48 Transportation Research Record 894

Seati ng Space Group trip Total trip turn 2 2 Turning maneuver C, 001 =I+ [0.0077 (63 -w) + 0.16 (30 -1) ] y, v Vertical direction (for 30 < w .; 63 and 18 < I .; 30) (C) z = Normal direction to cabin floor Appendix B: Nomenclature Appendix C: Flight Profile

a Acceleration An example of a more complex flight profile composed C = Comfort rating on a seven point scale of maneuvers and straight/level flight can also be dB(A) A-weighted noise level, dB obtained from the SEGMENT program. As will be E .. Event (a given ride situation) shown, there are 21 ride segments, 11 of which are g ,. Acceleration of gravity, 9.8 m/s' ::;;i:.Laiyht/l~vel J:.i..&.yuc.. This pLuL.L.u: Lepresents a h Rate of change of altitude, m/min typical flight with maneuvers at the beginning and 1 • Seat ley1oom, cm end for take-off and landing. There are also maneu­ p Roll rate, deg/s vers in the middle of the profile for the altitude s = Satisfaction change. The flight profile is summarized below: T = Temperature, °C V • Indicated air speed, knots Segment Segment Description w Seat width between armrests, cm 1 Maneuver--steady descent/climb y • Flight path angle, deg 2 Maneuver--compound maneuver 6 Kroneker delta 3 Maneuver--longitudinal dec.eleration e = Pitch angle, deg with pitchover cr ~ Standard deviation of acceleration, g 4 Maneuver--simple turn or s-turn ' = Roll angle, deg 5-10 Straight-level flight--terrain flati altitude = 5000 ft Subscripts 11 Maneuver--steady descent/climb 12 Maneuver--steady descent/climb cm = Compound maneuver 13-17 Straight-level flight--terrain de ,. Descent or climb maneuver wateri altitude = 5000 ft E .. Event 18 Maneuver--simple turn or s-turn env Environmental (factors other than maneuvers 19 Maneuver--longitudinal deceleration and seating) with pi tchover h Rate of change in altitude 20 Maneuver--steady descent/climb 1 = Longitudinal direction 21 Manuever--steady descent/climb man = Maneuver max Maximum Since all the segments must be of equal time dur­ mot = Motion ation, a common denominator for the time of a seg­ no = Noise ment must be found. For example, the maneuver or po " Pitchover straight/level flight with the shortest duration seat Seating space could be picked as the denominator if all other 'I' = TP.mperaturP. periods of flight character is tics are integer mul­ t Transverse direction tiples of this segment.

Structural Models of Attitude-Behavior Relations for Intercity Rail Travelers

PETER M. ALLAMAN AND TIMOTHY J. TARDIFF

The interrelationship of various attitude and behavioral measures of intercity traveler style, demographics, and trip characteristics in the formation of per­ rail travel within a simultaneous equation, multiattribute formulation is ex­ ceptions, the translation of perceptions into specific dimensions of satisfaction, amined. The overall structure of perception influencing preference, preference and the translation of these components of satisfaction into overall effect and influencing activity, and activity influencing perception has been established, frequency. The use of a methodology such as this can aid decisionmaking for although not for all dimensions of user evaluation. Satisfaction with the sched· service offerings, advertising campaigns, and design studies. ule, cost, speed, and physical design of the train seems to be the most impor· tant determinant of an overall positive evaluation of train travel, although the social environment inside the car and pretrip experiences also have an impact. Overall evaluation of train travel has a positive relationship to frequency, and Over the past few years, transportation planners and frequency in turn influences satisfaction with schedule, and (very slightly) per­ researchers have been exploring ways to integrate ception of the physical dimension of train travel. Satisfaction with schedule is consumer needs into the design process. A prime mo­ influenced not only by frequency of travel but also by travelers' general evalu­ tivating factor behind these efforts has been the ation of the train regarding cost, schedule, and comfort. Satisfaction with de­ desire to make cost-effective trade-offs among sys­ sign aspects of the train is influenced by perception of the physical qualities of tem features. Implicit in this approach is the as­ the ride, as well as perception of the train's comfort. The image of the train sumption that the quality of trade-off decisions re­ tends to influence evaluation of the food and facilities available on the train. quired in designing these systems may be improved by The major negative finding is the lack of significance of frequency in predict· a fuller understanding of which system features are ing more aspects of traveler perceptions or dimensions of satisfaction, in con­ important to users. There is a general consensus trast to its pivotal role for urban travel. These results indicate the role of among researchers that it is necessary to obtain Transportation Research Record 894 49

