Research Projects

Principal Investigator

University of California, Davis Sustainable Transportation Center, Year 3.  Activities Conducted while Traveling: An Examination of their Impact on the Value of Travel Time Savings.  October 2008 - September 2009.

University of California, Davis Sustainable Transportation Center, Year 2.  The Impact of the Internet on Shopping Behavior: Modeling the Choice of Pre-purchase and Purchase Modes.  October 2007 - September 2009.

Fully understanding the potential transportation impacts of new (internet) and old (going to the store, catalog) shopping alternatives requires investigating the adoption of the various alter­na­tives.  By now, numerous studies have analyzed (intended or actual) purchase or pre-purchase (search) behavior, but we are not aware of any empirical studies considering the combined choices of pre-purchase and purchase modes.  Yet it is important to understand those choices not as separate and independent, but rather as interrelated. This study continues the substantive analysis of data collected from an original survey of shopping attitudes and behavior, in this phase investigating the combined choices of pre-purchase and purchase modes (primarily store and internet, but also catalog and other pre-purchase modes).  We propose to (1) cluster cases based on a variety of possible pre-pur­chase/purchase mode patterns; (2) conduct descriptive analyses of the resulting clusters, using the large variety of personality, attitudinal, situational, and sociodemographic variables available to us; and (3) develop multivariate probit and/or multidimensional nested logit models of pre-purchase and purchase mode choice combinations.  The findings will have important implica­tions for transpor­tation planning as well as for the retail industry.

University of California, Davis Sustainable Transportation Center, Year 1.  Structural Equation Modeling of Desired Travel Amounts.  August 2006 - July 2007.

The tenet that “travel is a derived demand” is fundamentally entrenched in transportation models, policies, and planning.  Yet there is considerable reason to believe that to some extent, travel is desired and demanded for its own sake – i.e. has an intrinsic, or autotelic, utility.  In prior years, we designed and implemented a survey to empirically explore the circumstances under which that is speculation is supported, obtaining data from some 1,300 commuters in the San Francisco Bay Area.  In a series of papers, we found that positive travel-related attitudes are associated with more travel, even after controlling for the standard demographic generators of demand as well as some land use characteristics.  In single-equation-at-a-time modeling, we also found complex relationships among Objective Mobility (the amounts one actually travels, in various mode-, purpose-, and distance-based categories), Subjective Mobility (one’s perceived amount of travel in the same categories, on an ordinal scale), Travel Liking (how much one likes traveling in those categories, on a five-point scale), and Relative Desired Mobility (how much one wants to travel in those categories compared to now, on a five-point scale).  Those results pointed to the need for structural equation modeling (SEM) to more properly capture the multiple directions of influence among those variables.

The SEM results reinforce some key findings of the earlier work.  In particular, a liking for travel leads both to traveling more, and to wanting to travel (even) more.  We also found evidence (especially for commute travel) that Travel Liking changes the relationship between Subjective Mobility (SM) and Relative Desired Mobility (RDM).  For those who dislike commuting, SM strongly negatively (standardized coefficient of -0.28) influences RDM (the greater one’s commute amount is perceived to be, the less one wants to increase it), whereas for those who like commuting, the impact of SM on RDM is far weaker (-0.07) – suggesting that many of those who commute a lot but like it are not strongly motivated to reduce it.

The practical implications of these and related findings are important.  There is a segment of travelers whose behavior defies conventional model assumptions, and who will be resistant to policies intended to motivate them to reduce travel.  It is important to take travelers’ attitudes toward travel, perceptions of their current travel, and desires with respect to their travel into account when developing and implementing policies and forecasts.

U. S. Department of Transportation Region Nine Transportation Center, Years 17 & 18.  Modeling the Adoption of Teleshopping.  August 2004 - July 2006.

