Urban Design, Telecommunication and Travel Forecasting Conference: Summary, Recommendations and Compendium of Papers
Travel Impacts of Urban Form: Implications from an Analysis of Two Seattle Area Travel Diaries
G. Scott Rutherford
Professor of Civil Engineering
Edward
McCormack
Ph.D. Candidate-Geography
Martina Wilkinson
Graduate
Student-Civil Engineering
University of Washington
Department of
Civil Engineering
Box 352700
Seattle, WA 98195-2700
Introduction
Purpose of Research
Over the past 30 years, a notable change in land use has been the growth of residentially oriented suburban neighborhoods located some distance from employment and service centers. Linked with this growth are increasing levels of traffic congestion, air pollution, and general disenchantment with suburban life (Downs 1992; Langdon 1994). These negative impacts have focused on the potential transportation benefits of traditional oriented neighborhoods characterized by more diverse land use development patterns (Bookout 1992a, 1992b). Developers and planners have suggested that mixing land uses can reduce automobile dependency by making more goods and services available within walking and short driving distances. The new interest in mixed land use represents an about-face with regard to the basic assumptions that have shaped urban development patterns over the past 20 or 30 years.
While interest in mixed-use development is on the rise, only a handful of studies have explored the transportation implications of this type of development empirically. Existing studies typically contain only general information on the demographic characteristics and travel patterns of inhabitants of mixed-use areas. This research seeks to address at least part of this gap in the literature. The researchers used a two-day travel diary and demographic survey of 900 households in three greater Seattle area neighborhoods characterized by two or more distinct land uses. This detailed data set was then compared with similar household travel data collected by the Puget Sound Regional Council (PSRC). Both data sets used similar survey forms and were collected and coded by the same contractor. The data were compared to see whether the travel behavior of residents of mixed-use neighborhoods differed significantly from the travel behavior of residents in neighborhoods with more homogenous land use patterns.
One goal of the research described herein was to explore whether people in neighborhoods that provide goods and services travel less than people in other more homogenous neighborhoods. This study, unlike previous research on mixed-use, approached this question using detailed, empirical travel data collected specifically to explore the travel characteristics of mixed-use neighborhood residents. The travel data were designed to be compatible with similarly detailed regional level travel survey data from the PSRC. This effort resulted in the ability to compare and contrast the travel patterns of mixed-use neighborhoods to other areas.
In addition to the above comparisons, a second goal of this research was to explore the nature of weekend travel in the mixed-use neighborhoods through a variety of measures.
Analysis
This research had two major elements, the first of which was a county comparison of weekday travel. Insofar as the mixed-use travel survey was designed to be compatible with the PSRC county level survey (PSRC's survey was for four counties, this research only needed King County), it was possible to explore the differences in travel characteristics between non-mixed-use areas and mixed-use areas by comparing the PSRC's county level data with the neighborhood data.
The second element of this research was a descriptive examination of weekend travel. The mixed-use data used for the weekday research also collected travel information for weekends. Because weekend travel is little studied, the mixed-use survey results provided a welcome opportunity to consider this travel behavior separately.
Both elements of this research considered the following categories of analysis: travel times and distances, demographics, multi-purpose trips, and intra-neighborhood analysis.
Travel Times and Distances. The use of transportation modeling output, Geographic Information Systems (GIS) software (TransCAD), and a U.S. Census-derived computer file of the county street network allowed for calculation of a number of important spatially-oriented travel statistics. Most relevant was the ability to estimate respondents' trip mileage in the mixed-use neighborhoods, both individually and as households, from original survey data by estimating travel routes on the street network. The travel distance procedures also provided the ability to accurately calculate short trip distances. Travel times were reported directly in both data sets and provide valuable information.
Demographics. Household and individual demographic characteristics were compiled to identify possible correlation with observed travel patterns in each of the study neighborhoods.
Multi-Purpose Trips. Many people schedule their activities by combining several trips into a single, sustained journey or chain. Analyzing the number, length, and type of chains, as well as the characteristics of the trip maker, sheds light on how travel is organized for efficiency, especially around work trips.
Intra-Neighborhood Analysis. Each of the three mixed-use neighborhoods encompassed concentrations of retail and other service establishments. This research examined the travel patterns of households at various distances such concentrations, and explored the extent to which proximity to commercial outlets and services affected the mode choice other than the auto -- and in particular, whether walk trips replaced vehicle trips for short-distance travel.
Analysis Limitations
This analysis is limited in several respects. Since the research is based on survey data, it is possible that some variation in travel behavior is attributable to self-selection by certain types of individuals in different neighborhoods. It is difficult to measure this type of bias.
The research also compares the PSRC panel survey data with the mixed-use neighborhood data. While the research designs for the projects were similar, the two surveys were conducted two years apart, increasing the possibility of some incompatibility between the data sets. Additional incompatibility may result because the PSRC data were collected September through December, while the mixed-use data were collected only in November and December. As a result, the mixed-use data could be biased toward shopping trips because of the increase of retail activity during the Christmas season. The two data collection efforts also used slightly different sampling procedures, and the mixed-use survey form was more comprehensive, resulting in other possible limitations when comparing across data sets.
The PSRC panel survey instructions requested that the participants report all trips five minutes or longer. Since pedestrian, bicycle and short vehicle trips are important in studying mixed-use neighborhoods, the mixed-use data instruction requested that all trips be included. In spite of this difference in instruction, both data sets include similar percentages of trips less than five minutes long. The mixed-use data included 7.0 percent of the trips less than five minutes, while the PSRC data set had 6.5 percent of all trips. However, for the most accuracy, when possible, any comparison between the two data sets removed all trips under five minutes in length.
The PSRC panel data and the mixed-use data both constituted two-day travel diaries. Naturally, travel on one of the two days is not independent of the other; nor, for that that matter, are trips within a single day independent for a given person or between people. While this may cause problems for some types of analysis and for developing travel models, this study is merely comparing similar households across various types of geographic areas.
Report Structure
The remainder of this paper is divided into six sections.
Literature Review. The research literature concerning empirically oriented analyses of neighborhoods with mixed-use characteristics is reviewed, identifying the scarcity of quantitative analyses of neighborhood travel behavior, particularly as they relate to land use. The few empirical studies on weekend travel are also covered.
The Data Sets. This section discusses the two data sets used for the study. First, a brief review of the data collection methodology is given. Since the subsequent analysis of the data set requires specific knowledge of trip locations (origin and destination), the process by which those locations were derived from the survey responses and coded (i.e., the geocoding process) is discussed. The PSRC data set used for regional comparisons is also discussed. This discussion concludes with a comparison of the mixed-use and PSRC data sets; several key differences between the data sets are highlighted.
Research Methods. This chapter reviews the techniques and issues associated with processing and preparing the data for computer analysis.
Data Analysis -- Weekdays. The mixed-use data and the panel survey data are analyzed statistically and spatially. The major findings from the comparison of these data sets are then discussed and compared with those of other studies.