consumer ratings about the characteristics of sys­ al and specific, then determine a traveler's liking tems. However, there is disagreement about how to (or effect) toward the service offerings. This combine these ratings of features with user demo­ liking is then hypothesized as determining travel graphic characteristics and other data into a theory frequency. Travel frequency is assumed to determine of travel behavior, attitudes. This feedback relationship may be This paper presents one form of a multiattribute, stronger for some types of attitudes than others. non-recursive system of equations. This system The fundamental point, however, is that not only do links the perceptions, preferences, and behavior of attitudes influence the behavior of travelers, but intercity rail travelers with their demographic this behavior itself may also act to influence atti­ characteristics and trip characteristics. The ap­ tudes. Most commonly, this occurs when users gain a proach has its origins in similar studies performed greater familiarity with the service and modify in the urban travel field (1-3), These studies and their attitudes accordingly. Other mechanisms may others (4-6) have demonstrat';;d- the importance of at­ be involved as well. The purpose of this research titudinal variables for transportation research. In is not to clarify the sociopsychological mechanisms addition, the existence of feedback relationships that produce such feedback relationships, but simply between attitudinal variables and behavior has been to investigate whether they exist or not, and their investigated !!-l>· Not only are the users who rate relationship with other attitudinal constructs. features more favorably more likely to use the sys­ tem frequently, but frequent users of the system are DATA SOURCE likely to rate its features more favorably. This implies that attitudes affect behavior and, in turn, A series of focus groups was conducted to generate behavior affects attitudes. In addition, it has an understanding of the problem and to aid in ques­ been shown that user preferences constitute more tionnaire development. The data used for the model­ than a simply weighted combination of ratings of ing came from approximately 2200 questionnaires ad­ features . That is, user preference predicts fre­ ministered on various parts of the Amtrak system in quency of use over and above the predictive ability the summer and early fall of 1980. The question­ of perceptions Of a system's particular features. naire was one that respondents completed on board The present study extends this approach in two the train and implemented the major aspects of the principal directions. One emphasis is to establish modeling approach set forth above. more clearly the nature and determinants of user The trains sampled and the returns received are preference (termed "modal effect") and its relation­ given in Table 1. Six train types and potential ship to other attitudinal and demographic vari­ trip distances of between 3 and 50 h are included in ables. This will not only enlarge our understanding the sample. The interviewing was conducted by Am­ of this theoretically important variable, but will trak personnel. Interviewers boarded the trains at also enable us to focus on the design and evaluation prearranged points and passed out the questionnaires of transportation services in a manner consistent according to the sampling design. A set of instruc­ with the desires of users or potential users. tions was supplied in order to standardize the pro­ The second principal focus of this paper is on cedures. Questionnaires were gathered on six dif­ Amtrak intercity rail travel. The approach is to ferent types of equipment operating on the system: apply the theory and methods developed within the Superliner, Metroliner, Turboliner, long-distance context of urban travel to a study of intercity steam, headend power, and Amfleet equipment. Each travel by a particular mode and for an organization is distinctive in some way. that must design its service offerings to suit the The Superliners are bilevel cars that have many needs of its clientele. Such an approach will test amenities and boast a new image. The facilities are the robustness of the concept across domains of con­ such that one can envision taking a long-distance tent (urban versus intercity travel) and is intended trip across country purely for the enjoyment of the to develop conclusions relevant to the planning trip and not for some more practical purpose. The needs of an organization that is providing these Metroliner runs between Washington, D.C., and New services. York City in the heavily traveled Northeast Corri­ As in all marketing research, two approaches are dor. Its service characteristics and equipment possible to transportation research. One is a dol­ characteristics are distinctive, as is the image lars-to-sales approach that concentrates on inducing that is promoted. The Turboliner sampled is a tur­ current travelers to travel more and on attracting bine-powered train that runs between New York City new travelers who are most like current users. The and Syracuse, New York. The relatively short dis­ contrasting dollars-to-opportunity approach concen­ tances of these last two types bias the ridership trates on winning over the users of other modes of toward business travel. transportation. This study adopts the first ap­ The long-distance steam cars are the older cars proach--the emphasis is not on mode choice but rath­ that have been used on the system for several de­ er on dimensions of user satisfaction. To focus on cades. Their heating and cooling systems are pow­ mode choice is to examine aspects of service that ered by steam, which tends to be less reliable than are applicable to the comparative desirability of the more modern "headend power" arrangement. How­ alternative modes, whereas this study seeks informa­ ever, they are very comfortable cars, designed for tion on effective service designs for a particular long-distance travel. Headend power cars are com­ mode. This affects both the populations of interest pletely renovated and derive their heating and cool­ for study and the applications of the resulting in­ ing power from a power car at the head of the train, formation. Rather than focusing on the entire trav­ a .)!lore reliable technology than steam. Although eling population, this study emphasizes specific re~ovated from the same chassis as the long-distance groups of users and a specific mode. steam cars, they are modern cars. The Amfleet cars are newly developed and are also used on a good part MODELING ORIENTATION of the system. They have different ride character­ istics from the others, but offer modern amenities The conceptual structure from which this work draws and a new image. is discussed in detail elsewhere (7,8), The general structure is discussed briefly hei'°e. Demographic STRUCTURAL SYSTEM variables are seen as influencing traveler atti­ tudes; these attitudes (or evaluations), both gener- The overall concept previously described was imple- 50 Transportation Research Record 894

