the new real-time information system.

Individuals, who had prior or concurrent familiarity with other metropolitan public transportation systems (MPT), showed a higher appreciation towards CABS’s

68 contribution to promoting a green campus and reducing car traffic around campus compared to those who had never used any other MPTs.

Finally, regarding the effects of providing TRIP on value of CABS services and ease of access to CABS information, the following results are arrived at. A large majority of users believe CABS services provide a great value in meeting their travel needs, both before and after the implementation of TRIP. However, respondents appreciate the value of CABS services more in the second wave, in the presence of TRIP. Male and female respondents equally value CABS services. Undergraduate students (the main users of

CABS) recognize the value of CABS services to their travel needs more than the other three affiliation category groups (i.e. faculty, staff, graduate students). Undergraduate students are more dependent on CABS services since many of them do not have a personal vehicle available on campus and usually have to travel within the campus throughout the day to get to their destinations. Respondents who are using or have used other metropolitan public transportation systems (such as COTA) recognize CABS’s value to their travel needs more compared to others. Individuals with some prior or concurrent experience with other MPTs have a point of reference to compare CABS services with possibly leading to the observed effect. Travelers believe that accessing information about CABS services (e.g., routes, frequency of service, hours of operation) is easier in the second wave, after the implementation of TRIP, than in the first wave before the implementation. Respondents who believed accessing CABS information is easy also value CABS services more than others. Travelers who noticed the new real- time passenger information system (i.e., TRIP) agreed to the value of CABS services in

69 meeting travel needs more than those who did not notice it. Again, this result demonstrates the positive effect of providing real-time information to passengers on their perceptions of and attitudes towards the value of CABS.

6.2 Future Research

As already mentioned, many other questions regarding individuals’ choices and perceptions with regard to CABS services were asked in both waves of the survey. Some of these questions, such as parking location and whether passengers’ wait time are reasonable (Questions 15-18, 32, and 45 Table B.1 in Appendix B) could be used as either other potential explanatory variables or additional dependent variables in various models as exercise. For example, individuals who park at off-campus parking lots may value CABS services more and appreciate the provision of TRIP more than others. The real-time information provided by TRIP can potentially be more useful to individuals who park at off-campus parking lots give their dependence on CABS services to reach their campus destinations compared to those who take intra-campus convenience-based short trips on CABS. Similarly, since the off-campus parking locations tend to have poor lighting and empty at night times, travelers may feel less anxious in the presence of real- time information.

Moreover, responders who believe their waiting time is reasonable may, for example, feel safer or value CABS more than those who don’t. Therefore, such a variable could be used as an explanatory variable in models relation to safety and value of CABS.

In addition, this variable could also be investigated as a dependent variable. For example, individuals who noticed TRIP may believe their average waiting time for buses is reasonable compared to those who did not notice TRIP. The real-time information provided by TRIP may make the time spent waiting seem less unpleasant. In addition, the time-of-the day the respondent is usually on campus could be used as a proxy for socioeconomic characteristics and play a role as an explanatory variable for some models, especially for those associated with the safety dimensions. Including this factor can potentially explain why different affiliation categories of OSU had different safety concerns while walking to stops, waiting for buses, and riding the buses.

The first wave of the survey was conducted before implementation of TRIP and the second wave took place a year after this implementation. TRIP has now been in place for over three years and, therefore, it has reached a steady-state in terms of community awareness. Conducting a third wave survey would allow further investigations to be made on how TRIP has changed individuals’ perceptions of and attitudes towards CABS services and environmental and congestion impacts.

The estimations presented in this thesis are all carried out using a binary logit model. However, since the collected data are in a form of Likert scale responses, ordered logit (Ben-Akiva and Lerman, 1985), also known as proportional odds model, may be used instead. It would be interesting to compare the results found in this thesis with estimation results determined using an ordered logit model and assess whether there is a benefit capturing the higher weights associated with the “strongly agree” or “strongly disagree” responses.

References

American Public Transportation Association. Public Transpiration: Moving America Forwards. Washington, D.C. 2010.

