Intercity Passenger Rail in the Context of Dynamic Travel Markets (2016)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

1 1.1 The Purpose of This Research 1.1.1 Project Objectives “The overall objective of this research is to develop an analytical framework to improve understanding of how current or potential intercity travelers make the choice to travel by air, rail, bus, or private automobile for the majority of their trip. This framework should provide guidance for use by a diverse audience of practitioners and decision makers considering alternative planning, operating, financing, service, and capital investment strategies for intercity passenger rail service in existing and potential travel markets, and it should allow users to evaluate how mode choice is affected by a variety of changing and evolving parameters. A second objective is to produce this analytical framework on two levels: one that is designed to help the practitioner understand the interaction of key factors in the process of attitude formation toward intercity travel by mode and a second level in which these relationships are operationalized with methods designed to be ultimately incorporated into the transportation demand forecasting process.” [Unpublished request for proposals for National Cooperative Rail Research Program (NCRRP) Project 03-02] 1.1.2 Context and Work Plan Americans may be changing some fundamental patterns of their travel. For decades, trans- portation analysts have seen demand for automobile travel as almost inevitably rising with the continued growth of national economic activity. Today, analysts agree that overall rates of auto- mobile use are down since the new millennium, although there is debate about why rates are down. Ridership on rail and intercity bus, however, is consistently up. As travel demand analysts began examining why rates of automobile use are down, this NCRRP project sought to understand how a wide range of causal factors seemed to be coming into play at once, affecting both our present patterns of travel, and logically affecting the future patterns. Analysts of travel demand have focused on approaches based on economic theory— assuming that choices among modes are driven by modal service variables such as travel times and travel costs. Advocates of social psychology and market research look at values, attitudes, and preferences that influence our travel behavior. With original market-based data collection, this NCRRP project undertook the largest effort ever attempted to integrate traditional economic approaches with those used by market researchers and psychologists in order to answer these key questions about travel behavior. The work plan included a major, original survey of residents of the Northeast Corridor (NEC) and the Cascade Corridor. Early in the project, the research team proposed an analytical framework for the examination of values, preferences, and attitudes; this framework was key to the creation of a survey instrument, which broke out longer-term basic values from shorter-term attitudes toward the modes and trips. The results of the survey first provided the basis for the creation of attitudinal models C H A P T E R 1 Introduction and Major Conclusions

2 Intercity Passenger Rail in the Context of Dynamic Travel Markets for rail and bus markets, using structural equation modeling, which did not attempt to reflect detailed variation in trip characteristics (e.g., service levels and costs). The same survey results were then used in the creation of a single unified integrated choice/latent variable model that integrated both basic values and highly detailed information about specific trip choices. A sce- nario testing tool was then created to help the practitioner to interpret the very large amount of information from the unified model. Figure 1 diagrams the elements of the project work plan. 1.2 Need to Understand the Nature of Intercity Demand Travel behavior is a consumer behavior like other behaviors: it is affected by cold hard facts, such as those describing the travel times and costs of rail service versus competitive modes. But it may also be affected by human beliefs and attitudes that affect individuals on both rational and emotional levels. For example, people might be influenced by normative pressures, such as the belief that “others like me” choose to take rail, or by a personal desire for a pleasurable, stress-free experience. Several factors influence long-distance travelers to select rail, given the competition that exists from air, bus, and the private car. For decades, research into rail demand has focused on com- parative travel times and costs for rail against each of its competitors in a dynamic market. The purpose of this NCRRP project is to better understand a wider set of factors that seem to be also influencing the propensity to select rail. The additional factors include demographic factors and cultural factors involving values held over the long term and shorter-term attitudes that may vary by the day-to-day context in which they are formed. In sum, an individual’s propensity to choose rail may be influenced by demographic, geographic, and psychographic factors. • Demographic differences are easiest to spot. In this project, older males make long-distance modal decisions in a manner radically different from younger females, as shown in Figure 2. Model for rail The discrete choice model for four modes The scenario testing tool, from ICLV output Demographic details for each trip maker: Educaon, age, gender, income, employment status Trip specific costs, in vehicle mes, wait mes, and access mes Model for air Model for bus Model for auto Development of Attudinal Models for Rail and Bus Markets Updated Analyses of Competing Modes Conclusions and Recommen daons for Further Research Values Privacy In Travel Values Auto Orientaon Values Urbanism/ Sociability Values Informaon Communicaons Technology Figure 1. Conceptual diagram of the work plan for NCRRP 03-02.

