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.

105 As detailed in Chapter 1, this research project employed several research methods to explore the concept of integrating both “hard” and “soft” factors in the analysis of intercity travel demand into one, unified approach—an approach that assumes the existence of several alternative sce- narios for the background market conditions in the future. Central to this research is the concept that one, single vision of what the future holds cannot be predicted, but the research team can create alternative scenarios for different underlying markets—all of which directly acknowledges uncertainty as an integral part of the analysis of future travel behavior. This final chapter seeks first to bring together several of the threads of research reported in the previous chapters and second to suggest the implications of this integrated approach on the nature of further research needed to bring this about. 10.1 Bringing Together the Separate Themes: What Was Learned 10.1.1 Incorporating Uncertainty into the Planning Process In a recent article in the Journal of the American Planning Association, Noreen McDonald writes, “Failure to address uncertainty may also lead to increasing challenges to transport plans by the public and judiciary. For example, a federal court recently ruled that the North Carolina Department of Transpor- tation had failed to meet National Environmental Policy Act requirements because they did not consider uncertainty in growth forecasts (Catawba Riverkeeper Foundation v. NC Dept. of Transportation, 2015). . . . Planners concerned about uncertainty with millennial travel can use two strategies: Improve travel demand models and adopt a scenario planning approach to consider how different trajectories for the travel of millennials could influence infrastructure needs.” (McDonald 2015) Perhaps going beyond the recommendation of McDonald, this project has suggested a new direction for analysis of intercity travel demand by applying improved travel demand models directly into creation of alternative scenarios for the future. Chapter 6 presented four illustrative market settings for intercity travel created by the scenario testing tool, presented again as Table 36. In order to approach in the issue in cautious manner, those four scenarios reflected four possible resolutions of key unknowns in the future values of several cohort groups. Within Chapter 6, the scenarios included no variations in the quality of service characteristics nor variations in possible demographic mixes, for example. This section of Chapter 10 provides an illustration of how more categories of variation could be brought together in a later application of this integrated process. Given that Table 36 specifi- cally does not deal with variation in service characteristics of any kind, the interaction between such market scenarios now need to be explored with reasonable assumptions of candidate policies toward improved infrastructure, improved service levels, and alternative costs. C H A P T E R 1 0 Bringing It All Together: Where Do We Go from Here?

106 Intercity Passenger Rail in the Context of Dynamic Travel Markets 10.1.2 How Do the Factors Interact? How cultural values interact with more traditional indices of transportation service levels and costs should be explored in some detail in further research. For the purposes of this section, the research team took one, and only one, “cultural future” (holding it constant) and examined four service levels applied to that future market, just to establish some sense of scale of the two themes in the research. The reader should be aware that all scenarios have been designed to reveal char- acteristics of interactions; there is no attempt in this section to propose a most likely forecast for rail in the study area. The research team created a single, somewhat optimistic future cultural scenario for this exer- cise. It is optimistic in that the research team specifically assumed that with the passage of (say) 20 years, the millennial generation (now in a more senior age category) will have lessened its fear of traveling with others, and of traveling on rail, specifically. This level of concern for privacy is posited to drop to that of those presently in the 35–44 age group. Consistent with this optimism about fear of travel being dealt with, this scenario posits that attitudes about privacy for those with no college degree will come to mimic those with a college degree and that attitudes about privacy for those with less than full employment will come to mimic those will full employment. The research team further hypothesized that generation Z would also have less need for privacy in travel than the present occupiers of that age category, the millennials. The single future cultural scenario hypothesizes that the current high feeling of independence from the auto held by the millennials will drop to that of those currently in the 35–44 age group, as they face the child-rearing decades, resulting in a moderately less favorable climate for rail (by way of example, if the level of auto dependence mimicked that of the current over 65 age group, the impact on rail would be more negative than assumed using the attitudes of the current 35–44 age group). As shown in the first column of Figure 55, such a “culture” of values would be associated with a 12% increase in rail ridership, ipso facto in a world where all other variables were held constant. Moving to the next step (the second column from the left in the graph), the examination of factors not under the control of the rail sector is continued. Here, the exercise assumed that, over time, highway travel times would increase by 10% and auto costs by 5%. Because buses might adopt different operating patterns (e.g., more non-stops) and more use of high-occupancy-vehicle lanes, bus times were hypothesized to increase by 5%. Consistent with general patterns in the US airline industry, non-work air fares were assumed to decrease, in this case by 5% (the lowering of non-work 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 36. Four scenarios with cultural change but no service changes.