Ben-Akiva, Moshe E., and Steven R. Lerman. Discrete Choice Analysis: Theory and Application to Travel Demand. MIT Press, Cambridge, MA, 1985.

Campus Transit Lab, The Ohio State University. URL: http://transitlab.osu.edu/ campus-transit-lab. Accessed in February, 2013.

McCord, M.R., Mishalani, R.G., and Goel, P. Research and Education from a Smart Campus Transit Laboratory. Region V University Transportation Center, Final Report, Project No 006OY01, 2009.

Mishalani, R.G., McCord, M.R., and Goel, P. Smart Campus Transit Laboratory for Research and Education. Region V University Transportation Center, Final Report, Project No. 032OY02, 2011.

Appendix A: Wave 2 Questionnaire

This survey asks you questions about yourself, your transportation to and on campus, and your preferences and perceptions of transportation options during this quarter (Spring 2010). The intent is to develop general results. However, some questions relate to OSU transportation as a case study. What is your affiliation at OSU during this quarter (Spring 2010)? Student Staff Faculty (If “Student”) Are you currently an undergraduate or graduate student? Undergraduate Graduate (If “Student”) Do you currently live in an OSU campus dorm? Yes No (If “Staff”) What is your current job designation?  Administrative and Professional  Civil Service  Other (If “Administrative and Professional”, “Student”, or “Faculty”) Which academic group at OSU do you currently belong to?  Agriculture  Arts and Sciences  Business  Education and Human Ecology  Engineering and Architecture  Law  Medical and Health Sciences  Social Work  Centers not attached to any of the above groups  Other How many years have you been at OSU?  1 year or less  More than 1 year but not more than 2 years  More than 2 years but not more than 5 years  More than 5 years but not more than 15 years  More than 15 years Are you currently enrolled/employed on the main OSU campus? Yes No What is your gender? Male Female 73

During this quarter, how many days per week are you usually on campus?

During this quarter, which times of the day are you usually on campus? Please select all that apply.  Before 8:00 am  Between 8:00 am and 5:00 pm  Between 5:00 pm and 7:00 pm  After 7:00 pm On a typical day during this quarter, what is the total time in hours you spend engaged in browsing the web, writing or reading e-mail, or sending and receiving instant messages? Please select only one.  1 hour or less  More than 1 hour and less than 3 hours  More than 3 hours and less than 5 hours  5 hours or more During this quarter, did you have a car available to you in the Columbus area? Yes No Other than OSU’s Campus Area Bus Service (CABS), do you currently use metropolitan public transportation or have you regularly used it in the past? Yes No During this quarter, what is the most common mode of transportation you used to travel to campus? Please select only one.  Drove a car alone  Shared a car as a driver or passenger  COTA bus  CABS bus  Motorbike as a driver or passenger  Bike  Walk  Other Please select all the modes of transportation you use to travel to campus during this quarter.  Drove a car alone  Shared a car as a driver or passenger  COTA bus  CABS bus  Motorbike as a driver or passenger  Bike  Walk  Other 74

(If “Drove a car alone” or “Shared a car as a driver or passenger” in above question) When you currently travel to campus by car, where do you usually get out of the car (either as a driver or passenger)? Please select only one.  West campus parking lots  North campus parking lots  Main or medical campus parking lots or garages  Main or medical campus on-street parking

How familiar are you with OSU’s CABS?  I do not know of its existence  I know it exists, but I am not familiar with any CABS route  I am familiar with one CABS route only  I am familiar with more than one CABS route

For each of the following statements, please indicate the degree to which you agree or disagree regarding CABS based on your current impressions.

ee N

u /

r b

ongly

s e

rong

lica

1. Having CABS reduces the amount of car traffic on campus. 1 2 3 4 5 NA

2. Providing bus service around campus should be part of 1 2 3 4 5 NA OSU’s 3.efforts CABS to offerspromote service a green that campus. is valuable to my travel needs. 1 2 3 4 5 NA