Data Analysis -- Weekends Travel in the mixed use neighborhoods for weekends is analyzed and numerous descriptive statistic regarding weekend travel are presented. General travel characteristics, day and time variations, as well as a separate look at walking trips provide insight into the weekend travel patterns of the mixed land use neighborhood residents.
Summary and Future Research. The results of this research are summarized and the conclusions are presented.
Literature Review: Travel and Mixed-Use Neighborhoods, Weekdays and Weekends
General Travel
The Nationwide Personal Transportation Survey (NPTS) is a randomly sampled telephone survey collected every few years designed to provide a comprehensive look at personal travel in the United States. The 1990 survey and the three earlier surveys provide data (although no origin and destination data) useful in examining the relationship among demographic, land use, and transportation changes. The following three summaries are from studies that utilize these data.
According to Comsis (1994) vehicle miles of travel (VMT) increased nationwide by 37 percent between 1983 and 1990, although the population only increased by 4 percent. The report indicated that higher residential and employment densities can promote less reliance on private vehicle trip making: "Persons residing inside the central cities of urbanized areas make more shorter trips than persons living outside central cities."
Pisarski (1992) used the weekday data from the Nationwide Personal Transportation Survey (NPTS) and found that "the geographic distribution of population is far more crucial than population growth in creating dramatic changes in travel in individual locations." He indicated that one of the most significant factors in trip growth is the population shift to large metropolitan areas, and subsequently to these areas' suburbs.
Gordon and Richardson (1994) published another NPTS-based study. They sought to explain changes in work trip length and determined that although trip lengths have increased, so have travel speeds. Their findings support the view that suburbanization allows people to live farther from activity centers at a modest marginal cost in terms of extra time traveled, due to higher speeds. In contrast to almost every other researcher, they conclude that "urban sprawl is a transportation solution, not a problem."
Trip Characteristics and Travel Patterns of Suburban Residents by Prevedouros and Schofer (1991) analyzed weekday travel behavior based on a 1989 mail-back survey of individuals residing in selected Chicago suburbs. One of four factors analyzed were two classes of suburbs: inner-ring, high density, stable suburbs; and outer-ring, low-density, growing suburbs. Key general findings indicated that residence location in outer-ring suburbs implies longer trips and more frequent local trips. Although the average travel speed by automobile is higher for residents of growing suburbs, they still stay in traffic 25 percent longer and have a 40 percent longer total daily distance compared with stable-suburb residents.
In a more general study, the Puget Sound Council of Governments (PSCOG) published a report in 1990 on household travel surveys from the counties in the Seattle metropolitan area. One- to three-day travel diaries were collected from 4,500 households between 1985 and 1988. The survey's purpose was to update earlier survey research for use in travel demand forecasting and planning. Results indicated that while household size is decreasing, the smaller households have more vehicles. The surveys confirmed that trip making per person and per household have increased substantially (Table 1) and that nearly 90 percent of all trips are made by private vehicle. Average vehicle occupancy in the region declined from 1.25 persons per vehicle in 1961 to 1.1 in 1987.
Table 1. Person Trips in Puget Sound Region
| Year | Average Number of Person Trips per Person in Puget Sound Region |
|---|---|
| 1971 | 2.6 |
| 1987 | 4.3 |
| Source: PSCOG 1990 | |
Summary of General Travel
Empirically based travel studies generally indicate that travel is
increasing, and that residents outside the central city travel longer and
farther, although at higher speeds, than inner ring residents (Table
2).
Table 2. Summary of General Travel Studies
| Author | Key Findings |
|---|---|
| Comsis | Travel increasing due to increased trip frequency and length higher densities and living inside central cities promote shorter trips |
| Pisarski | Geographic distribution of population is more important than population growth in travel patterns changes. |
| Gordon et al. | General travel trends indicate longer trip lengths but also higher travel speeds |
| Prevedouros et al. | Outer ring growing suburb residents make 40 percent longer trips (but have higher travel speeds), spend 25 percent more time in the car, and make more frequent local trips than residents from inner ring stable suburbs. |
| PSCOG | Household size and vehicle occupancy decreasing, while there are more vehicles per households, and more trips per person. |
Travel in Mixed-Use Neighborhoods
After studying southern California, Giuliano (1995)contended
that the connection between land use and transportation is negligible
because urban areas in the U.S. are already so accessible, because
settlement patterns are well-established, and because privacy is so
important to most people. As such, transportation plays an ever-decreasing
role in the locational decisions of households and businesses. Her essay
implies that the land-use transportation connection is too weak to provide
of public policy direction.
Cevero and Landis (1995) rebutted Giuliano's article. Although they agreed that the connection is much weaker today than a century ago, they argued that the relationship remains important. In support of this view, they cited studies showing how land prices have gone up around new transit stations and commute trips that tend to be shorter for those living in areas with balanced housing and jobs. They conclude that land use can be an important contributor to transportation trends and vice versa. The authors expressed belief that, in the land use-transportation connection, considerable elasticity remains.
Another study to examine the land use connection to transportation was Frank and Pivo (1994), the first in a series of projects seeking to identify which land-use patterns reduce auto use. The authors studied 1989 travel in the greater Seattle/Tacoma region and found that commute distances and times tended to be shorter for those inhabitants of balanced areas. More specifically, the average length of work trips ending in a balanced census tract was 29 percent shorter than work trips that end in unbalanced areas (Table 3).
Table 3. Work Trip Length in Puget Sound Area
| Balance of Census Tract where Work Trip Ends (jobs to household ratios) | Distance of Work Trip (miles) |
|---|---|
| Balanced area (ratio = 0.8 -- 1.2) | 6.9 |
| Unbalanced area (ratio < 0.8 or > 1.2 ) | 9.6 |
| Source: Frank and Pivo 1994 | |
In a later study, Pivo et al. (1995) examined the market for less auto dependent land use by studying 1970 - 1990 data on the population density, housing density, employment density, jobs-housing balance, and retail-housing balance of both metropolitan cities and unincorporated areas in Washington state. Through examination and comparison of statistical distributions, relationships between land use variables were found, and associations between both density and balance and less auto use were confirmed. The report recommended promoting greater density and balance to communities whose land-use patterns are capable of supporting greater transit use and less out-commuting.
In 1992, a series of articles by Bookout in Urban Land explored neotraditional development. In the first article (Bookout 1992a) the author argued that the 45 percent of the population that moved to the suburbs after World War II never really realized the American dream due to traffic snarls, inadequate social services, etc. One of the major flaws with the suburban vision is excessive travel needs brought on by low-density development. The author recommended neotraditional communities, with (among other things) more through streets instead of cul-de-sacs, to give drivers alternate routes between points, which may result in shorter and less congested travel.
A second article, "Cars, Pedestrians, and Transit" (Bookout 1992b), asked whether people must continue to drive between each and every one of the places they visit regularly. The author suggested that with the building of more neotraditional communities, the answer should be "no." He advocated three items to reduce either the number of vehicle trips or trip distances:
- A return to the grid pattern for streets, or at least an effort to provide more direct connections between any two points within a community.