Bringing It All Together: Where Do We Go from Here? 107 air fare is also consistent with patterns reported in Chapter 9). Figure 55 shows that the combination of the culture change and the change from the competition together would be associated with about a 20% increase in rail ridership. Finally, in the next three steps of the exercise, traditional improvements were assumed for rail: first with a 10% reduction in rail IVTs, then a 20% reduction, and finally with a 5% decrease in access and egress times to/from the rail stations. There are several key implications for the results shown on Figure 55. The ability of the modeling procedure to deal with several kinds of input data at once, including the specified changes in basic values and the travel characteristics of the competition, reveals the extent to which those two input categories impact volumes on bus simultaneously with rail. Column 1 shows how the cultural changes in the importance of privacy and auto orientation impact bus and rail about the same amount. The model very clearly shows how an assumption of increases in costs for the auto trip will move some travelers to bus, which explains why the bus ridership in column 2 increases even when the travel times are hypothesized to get worse over time. Columns 3, 4, and 5 then show a more predictable subsequent decrease in bus ridership as the travel times of rail are hypothesized to improve significantly. However, bus ridership still registers a strong net increase (over the base) in column 5, which was designed to illustrate a fairly significant capital investment commitment to rail. The independent strength of the bus market, and the power of its lower fares for its loyal customers, is suggested in Figure 55. (The research team has explored the role of lower rail fares in the scenario testing tool, and the results are very powerful. These implications need to be addressed in appropriate detail in further research.) Columns 4 and 5 shows that a future with much better travel times might result in a roughly 40% increase in rail ridership. Of that increase, it seems that about half of the growth would stem from factors other than the times and costs of the rail operation itself (e.g., from assumed changes in the environment not influenced by the actions of the rail managers). Although the results of this particular exercise are only illustrative in nature, they tend to support a basic premise of this study, that factors above and beyond the control of the rail sector must be integrated into the analysis of rail in the context of dynamic travel markets. Elements of the Combined Scenario Exercise Note: The set of columns on this chart were not ordered to represent progression over time, as the full values of Column 1, “Change in Basic Values” would presumably phase in slowly over the remaining four columns. 0% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50% 1. Change in Basic Values 2. That, plus change in compeon 3. That, plus 10% reducon in rail me 4. That, with 20% reducon in rail me 5. That, plus be er rail access In cr ea se in Ri de rs hi p ov er Ba se Rail Bus Figure 55. Changes in basic values and competition affect both rail and bus volumes.