Table 1. Returns by equipment type and route.

Distance Scheduled Avg Speed Returns Equipment Type Train Name Route (miles) Time (h) (mph) Obtained

Superliner Empire Builder Seattle-Chicago 2281 49 :348 46 664 Amfleet Palmetto Savannah-New York 828 15:45 53 212 Inter-American Houston-Chicago 1333 20:05b 46 208 Total 420 Steam cars Silver Meteor New York-Miami 1379 26:25b 52 122 Crescent New Orleans-New York 1378 30: 59b 44 112 Southwest Limited Los Angeles-Chicago 2240 43 :35" 51 ___!.!.L Total 34~ Headend power Broadway Limited Chicago-New York 909 20:48b 44 213 · Lake Shore Limited New York-Chicago 939 19 :55 48 170 Total 383 Turboliner Salt City Express New York-Syracuse< 286 5:10 55 190 Metroliner Metroliner New York-Washington< 224 3:45 60 182 Total returns 2184

3 Two-nieht trir. bnvernight trip . cSampled in both directfons.

mented as a set of 12 simultaneous equations. The GENERAL PERCEPTIONS general perceptual attitudes are represented by four scales. Specific perceptions and components of Perceptions of train travel for trips like the cur­ preference are represented by the six satisfaction rent trip were assessed by means of a factor analy­ scales. Effect is represented by the answer to the sis of 15 semantic differential items. The first question: "Overall, how satisfied are you with your factor relates to physical aspects of the trip, such train trip today?" Frequency of travel is a func­ as cleanliness, noise, comfort, and amount of space, tion of preference and in turn is a predictor of the as well as evaluative items, such as the trip's be­ general perceptual attitudes. Two-stage least ing pleasant and impersonal. The second factor squares were used to estimate the interrelationships dealt more with the image of the trip, such as being among the jointly dependent variables. interesting, smooth, uncrowded, and fashionable. The third factor related to aspects of service, such SATISFACTION SCALES as reliability, expense, convenienc~, and speed. The three general perceptions were thus physical The six dimensions of user satisfaction were devel­ amenities, image, and cost/schedule aspects. Fi­ oped from 32 items by factor analysis. Two factors nally, a comfort score was developed from the stan­ measure aspects of service provided before the trip dardized scores to seven comfort items. er ;::t the beginning of the trip.. One dealt oe:-!clu­ There i".:re a la:r']e munb"r of innepenil<>nt variables sively with pre-trip assistance, correlating highly used in the system. These include variables dealing with luggage assistance and assistance in boarding. with aspects of service such as trip time (total The other related to trip cost and terminal access. trip time as well as the period of time from the be­ Clean and safe terminals had a marginal correlation ginning of the trip up to the time of the inter­ with both of these factors. view), trip and terminal access mode, time, and The third factor had more to do with the speed problems. The alternative mode (plane or car) for and performance of the train, as the highest load­ this trip is also represented, as is the equipment ings were obtained for on-time performance and speed type on which the traveler is making the trip. At­ of the trip. The fourth factor correlated highly tributes of the trip include trip purpose, number in with the measures relating to food service and on­ the party, and type of party, including whether the board facilities. The price of food, the quality of traveler is alone or traveling with children. The food service, clean restrooms, and facilities for traveler's age, family income, employment status, washing up and comfortable sleeping correlated educational attainment, and household type are also highly with this factor. included in the models. Except for length of time Two factors represented the on-board environ­ on the train, number of pre-trip troubles, access ment. One related primarily to the social environ­ time, and number in party, all ot the above vari­ ment on board, as well as the service provided by ables are dummy variables for the presence or ab­ on-board personnel. Indeed, the highest loading sence of the characteristic. variable was service to passengers, followed by friendliness of other passengers, friendly and help­ TRAVEL STYLE ful train personnel, as well as passing interesting scenery, and information provided on board. The Another aspect of travelers is their travel style, final factor related to the physical environment and or how they approach the travel experience; these design of the cars. Variables with high loadings are represented as continuous variables. Riders' included amount of space per passenger, comfort of approaches to and views of the travel experience the seating area, choice of places to sit, interior were explored with a series of 26 items in a Likert design of the car, luggage storage areas, a feeling scale format. The set of questions was developed of luxury, and amount of privacy available. Other from analysis of focus group responses and, to a variables that had moderately high loadings included limited degree, from other questionnaires. The set temperature and ventilation, separation of smoking of questions is similar in purpose to "attitude, in­ and non-smoking areas, and provisions for comfort­ terest, and opinion" batteries used in psychographic able sleeping. research; however, the set of questions used here is oriented specifically to dimensions of travel style and is not related to other aspects of consumption, as are the typical psychographic batteries. These Transportation Research Record 894 51

Figure 1. Structural system: interaction of jointly dependent variables.