Introduction and Major Conclusions 3 The two groups face the same set of times and costs, but younger females choose the bus at a rate 3.5 times that of older males. This prompts the question: Do these groups have a different set of values, attitudes, or preferences? • Geographic differences merit examination. Intercity trips with higher activity density at the point of destination attract far more riders to rail than do destinations with lower density. In many cases in other domains of travel demand, people from higher-density residential areas make more trips by public mode than people in lower-density residential areas. • Psychographic differences may be the hardest to examine, as there is little agreement in the literature about which of these are relevant in affecting the choice of rail. There is a body of research literature that relates travel to deeply held psychological beliefs, but the application of such concepts to public policy has been problematic, at best. 1.2.1 The Influence of Theory from Social Psychology This NCRRP project attempts to integrate all three of these categories in examination of factors beyond times and costs that are influencing demand for intercity rail today, and presumably will be influencing demand in the future. By far, the most difficult category to incorporate into the research is the third, the psychographic factors. In this work, they will be examined in terms of long-term preferences, referred to as “values,” and shorter-term preferences, referred to as “attitudes.” Longer-term values tend to influence long- and medium-term decisions, such as the location of one’s residence or the number of cars owned. Shorter-term attitudes might concern one’s feelings (1) that a mode (e.g., rail) is safe, convenient, or stressful or (2) that going from 0 to 60 miles per hour in 3 seconds in a new car makes one feel happy. Considerable research in this area has been based on the value–attitude–behavior hierarchical model, as illustrated in Figure 3 (Note that the values explored in this NCRRP project were not designed to reflect such psychological concepts as hedonism, security, and power, but rather more direct alternative influences on transportation behavior 25 years in the future). There are several methods for applying and implementing the basic theory that suggests behavior is influenced by salient short-term attitudes, which in turn have been influenced by longer-term values and preferences. The dominant theory applied in the analysis of transportation Female Male Female Male Millennial Group Older Group Bus 17% 12% 9% 5% Rail 11% 12% 11% 11% 0% 5% 10% 15% 20% 25% 30% Pe rc en t M od e Sh ar e Figure 2. The influence of age and gender on choice of rail and bus.

4 Intercity Passenger Rail in the Context of Dynamic Travel Markets issues is the Theory of Planned Behavior (TPB), developed by Dr. Icek Ajzen based on earlier work he undertook with Dr. Martin Fishbein, the Theory of Reasoned Action. The survey instru- ment developed by the research team was undertaken with the advice of Dr. Ajzen and was utilized in many phases of the work plan, including the creation of a new discrete choice model, called the “Integrated Choice/Latent Variable (ICLV) Model,” presented in Chapter 5. According to Dr. Ajzen, this project represents the largest application yet of the TPB to transportation, but not the largest in other domains. 1.2.2 A Framework for Long- and Short-Term Preferences in the Choice of Rail Prior to beginning the analysis of highly diverse and multifaceted attitudinal data, it is important to first establish a unifying theory of how the pieces fit together. In this NCRRP project, the research team applied the basic logic of the value→attitude→behavior concept to years of applied research undertaken in the analysis of transportation. Figure 4 shows a simple approach to assigning preferences to a logical sequence of experience. Although conceptually part of an approach, this analytical framework for preferences was specifically created to serve the needs of this research project, in which several alternative scenarios for the future will be created. By way of example, in one scenario one could hypothesize that in 25 years a given cohort group will (1) hold positive values about urbanism, (2) be located in a dense urban area, (3) have near- term attitudes based on their experience at that location, and (4) have their mode choice affected by some combination of the three aforementioned explanatory factors. By the midpoint of this Source: Paulssen et al. (2014, p. 875, Figure 1: An illustrative representation of the value–attitude–behavior hierarchical model). © Springer Science+Business Media New York 2013, used with permission of Springer. Figure 3. The NCRRP analytical framework builds directly upon the value–attitude–behavior model now prominent in the literature. Figure 4. The NCRRP analytical framework for attitudes and preferences. Figure redacted for on‐line use; available in printed version. 