108 Intercity Passenger Rail in the Context of Dynamic Travel Markets 10.2 Major Conclusions from the Project 10.2.1 Major Themes Revealed Are Millennials a Market That Can Be Taken for Granted? The results of this research should pose an immediate warning against any broad assumption that the future preferences of the millennial generation will automatically facilitate a growth in rail as this cohort group proceeds through the life cycle. Far from it. Analysis of the direct results from the survey reported in Chapter 3 show that this group has a higher negative feeling about “traveling with people they do not know” than the older groups. Further, they report greater fear of crime and disturbing activity than older groups. For example, in our rural sample, 28% of the millennials agreed “It might be unsafe to make this trip by bus or train,” compared with 17% of the older respondents; 23% of the millennials agreed that “The experience at the New York City bus or train station would be so unpleasant that I would try to avoid it,” compared with 18% of the older respondents. (Both of these differences were significant at p < 0.05; analyzed separately, all four responses are causes for policy concern.) As reported in Chapter 6 (the pessimistic scenario), if the millennial cohort group retains this set of values over time, and if their independence from the auto is diminished as they grow further into the child-educating years, the underlying market setting for rail growth becomes distinctly negative, not positive. It is clear that those who set policy for investment in intercity public mode services (both bus and rail) should see the millennial generation as a positive market, but one with its own standards and its own demands. Issues of cleanliness, rowdiness, and general fear of unwanted behavior need to be addressed. The importance of investments that deal with human values not associated with faster rail travel times such as the new New York City Moynihan Station, an expanded Washington Union Station, or Boston South Station should not be underestimated. Role of Auto Orientation The ICLV model developed in Chapter 5, and applied in Chapter 6, suggests that the preferences expressed concerning alternatives to car ownership are directly related to variation in rail demand. This suggests that further research should be undertaken on this kind of connection—for instance, can we prove that new mobility solutions lead to lower auto ownership, which might lead to greater propensity to take the bus or rail? Would a world with more Uber and car-sharing support pro- sustainable intercity mode choices? Or, is this a statistical fluke involving this project only? Such future research would need to explore the (assumedly) negative impact on rail of long-distance trip sharing, such as the BlaBlaCar program (www.blablacar.com) in France. Fundamental questions exist about the role of preference for auto ownership in a (future) world that includes driverless vehicles. Will the autonomous vehicle of the future be seen as something we “own” or something we “borrow”? Assuming that love of the private auto is retained in its new incarnation, will the convenience of auto automation undermine key market support for rail and bus? Role of ICT Although perhaps not as dramatically as with certain other factors, the research does link an orientation toward ICT devices in general, and being connected in specific, to the choice of rail and bus for intercity travel. In many of the future scenarios, the research team assumed that as present generations get older, they will not lose this orientation, which is seemingly so ubiquitous for the younger generations at this time. Perhaps more relevant is the question of the manner in which information technology will impact the modal patterns of the future, which is explored in the discussion of future research, in the following subsection.

Bringing It All Together: Where Do We Go from Here? 109 Future Markets for Intercity Rail A major conclusion of this NCRRP project is that the next market for growth in intercity rail would come from diversion from the automobile—more than any diversion from bus, and more than any diversion from air.2 As noted previously, this stems from the dominance of auto trip making in the study areas. The conclusion is supported by a wide range of early sensitivity testing exercises conducted with the scenario testing tool, which are not reported here. This strongly suggests that the next theme for future research in this area should focus on those factors that aid in understanding the diversion from auto to rail. Given the large share of intercity trips made by car, it is not surprising that the largest “new market” for intercity rail is those who currently drive. Results from studies undertaken for the I-95 Corridor Coalition agree with others that congestion on the major alternative routes to rail will render them more and more degraded over time. Somewhat more surprising are the results of the illustrative exercise summarized in Figure 55 that, with the specified pricing assumptions, rail ridership could grow without a net decrease in bus volumes compared to the base case. As noted in Chapter 8, air outcome volumes vary with input assumptions; in all the scenarios tested, the major diversion is from the auto. 10.2.2 Understanding Market Factors and Segments What Are the Factors That Influence Intercity Mode Choice and How Do They Interact Without question, the dominant factor in the explanation of mode choice revealed in the three attitudinal models created for this project concerns the inconvenience of the trip by the specified mode (once for rail, once for bus, and once for both). The factor Inconvenience produced the most explanatory power for variation in mode choice for all three of the attitudinal models, and in the TPB model, which had a somewhat different structure. Because these models were not designed to reflect actual trip times and costs, this factor stood as a surrogate for travel times being competi- tive. In short, if rail (or bus) is not a convenient alternative, there is little chance it will be chosen. Table 37 shows the top five factors for each mode from tables presented earlier in the text. After inconvenience, basic orientation toward the automobile, perceived stress of taking the car, valuing privacy, and fear for personal safety provide the most important predictors for mode 2 The research team has explored the role of greater access and egress times to airports, and greater congestion in the air system; these factors tend to lower air volumes while other assumptions (e.g., degradation of the auto trip) tend to increase air volumes. These implications need to be addressed in appropriate detail in further research. Bus Rail STE Factor* STE Factor* 0.46 Bus Trip Inconvenient 0.73 Rail Trip Inconvenient 0.42 Less Stressful than Car 0.34 Value Auto Orientaon 0.39 Value Auto Orientaon 0.33 Less Stressful than Car 0.13 Value Privacy in Travel 0.29 Value Privacy in Travel 0.10 Bus Trip Unsafe 0.22 Rail Trip Unsafe *The basic values are shown in italic bold and short term atudes are shown in roman. Table 37. Five most important factors in the explanation of rail and bus choice from the attitudinal models.