COST/ + SCHEDULE SCHEDULE

PRE-TRIP ~---- EXPERIENCE

+ FOOD/ FREQUENCY COMFORT IMAGE FACILITIES OF USE

COST/ ACCESS

+ SOCIAL ENVIRONMENT

'------! PHYSICAL PHYSICAL AMENITIES ENVIRONMENT

GENERAL PERCEPTIONS COMPONENTS OF OVERALL BEHAVIOR SATISFACTION SATISFACTION LEGEND Significant at .05 Level or Better Significant at .10 Level or Better questions, with some modification, could be used to preference. These include satisfaction with cost study the travel style of users on any mode. In­ and access, satisfaction with schedule, and satis­ deed, the focus group results suggested that users faction with the physical environment in the cars. of different modes would score differently on these Pre-trip experiences and the social environment on dimensions, although that is a topic for future re­ board are also marginally relat.ed to overall satis­ search. faction, and satisfaction with food and facilities The first factor appears to be a sociability fac­ bears little relationship. tor, having high correlation ~ith enjoyment of being Preference has a strong positive relationship to seated next to a stranger and, usually, starting a frequency of taking the train in the past year for conversation when traveling. The second factor, this type of trip. Frequency itself is signifi­ labeled "destination orientation," consists of cantly related (negatively) only to satisfaction strong agreement with the statement, "Getting to my with the schedule. Whether this lack of relation­ destination quickly is very important to me." It ship is due to statistical reasons inherent in the also encompasses a tendency to be in a hurry or to estimation procedure or to an actual lack of rela­ become upset when the schedule is not met. Third tionship could not be determined in the research. was a family orientation factor, accounting for peo­ P.lthough frequency of use is not related to the ple with travel plans dependent on others in the general perceptual factors, perception of comfort household and leading lives centered around others is. A high evaluation of comfort is associated with in the household or family. The fourth factor rep­ a positive evaluation of physical amenities. Com­ resented a Spartan travel style, characterized by fort is positively related to the second factor, packing only essentials and not demanding much when which represents a positive image perception of traveling. The fifth factor measured the amount of train travel as being interesting, smooth, un­ peer support or pressure experienced in the respon­ crowded, and fashionable. Comfort is related to a dent's social environment either through friends' positive evaluation on the third dimension, which behavior or reactions to train travel. The sixth measures the reliability, expense, convenience, and factor related to the dining style that respondents speed of train travel. Comfort thus plays a pivotal preferred, either eating a hot, sit-down meal on the role in these models. train, or bringing one's own food on board the The general factors are related to several of the train. The seventh factor indexed the respondents' components of satisfaction as well. Evaluation of travel orientation, involving a willingness to enjoy comfort and the general evaluation along the cost the trip for its own sake, rather than as a means to and schedule dimension are both related to satisfac­ an end. The eighth factor related to a tendency for tion with the schedule. The positive image percep­ the respondent to spend the time alone when travel­ tion of the train as being interesting, smooth, un­ ing, not traveling in a group, and leading a hectic crowded, and fashionable is positively related to life. The final factor had a high correlation with the satisfaction with food and facilities. Finally, the variable, "The best kind of trip is one without the perception of the train as possessing physical many plans." This measures orientation toward trip amenities is significantly related to satisfaction planning, with the physical design and environment dimension of satisfaction factors. Jointly Dependent Variable Interrelationships It would appear that on-time performance is a very important component of the attitude-behavior The interrelationships among the jointly dependent feedback relationship: however, this component is variables are displayed in Figure 1. This figure supplemented by aspects of the physical and social summarizes the coefficient estimates. environment aboard the train. It is natural to begin with the determinants of the overall liking of the train trip (effect). Three of the factors are strongly related to user 52 Transportation Research Record 894

Contributions of Indepe ndent Variables 3, The variables for the jointly dependent vari­ ables are represented in Figure l and have been dis­ A listing of significant variables for the satisfac­ cussed in the previous section, but they are also tion component equations is reported in Figure 2; presented in Figures 2 and 3. This section de­ those for the other variables are found in Figure scribes primarily the other influences identified

Figure 2. Significant variables in satisfaction equations. Food & Social Physical Schedule Cost-Access Facilities Environment ~esi gn

Frequency (-) Amfl eet; Headend Alternative Mode Perception of Time onboard ( + l Perception of Cost/Schedule Power; Steam (+) Plane (+) , General ( +) Superl i ner ( - ) Phy5ical General Age over 64 (+) Superliner (+) Image Am fleet; Headend Amenities (+) Perception Pretrip troubles(-) Amfleet, Headend Time onboard (+) Power; Steam(-) Time (+) Power; Steam (+) Dining Style (+) Female (+) Onboard ( - ) Time Age over 34 (+) Family with Superl iner (-) Onboard (-) Number of pretrip Children(-) Number in Superl i ner troubles ( - ) Number in Party (-) (+) Terminal Access party ( +) Age 25-54 ( - ) Time ( - ) Sociability Age over Income under Travel Style (+) 64 (-) $15,000 (-) Dining Style (+) Some Respondent employed Spartan Travel College (+) full or part time (+) Style (+) Spartan Travel Travel Style (-) Orientation (+)