Fully understanding the potential transportation impacts of new and old shopping alternatives requires investigating the adoption of the various alternatives.  This multi-year study proposes to design, administer, and analyze an original survey of shopping attitudes and behavior, leading to a model of shopping mode choice.  To reduce the heterogeneity of shopping behavior, we focus on one or two frequently-purchased product classes.  We define alternatives in terms of the dimen­sions of pre-purchase behavior (with store, catalog, and Internet modes) and transaction behavior (store, phone, mail, and Internet modes, distinguishing auction sites from conventional e-tailers).  Research questions include:  (1)  For the selected product class(es), what are the ad­van­tages and disadvantages of each shopping mode? (2)  Can market segments with different propensities to use alternative modes be identified? (3)  To what extent are there perceived to be viable alternative modes for a given shopping occasion?  (4)  Are the various shopping modes substitutes, or complements?  Offering the option of paper or web-based surveys, we plan to obtain about 2,000 responses.  The first year of the study is mostly devoted to survey design, data collection, and cleaning, with some preliminary descriptive analyses.  Future years will involve various multi­var­iate statistical analy­ses and multidimensional discrete choice modeling.

DaimlerChrysler.  Determinants of Travel Liking:  Who has a Positive Utility for Travel? September 2002 - August 2003.

We have identified three components of a positive utility for travel:  1. The utility of reaching the destination, which is the basis for the conventional view of travel as a derived demand; 2. The utility of activities that can be conducted while traveling (listening to music, transitioning between roles, making phone calls); and 3. The utility of travel itself (fulfilling needs for movement, variety, adventure, escape, status, satisfaction of curiosity).  To the extent that the second and third forms of utility are valued (and this will vary by mode, purpose, person, and circumstance), policies directed at reducing travel may not achieve the expected results.  We have obtained empirical evidence of the positive utility of travel from a sample of more than 1,900 San Francisco Bay Area residents.  For example, more than 3/4 of the sample reported sometimes or often traveling "just for the fun of it", and more than 2/3 disagreed that "the only good thing about traveling is arriving at your destination".

In a study of people's ideal commute times using the same data, we found that most people had an optimal time greater than zero (16 minutes on average), that a few (7%) people wanted to commute more than they currently were, that the actual commute time for a large minority of the sample (42%) was within 5 minutes of their ideal, and that a small majority (52%) wanted to reduce their commute times (although generally not to zero).  Higher values of attitudinal variables measuring the liking for commute travel and perceived benefits of commuting were associated with longer desired commute times, confirming the existence and value of positive attitudes toward commuting in influencing commute-related choices.

We built models of Objective Mobility (distance traveled in various categories), including as explanatory variables not only the conventional demographic characteristics, but also attitudinal, personality, and lifestyle measures – the first such “trip generation” models to do so, to our knowledge.  We found demographic variables (especially income) to be significant as usual, but also found the other variables to be highly important in every model except the one for commuting.  In particular, the adventure-seeker personality factor or the “excess travel indicator” appeared in 10 of the 11 models, with a positive influence on distance traveled.  For example, compared to those who are average adventure-seeking, those who are one standard deviation above average (all else equal) travel 13% farther for short-distance total; 26% farther for long-distance total; 21% farther for short-distance work-related purposes, 88% farther for long-distance work-related purposes, and so on.  Thus, rather than being purely mechanically derived from demographically-driven needs, the amount of travel demanded on that basis – even for mandatory purposes – can be stretched or shrunk by non-trivial amounts, depending on one’s travel-related attitudes.

We built models of Subjective Mobility (the ratings of various categories of travel on a 5-point semantic differential scale ranging from “none” to “a lot”) to analyze how people differ in the way they perceive the amounts of travel they do.  As expected, Objective Mobility – including measures of both trips and distance – heavily influences Subjective Mobility.  But attitudinal, personality, and lifestyle variables act to magnify or diminish the cognitive weight of the objective amount traveled.  We found that this weight can be magnified either because travel is burdensome or because it is enjoyable.  A number of cognitive mechanisms either magnifying or diminishing the effect of Objective Mobility on Subjective Mobility were identified.

We built ordinal probit models of Relative Desired Mobility (how much people want to travel compared to currently, on a 5-point scale ranging from “much less” to “much more”), as indicators of the dissonance between actual and preferred travel choices, and hence as potential indicators of latent demand if constraints were to be relaxed.  We found, as expected, that RDM is negatively related to Subjective Mobility (the more one thinks she travels, the less she wants to increase her travel), and positively related to Travel Liking (suggesting that a love of travel for its own sake engenders the desire to do more of it).  A number of recurring effects illustrated by other variables were identified, such as complementarity (a positive orientation toward one type of travel accompanies a desire to increase a related type of travel) and competing preferences (liking for one kind of travel accompanies a desire to reduce a different kind of travel, in order to increase the time available for the preferred form).