- Communities that are pedestrian and bicycle friendly.
- Increased transit viability.
Bookout cited Kulash's study (1990) to substantiate his recommendations. Kulash's study used simulation modeling to compare the traffic patterns of developments with densely gridded streets (called "traditional neighborhoods" but referred to as "neotraditional" by most authors and in the remainder of the paper) to communities with partially connected streets and cul-de-sacs (called "conventional suburban developments"). The author analyzed the travel performance of the theoretical developments and found that a traditional neighborhood design could produce fewer total vehicles miles traveled than a comparable conventional suburb (although much higher travel on local streets) (Table 4). Traditional neighborhoods did have lower travel speeds, but trips were also shorter. He concluded that traditional street networks function more efficiently than do conventional networks. However, this study did not measure trips beginning or ending outside the community. Nor did he indicate whether a traditional development would actually generate fewer trips than a conventional development.
McNally and Ryan (1993) used modeling to explore potential transportation benefits of neotraditional neighborhood design. They compared the traffic performance of a conventional suburb (with a hierarchical street network) to that of a neotraditional community (with highly connected gridded streets). All aspects of the theoretical neighborhood including land use, were held constant except for the actual configuration of the networks. The models indicated 10 percent fewer vehicle-kilometers traveled in the neotraditional network for the same level of trip generation. Total vehicle-hours traveled in the neotraditional network were reduced by 27 percent and the average trip lengths were 15 percent shorter than in the conventional network. The authors concluded that with the same level of activity, neotraditional networks operate more effectively, experience less congestion, and require less travel than conventional networks. They also indicated the drivers in neotraditional networks may choose more direct routes.
Table 4. Vehicle Miles of Travel in Theoretical Communities
| Vehicle Miles of Travel (internal travel only) | Difference Between Traditional Development (TND) and Conventional Suburb (CSD) |
|---|---|
| Arterial Streets | TND is 25% of CSD |
| Collector Streets | TND is 15 % of CSD |
| Local Streets | TND is 400% of CSD |
| Total Vehicle Miles Traveled | TND is 57% of CSD |
| Source: Kulash 1990 | |
Friedman, Gordon, and Peers (1992) compared 1980 travel data from traditional communities to suburban tract developments in the San Francisco bay area in order to investigate any differences in trip generation and mode choice. The older communities developed prior to World War II had gridded street networks and were characterized by a mixture of residential and non-residential uses. In contrast, the suburban developments tended to contain many cul-de-sacs and segregated land use, and a hierarchical roadway. In order to control for income differences, the wealthiest and poorest households in each neighborhood were eliminated from the study. The results, which provide a basis for measurement of the potential impacts of different land use patterns, showed that suburban areas generated 23 percent more trips, had higher drive alone rates (68 percent in suburban neighborhoods versus 49 percent in traditional communities), and had half the transit share of traditional communities. The authors concluded that traditional neighborhoods have characteristics that result in fewer automobile trips than do newer suburban developments.
Although the Friedman et al. study is widely cited as "proof" of mixed-use neighborhood transportation advantages, others counter that it is impossible to separate out the relative importance of the many differences between suburban and traditional communities. Crane (1996), for example, praised neotraditional town planning for its thoughtful and functional design, but he questioned its actual transportation benefits. He pointed out that transportation problems may, in fact, worsen -- while it is likely that many elements of the new designs discourage driving for some kinds of trips, the aggregate effect is uncertain. Here's why:
The rectilinear grid street pattern is the easiest transportation feature to implement in neotraditional town planning and is widely encourage by many observers (e.g., Kulash, McNally, and Ryan). However, these authors assumed that trip frequencies are fixed; they never analyzed the potential change in demand for trips due to the new street pattern. Crane agreed that the grid pattern creates more access and thus shorter average trips than a cul-de-sac pattern. However, he countered that the increased access reduces the cost of travel, thus encouraging people to take more trips. He concluded that a change in street configuration may or may not reduce auto travel, that the transportation benefits of neotraditional designs have been oversold, and that each development must be evaluated on a case-by-case basis to predict the net use.
Another study skeptical of the supposed transportation advantages of mixed land use is Kitamura, Mokhtarian and Laidet (1994). This study used travel diaries and attitude surveys to explore travel behavior in five diverse neighborhoods in the San Francisco area. Initially they showed that neighborhood characteristics were significantly related to travel behavior. Measures associated with lower rates of travel included higher residential density and more mixed land use. The next step of the study attempted to demonstrate that attitude, not land use, was the primary determinant of travel behavior. By showing that attitudes were more strongly correlated to travel behavior than neighborhood characteristics, the authors argued that the land-use - travel relationship was an artifact of an association between land-use and variety of social and demographic characteristics associated with travel. They suggested that land-use determined attitudes; higher density, for example, means smaller houses, lower incomes and other factors that affected one's attitude. Attitude in turn influenced travel behavior. The authors concluded that land-use policies promoting high densities and more mixed use may not influence travel behavior unless residents' attitudes are also changed.
A study centered around household based trip statistics is by Holtzclaw (1991), who studied data from several types of communities with varying densities and land use mixes in the San Francisco region. Odometer readings and trip logs were used to determine reduction in automobile mileage due to higher residential density, neighborhood businesses, and improved transit service. The conclusion is that as the housing, population, and commercial densities decrease, and the transit service decreases, the vehicles miles traveled (VMT) per capita and per household increase. Doubling residential or population density reduces the annual auto mileage per capita or per household by 20 to 30 percent.
A detailed travel survey study was documented by Ewing et al. in 1994. Six communities in Palm Beach County, Florida, were chosen for study based on their diverse development. Household travel data including trip frequency, mode choice, trip chaining, trip length, and overall vehicular travel were used to study the relationship between household travel, location and land use. The researchers concluded that households in the "sprawling" non-gridded suburban community (composed mainly of single-family homes) had almost 66 percent more vehicle-hours than did a traditional gridded community with varied land use. Other communities fell between the extremes. The authors concluded that higher density, mixed land use, and central location tended to be associated with reduced vehicle-hours of travel.
An article that encouraged further research was by Steiner (1994). This study documented literature on residential density and travel patterns. The author concluded that decreased usage of the automobile is possible in higher-density residential areas because of several factors:
- High density puts destinations close together, making it possible to walk.
- The greater number of people in an areas, the more an area is perceived to be safe for walking.
- Certain types of people and households may be more likely to live in high density residential areas.
However, like Crane, Steiner cautioned that often assumptions are made concerning the relationship between high-density neighborhoods and the residents' transportation choices which may or may not be true. She indicated that many studies have not separated out factors such as income, household size, life-cycle characteristics, etc., which also affect transportation choices. Steiner advocated further research to sort out the importance of the pattern of travel based on the above characteristics; only then can conclusions be drawn on which households might be willing to live in high-density areas and the extent to which changes in land-use patterns reduces travel.