110 Intercity Passenger Rail in the Context of Dynamic Travel Markets choice, though the exact ranking is somewhat different between modes. In the application of the scenario testing tool in Chapter 6, issues of privacy and issues of auto orientation dominated in the creation of the four underlying scenarios. (In the ICLV model, issues of convenience are dealt with through the modeling of actual comparative times and costs, and thus not included in the initial definition of the four background scenarios.) The interaction between and among factors is expressed visually in the diagrams for the Attitudinal Models for Rail and Bus and the Attitudinal Model applied to a rural market in Figures 34, 47, and 44, respectively. What Are the Key Market Segments? Based on the conclusions of Chapter 3 (summarized in Figure 56), it is clear that a small minor- ity of the population comprise the most supportive segment for intercity rail and bus use. In some of the popular press, young, sophisticated urbanites would represent the prime market for public modes; the problem with this loyal group of Young Urbanites is that they make up only about 11% of the study area population. At the other end of the market support spectrum, the combination of Rail Rejecters and Cars for Convenience show that 44% of the market may well be out of the reach of rail. The mid-section of the market comprises 30% of the total population labeled Open to Rail, with another 15% Curious but Cautious. The latter group needs to see that their concerns with privacy and personal safety are being addressed before they become positive about rail—a scenario that this study finds plausible and worthy of policy attention, but not until their concerns have been addressed. This allows the focus to be on the Open to Rail segment. As noted in Chapter 3, this group perceives rail to be superior to the car in that train travel makes for a less tiresome and stressful travel experience. Additionally, they reveal normative and social influences, as friends and family members of this segment use the train and approve of them doing the same. In short, they are not predisposed to dislike the rail choice, if and when it provides acceptable levels of convenience, etc. Perhaps this market segment best reflects the earlier conclusion that the next additional market for intercity rail must come from those who currently drive. Thus, this quick review finds 41% of the population with generally positive predisposition to choosing the train, if its service is good enough. That same review finds 44% of the market Open to Rail 30% Cars for Convenience 29% Curious but Cautious 15% Rail Rejecters 15% Young Urbanites 11% Figure 56. Definition of market segments from Chapter 3.