Comfort ( +) Alternative Mode Alternative Mode Supe rl i ner ( +) Work trip ( - ) Travelling Family income Plane(-) Car ( +) Traveling Alone(-) Vacation al one ( +) under Superliner (+) Traveling with Income under trip ( - l Age 25-34 (-) $15,000 (+) Female (-) Preschoolers (+) $15,000 (-) Family Destination Terminal Number in party ( +) Respondent employed Orientation (+) Orientation Access Time (+) Dining Style (+) part or full ( +) Access by Public time (-) Dining Transit (-) Spartan Travel Style(-) Spartan Travel Style (+) Peer Style ( +) Pressure (+)

NOTE : "+"denotes positive relationship and"-" denotes negative relationship. Variables above the dashed line are s1gnlt1cant at p < .u~ anc:I those t>e 1ow the line are s1gn1 .t1cant at .IJ~Z p < .~.

Figure 3. Significant variables in other equations. General Perception Factors Cost­ Imag e Ph -y-s~i -c ~a,.----- Com fo rt Modal Affect Frequency Schedul e Perception Amenities Perception

Satisfaction with: Modal Comfort ( + l Comfort (+) Comfort (+) Am fleet; Steam; Cost-Access ( + l Affect (+) Total Trip Time (-) Time Onboard (-) Time Onboard (-) Headend Power (+) Schedule (+) Superl i ner (-) Age over 64 (+) Female (-) Amfleet; Steam; Female (+) Design (+) Am fleet; Steam; Inc001e under Family Ori en- Headend Power Sociability(+) Non-work Headend Power (-) $15,000 (-) tat ion (-) ( - ) Destination Trip (-) Work Trip (+) Income greater than Retiring Travel Peer Pressure (-) Orientation (-) Age 55-64 (-) Recreation - $30,000 (-) Style (+) Family De~Li11dliu11 Vdt:dliU11 Tr'iµ (+) Di rii 11y S Lyle ( - ) Orientation (+) Orientation (-) Visiting Trip (+) Dining Style (+) Travelling Alone(+) Peer Pressure (+) Income more than Spartan ( +) $30,000 (+) Travel Destination Orientation ( +) Orientation (+) Peer Pres sure ( + l

Pre-trip Female(-) Superliner (-) · Superliner (+) Frequency ( +) Preschooler Satisfaction (+) Terminal Access Amfleet; Steam; Age over 54 (+) Superl i ner ( +) Along ( +) Social Time (-) Headend Power (-) Sociability Age 18-24 ( +) Child Under 18 Environment ( +) Family Pretrip Troubles ( -) Travel Style Along ( - ) Travel ling with Orientation ( - l Access Time(-) ( - ) Age over 54 (-) Child under Travel Spartan Travel ~lumber in Party 18 (+) Orientation (-) Style ( - ) (-) Retiring Travel Income under Style (+) $15,000 (+)

NOTE: "+" denotes positive relationship and "-" denotes negative relationship. Variables above the dashed line are significant at p < .05 and those below the line are significant at .052 p < .2. Transportation Research Record 894 53