We studied the adoption and consideration of 17 travel-related strategies similar to those described in the “Responses to Congestion” project below.  We have related adoption and consideration to other explanatory variables, including travel attitudes and mobility indicators, and are analyzing binary logit models of consideration of strategies, both individually and bundled with related strategies.  We have found that those who travel a lot are more likely to consider not only travel-reducing strategies, but travel-maintaining/increasing ones as well, suggesting that such people seek to reduce their travel when they can, but make it more comfortable or productive when they cannot.

We then modeled Travel Liking.  We have found evidence for most of the hypothesized sources of an affinity for traveling, such as adventure- or variety-seeking, curiosity, escape, status, need for a buffer between activity types (e.g. home and work), desire for scenery or other amenities, exposure to the environment, synergy with other activities, independence, and a need for control.

We also explored the role of residential neighborhood dissonance on travel behavior.  We found that about 25% of residents in suburban neighborhoods preferred high-density environments, and a similar proportion in the urban neighborhood preferred lower-density environments.  With respect to commute mode choice, dissonant suburban residents were relatively similar to consonant suburban residents, whereas the auto mode share of dissonant urban residents was higher than that of consonant urban residents’, and nearly equal to that of consonant suburban residents.  The implication is that the built environment has a relatively stronger influence on commute mode choice in the suburbs (leading to auto use despite predispositions otherwise), whereas in urban neigh­bor­hoods people’s predispositions toward auto use still tend to be realized despite a less supportive built environment.

California Department of Transportation.  Alternative Fleet Scenarios.  July 2002 – June 2004. 

U. S. Department of Transportation Region Nine Transportation Center.  An Input/Output Analysis of the Re­la­tionships between Communications and Travel for Industry.  August 2002 - July 2003.

This study explores the aggregate relationships between transportation and communications as industrial inputs in the U.S., using input-output accounts produced by the U.S. Department of Commerce, and compares the results across time.  We analyzed Spearman correlations of transportation and communications inputs (both utilities and manufacturing) using ten direct coefficient tables from benchmark input-output years (1947-1997).  To properly represent the overall economy-wide relationship, each industry in the correlation was weighted according to the monetary value of its total output.  We found a pattern of predominant complementarity between transportation and communications manufacturing, and substitution between transportation and communications utilities.  There are intriguing indications, however, of a shift from substitution to complementarity in the latter case, beginning around 1987.  We believe this study offers a more informed view of the extent to which it is realistic to expect telecommunications to substitute for travel, at least in the industrial context, which constitutes a sizable proportion of the total demand for telecommunica­tions and transpor­tation.

U. S. Department of Transportation Region Nine Transportation Center.  Telecommuting over the Long Term:  Patterns of Engagement and Impacts on Residential Location.  August 2001 - July 2002.

This study analyzes 10-year retrospective data on telecommuting and residential/job location histories for a sample of 227 employees of the State of California (including non-telecommuting control group cases), many of whom have been telecommuting for at least five years.  We evaluated telecommuting frequencies and durations for the current and former telecommuters, and compared their one-way commute lengths and commute distance traveled to those of the non-telecommuters to ascertain whether the long-term net impact of telecommuting is to reduce or increase vehicle-miles traveled.  We found that although telecommuters live farther from work than non-telecommuters, they more than compensate for that by telecommuting frequently enough that their total (per capita) commute distance traveled is slightly less than that of non-telecommuters.  Further study investigated whether telecommuting is a cause or an effect of the longer-than-average one-way commute lengths.  The predominant role of telecommuting appears to be that of an effect (people first move for other reasons, and then telecommute), but in either case the net impact on travel appears to be benign.

California Energy Commission.  An Aggregate Evaluation of the Impact of Teleworking on Vehicle-Miles Traveled.  June - December 2001. 