Handy in a 1991 article summarized the issues surrounding the concept of travel in mixed-use neighborhoods. Proponents claim fewer and shorter auto trips, more walking trips, and a greater sense of community in these developments. Yet critics and skeptics indicate these claims are not proven, that people may not want to live in these neighborhoods, and that the entire concept is simply not feasible. The author articulated a need to answer the underlying question of how neotraditional developments will relate to the larger settlement patterns. She concluded that the entire debate over the neotraditional issue "is greatly in need of substantive arguments, of testing and exploration of issues at a much greater depth than has occurred to date."
Summary of Travel in Mixed Use Neighborhoods
All of the studies in the above section detail travel in mixed-use
neighborhoods. Table 5 summarizes the key findings for
each study and identifies relevant items to the research in this paper.
Table 5. Summary of Travel Studies on Mixed Use Neighborhoods
| Author | Key Findings | Relevant Notes |
|---|---|---|
| Giuliano | Land use -- transportation connection too weak to matter in terms of public policy | Based primarily on commuting research |
| Cevero and Landis | Land use can be important contributor to transportation trends. | Includes research on rail transit |
| Frank and Pivo | Average work trip length ending in a balanced area 29 percent shorter than work trips that end in unbalanced areas | |
| Pivo et al | Association between both density and balance and less auto use confirmed | |
| Bookout | Fewer cul-de-sacs, return to grid street pattern, and ped. and transit friendly neighborhoods will reduce vehicle trips or trip distances | Based on Kulash Study |
| Kulash | Traditional street network produces 57% less total VMT, shorter trips and works more efficiently than conventional suburb | Internal trips only, trip frequencies are fixed and local street traffic much higher for traditional network |
| McNally and Ryan | Neotraditional network has 10% less veh.-km and 27% less veh.-hrs traveled, and 15% shorter ave. trip length than conventional network | Trip frequencies are fixed |
| Friedman, Gordon and Peers | Suburban areas generated 23 % more trips, had higher drive alone rates and had half the transit share of traditional communities | |
| Crane | Contradicts Kulash, McNally and Friedman. Says one cannot separate out many differences between suburban and traditional communities. Transportation problems may worsen in traditional communities because trip demand may go up. Therefore trip frequencies in network studies should not be fixed. | |
| Kitamura et al. | Land policies promoting high densities and more mixed land use may not influence travel behavior unless resident's attitudes were also changed. | |
| Holtzclaw | Doubling residential or population density reduces the annual auto mileage by 20 to 30% | Did not correct for income |
| Ewing | Households in suburban community had 2/3 more veh.-hrs than a traditional community with gridded streets and varied land use. | Controlled for income & included chaining analysis |
| Steiner | Higher density residential areas make decreased usage of auto possible. Household and life -cycle characteristics need to be factored out | Advocates further research |
| Handy | Need to answer how mixed use developments will relate to larger settlement patterns | Advocates more research |
Most of these studies find some sort of association between mixed-use neighborhoods and less auto travel. However, some authors (e.g., Kitamura; Crane; Steiner; and Handy) urge caution because the issue is complex. They contend that studies need to carefully factor out household and life-cycle characteristics before relevant comparisons can be made.
Weekend Travel
Weekend travel in the Puget Sound region was the topic of a 1971 PSCG
report which indicated that "there is increasing concern that proper
attention has not been given to recreational travel (primarily done on the
weekends) as a factor in transportation planning." The study proposed
a multi-phased concept for long-range planning of urban transportation
facilities to serve the weekend travel demands of metropolitan areas.
Recommended methods included a variety of modeling (due to limited
availability of empirical data). The study did not report any results and
only travel to major recreational areas were addressed.
A recent study from Japan (Yai et al. 1995) indicates that the volume of passenger vehicles for recreational traffic on weekends can be equivalent to that of weekday commuter traffic. This study, like the previous one, will develop recreational travel demand models. It will be used for trip generation and trip distribution using an aggregate regression model and a disaggregate model.
Voorhees and Associates (1974) also proposed modeling to analyze weekend vehicular travel. The scope of the study was Sunday afternoon traffic on rural highways returning from recreational destinations to urban areas. Some state and national data were used to calibrate the model, some output of which shown in Table 6. The authors concluded that weekend travel demand must be linked with weekday travel estimates for adequate highway design. They added that any "model is only as good as the input data. Therefore, a large amount of empirically derived data is necessary to simulate present travel patterns." They cautioned that good travel data for weekend analysis is lacking in many state planning agencies.
Table 6. Weekend Travel Characteristics
| Topic | Finding | |
|---|---|---|
| Primary Weekend Trip Purposes: | Recreation: 33 % Social: 34 % Shopping: 10% |
|
| Average Trip Lengths: | 70-100% longer for weekend trips than for weekday non-work trips | |
| Source: Voorhees and Associates 1974 | ||
A more recent study on vacation travel was entitled "Weekend Travel: America's Growing Trend" (US Travel Data Center 1990). Its focus was for round trips of at least 200 miles, multiple day (one to five night) trips taken over a weekend. Although geared toward the travel industry, some findings are illustrative of weekend travel in general. Information was obtained from the Data Center's National Travel Survey and indicated that between 1984 and 1989 total trips increased by 26 percent, while weekend trips increased by 34 percent. Table 7 shows trip characteristics and demographics for weekend vacation travel.
Table 7. Weekend Vacation Trips
| Topic | Response | Value |
|---|---|---|
| Main Purpose of Trip | Visit Friends and Relatives | 45 % |
| Outdoor Recreation and Entertainment | 45 % | |
| Travel Party Size | One | 41 % |
| Two | 33 % | |
| Three | 12 % | |
| Presence of Children | Parties without children | 78 % |
| Parties with children | 22 % | |
| Income | Less than $35,000 per year | 45 % |
| More than $35,000 per year | 39 % | |
| Household Structure | Single adult, no children | 19 % |
| Single adult, with children | 4 % | |
| Two or more adults, no children | 40 % | |
| Two or more adults, with children | 37 % | |
| Household Size | One | 19 % |
| Two | 29 % | |
| Three | 22 % | |
| Four | 19 % | |
| Five or more | 10 % | |
| Age | Average | 39 years |
| Sex | Male | 51 % |
| Female | 49 % | |
| Source: US Travel Data Center 1990 | ||
An empirical study exploring weekend traffic volumes was done in the Santa Monica Mountain area of southern California (City of Los Angeles 1978). Although most of the results are specific to that area (e.g., volume percentages for certain intersection approaches), this study introduced the concept of temporal distribution. The study demonstrated that both Saturday versus Sunday and time of day distributions would be interesting variables to explore.
Hu (1996) used the Nationwide Personal Transportation Survey for a study on travel behavior by day of week. Multitudes of figures and tables describe household characteristics, person characteristics, and trip characteristics for Saturday and Sunday travel. This information can serve as a benchmark for weekend travel in typical urban areas.