Bringing It All Together: Where Do We Go from Here? 111 satisfied with their auto-oriented lifestyles, and not very likely to become consistent rail customers in the future. The remaining 15% might become rail users if important issues about the perceived lack of privacy and personal safety were solved to their satisfaction. Aspects of Rail Service to Improve or “Sell” to Targeting Market Segments Given these market segmentation results, the research team believes that, for a great percentage of the population, the image of rail is already positive. The ICLV model in Chapter 5 documents in some detail the same argument–that when looking at modal constants, rail would be more preferred than the auto for most trip purposes, if and when its comparative service levels justify it as the final mode choice. Finally, a group such as the Open to Rail segment can be observed through the structure of the TPB. As reviewed in Table 38, the application of the TPB model suggests that a member of this segment wants to make the conclusion that taking the train would be pleasurable for him/her; that others like him/her would take the train; and that he/she has the self-efficacy to actually do it. Areas to Target to Grow Market for Rail The research team believes that the results of the scenario creation exercise in Chapter 6 are remarkably clear regarding target areas to grow the market for rail. As noted consistently in this report, as the millennial generation gets older, additional personal experience may lessen the extent of disliking travel with others, and the sense of fear of the public trip associated with it. Inversely, as the millennial generation gets older, the pressure will build to become more auto- dependent. In this chapter, the research team proposes that more research should be undertaken on both these markedly different social patterns, and on what could be done to better target strategies to deal with them. Results That Challenge Conventional Wisdom Prior descriptions that have appeared in the popular press about millennials uniformly liking urbanism, disliking suburbs, and wanting to act in order to improve the environment may have been oversimplified or misinterpreted. Ironically, these oversimplifications may ultimately be irrelevant as relates to intercity rail travel, as the current analyses suggest that having urban values, or pro-environmental values, has little or nothing to do with the propensity to choose intercity rail (The reader is reminded that for metropolitan travel, both attitudes toward urbanism and the density of the residence are powerful predictors of local travel mode). The research team strongly prefers the definition of relevant groups through advanced market segmentation techniques over the use of such a priori methods as grouping people together on the basis of their age, gender, Rank STE Explanatory Factor* 1 0.59 Train Trip Inconvenient 2 0.43 Taking train would be pleasant (ATB) 3 0.33 I have the power to take the train (Control) 4 0.30 ‘People like me’ would take the train (Normave) 5 0.29 Value Privacy in Travel 6 0.28 Value Auto Orientaon *The basic values are shown in italic bold and short-term atudes are shown in roman. Table 38. Explanatory factors on rail choice from the TPB model.

112 Intercity Passenger Rail in the Context of Dynamic Travel Markets or race. Market segmentation reveals that within such a category (e.g., age), there are definable groups with positive orientation to rail and those with negative orientation to rail. The research team would encourage further attention be paid to the Open to Rail market segment rather than to generic categories such as age group. 10.3 Recommendations for Further Research 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 make prudent decisions about investing in intercity transportation. As FRA considers a next generation of policy analysis concerning the future of rail, how might the direction of this study affect future efforts to better understand intercity travel behavior? 10.3.1 Conclusions About New Directions for Modeling A key question concerning priorities for future research involves the next steps in the applicabil- ity of the hybrid/ICLV model to the day-to-day transportation demand forecasting profession. The question turns to the best use for this kind of model. Vij and Walker (2015) argue “that studies that use ICLV models need to show . . . that the greater insights into the decision-making process offered by the ICLV model can be used to inform policy and generate forecasts in unobvious ways that would not be possible using choice models without latent variables.” Based on its experience, the research team would agree with the position taken by Vij and Walker that analysis of the quantitative complexity undertaken here is of most value when such “model can be used to inform policy.” In this report, the new modeling process has flagged the importance of safety and privacy concerns for the millennial generation, and the overall need to monitor cultural patterns concerning attitudes to car ownership. In short the more complex model has shown its value in the improved understanding of public policy toward rail. A separate question concerns the “need” to incorporate such additional complexity into the day-to-day travel forecasting process. The research team would propose a cautious, step-by-step process to transform what has been the subject of academic exploration into the main stream of demand modeling, taking into consideration the importance of the parsimonious approach– additional complexity should not be worshipped for its own sake. The ICLV models that were developed to capture the effects of attitudes on mode choice were estimated using techniques that, for a variety of reasons are very computationally demanding. Estimating coefficients for a single model took multiple days, limiting the team’s ability to do the type of extensive specifica- tion testing that is typically done for travel forecasting models. There are alternative methods