4. I feel safe walking to CABS stops. 1 2 3 4 5 NA 5. I feel safe waiting for CABS buses. 1 2 3 4 5 NA 6. I feel safe riding CABS buses. 1 2 3 4 5 NA 7. CABS bus drivers are professional. 1 2 3 4 5 NA 8. CABS buses are comfortable. 1 2 3 4 5 NA 9. CABS routes are reasonable. 1 2 3 4 5 NA

10. My travel time to reach my destination using CABS is 1 2 3 4 5 NA reasonable. 11. My waiting time for CABS buses is reasonable. 1 2 3 4 5 NA

12. Accessing information about CABS service (e.g., routes, 1 2 3 4 5 NA frequency of service, hours of operation) is easy. 13. CABS is reliable. 1 2 3 4 5 NA

14. Overall, I am satisfied with CABS. 1 2 3 4 5 NA

When traveling on campus during this quarter, how often do you use CABS instead of

using a car or motorbike (as a driver or passenger)?  Using a car or motorbike on campus is not an option for me (for example, I have no access to a car or motorbike or I am unable to drive)

 I never use CABS when I could use a car/motorbike  I sometimes use CABS when I could use a car/motorbike  I often use CABS when I could use a car/motorbike  I always use CABS even when I could use car/motorbike When traveling on campus during this quarter, how often do you use CABS instead of a mode other than a car or motorbike (e.g. walking, biking, skating, etc.)?  I never use CABS when I could use a mode other than car/motorbike  I sometimes use CABS when I could use a mode other than car/motorbike  I often use CABS when I could use a mode other than car/motorbike  I always use CABS even when I could use a mode other than car/motorbike During this quarter, how often do you use CABS (whether to travel to campus or on campus)?  I never use CABS  I use CABS occasionally  I use CABS regularly (If “occasionally” or “regularly”) During this quarter, how frequently do you ride CABS (whether to travel to campus or on campus)? Consider each boarding of a bus as a single ride.  I do not ride every day I am on campus  I ride 1 to 2 times each day I am on campus  I ride 3 or more times each day when I am on campus

(If “occasionally” or “regularly” two questions above) During this quarter, how many minutes do you usually spend waiting for a CABS bus to arrive (even if you are occupied with other activities during your wait time)? Please select only one.  0 to less than 2 minutes  2 to less than 5 minutes  5 to less than 10 minutes  10 to less than 15 minutes  15 minutes or more  My wait time is too variable to determine usual waiting time (If “occasionally” or “regularly” three questions above) During this quarter, what activities are you typically engaged in while waiting for a CABS bus to arrive? Please select all that apply.  Read  Study  Talk with friends standing or sitting with me  Talk on the phone  Use my PDA, smart-phone, or computer  Just wait

(If response to “How many years have you been at OSU?” is “1 year or less” AND If response to “During this quarter, how often do you use CABS (whether to travel to campus or on campus)?” is “I use CABS occasionally” or “I use CABS regularly”) Over the period starting from this past Autumn 2009 quarter to the present, please indicate the degree to which

r you agree or disagree with the following statements: r

g lica

ongly

s e

r rong

i S

t t

S S

N 1. I have noticed changes in the information CABS provides 1 2 3 4 5 NA regarding bus arrivals at stops. 2. I have noticed changes in the areas on campus served by CABS. 1 2 3 4 5 NA

(If response to “How many years have you been at OSU?” is OTHER THAN “1 year or less” AND If response to “During this quarter, how often do you use CABS (whether to travel to campus or on campus)?” is “I use CABS occasionally” or “I use CABS regularly”) Over the period starting from Autumn 2008 quarter (approximately a year and a half ago) to the present, please indicate the degree to which you agree or disagree with the following statements: 1.