Finally, some useful weekend travel data for this project were obtained from Murakami (1996) who used data from the Nationwide Personal Transportation Survey. Several tables listing weekday travel were redone for weekend travel only and served as an excellent reference for the research summarized in this paper. Table 8 shows general information about the three variables used in this research.
Table 8. Data Comparison Topics from Murakami
| Table Title | General Information |
|---|---|
| Average Weekend Trip Length by Purpose | 4-18 miles |
| Average Weekend Trip Length by Mode | 0.5-14 miles |
| Daily Person Trips per Household by Household Size | 2-11 trips |
| Source: Murakami 1996 | |
Summary of Weekend Travel
Table 9 shows the major findings of weekend travel
studies. The focus is mostly on long distance recreational travel not
influenced by urban form. Hu and Murakami will serve as excellent base
points from which to compare the characteristics of typical urban areas
with those of mixed use neighborhoods. However, none of these studies
addressed the issue of walking trips.
Table 9. Summary of Weekend Travel
| Author | Key Findings | Relevant Notes |
|---|---|---|
| PSGC, and Yai | Highlight need for weekend travel studies. Propose modeling to study weekend recreational travel demands | No results |
| Voorhees et al | Primary weekend trip purpose: social, recreation and shopping. | Recommend more empirical studies |
| US Travel Data Center | Provides trip characteristics and demographics. | Based on longer vacation trips only |
| City of LA | Not relevant | Use temporal distributions |
| Hu | Household, person, and trip characteristics listed for more typical urban areas | Good source for data comparison |
| Murakami | Average trip length by purpose and mode as well as trip frequency given for typical urban areas. | Good source for data comparison |
Literature Review Summary
This literature review has shown that travel is increasing, and that
mixed-use neighborhoods may offer some transportation benefits. Many of
these studies have shown that mixed-use or neotraditional neighborhoods
are associated with less auto travel.
On the other hand, several authors urge caution and more research because of the issue's complexity. Household and life-cycle characteristics need to be carefully factored out before relevant comparisons can be made between mixed-use neighborhoods and more suburban areas. Additional measures such as trip frequency and travel speed must be analyzed to portray travel patterns.
Non-work travel is gaining in magnitude and complexity. Trip chains are becoming increasingly important, and trip counting techniques (such as number of trips) must be modified to reflect the new transportation trends more accurately. Short walking trips are important in non-work travel in mixed-use neighborhoods, as such, they should be included in the analysis.
Finally, weekend travel is an area that has received little research attention. Most modeling studies suffer from a lack of data upon which to calibrate the models. Extant empirical studies have not addressed mixed-use neighborhoods explicitly (including short walking trips).
The Data Sets
Introduction
This research was based on two data sets. First, a mixed-use neighborhood data set was collected by the Washington State Transportation Commission's Innovations Unit in November and December of 1991 as part of this study. Second, the Puget Sound Regional Transportation Panel Survey, conducted from September through November 1989 and obtained from the PSRC, was used as a reference data set. To enhance the validity of comparisons between the two data sets, the mixed-use data collection effort was designed for compatibility with the PSRC's panel survey methodology.
While this section focuses on the data collected from the mixed-use neighborhoods, the data collection methodology for both data sets is discussed briefly. The mixed-use data required considerable preparation for analysis, and the steps of this process are documented herein. Since both data sets are compared, differences between the mixed-use data set and PSRC data set are discussed.
The Mixed-Use Data
The mixed-use neighborhood data set was obtained from a series of two-day travel diaries completed in November 1992. Over 1,620 individuals in 900 households in the Kirkland, Wallingford, and Queen Anne neighborhoods in the greater Seattle region responded. A project report (Zemotel et al. 1993) details the data collection methodology, characteristics of the study neighborhoods, and preliminary data analysis.
Neighborhood Descriptions
Neighborhoods were selected for study because they had more than one
distinct land use (residential as well as other uses), and because each
was located in an area offering a range of mode choices. The location of
each neighborhood is shown in Figure 1.
Figure 1. Study Area Vicinity Map
Queen Anne, located a few miles north of downtown Seattle, was the smallest of the three study areas. The study area was roughly 0.5 mile by 0.7 mile, centered on Queen Anne Avenue, a busy shopping street with supermarkets, banks, restaurants and retail shops. The rest of the study area was residential with a few scattered retail and office facilities. Queen Anne's streets form a grid pattern.
Wallingford is west of Interstate 5, a few miles north of downtown Seattle, and west of the University of Washington. The study area was approximately 0.75 mile by 1.25 miles long. The neighborhood's land use is diverse with parks, residential uses, and a variety of retail and commercial buildings. The main shopping area is along Northeast 45th Street and, to a lesser extent, along Stoneway Avenue North. The street pattern forms a grid.
Kirkland is a suburban neighborhood bordered by Lake Washington on the west and Interstate-405 on the east. The study area was the largest and was approximately 2.0 miles by 1.2 miles. The area includes a renovated downtown and a mix of housing types. Kirkland's shopping and commercial facilities are somewhat more scattered than those of the other study neighborhoods, but there are concentrations along Central Way and at the downtown 'core' where Central Way meets Lake Street Kirkland has a combination of a grid street pattern and curvilinear streets with cul-de-sacs, which is different from the strictly gridded streets of Wallingford and Queen Anne. Kirkland's land use pattern in many ways represents a transition between a mixed-use area and other suburban development.
Data Collection Process
Individuals in each neighborhood were initially contacted through a random dialing phone survey. First, a range of demographic information was collected from each respondent. This information included the number of vehicles owned, family size, and income. Information was also collected on each person (over the age of 15) surveyed. This information included age, sex, and whether the respondent was employed, a student, or neither. Respondents were then asked to participate in a travel diary survey. Those who agreed to participate were then sent a travel diary packet. Forty-three percent of the people contacted agreed to complete the travel diary. Among this group 76 percent returned a completed diary resulting in an overall response rate of 33 percent.
Each family member over the age of 15 in the survey household was asked to fill out a two-day travel diary describing every trip taken over that period. Information on each trip was to include purpose, travel mode, number of people in the vehicle, trip duration, and amount of time spent at the destination.
The Location Data
The travel diary data focused on respondents' travel patterns. However, as collected on the diary, travel origins and destinations were listed as only a set of addresses, an intersection, or the name of a landmark. To make these data usable the information was geocoded. The resulting data set contains more than 24,000 addresses, intersections, and landmarks. The Census' TIGER line file for all of King County was used for address matching . Computer software successfully geocoded 65 percent of all the location with the remaining location coded by hand. Ultimately over 96 percent of the locations were successfully geocoded. The few locations that could not be coded involved a trip that was outside King County, bad information, or incomplete survey responses.
Panel Survey Data
The PSRC transportation panel survey was used as the source of comparative county-level travel characteristics. Since the PSRC data collection effort was started before the mixed-use survey project was initiated, the PSRC survey was used as the basis for the design of the mixed-use survey.