from the models. Available space does not permit a travel more highly. The Super liner is rated presentation of the coefficient estimates them­ slightly higher than other equipment types for this selves: these are available from the authors or in dimension, while those traveling alone rated it the report (_!!) • slightly lower than those not traveling alone. The food and facilities are rated somewhat lower by low­ SATISFACTION COMPONENTS income people, perhaps because high-income people are not constrained by price. Employed persons rate This discussion begins with the six satisfaction the service low. Those who favor a hot, sit-down factor scores. Each was modeled separately: the re­ meal rate the train very highly on this dimension, sults are displayed in Figure 2. as do those favoring Spartan travel style. Satisfaction with speed and schedule is a nega­ Travelers' satisfaction with the social environ­ tive function of frequency of travel: that is, the ment aboard the train is not affected by frequency more experience the traveler has with train travel, of travel, or by perceptions of the physical envi­ the less satisfied he or she is with this dimen­ ronment. However, this factor is positively related sion. However, the more comfortable the ride, and to length of time on the train, perhaps because of a the greater the general perception of cost and process of adjustment and getting to know other pas­ schedule, the greater the satisfaction with this sengers. Train types other than Metroliner or specific dimension. Satisfaction with speed de­ Turboliner scored lower on this variable. Females clines as a function of length of time on the train were more satisfied with the social environment but is higher for travelers on the Superliner than aboard the train than males. Those in a larger for other types of equipment. Low-income persons party were more satisfied. However, those traveling are more satisfied with this dimension, as are per­ with children or those traveling on business or va­ sons who have attended college. With other factors cation were less satisfied. A number of travel controlled, however, older people are less satisfied style variables were positively related to satis­ with this dimension than younger people. faction with the social environment, as might be ex­ Pre-trip satisfaction tends not to be related to pected. The stronger variables included the socia­ frequency or aspects of the on-board experience, as bility, family orientation, eating, Spartan style, one would expect. The best pre-trip experience and travel orientation variables. seems to be associated with Amfleet, steam, and Satisfaction with the physical design of the headend power cars and Superliner cars to a lesser cars, again, was not related to frequency of travel extent. Females are very slightly less satisfied but to the general evaluation of the physical envi­ with this aspect of travel. However, senior citi­ ronment aboard the train. With regard to train zens are well satisfied. Travelers with the plane type, satisfaction with Superliners was lowest. as an alternative mode are slightly less satisfied Satisfaction with car design deteriorated with with the pre-trip experience. Several variables length of time on board, but was somewhat more posi­ were included to measure aspects of the pre-trip ex­ tive for those traveling alone. The larger the num­ perience directly as related to pre-trip satisfac­ ber in the party, the lower the satisfaction with tion. The number of troubles encountered before the car design; perhaps this is because of the two-by­ trip is negatively related to pre-trip satisfaction, two