such as maximum approximate composite marginal likelihood (MACML) estimation and alternative model formulations that could more efficiently accomplish the same objectives. Further research is needed to explore the application of these approaches to this type of model application. In sum, the research team believes that the merging of analysis involving cultures and values with analysis based on detailed trip characteristics (including times and costs) should be further developed for the purpose of understanding public policy without necessarily assuming that it would replace more conventional methods of travel forecasting in the near term. 10.3.2 Implications from the Project for Further Research As discussed, the purpose of this project was to develop and demonstrate the feasibility of a comprehensive analytical framework to improve understanding of intercity travel demand. The continued application and further refinement of the tools developed will require more time and

Bringing It All Together: Where Do We Go from Here? 113 effort, and focused attention. For instance, the single strongest elasticity for rail share comes from a change in the price of the rail ticket, specifiable by work and non-work trip purpose. In this NCRRP project, the research team has focused on the interaction between cultural assumptions for the underlying characteristics of the market, and improvement in travel times, with fare not yet added as an additional dimension for analysis (This leads almost directly to the creation of three dimensional models, with background assumptions separate from service characteristics which are separate from pricing assumptions. Multidimensional models, are, by their nature, difficult to explain, diagram, and interpret). The same is true for alternative assumptions about the distribu- tion of demographics, such as income, education, and employment status (e.g., comparing a future with higher income growth against one with lessened income growth). 10.3.3 Relationship to Issues in Rail Policy Planning With the conclusion of this NCRRP project, the research team demonstrated that, with careful continued work, preferences of the customer(s) can be integrated into the most accepted for- mat for travel demand forecasting (discrete choice modeling). Given the conclusion that a set of methods can be developed to specially incorporate values and attitudes into an integrated analysis process—one which emphasizes the appropriateness of scenario futures as a method to incorpo- rate uncertainty—the research team has developed a list (discussed in the following subsections) of possible applications of such a process in the ongoing effort to improve rail demand analysis. (The order of the following list discussion is random.) Research About Refining Future Scenario Applications Dealing with Extremely Long-Range Time Frames. In a possible policy context, we might assume that one package of infrastructure investments is required first, and could reasonably take 10 to 15 years to complete (Quite possibly the first decade of expenditures might largely deal with situations of safety and state of good repair, rather than radical improvement in travel times, etc.). We could assume further, that a second decade of facility/service improvements might follow that make highly predictable incremental improvements in travel times, and capac- ity at key bottlenecks. For both of these planning settings, the analyst might possibly conclude that the addition of complicated alternative future scenarios is not needed to produce reliable, documentable forecasts. However, as soon as the policy analyst is asked to consider infrastructures and services which are different both in the time of the implementation and in the qualitative details of their services, it would become entirely appropriate to insist on an approach based on alternative future scenarios (see, for example, Catawba Riverkeeper Foundation v. NC Department of Transportation 2015). Such an approach might require exploration of the concept that, over a several-decade planning horizon, the values and preferences of the market almost certainly will have shifted, in some direc- tion and to some extent (both potentially unknowable). By way of simple illustration, if society spends two decades actually improving the terminals, the “culture” of attitudes toward terminals would be expected to evolve over time (Over such a multidecade time frame, it could be argued that the cultural attitude about travel behavior between Paris and Lyon actually changed over time). Understanding the Mechanisms of Cultural Change. Such a future research effort might seek to better explain just how such an evolution of preferences might actually occur. Such a theory of change might start with the base case showing the rail rider skeptical of an attribute; in an early step, the customer begins to show credulity in the observed good intentions of the service provider; a key element in the evolution might find the customer routinely pleased about the attribute, moving toward final steps of sharing their enthusiasm with their peers and equals [Others have attempted the application of Prochaska’s Transtheoretical Model to transportation. See Mundorf et al. (2013)]. In short, further research might explore just how these changes in