The PSRC panel survey was a major effort aimed at collecting data on the effect of transportation conditions and demographic characteristics on household travel behavior in urban areas. Only part of the PSRC survey effort (the first wave conducted in 1989) was used for this study. The data used for this study involved 663 households in King County making almost 12,000 trips (see Murakami and Watterson (1992) for detailed information on the survey methodology).
Identification of Trip Chains
Since the 1970s, the emphasis in many studies of transportation behavior has shifted from analysis of individual trips to that of multipurpose trips or chains. This shift is due to the recognition that understanding chained travel is crucial in understanding most individual travel behavior (Alder and Ben-Akiva 1979). A more accurate view of urban travel accounts for sequential, multipurpose travel and assumes accessibility changes as a person moves from one trip origin to another.
The methodology used to organize the mixed-use and PSRC data into trip chains borrowed from previous trip chaining research. Examination of the literature suggests a chaining definition on the use of home or work as an anchor point. Adler and Ben-Akiva's (1979) widely cited model of chain behavior was based on chains defined as trips to or from home. A link (which they called a sojourn) is a visit to any place remote from home. A combination of trips away from home defined a trip tour (or chains). Southworth (1985) divided chains into five types based on trips that started from home or work. Strathman and Dueker's (1994) analysis of the National Personal Transportation Survey (NPTS) used a typology based on chains that started and ended at home. Hodge, while exploring multi-purpose travel in King County (the same area as this study), considered a chain to be any set of trips that had home or work as an endpoint (1991). The trip chain was considered broken if an individual stayed at a location longer than 90 minutes.
Each of the studies listed above started and stopped (that is anchored) trip chains at a home location and sometimes at a work location. For this research, chains were also anchored at home or work. However, this study, like Hodge's, also broke chains after an individual remained at a stop longer than 90 minutes. Breaking a chain after a time threshold served as a mechanism to clearly delineate the importance of the home and work trip anchors in determining trip chains. In addition, Richardson and Young argued that the use of temporal constraint serves to reduce the number of unrealistically long chains and could make the process of exploring travel more tractable (1982).
Data Analysis -- Weekdays
Overview
In this section, travel characteristics of the inhabitants of the mixed-use neighborhoods and the PSRC survey are explored. The measure of travel most commonly used in this paper is average daily travel mileage per person (over age 15). This figure expresses the average per-person mileage of all trips made in one day, based on all the survey respondents fitting into the category of interest.
The analysis begins by examining the geographical areas and general travel characteristics of the survey respondents. The relationship between household income, household category, respondent's age and sex, and the average daily mileage traveled is explored. The section also looks at transit, walk and bicycle trips.
Since most urban travel involves multi-purpose trips, there is also some focus on trip chaining behavior. Given the importance of nearby destinations to the neotraditional concept, an identification of trip stops that were close to each respondent's household is also completed. Work travel is given separate consideration. This analysis looks at work chains, chain lengths, and work locations.
The data analysis then looks at the neighborhood-level travel patterns of the mixed-use respondents. This section examines the pattern of trips generated by local commercial establishments and bus stops. The trip length and travel characteristics of the mixed-use households and PSRC's King County households are directly compared. The analysis involves a number of household and income categories and analysis zones.
It should be reiterated that analysis in the weekday portion of this study compares the mixed-use data set with the PSRC data set and that both data sets were adjusted for compatibility. Since the PSRC respondents were asked only to include trips five minutes or longer, only mixed-use weekday trips of more than five minutes duration are included in comparisons.
Geographical and Household Characteristics
Geographical Variables
Because this study was driven by the geographical location of
households, analysis required development of a number of distance and
zonal variables. Several geographic zones were created based on when the
cities or census places in the county were initially developed (Figure
2). The first zones were the three mixed-use neighborhoods of Queen
Anne, Wallingford and Kirkland. The city of Seattle is divided into
north Seattle. Since north Seattle encompasses the Queen
Anne and Wallingford study areas, these areas were frequently compared. In
the PSRC data sample 176 households were randomly sampled in north
Seattle. The next zone is an inner ring, and about 30
cities surrounding Seattle that were developed in the 1940s, '50s and
early '60s, and sampled 163 households. The outer ring
includes both newer suburban developments and the remaining rural and
unincorporated portion of King County, and sampled 248 households.
Household Characteristics
A summary of demographic characteristics of the mixed-use neighborhoods
only and several King County analysis zones are shown in Table
10. The two mixed-use neighborhoods within Seattle are similar. The
third mixed-use neighborhood, Kirkland, has a higher median age and
considerably lower residential density. With the exception of income,
North Seattle is much like Queen Anne and Wallingford. Inner and outer
King County are also similar to each other and have larger household sizes
and higher auto ownership levels than areas in Seattle.
Table 10. Summary of Household Characteristics
| Location | Average Household Size | Average Number Employees per Household | Average Number of Vehicles per Household | Median Age of Persons over 15 | Percent Income over $35.000 | Gross Density hh per Acre |
|---|---|---|---|---|---|---|
| Queen Anne | 2.2 | 1.4 | 1.7 | 39 | 67% | 7.6 |
| Wallingford | 2.1 | 1.3 | 1.6 | 37 | 56% | 7.2 |
| North Seattle | 1.9 | 1.2 | 1.8 | 37 | 41% | 5.4 |
| Kirkland | 2.0 | 1.0 | 1.9 | 47 | 61% | 3.1 |
| Inner | 2.5 | 1.4 | 2.1 | 35 | 56% | 1.2 |
| Outer | 2.7 | 1.4 | 2.2 | 37 | 55% | 0.2 |
| Urbanized King Co. | 2.5 | 1.3 | 2.1 | 37 | 51% | 2.0 |
General Travel Characteristics
Age
Both the mixed-use and PSRC surveys elicited respondents' ages.
Table 11 compares average daily travel mileage per
person for each survey in eight age categories.
Table 11. Average Daily Mileage Per Person by Age Group (Weekdays only)
| 15-17 | 18-24 | 25-34 | 35-44 | 45-54 | 55-64 | 65-98 | All Age | Total. (n) | |
|---|---|---|---|---|---|---|---|---|---|
| Queen Anne | 19.0 | 21.2 | 18.0 | 17.7 | 22.4 | 14.3 | 14.5 | 18.2 | 670 |
| Wallingford | 9.5 | 13.6 | 18.1 | 16.0 | 18.8 | 19.8 | 16.9 | 16.9 | 594 |
| North Seattle | 13.3 | 20.0 | 24.5 | 24.2 | 23.4 | 21.6 | 16.6 | 22.4 | 581 |
| Kirkland | 12.7 | 32.6 | 31.7 | 29.8 | 26.4 | 27.2 | 21.9 | 27.1 | 589 |
| Inner | 18.4 | 31.0 | 33.3 | 35.2 | 29.5 | 28.8 | 22.0 | 30.3 | 659 |
| Outer | 26.0 | 43.4 | 40.3 | 42.4 | 37.0 | 34.6 | 36.9 | 38.5 | 924 |
| n = number of person days, italic = n less than 25 | |||||||||
Across the two data sets, the King County respondents generally traveled more miles per day than did their counterparts in the mixed-use neighborhoods. Individuals from the outer area groups tended to have the highest mileage, followed by the inner areas. The Kirkland neighborhood tended to fall between the other two mixed-use neighborhoods and the King County areas. Among age groups, the youngest and oldest groups had lower mileage than did those in the more middle-age categories. The higher mileage groups in the neighborhood of Queen Anne and Wallingford tended to be older than those in the other areas.