Introduction and Major Conclusions 5 project, the research team had designed the survey instrument to gather information in all four subject areas—values, location, attitudes, and travel—shown in Figure 4. 1.2.3 What Are the Long-Term Values in the Analytical Framework? The NCRRP analytical framework incorporates four basic, longer-term values (i.e., under- lying factors) that might influence the choice of transportation mode: 1. Values Privacy in Travel. This factor represents a propensity to not want to travel with other people and to be uneasy and uncomfortable with the idea of being with people that one does not know. 2. Values Auto Orientation. This factor represents a propensity to value the freedom and independence gained from owning cars and to disagree with the concept that borrowing or sharing a car is just as good as owning one. It also reflects a feeling that one is (or is not) less dependent on the car as their parents were at that age. 3. Values Urbanism and Sociability. This factor represents the idea that one likes being in an urban setting, likes being able to meet and greet people, and feels that, when people work together, improvements to the world around them can result. 4. Values Information Communications Technology (ICT). This factor represents the propensity to value being productive with one’s time and desire to stay connected and to own/use mode- connected information technology. 1.2.4 Why Create a Unified Model? The analytical framework for understanding attitudes does not include much about specific times and costs (a diagram of the NCRRP 03-02 work plan that does integrate all elements was presented as Figure 1). Meanwhile the discrete choice modeling does not include much about attitudes. Therefore, the challenge turned to the integration of the two methods. A highly valu- able (and critical) summation of the logic behind the concept of integrating the strong points of discrete choice mode modeling with that of attitudinal-based research is provided by Vij and Walker (2015): “Traditional models of disaggregate decision making have long ignored the question of why we want what we want. Human needs have been treated as given, and attention has largely centered on the expression of these needs in terms of behavior in the marketplace. As a consequence, traditional models of dis aggregate decision making have focused on observable variables, such as product attributes, socio- economic characteristics, market information and past experience, as determinants of choice, at the expense of the biological, psychological, and sociological reasons underlying the formation of individual preferences (McFadden 1986). This idealized representation of consumers as optimizing black boxes with predetermined wants and needs is at odds with findings from studies in the social sciences that have attempted explicitly to map the cognitive path that leads consumers from observable inputs to their observed choices in the marketplace. These studies have consistently shown that latent constructs such as attitudes, norms, perceptions, affects, and beliefs can often override the influence of observable variables on disaggregate behavior. . . .” The NCRRP 03-02 work plan called for the collection of original data that would support (1) the analysis of the value of times and costs for intercity modes as expressed in stated preference models, (2) the creation of attitudinal models that seek to reveal roles of both longer-term values and shorter-term preferences by the consumer toward intercity modes, and (3) the creation of an ICLV hybrid model, generally as advocated by Vij and Walker. 1.2.5 Can the Future Market Context for Rail Be Predicted? The future market context for rail cannot be predicted. However, a small number of scenarios can be created to represent several possible futures for American society. These scenarios can

6 Intercity Passenger Rail in the Context of Dynamic Travel Markets explore how several alternative futures might affect the propensity to choose rail. The concept of uncertainty must be addressed in long-term planning; these new tools help decision makers understand numerous potential scenarios, thereby mitigating the inevitable uncertainty of the future. Without doubt, even after the careful examination of possible alternative scenario futures, uncertainty remains a key concept. The NCRRP work plan called for the creation of four separate scenarios for the future market context for future rail travel. These four scenarios assumed the same level of rail service, in terms of both times and costs relative to competing modes. 1.3 Major Conclusions 1.3.1 What the Scenarios Revealed The scenarios revealed how certain demographic, geographic, and psychographic factors in the future might impact rail ridership more powerfully than other factors and why some patterns might be important to start tracking more carefully than before. Specifically, work undertaken in several parallel research efforts of this project suggests that the millennial generation seems to have a serious concern with what it perceives as a lack of personal safety making trips by both bus and rail (a theme which will be developed later in this report). The research team’s analysis suggests that—over time—this could seriously imperil the growth of rail and bus markets. On the other side of the spectrum, the research team’s findings suggest that the millennial generation holds certain values and attitudes concerning the lack of need for traditional forms of auto ownership that differ distinctively from older cohort groups. The research team’s scenario testing exercise suggests that if, over time, this cohort abandons this orientation as it proceeds through the life cycle (e.g., as this cohort enters the child-rearing and child-educating age category), then any evident positive orientation of this group to rail and bus would be in jeopardy. These concepts are developed in some quantitative detail in Chapter 6, with the implications explored in Chapter 10. 1.3.2 Differences Between Millennials and Older Groups in the Four Values There were statistically significant differences between the mean scores of millennials and the older groups for privacy in travel, auto orientation, and ICT. The differences between groups for urbanism values were insignificant. While the comparative mean values for auto orientation and desire for ICT were consistent with conventional wisdom, the fact that millennials had more interest in privacy in travel than their elders was perhaps surprising. Basic relationships between demographic category and survey responses are presented in Chapter 3. 1.3.3 How Important Were Each of the Factors in the Explanation of Present Rail Choice? The Attitudinal Model for Rail was applied to the question of relative importance of factors contributing to rail choice. As shown in Table 1, the strongest explanatory factor in rail choice is the idea that the rail service is inconvenient, which is influenced by finding the schedule fre- quency unacceptable. This factor is similar to the role of travel times in more traditional demand modeling; thus, its high ranking is not unexpected. The concept that a car trip is more stressful than a rail trip is also a key explanatory factor. The methods and assumptions underlying these conclusions are presented in Chapter 4, with further documentation in Technical Appendix: Documentation for the Structural Equation Models in NCRRP Web-Only Document 2.