U. S. Department of Transportation Region Nine Transportation Center.  The Impact of Attitudes toward Mobility, Adoption of Previous Strategies, and Demographic Characteristics on Responses to Congestion.  August 2000 - July 2001. 

Daimler-Benz.  Is there a Drive to Travel?  December 1998 - December 2000. 

Daimler-Benz.  Modeling Mobility Budgets.  December 1998 - September 1999. 

Bayerische Motoren Werke AG (BMW).  Long-Term Effects of Telecommuting on Travel Behavior and Residential Location.  August 1998 - March 1999.

Combined with later project.

Improvements in accessibility are increasingly suggested as strategies leading to a reduction in vehicular travel, congestion and their related impacts. This approach assumes that individuals, if offered an opportunity, are likely to reduce their travel. It also assumes that such land-use changes will increase non-motorized trips in lieu of automobile usage. However, there are numerous indications that people engage in excess travel and are not necessarily inclined to reduce it. The proposed study will test a number of hypotheses on the reasons for excess travel and the relationships among attitudes toward travel and responses to accessibility-enhancing strategies. It will investigate these relationships across different travel categories (work, maintenance and discretionary). The study assumes that different market segments exhibit different relationships and hence are likely to respond to policy measures in different ways. It is suggested that if a large segment of the population prefers mobility over the reduced travel offered by accessibility improvements, then such policies will be less effective than anticipated. The analysis will be based on data collected for this purpose in two or more communities in a metropolitan area of California.

Partners for Advanced Transit and Highways (PATH). Beyond Telecommuting: The Travel/Communications Impacts of Advanced Telecommunications Services. September 1996 - July 1997.

The travel impacts of telecommunications applications such as telecommuting have been the subject of several studies to date; other applications have received considerably less attention. To our knowledge there are no other studies of the travel impacts of recently-developed applications such as community networks and other on-line information- and transaction-oriented services - the subject of this proposal. These services can provide information about activity opportunities as well as about the transportation system, and as such may generate or modify travel as well as replace it. It is important to learn much more than we now know about how the activity and transportation information provided by these services will affect demand, with its consequent effects on network performance. Thus, the proposed project represents a valuable extension of existing ITS research into the impacts of transportation system provision (such as real-time congestion information or carpool formation opportunities) on the demand for travel.

This project will examine the behavioral impacts of telecommunications-based information/ transaction services. Part of the project continues the analysis of data collected under the Davis Community Network (DCN) Project funded by Caltrans. These unique data were collected from multiple, complementary instruments, including before and after logs of communications activities over a four-consecutive-day period, an activity diary which records information on likely impacts of specific uses of the DCN system, system utilization statistics, and a background demographic survey. In addition, the project will gather information on existing and planned traveler information systems, and examine policy issues relevant to the questions of adoption and travel impacts of such systems.

This exploratory study will lead to refinements in our conceptual and methodological approach to understanding the impacts of similar telecommunications systems. Further, it will inform the policy question of whether (and if so, how) such systems should be promoted as congestion management measures.

Researchers: Ravikumar Meenakshisundaram and Ilan Salomon

Partners for Advanced Transit and Highways (PATH). Is there a Case for Public Investment in Telecommuting? The Cost/Benefit Analysis. (Co-PI with Debbie Niemeier). September 1996 - July 1997.

While much has been learned to date about the specific transportation impacts of small-scale telecommuting projects, important questions about the degree to which telecommuting is a cost-effective transportation policy remain unanswered. The proposed project would redress that deficiency by conducting a formal economic evaluation of telecommuting and assessing the resulting transportation policy implications. The study will assess the costs and benefits of home-based telecommuting, develop a suitable framework for the future economic analysis of telecommuting, and offer policy guidance on the conditions under which benefits may be optimized. The information provided by such a study is fundamental to proper placement of telecommuting on the public policy agenda, and has been notably lacking from telecommuting research and policy discussion to date. The methodology developed for this project will also be highly relevant to the assessment of the cost-effectiveness of other ITS strategies.