Income
Table 12 shows the daily average mileage per person
related to annual household income. The households were classified by low
or high income with an income of $35,000 as the cutoff point.
Table 12. Average Daily Mileage Per Person by Income (Weekdays)
| Household Income Less Than $35,000 a Year | (n) | Household Income More Than $35,000 a Year | (n) | % Diff. | |
|---|---|---|---|---|---|
| Queen Anne | 14.5 | 181 | 19.7 | 475 | 26.2% |
| Wallingford | 16.1 | 231 | 17.2 | 353 | 6.4% |
| North Seattle | 20.3 | 290 | 24.3 | 263 | 16.5% |
| Kirkland | 22.0 | 184 | 29.7 | 386 | 25.9% |
| Inner King Co. | 27.6 | 240 | 32.2 | 397 | 14.3% |
| Outer King Co. | 36.7 | 346 | 37.4 | 549 | 1.9% |
| n = number of daily person trips | |||||
For both the mixed-use neighborhoods and in King County suburban areas, individuals from the lower income households traveled less per day. Differences between lower and higher income individuals ranged from 1.9 percent (less than a mile a day) in the outer zone of King County to 23 percent (almost 8 miles day) for the Kirkland neighborhood. The PSRC survey respondents who lived in outer King County had a daily mileage that was high regardless of their income category.
The two mixed-use neighborhoods in Seattle (Queen Anne and Wallingford) also had considerably lower daily mileage per person than did the north Seattle households. The Kirkland respondents' mileage was greater than the other mixed-use neighborhoods but less than that of the inner and outer areas of King County. This perhaps reflects Kirkland's combination of mixed-use and suburban characteristics.
Household Category
A detailed analysis of mileage was completed by examining travel as
related to household category. The use of household categories attempted
to remove any effect that household size and type may have on daily travel
patterns (Table 13).
Table 13. Average Daily Travel per Household Category (Weekdays)
| child(ren) under 6 | child(ren) 6-17 | one adult <35 | one adult 35-64 | one adult 65+ | two adults < 35 | two adults 35-64 | two adults 65+ | Total (n) | |
|---|---|---|---|---|---|---|---|---|---|
| Queen Anne | 19.9 | 20.0 | 6.5 | 10.7 | 19.5 | 19.6 | 16.9 | 18.2 | 671 |
| Wallingford | 16.9 | 17.9 | 21.1 | 16.4 | 13.9 | 15.5 | 17.5 | 17.1 | 595 |
| North Seattle | 29.0 | 21.7 | 19.2 | 19.9 | 14.4 | 23.0 | 22.2 | 17.1 | 636 |
| Kirkland | 28.2 | 29.3 | 31.2 | 23.4 | 24.2 | 32.6 | 30.4 | 21.0 | 591 |
| Inner King | 32.4 | 32.5 | 46.3 | 28.6 | 21.1 | 31.7 | 30.0 | 22.2 | 712 |
| Outer king | 45.2 | 37.1 | 36.7 | 33.4 | 42.5 | 36.6 | 37.9 | 34.0 | 998 |
| n = number of person days, italics/emphasis = n less than 25 | |||||||||
Several patterns are visible in Table 13. For both data sets, households with young children showed higher rates of daily travel. In the King County data, households with older children also traveled a larger number of miles per day. In the mixed-use neighborhoods, individuals from households with two middle-aged adults traveled as many miles per day as did individuals from households with small children. In both data sets, the lowest mileage was found in households with individuals 65 years or older. Across the data sets, King County respondents traveled more per day than did those from the mixed-use neighborhoods.
Sex
Table 14 shows the average daily trip mileage by sex
for both automobile and bus modes. The mileages for the two modes is the
averaged total mileage traveled per day by a survey respondent on either
transit or automobile. Some of the transit information should be
interpreted with caution because of small sample sizes.
Table 14. Average Daily Mileage Per Person by Mode and Sex (Weekdays)
| Automobile | Bus | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Male | (n) | Female | (n) | % diff | Male | (n) | Female | (n) | % diff | |
| Queen Anne | 20.5 | 297 | 16.9 | 279 | 19.7% | 6.4 | 48 | 6.5 | 80 | -1.6% |
| Wallingford | 18.3 | 217 | 16.9 | 293 | 7.6% | 6.6 | 58 | 6.6 | 69 | 0% |
| North Seattle | 24.2 | 250 | 24.8 | 69 | -2.4% | 11.2 | 43 | 8.8 | 63 | 21.4% |
| Kirkland | 28.8 | 256 | 24.2 | 308 | 16.0% | 23.4 | 14 | 14.0 | 33 | 40.2% |
| Inner King | 31.0 | 299 | 27.6 | 352 | 11.0% | 16.1 | 24 | 17.7 | 23 | -9.9% |
| Outer King | 38.9 | 444 | 35.4 | 455 | 9.0% | 22.2 | 23 | 25.0 | 30 | -12.6% |
| n = number of person days | ||||||||||
As seen in Table 14, men typically traveled more miles per day by automobile than did women. Among the various areas, Queen Anne and Kirkland saw the greatest difference in automobile between men and women. For transit mileage, the Queen Anne and Wallingford neighborhood showed minimal differences between the sexes. The North Seattle and Kirkland areas, on the other hand, had notably higher transit mileage.
Transit Use
Table 15 shows the relationship between transit and
non-transit users in terms of daily mileage using several modes. A survey
respondent is considered a transit user if they used transit for any trip
during a day.
Table 15. Transit and Non-Transit Users Average Daily Mileage (Weekdays)
| Non-transit User | (n) | Transit User | (n) | Difference | |
|---|---|---|---|---|---|
| Queen Anne | 19.6 | 514 | 13.3 | 127 | -47.4% |
| Wallingford | 18.0 | 423 | 14.0 | 126 | -28.6% |
| North Seattle | 23.1 | 490 | 17.1 | 106 | -35.1 |
| Kirkland | 27.3 | 465 | 27.8 | 45 | 1.2% |
| Inner | 29.7 | 618 |
32.3 |
47 | 8.0% |
| Outer | 37.8 | 874 | 37.2 | 53 | -1.6% |
| n = number of person days | |||||
Table 15 shows that in Queen Anne, Wallingford and North Seattle transit riders traveled less miles per day than non-transit users. In the other areas the difference between transit and non-transit user was minimal. One interesting finding is the that non-transit users in the inner suburbs of King County traveled 8 percent less per day than transit users. Since the data is for weekdays, one possible reason for this situation is a long transit commute to the Seattle CBD.