114 Intercity Passenger Rail in the Context of Dynamic Travel Markets preferences could actually come about [Others suggest that, by contrast, raising the customer’s expectation about an attribute, and then failing to live up to that expectation, may be among the worst of possible outcomes. Personal communication with W. Brog. See Amrosell (2008)]. Research About the Tradeoffs in Rail System Design Attitudes and Preferences About Directness or Frequency. In some cases, rail system design- ers must deal with design tradeoffs that pose a challenge to the rail demand analyst. Within an overall system design, one strategy might offer directness of service coverage: a small number of trains per day might provide no-change-of-train service from point A to point B. An alternative system design might emphasize service frequency, almost ensuring that transfers will be required between trip segments. Is it possible that travelers have strong pre-set preferences on the subject of transfer? Is the transfer assumed to be a pleasant experience, or a virtual guarantee that one cannot find good seats on the second segment? Would an analytical approach that integrates real times and costs with very focused attitudes be of help in such a question of system design? [A somewhat rare example of analysts having access to longitudinal data in such a focused study is provided in Hess and Adler (2009).] Research About ICT and Rail Service How ICT Might Be Used in the Service of Rail. Research could be undertaken to use analytical methods that integrate “hard” and “soft” variables to better understand the need for information technology to support the rail and bus user through all segments of the complex multimodal and intermodal trip. Technology currently being applied in Germany helps the public mode traveler with step-by-step (literally) information in support of the complicated transfer movements often required in the intermodal trip. Much of the current work in this area is simply based on assump- tions that, of course, rail travelers want better information and would benefit from having that information. The new research tools could help in both exploring the need for such applications and making the case for the improved ICT systems needed. How ICT Might Be Used to the Detriment of Rail. Without question, the private car will become more and more the location of a vast set of information and entertainment technologies. Research could help us understand the implications of these patterns on public mode use. In the nearer term, long-distance ride sharing will soon become a major mode of travel, as it has already become in Europe. In the long term, automation of private automobiles could alter some of the most basic factors that differentiate the auto experience from the rail and bus experience. Research into these areas will require the best tools of travel demand analysis that can be developed, over time. 10.3.4 Relationship with the Larger Question of Transportation Demand Perhaps most importantly, it should be restated that this NCRRP project focused only on inter- city transportation behavior. As is well documented, the number of miles of long-distance travel for a given person is a small percentage of her/his yearly travel, most of which takes place in or near the metropolitan area of residence. Future research into the question of integration of long-term cultural values, near-term attitudes, and mode choice needs to deal with person travel in general, in addition to further research about intercity travel. If and when this is done, this NCRRP research may suggest some direction. Dealing with the Automobile The key issue in understanding any mode of travel in the United States is its relationship with travel by private car. This is not an issue of preference or bias, but simply due to the fact that

Bringing It All Together: Where Do We Go from Here? 115 most trips are made by car in this country. There is a significant body of research on alternatives to the automobile in the European literature, much of which assumes that concerns about the environment are a near-universal motivation—a dimension which the research team was not able to replicate or confirm in this American study. Given this, the question turns to how Americans make the decision to make trips by modes other than the private automobile. What are the key factors involved in such a mode choice? With the presence of newly improved research tools which combine “hard” and “soft” factors in the influence of travel behavior, these methods can be applied to the more general question of the choice to replace (single-occupancy vehicle) auto travel with more sustainable modes and patterns [Of course, increased levels of auto occupancy (by whatever strategy) falls into the latter category]. In this pursuit, a wide range of research methods should be applied. A considerable body of European research explores the decision to choose a mode other than the car by applying values such as hedonism, egoism, security, and need for power (see Figure 8). While in this project, the research team defined the basic values to support the development of certain future scenarios, in subsequent research, exploring motivations associated with basic psychological needs should not be ruled out. There are many theories currently being developed in the literature to guide the examination of the shift away from the car, which deserve future research attention. Acknowledging the Bus The research has provided considerable support for the concept that bus markets are largely separate from rail markets, at least with the pricing assumptions made in this research project. In fact, given that it may not be in the best interests of Amtrak to lower its fares, a vast market exists for bus travel even in scenarios where rail running times are significantly improved (Figure 55). Arguably, given the severe capacity constraints that impede increased train service, intercity bus could be seen as a positive complement to the rail system, and should be treated as such in public policy. Compared with any other major mode, it is the intercity bus which has lagged behind in terms of basic research about its potential. Going Beyond Planning—Understanding Marketing and Advertising And finally, connections between marketing strategies and the kind of integrated analysis approach used here should be better documented. In many instances throughout this project, it becomes clear that a given issue might be more related to the marketing of a set of services than to the service and infrastructure planning for them. Just how marketing and advertising relate to the series of serious challenges revealed in this research needs to be explored in more detail.