The proposed research involves first, a review of relevant cost-benefit and telecommuting literature. This review will identify the important components of costs and benefits, and the pros and cons of different economic analysis methodologies. Specific scenarios of home-based telecommuting adoption will be developed, from both public- and private-sector perspectives. A base case of background conditions (such as traffic levels) and their evolution over time will be developed. The costs and benefits of each specific telecommuting scenario - to the public sector, the employer, and the individual - will be measured. Primary and secondary data will be used to estimate the specific costs and benefits of the selected scenarios over the project life. This project will explore a novel methodological approach to the measurement and valuation of qualitative benefits (particularly to the individual), through the use of attitudinal factor scores and logit models of choice developed in a previous study. Alternative evaluation techniques such as net present value and benefit/cost ratio will be analyzed and the most appropriate one(s) selected. Sensitivity analysis will be conducted on key inputs and assumptions. The policy implications of the empirical findings will be examined, and recommendations made.

Researchers: Kevan Shafizadeh and Ilan Salomon

U. S. Department of Transportation Region Nine Transportation Center. Behavioral Adjustments to Congestion. August 1995 - July 1996.

This study evaluated commuters' responses to congestion, which may differ in time frame, generalized cost of implementation, and distributional impact.  Using a data set pre­viously collected from a sample of 628 people, this study em­pir­i­cally tested the hypotheses that responses are a function of previously adopted adjust­ments, and that the effectiveness of congestion-reduction policies is distributed differently across various socio-economic segments.  We have identified three tiers of responses, ranging from travel-maintaining responses through travel-reducing strategies to changes in location and life style adjustments (which may also reduce travel).  We have empiri­cally supported the hypothesis that people tend to try the less costly measures first, and if dissatisfaction persists, then proceed to try more costly measures.  In examining distri­bu­tional effects, we have found that the adoption of most types of strategies, especially the more costly ones, appears to fall disproportionately to women.  Additional differences were identified by family status, income level, employment status, and household type.

Further work has involved developing models of consideration behavior.  Among other results, these models confirm the expectation that individuals consider the strategies studied for many reasons other than congestion relief.  One implication is that policies designed to change transportation behavior may be less powerful than expected, because reactions are filtered through a variety of other motivations and constraints.  An improved understanding of the response to these policies must acknowledge and incorporate the complexity of the choice situation facing the typical individual in modern society.

Researchers: Liz Raney and Ilan Salomon

U. S. Department of Transportation Region Nine Transportation Center. An Empirical Analysis of the Travel Behavior of Telecommuters and Home-Based Business Owners. August 1994 - July 1995.

Two existing data sets are used to compare travel behavior across four groups:  home-based telecommuters, home-based business owners, center-based telecom­muters, and conventional workers.  Groups 1, 2, and 4 are analyzed using the 1991 Cal­trans Statewide Travel data base, and Groups 1, 3, and 4 are analyzed (for emissions as well as travel) using the Puget Sound Telecommuting demonstration project data.  The trans­portation behavior of Groups 2 and 3 has been little-studied to date.  A key finding from the Puget Sound data is that the VMT reduction of center-based telecommuters on their telecommuting days is comparable to that of home-based telecommuters, but there is no significant change in their number of trips.  A key finding from the Caltrans data is that total trip rates for home-based telecommuters are statistically equivalent to those of conventional workers (in marked contrast to previous specialized-sample studies).  However, differences between the two groups in drive alone trip rates are similar to those of other studies, with home-based telecommuters making 0.6 (18%) fewer (telecom­muting-day) trips than conventional workers.

Researcher: Dennis Henderson
Publications: TRP21, TRP22

Washington State Energy Office. Puget Sound Telecommuting Demonstration Transportation Emissions Analysis. June 1994.

An analysis of the emissions impacts of telecommuting was conducted using travel diary data from telecommuters and non-telecommuters in the Puget Sound area. Characteristics of individual vehicles and trips and region-specific temperature data were used as input to the California Air Resources Board's EMFAC/BURDEN 7F models. This and a similar study described below represent novel applications using EMFAC and BURDEN with travel diary data to evaluate a specific transportation control measure (as opposed to the more conventional use of these models to estimate regional emission inventories).

Researchers: Dennis Henderson and Brett Koenig
Publications: TRP19, R9

U. S. Department of Transportation Region Nine Transportation Center. Market Segmentation of the Demand for Telecommuting. August 1993 - July 1994.