Introduction and Major Conclusions 7 At the other end of the rank ordering, residential density turned out not to be a significant factor and values toward urbanism is on the border of being statistically insignificant, and certainly inconsequential, in the explanation of long-distance mode choice. 1.3.4 How Did the Four Future Scenarios Address These Issues? As described in more detail in Chapter 6, four long-term scenarios for the future market con- text were developed by the research team. Each scenario assumed the same level of service from the transportation system, and the same set of demographics for education and employment. Highlights of the four long-term scenarios included the following: • Pessimistic interpretation of trends extended. Each of the relevant policy input assumptions was set at a level not supportive of rail ridership. Millennials were assumed to drop their “anti- auto” attitudes as they transitioned into the prime child-rearing years (35 to 55 years old). Millennials were assumed to keep their “pro-privacy” values as they moved into middle age. No increase in education level or employment level was assumed. As groups get older, they were assumed to lose interest in their ICT devices. • Mixed Scenario A with continued preference for privacy in travel. This scenario assumed that millennials would adopt the same attitudes toward the auto as those currently in their child-rearing years. However, this scenario assumed they would outgrow their fear of and discomfort with travel with others, even as they gain more experience in life. As they grow older, the need for ICT was assumed to continue. • Mixed Scenario B with continued anti-auto pattern. As the mirror image of Mixed Scenario A, this scenario assumed that millennials would keep their anti-auto patterns throughout the child-rearing years. At the same time, it was assumed that, even with more experience, this cohort retains its fears about traveling with people they do not know. With increasing age, they were assumed to lose interest in their ICT devices. • Optimistic interpretation of future trends. To test the upper levels of societal support for rail, this scenario assumed millennials would keep their anti-auto patterns. They would lose their pro-privacy concerns about traveling with others, to mimic the attitude of present senior citizens who have little fear of traveling with others. ICT devices would be important to everyone. Rank Order Standardized Total Effect* Factor** 1 0.73 Train Trip Inconvenient 2 0.34 Values Auto Orientaon 3 0.33 Train Trip Less Stressful 4 0.29 Values Privacy in Travel 5 0.22 Train Trip Unsafe 6 0.19 Values ICT 7 0.15 Train Trip Expensive 8 0.09 Educaon 9 0.03 Employed 10 0.03 Values Urbanism/Sociability 11 Not Significant Density * Standardized total effect is defined and discussed in Chapter 4. **The four basic values are shown in italic bold; four short term atudes are shown in roman; and demographics are shown in italic. Table 1. Ranking of importance of factors in the explanation of rail mode choice.