N 2.

g lica

ongly

s e

3. r rong

i S

t t

S S

4. N 5. I have noticed changes in the information CABS provides 1 2 3 4 5 NA regarding bus arrivals at stops. 6. I have noticed changes in the areas on campus served by CABS. 1 2 3 4 5 NA

(If response to “How many years have you been at OSU?” is

“1 year or less” AND If response to “During this quarter, how often do you use CABS (whether to travel to campus or on campus)?” is “I never use CABS”) Over the period starting from this past Autumn 2009 quarter to the present, please indicate the degree to which you agree or disagree with the following statements:

r you agree or disagree with the following statements: r

g lica

ongly

s e

r rong

i S

t t

S S

N 3. I have noticed changes in CABS’ use of modern 1 2 3 4 5 NA technologies. 4. I have noticed a change in the number of CABS buses on Campus 1 2 3 4 5 NA

((If response to “How many years have you been at OSU?” is OTHER THAN “1 year or less” AND If response to “During this quarter, how often do you use CABS (whether to travel to campus or on campus)?” is “I never use CABS”) Over the period starting from Autumn 2008 quarter (approximately a year and half ago) to the present, please indicate the degree to which you agree or disagree with the following statements:

g lica

ongly

s e

r rong

i S

t t

S S

N 7. I have noticed changes in CABS’ use of modern 1 2 3 4 5 NA technologies. 8. I have noticed a change in the number of CABS buses on 1 2 3 4 5 NA Campus

Appendix B: Survey Data Variable Names and Descriptions

Question number Field Description

1 Record id 2 Completion date 3 Affiliation at OSU 4 Undergraduate or graduate student 5 Live on campus dorms 6 Job designation 7 Academic group 8 Academic group 9 Academic group 10 Academic group 11 Years had been at OSU 12 Currently employed 13 Gender 14 Number of days on campus 15 Usually on campus before 8 am 16 Usually on campus between 8 am and 5pm 17 Usually on campus between 5 pm and 7 pm 18 Usually on campus after 8 pm 19 Total time spend browsing internet 20 Have a car available 21 Using other Metropolitan public transportation 22 Most common mode of transportation to campus 23 Most common mode of transportation to campus 24 Traveling to campus-Drove a car to campus alone 25 Traveling to campus-Shared a car as a driver or passenger 26 Traveling to campus-COTA bus 27 Traveling to campus-CABS bus 28 Traveling to campus-Motorbike as a driver or passenger

Table B.1 Survey Data Variable Names and Description, (Continued on the next page) 80

Table B.1 Survey Data Variable Names and Description (Continued)

Question number Field Description

30 Traveling to campus-Walk 31 Traveled to campus-other modes 32 Where you get out of car on campus 33 Where you get out of car on campus 34 Familiarity with CABS 35 CABS role in traffic reduction on campus 36 CABS role in promoting green campus 37 CABS services are valuable to travel needs 38 Feeling safe walking to CABS stops 39 Feeling safe waiting for CABS buses 40 Feeling safe riding CABS buses 41 Professional bus drivers 42 Comfortable CABS buses 43 Reasonable bus routes 44 Reasonable travel time 45 Reasonable waiting time 46 Ease of access to CABS information 47 CABS is reliable 48 Overall Satisfaction with CABS 49 Using CABS instead of a car or motorbike 50 Using CABS instead of a mode other than a car or motorbike 51 CABS usage 52 Frequency of CABS usage 53 Waiting time for bus arrivals 54 Reading while waiting for the CABS buses 55 Studying while waiting for the CABS buses

(Continued on the next page)

Table B.1 Survey Data Variable Names and Description (Continued)

Question number Field Description

56 Talking with friend while waiting for the CABS buses 57 Talking on the phone while waiting for the CABS buses 58 Using computer while waiting for the CABS buses 59 Just waiting while waiting for the CABS buses 60 Other activities while waiting for the CABS buses CABS users who have been at OSU for 1 year or less, noticed 61 changes in information CABS provides CABS users who have been at OSU for 1 year or less, noticed 62 changes in areas on campus served by CABS CABS users who have been at OSU for more than a year, 63 noticed changes in information CABS provides CABS users who have been at OSU for more than a 1 year, 64 noticed changes in areas on campus served by CABS CABS nonusers who have been at OSU for 1 year or less, 65 noticed CABS use of modern technology CABS nonusers who have been at OSU for 1 year or less, 66 noticed changes in number of CABS buses on campus CABS nonusers who have been at OSU for more than a year, 67 noticed CABS use of modern technology CABS users who have been at OSU for more than a 1 year, 68 noticed changes in number of CABS buses on campus