This is a continuation of the study titled "Modeling the Choice to Telecommute", described below. Here, we focused on potential market segments.  One basis for segmentation is the type of telecommuting involved -- home-based or center-based.  We developed logit models of the preference among "home", "center", "either", or "neither" alternatives.  In another direction of segmentation, we examined the differences in perceived advantages and disadvantages of telecommuting by gender and occupation.  Finally, we studied variations in response rate and content across two state agencies in different metropolitan areas.

Researcher: Michael Bagley
Publications: TRP23, R10

California Energy Commission. Present Status and Future Directions of Telecommuting in California. November 1992 - September 1993.

This study provided input to the CEC's forecasts of future statewide energy consumption. The current (1991) aggregate level of telecommuting in California was estimated from a variety of national, state, and regional sources. Transportation-related impacts of the current level of telecommuting were estimated through a synthesis of a number of published empirical studies. Future levels of telecommuting and its transportation impacts were estimated under the assumed continuation of present trends. Factors and policies likely to affect future levels of telecommuting were described, and suggestions for incorporating telecommuting into regional and statewide transportation and energy models were made.

Researcher: Dr. Susan Handy
Publications: TRP11, TRP12, TRP17, TRP18, R3, R4, R5, R6

California Department of Transportation (Caltrans). Residential Area-Based Telecommuting Work Centers. July 1992 - January 1997.

A number of studies have now demonstrated the transportation-related benefits of home-based telecommuting. But for a variety of reasons, telecommuting centers may be a preferred form of telecommuting for some employees or employers. It is important to analyze the effectiveness of telecommuting centers both as a travel reduction strategy and as an alternative work arrangement. This project involves opening and evaluating 12 telecommuting centers in metropolitan areas of the state. The centers are located in or near residential areas so as to maximize the potential for eliminating vehicle commute trips. Before and after travel and attitudinal information is being collected from center-based telecommuters, and from home-based and non-telecommuting control groups. Before and after attitudinal information is being collected from managers of the employees in each of these groups. Data on center description, operation, funding, marketing, and occupancy are also being collected.

Administrative Staff: Carol Buckinger, Michelle Derr, and Francisca Mar
Researchers: Prashant Balepur, Sally Ho, David Stanek, and Krishna Varma
Publications: TRP25, R8, R11

California Department of Transportation (Caltrans). Davis Community Network Demonstration Project. July 1992 - June 1995.

This project is intended to evaluate the transportation and communications impacts of providing advanced telecommunications services to homes and businesses in Davis. These services could result in (1) the substitution of trips or other communications (e.g. the replacement of paper mail with electronic mail), (2) generation of new trips or communications, and/or (3) modification of existing travel or communications (e.g. changing mode in response to the provision of real-time transit system information). Original data collection instruments have been developed, and before and after data are being collected to quantitatively and qualitatively assess the degree to which each of these outcomes occurs.

Researchers: Dr. Prasuna Reddy, Prashant Balepur, and Krishna Varma (formerly Dr. Susan Handy, Michael Bagley, and David Stanek)
Publication: R7

U. S. Department of Transportation Region Nine Transportation Center. Modeling the Choice to Telecommute. August 1991 - July 1992; August 1992 - July 1993.

This project was designed to achieve greater understanding of an individual's choice to telecommute or not, to provide a behavioral foundation for forecasts of future levels of telecommuting. A conceptual model of the individual's decision process was developed, involving the identification of constraints or facilitators that (respectively) inhibit or support telecommuting, and drives or motivations to telecommute. Most people currently do not have the choice to telecommute, due to one or more constraints such as management unwillingness or job unsuitability. However, removal of constraints is only a necessary but not sufficient condition for telecommuting to be adopted; it is also necessary that one or more drives be active. A questionnaire was designed to obtain data on the identified drives and constraints, and was administered to a sample of more than 800 people in three organizations. These data permit analysis of the extent to which various constraints are active, and the development of quantitativ e models of the preference for and choice of telecommuting.