8 Intercity Passenger Rail in the Context of Dynamic Travel Markets 1.3.5 Do the Scenarios Suggest a Wide Range of Possible Future Market Contexts? Accepting the logic that the scenarios were indeed designed to emphasize variation, the scale of the variation based on alternative future cultural settings is significant. Holding constant the quality of rail services (and that of its competitors), the optimistic scenario predicts rail ridership to be 22% higher than ridership in the pessimistic scenario. These analyses are made possible by a combination of the hybrid travel demand model described in Chapter 5 and the scenario testing tool described in Chapter 6. Table 2 shows a summary of the scenario testing. The predictions for each of the four future scenarios were undertaken for four trip purposes— business, vacation, visiting friends/relatives, and other. The “other” trip purpose, which is the smallest segment, is the most volatile, while the other three trip purposes all show fairly similar trends across the scenarios (see Table 3). The NCRRP analysis process illustrates how differences between demographic groups can explain the differences in propensity to choose intercity rail, as shown in Table 4. To provide some inter- pretation: If all the members of the population sample were to adopt the values and attitudes of millennials concerning their need for privacy in travel, rail ridership would go down by 3.4%. If all the members of the population were to adopt the values and attitudes of those over 65 years of age concerning their need for privacy in travel, rail ridership would go up by 10.4%. Looking at the basic value of auto orientation, if all the members of our population were to adopt the millennials’ view of a decreased need for a private car, rail ridership would increase by 17.9%. If the values of the total population mirrored those of the population over 65 years of age concerning Decreasing Role of Auto Orientaon in Future Decreasing Concern forPrivacy in Travelin Future Pessimisc for Rail Bad future for auto rejecon Bad future for privacy tolerance ICT need will decrease with age Rail decreases by 4% Mixed Scenario B Good future for auto rejecon Bad future for privacy tolerance ICT need will decrease with age Rail increases by 4% Mixed Scenario A Bad future for auto rejecon Good future for privacy tolerance ICT need will connue with age Rail increases by 10% Opmisc for Rail Good future for auto rejecon Good future for tolerance ICT need will connue with age Rail increases by 18% Table 2. Summary of change in rail ridership by scenario. Trip Purpose Pessimisc Mixed ScenarioA Mixed Scenario B Opmisc Business 5% 13% 3% 22% Vacaon 4% 8% 2% 14% Visit friends/relaves 2% 9% 4% 15% Other 11% 14% 9% 35% Total 4% 10% 4% 18% Table 3. Change in rail trips under the four scenarios by trip purpose.

Introduction and Major Conclusions 9 the need for a private car, rail ridership would go down by 11.9%. Table 4 also illustrates the inter- relationship between the demographic category and the four latent factors by allowing the (unlikely) assumption that a given demographic shift would apply to all four latent factors at once, which may or may not be a realistic scenario future (see column “All Values at Once”). 1.3.6 Key Factors Identified in this Report A major theme of this report is the need to undertake further research around both the latent factor concerning decreased car orientation (here labeled Auto Need) and the latent factor con- cerning need for privacy in travel. These two factors have the greatest impact on the change in rail ridership across the research methods employed in this NCRRP project. The results of the project’s scenario testing task suggest there are elements of the choice of longer-distance mode that are based on market reaction to improved services at competitive costs. However, beyond this important component of choice, the decision to choose a mode is associated with highly personal preferences about privacy and personal safety, and devotion to or freedom from the personal automobile. To a lesser extent, these decisions may be associated with the desire for travel time to be productive time, connected to the electronic world. In the scenario analyses, attitude toward the urbanity of dense cities seems to predict less than originally hypothesized (this observation, drawn from the Chapter 6 scenario analysis, is consistent with the Chapter 4 conclusion that density is not a significant factor in the explanation of rail mode share). 1.3.7 Why Are These Alternative Future Scenarios Important? This NCRRP project is demonstrating the logic of creating a set of alternative background futures, in a disciplined manner. The futures must reflect what is known about the key variables affecting background market conditions. Then, these alternative futures can be used to examine various policies and facilities characterized by the relative times and relative costs. A particular vision of new rail facilities, better headways, more coverage, and improved travel times might be tested under four possible futures, rather than one. Recently, a federal court in North Carolina struck down an environmental impact statement simply because it was based on one assumed future, rather than several possible futures (McDonald 2015). Most importantly, incorporation of values, attitudes, and preferences in the travel demand analysis process may help policy planners understand which strategies and practices might be most effective in changing travel patterns. Demographic Shis in Atude Latent Factors for Four Basic Values Privacy Auto Need Urbanism Technology All Values at Once Female to male 0.4% 2.3% 0.3% 0.4% 1.2% Male to female 0.4% 1.8% 0.2% 0.3% 1.0% All age groups to under 35 3.4% 17.9% 0.0% 2.5% 16.4% Under 35 to 35–44 0.0% 1.7% 0.0% 0.0% 1.7% All age groups to over 65 10.4% 11.9% 0.0% 3.4% 5.7% No college to college 2.7% 1.2% 0.1% 0.1% 4.2% College to no college 7.5% 3.6% 0.2% 0.3% 11.4% Unemployed to employed 1.3% 0.6% 0.0% 0.2% 0.9% Employed to unemployed 2.5% 1.2% 0.0% 0.4% 1.7% Table 4. Variation in rail use by basic value, expressed as shift in demographics.