arrangement on most trains that prohibits large as would be expected. Somewhat paradoxically, the parties from sitting together as a group. People length of time it took travelers to get to the sta­ aged 25-34 were slightly less satisfied with this tion, the greater the pre-trip satisfaction, al­ dimension. Three travel style variables were sig­ though this may signify trouble-free access by auto­ nificantly related to design satisfaction. Those mobile. To a slight extent, users of public who were destination-oriented were more satisfied transportation to the terminal are less satisfied with the design than those who were not. Those who than others, and those who score highly on the Spar­ preferred a hot, sit-down meal were less satisfied tan travel scale are somewhat more favorably dis­ with the physical design of the car. Those who per­ posed toward the pre-trip experience. ceived support from peers were more satisfied with As with pre-trip experience, the cost and access physical design. satisfaction score is not related to either fre­ quency or the general satisfaction variables. Those OVERALL SATISFACTION whose alternative mode of travel is by plane rate the train much more highly on this dimension and The other components of the system are detailed in those whose alternative mode is by car have a slight Figure 3. The most important of these variables is tendency to rate the train higher. The Super liner, effect, the overall satisfaction with the trip it­ Amfleet, steam, and headend power cars are rated self. This was modeled as a function of all of the quite highly on this dimension compared with other satisfaction variables as well as other independent types of equipment. Those who have a pre-schooler variables. As noted in the preceding section, sat­ along on the trip also evaluate this dimension isfaction with the schedule, cost, and the physical highly, as do persons of middle age and above. design were the most important determinants of over­ Again, the number of troubles encountered is nega­ all satisfaction. There are weaker contributions tively related to satisfaction on this dimension, as from the social environment on board the train and is access time. The larger the party, the greater the pre-trip experience as well. In general, satis­ the satisfaction with cost and access. High-income faction regarding recreation, vacation, visiting, individuals are more satisfied with cost and access and personal business trips was lower than for work than are low-income persons, however, and those who trips or other types of trips. Families traveling are employed rate this dimension more highly than with children under 18 tended to be slightly more the unemployed. Of the travel style variables, the satisfied overall. Persons aged 55-64 seemed to be higher the Spartan travel score, the lower the sat­ significantly less favorably disposed toward the isfaction: the higher the preference for a hot, sit­ train, other considerations being equal. Those who down meal, the greater the satisfaction with cost. were destination-oriented were less satisfied with Although frequency is not related to satisfaction the train trip overall than those who were not. For with food and facilities, the general evaluation di­ the other travel style variable tested, those who mension relating to the trip image is strongly and scored most highly on the retreatist orientation positively related. Those who have traveled a (spending time alone when traveling) had a weakly greater distance tend to rate this aspect of train positive relationship with overall satisfaction. 54 Transportation Research Record 894