Researchers: Prof. Ilan Salomon and Michael Bagley (formerly Laura Laidet and Jill Mannering)
Publications: TRP8, TRP13, TRP14, TRP15, TRP16, TRP20, NRP9

Co-Principal Investigator or Collaborator

This research investigates the self-selection issue in the most sophisticated way seen to date, using survey data col­lected from about 1,600 residents of eight different neighborhoods in northern California.  We modeled a number of different measures of travel behavior (driving, walking, biking, transit, auto ownership), and dealt with self-selection in several ways:

·       by incor­pora­ting quanti­tative measures of perceptions and preferences using factor analysis of attitudinal measures,

·       by using structural equations modeling (i.e. estimating the parameters of several interconnected equations simultaneously) to account for multiple causal relationships, and

·       by accounting for the dynamic nature of the effects through the use of quasi-longitudinal data (retrospectively assessed changes in certain key variables).

In general, the results repeatedly indicate that the built environ­ment does exert a separate influence of its own, after attitudes and preferences are accounted for.  The relative magnitude of that influence appears to be non-negligible, although an exact quantification is elusive.

National Science Foundation.  Decentralized Decision-Making in Complex Network Systems. 2000-2003 (in collaboration with Prof. Anna Nagurney of University of Massachusetts, Amherst, and Prof. June Dong of State University of New York, Oswego). 

A number of studies have shown an associative relationship between land use and travel behavior. Specifically, low-density single-purpose land uses and low levels of transit service tend to be associated with greater numbers of vehicle-trips and more vehicle-miles traveled. It is tempting then to draw the conclusion that a land use policy promoting greater densities, mixed uses, and higher transit levels of service will be an effective strategy in reducing the demand for vehicular travel. However, evidence of the direction of causality, or even the existence of causality at all, has not been obtained. For example, do people make fewer trips because they live in higher-density neighborhoods, or do they live in higher-density areas because they prefer to make fewer trips? If the latter situation is the case, then putting a different type of person into a high-density development may not have the desired effect on travel. This study makes a contribution to this complex issue by examining the role of attitu des in travel behavior. Travel diary, attitudinal/lifestyle, land use, and socioeconomic data were collected for respondents in five diverse neighborhoods in the San Francisco Bay Area. Regression models of trip-making behavior (total number of trips, number of auto trips, and so on) were developed with socioeconomic, land use, and attitudinal explanatory variables. The attitudinal variables (including orientations toward various residential and travel lifestyles) contributed significant explanatory power to the models, suggesting that they have a stronger and more direct link to travel behavior than objective characteristics alone.

Researchers: Carol Buckinger, Fred Gianelli, Laura Laidet, and Francisca Mar
Publication: ORP3

U. S. Department of Transportation Region Nine Transportation Center. Impact of Telecommuting on Travel: Accessibility Implication of Working at Home. August 1990 - July 1991 (co-PI with R. Kitamura, P. P. Jovanis, and D. S. Bunch).

This project focused on changes in non-work travel-related behavior due to telecommuting. The travel diary data from the State of California Telecommuting Demonstration Project were geo-coded (latitudes and longitudes of recorded locations were identified and appended to the file) and used to analyze changes in the spatial distribution of activity locations. Models of time spent in-home versus out-of-home were also developed.

Researchers: Kostas Goulias, Ram Pendyala, Srikanth Sampath, and Somitra Saxena.
Publications: TRP24, NRP4

University of California Energy Research Group. Telecommuting, Energy and Air Quality. July 1990 - June 1991 (co-PI with R. Kitamura and P. P. Jovanis).

The energy and air quality impacts of telecommuting were evaluated for the State of California Telecommuting Demonstration Project. Travel diary data were collected before and after telecommuting began, from telecommuters, a non-telecommuting comparison control group, and driving-age household members of both groups. From these "person" diaries, "vehicle" diaries were prepared that tracked movement of individual household vehicles across the three-day travel diary period. Only in this way could a specific vehicle trip be confidently identified as a hot start or a cold start. Sample-specific data on vehicle types and trip-making activity were developed and input to the California Air Resources Board's EMFAC/BURDEN 7E models to quantify emission levels associated with the travel behavior of the sample. This constitutes the first known integration of travel diary data input to emissions inventory models in order to evaluate the impact of a specific transportation control measure.

Researchers: Srikanth Sampath and Somitra Saxena
Publication: TRP5