10 Intercity Passenger Rail in the Context of Dynamic Travel Markets 1.3.8 Additional Factors to Explore The Influence of One’s Peers on the Selection of Intercity Mode The modeling of the attitudes through the format of the TPB suggests two more possible areas for future research. First, the choice of rail may be—perhaps more than anticipated—influenced by normative pressures. Even the novice student of human behavior would expect that issues like “what clothes I wear” or “what dancing I like” to be influenced by the views of peers and friends; however, the research team’s findings suggest that the choice of intercity mode may be influenced by a person’s sense that “others like me” are also taking this mode. Figure 5 shows that millennials are more likely than older groups to report that “people who are important to my life” or “my friends and coworkers” take the train. And yet, millennials have the same overall propensity to take the train as the older groups, as shown in Figure 2. In the present literature in social psychology, this concept of descriptive norm is being explored as a strong explanation for propensity to change behavior. Interestingly, in the development of the TPB model (see Chapter 4), in the millennial-only sample, little relationship was observed between choice of rail and belief that others would approve of one’s behavior; however, a strong relationship was observed with the belief that rail was being taken by “people who are important in my life.” This suggests that one’s personal attitudes may be intertwined with the perceptions of those of his/her peers and equals. Further research should examine the possibility that younger groups might be more receptive to information seen in the context of their relationship to their personal social networks. Hedonic Considerations in the Choice of Mode This research suggests that choice of intercity mode is strongly influenced by what psycholo- gists call hedonic considerations [see, for example, Steg et al. (2014)]. Consistently, and over several studies, the analyses show that a clear explanation for the propensity to choose rail or bus is the desire to avoid the sheer stress of the automobile trip. Work on the TPB showed this con- cern with the stress of the car trip was highly correlated with reporting that the overall rail trip was more “pleasant.” Both of these concepts are reflected in the literature of social psychology as being part of a larger category referred to as “hedonic” or “hedonistic,” which refers to desire to maximize personal pleasure (rail managers might consider this when pressured by political forces seeking to end food and beverage services on intercity trains). 10% 15% 20% 25% 30% 35% 40% My friends and coworkers usually take the train when they travel to Boston If they had to make this trip, most people who are important in my life would take the train Millennials Older Groups Figure 5. Reported descriptive norms for taking rail by age cohort.

Introduction and Major Conclusions 11 1.3.9 Rail in a Competitive Dynamic Market— What about the Competing Modes? This project provides a major update on the status of competing modes in the United States. The research also examined bus and rail together in smaller markets between rural areas and major urban destinations. In addition, the research team has concluded that the competition between rail and air may be fundamentally different in nature, as the comparative variation in both travel times and costs is so dramatic. Given the evident importance of both rail travel times and variation in air fares, a somewhat different analysis has been undertaken to explore the interaction between these two factors, including the following investigations: • The research team applied the basic structure of the NCRRP Attitudinal Model for Rail to the analysis of bus travel in the study area. The research team found that the bus industry has, compared to rail, an even more serious challenge in terms of concerns for personal safety, even among the millennials that use it. The team also found that, compared to rail choice, the selection of bus may be more related to the perceived level of stress from the auto alternative. • The research team applied the basic structure of the NCRRP Attitudinal Model for Rail to a rural market that offered both rail and bus services and found that the majority of this market has fears for their personal safety in the bus or rail trip to the urban area. • The research team further developed a separate model for the competition between rail and air. This revealed that higher-speed, faster trains may be specifically vulnerable to changes in the price of the competing air service. 1.3.10 What Is Happening to Travel Behavior as a Whole? In areas where good services are provided, Amtrak ridership is consistently up. Is this consistent with broader trends affecting travel? For decades, the expectation has been that automobile vehicle miles traveled (VMT) per person would continue to grow in parallel with economic growth. However, even this assumption regarding the pattern of auto dependence is now being challenged around the developed Western world. As will be discussed further in the following chapter, any assumption that the historical rise in personal VMT would continue the unbroken pattern witnessed in the previous century is unjustified. Several calculations have shown that personal auto use seems to have peaked around 2004–2005, as shown in Figure 6; however, this pattern is highly influenced by age. Figure 7 Source: FHWA for VMT, US Census for Total Population. 8,600 8,800 9,000 9,200 9,400 9,600 9,800 10,000 10,200 10,400 19 95 19 96 19 97 19 98 19 99 20 00 20 01 20 02 20 03 20 04 20 05 20 06 20 07 20 08 20 09 20 10 20 11 20 12 20 13 20 14 U S VM T pe r P er so n Figure 6. Change in vehicle miles traveled per person in the US population over the last two decades.