TRIP FREQUENCY tic life pace viewed the train more positively along this dimension. Overall satisfaction was used with other variables The comfort dimension, so important in predicting to predict a frequency of travel for this purpose in other evaluations of train travel, is not itself in­ the previous year. Overall satisfaction had a fluenced by trip frequency. Amfleet, steam, and strong positive relationship to trip frequency. headend power cars are seen as somewhat more com­ Those persons on trains outside the Northeast Corri­ fortable than other types. Females view the train dor had significantly lower trip rates than did as more comfortable than males. There is a slight those riding Metroliner or Turboliner equipment. tendency for those with a pre-schooler along to view Likewise, those taking work trips had a signifi­ the train as comfortable, but there is also a slight cantly qreater number of trips than others, although tendency for those traveling with family members un­ both recreation/vacation and visit/personal business der 18 to view the train as relatively uncomfort­ trips were taken with greater frequency than school able. Older persons view it more negatively, as do travel, the omitted category. Females had a those in larger parties. Low-income persons view slightly lower trip rate than males, and those trav­ the train as comfortable. Although those with a eling alone a significantly higher trip rate. There destination orientation view the train as less com­ was a slight tendency for those with longer access fortable, those scoring high on sociability, family times to travel less frequently. High-income indi­ orientation, eating style, peer pressure, Spartan viduals tended to travel more frequently than oth­ style, and travel orientation score the train highly ers. Regarding travel style variables, persons with on comfort. a destination orientation and those with peer sup­ port traveled more frequently than others. Family ACKNOWLEDGMENT orientation had a negative but very weak impact on travel frequency. The work reported here was performed under contract to the Transportation Systems Center (TSC), U.S. De­ GENERAL PERCEPTIONS partment of Transportation. Donald Sussman and Joseph Dumas were the TSC contract monitors for this The evaluation of cost and schedule aspects of train research. Helpful suggestions on questionnaire de­ travel in general is not significantly affected by sign and analysis were received from them, as well frequency of travel, but it is affected by perceived as consultants to the project: Christopher Love­ comfort. The longer the trip overall, the more neg­ lock, Harvard Business School; David Reibstein, ative the evaluation of cost and schedule aspects of Wharton School of Management, University of Pennsyl­ the train. Likewise, non-Northeast Corridor trains vania; and Alice Tybout, Kellogg Graduate School of are viewed in a slightly negative light. Senior Ma»ugement, Northwestern University. Pamela Labeaux citizens rate the cost and schedule aspects of train provided substantial input into the questionnaire travel more highly than do other groups. Not sur­ development, as did Kim Honetschlager and Robert prisingly, a larger number of pre-trip troubles and Hirshey into the data processing. Harrison Campbell greater access time promote a slightly more negative was the CRA officer in charge of the project. Data evaluation of train travel. Middle-income persons collection was handled by Leslie Adams, Marketing have a more positive evaluation than do lower- or Research Analyst, Amtrak. Robert Gall, Assistant n1gner-income persons. Of the travel style va