12 Intercity Passenger Rail in the Context of Dynamic Travel Markets shows that—between 1995 and 2009 [the analysis years available from the National Household Travel Survey (NHTS)]—for those under 36 years of age, VMT per person was down and, for individuals 51 years old or older, the VMT per person was up. Thus, the decline by the millennial generation in those years was strong enough to overpower the increase for those in the post child-rearing decades. The answer to the question of how the millennials will either retain their unique patterns, or adopt the patterns of older Americans, will be addressed in some detail in the scenario analysis of Chapter 6. 1.4 Report Organization NCRRP Project 03-02, “Intercity Passenger Rail in the Context of Dynamic Travel Markets,” was designed to help practitioners understand the forces influencing the choice of the rail travel mode and, once these forces are understood, to predict alternative contexts for travel behavior in the future. The remaining nine chapters document the research team’s actions and findings in conducting NCRRP Project 03-02: • Chapter 2: Previous Research and the Collection of New Data. This chapter reviews the existing literature. It describes the process of collecting the data used in the project and presents the basic descriptions for much of these data. • Chapter 3: Survey Results by Demographics, Region, and Market Segment. This chapter presents several relationships between basic demographic categories and key results of the NCRRP 03-02 survey. Market segmentation identifies groups that are bonded to rail; groups that will never come to rail; and a dynamic, malleable portion of the market between these two extremes. Source: NHTS data provided through the Oak Ridge National Laboratories website. 4,000 6,000 8,000 10,000 12,000 14,000 16,000 18,000 16-20 21-25 26-30 31-35 36-40 41-45 46-50 51-55 56-60 61-65 66-70 71-75 75+ An nu al V M T pe r D riv er Earlier 2009 Figure 7. Effect of age on annual VMT per driver, 2009 versus 1995–2001 average.

Introduction and Major Conclusions 13 • Chapter 4: Understanding Values, Preferences, and Attitudes in the Choice of Rail. This chapter presents the concept of an “analytical framework,” which develops the concept that long-term values might influence residence, that the geographic characteristics of residence might influence short-term attitudes, and that all these factors might converge to facilitate understanding variation in rail ridership. • Chapter 5: Merging Economic Modeling Theory with Analysis of Attitudes and Prefer- ences. This chapter provides the reader with a unified behavioral model that explicitly adds “softer” factors into the same modeling process used by the economists. • Chapter 6: Model Application for Scenario Analysis. This chapter illustrates how the results of the modeling process were combined to create four future scenarios, easily understandable by planners, policy makers, and all those interested in the future of rail. • Chapter 7: The Role of Rail in a Rural Market. This chapter reviews what can be learned about rail ridership patterns in markets where the more sustainable modes (rail and bus) must attract riders from more rural settings, such as those surveyed in northern New England by the research team. • Chapter 8: Competition to Rail from Intercity Bus. Given the intense competition now coming from the intercity bus industry for some demographic groups, bus and rail may have much in common in their need to compete with the private car. • Chapter 9: Competition Between Rail and Air. This chapter provides a brief summary of new research commissioned by NCRRP to better understand how the competition from air may differ from the competition from other modes; it examines a newly documented interaction between an improvement in rail travel time with variation in air price. • Chapter 10: Bringing It All Together: Where Do We Go from Here? This report concludes with a discussion of the way in which these findings suggest the need for further research in a manner most conducive to helping policy makers decide wisely about investing in intercity transportation. As FRA proceeds with follow-up work to its NEC Futures study, how might the direction of this NCRRP project affect future efforts to better understand rail demand?