Bicycle Use
In the mixed-use neighborhoods 94 weekday trips (0.9 percent) were by
bicycle, while in King County 40 trips were by bicycle (0.3 percent).
Because these numbers were so small a further breakdown of the bicycle
trips was not completed.
Pedestrian Trips
In the mixed-use neighborhoods 7,474 trips (11.3 percent of all trips)
were by pedestrians while King County had 332 trips by pedestrians (3.6
percent of all trips). It must be recognized that these figures may
underestimate the number of daily walk trips since they include only trips
greater than five minutes in duration. If short trips are included the
number of walk trips increases. For example, for the mixed-use data,
including all trips both above and below five minutes increased the number
of walk trips from 11.3 percent to 15.9 percent. A distribution of walk
trips by geographic area is shown in Table 17.
Table 17. Walk Trips as a Percent of All Trips (Weekdays)
| Percent of Walk Trips | n | |
|---|---|---|
| Queen Anne | 18.1% | 610 |
| Wallingford | 17.7% | 529 |
| North Seattle | 8.8% | 246 |
| Kirkland | 7.8% | 227 |
| Inner | 2.8% | 90 |
| Outer | 2.0% | 84 |
| n= number of trips (links) | ||
Table 17 clearly shows that the mixed-use neighborhoods of Queen Anne and Wallingford had the highest level of walking with around 18 percent of all trips on foot. North Seattle and Kirkland had fewer walking trips with 7 to 9 percent of all trip on foot. In the suburbs of King County less than 3 percent of all trips were by foot.
The distribution of weekday pedestrian trips by trip purpose is shown in Table 18.
Table 18. Walk Trips by Purpose (Weekdays)
| Work | Shop | School | Personal | Appointment | Home | N | |
|---|---|---|---|---|---|---|---|
| Queen Anne | 21.5% | 13.9% | 2.1% | 34.1% | 1.0% | 27.4% | 610 |
| Wallingford | 13.6% | 17.6% | 4.9% | 31.6% | 1.5% | 30.8% | 529 |
| North Seattle | 17.5% | 10.2% | 6.9% | 35.0% | 3.7% | 26.8% | 246 |
| Kirkland | 17.2% | 12.3% | 0.9% | 42.3% | 0.4% | 26.9% | 227 |
| Inner | 23.3% | 12.2% | 1.2% | 40.5% | 3.6% | 21.4% | 90 |
| Outer | 22.6% | 10.7% | 1.2% | 40.5% | 3.6% | 21.4% | 84 |
| n= number of trips (links) | |||||||
Table 18 shows that the most common purpose for walk trips is personal. This is reasonable since many personal trips include walking and running for exercise as well as simply recreational walking. Not including trips that return to the home, the most common purpose for walk trips, with the exception of Wallingford, was for work. For Wallingford, shopping saw more pedestrian trips than did work.
Trip Chains
Each trip in the mixed-use and PSRC data set was assigned a chain and
link variable. The first trip of the day for any respondent was
always chain 1, link 1. If the next trip for that person
started after a stay of less than 90 minutes and did not start from home
that trip would be chain 1, link 2. Otherwise the next
trip would be chain 2, link 1. This process continued
until the next respondent or next day occurred in the data set. During
this process, trips of all duration were included; removing trips of under
five minutes, as occurred in other parts of this analysis, could have
influenced the continuity of some of the chains. This probably had minimal
impact on the analysis of chains since the PSRC respondents tended to
include all trips, including those of five minutes or less. The percentage
of chains by the number of stops is shown in Table 19.
Table 19. Distribution of Number of Links (%) in a Trip Chain (Weekdays)
| 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9+ | |
|---|---|---|---|---|---|---|---|---|---|
| Queen Anne | 61.1 | 26.0 | 7.8 | 2.9 | 1.2 | 0.6 | 0.3 | 0.1 | 0.2 |
| Wallingford | 61.1 | 26.0 | 7.9 | 2.9 | 1.1 | 0.5 | 0.3 | 0.1 | 0.1 |
| North Seattle | 69.3 | 20.2 | 6.2 | 2.5 | 1.0 | 0.5 | 0.2 | 0.1 | -- |
| Kirkland | 58.1 | 26.4 | 9.1 | 3.6 | 1.6 | 0.6 | 0.3 | 0.2 | -- |
| Inner King | 69.5 | 18.8 | 6.6 | 2.9 | 1.4 | 0.6 | 0.2 | -- | -- |
| Outer King | 68.2 | 18.8 | 7.3 | 2.9 | 1.2 | 0.6 | 0.4 | 0.2 | 0.4 |
| All King County | 72.2 | 17.7 | 5.6 | 2.3 | 0.9 | 0.4 | 0.2 | 0.1 | 0.1 |
The three mixed-use neighborhoods showed similar chaining behavior. About 60 percent of all chains contained a single trip. These were mainly trips connecting home and work, or trips wherein travelers arrived at a stop and spent more than 90 minutes there. About a quarter of the chins were two-link trips. This included common trips, such as dropping a child off at day-care on the way to work, as well as going from home to do some quick grocery shopping and then returning. This indicates that a significant number of the trips taken by the mixed-use respondents involved multi-purpose travel.
The data for north Seattle and the inner and outer suburban area of King County indicated that about 70 percent of all chains were single-purpose trips that traveled directly from home or work locations without any intervening stops. This suggests that these residents have a lower rate of multi-purpose trips than do those living in the mixed-use neighborhoods.
The distribution of stops found in Table 19 can be examined in more detail by looking at the average number of links (trips) per household per day. Table 20 shows the average number of links (trips) per household, while Table 21 shows the average number of chains (under the definition used here, a single trip with an anchor at home or work is a chain).
Table 20. Average Daily Trip Links (Trips) per Household (Weekdays)
| Household Type | ||||
|---|---|---|---|---|
| With child(ren) | 1 Adult | 2+ Adults | Senior | |
| Queen Anne | 12.9 | 5.2 | 10.8 | 6.9 |
| Wallingford | 11.5 | 5.3 | 10.4 | 6.9 |
| North Seattle | 10.7 | 4.7 | 10.6 | 7.2 |
| Kirkland | 11.4 | 5.2 | 11.6 | 7.0 |
| Inner | 12.0 | 4.7 | 9.6 | 6.8 |
| Outer | 11.3 | 4.1 | 9.2 | 7.7 |
Table 21. Average Daily Trip Chains per Household (Weekdays)
| Household Type | ||||
|---|---|---|---|---|
| With Child(ren) | 1 Adult | 2+ Adults | Senior | |
| Queen Anne | 7.8 | 3.5 | 7.1 | 3.5 |
| Wallingford | 6.7 | 3.4 | 6.6 | 3.7 |
| North Seattle | 7.6 | 3.3 | 7.5 | 4.5 |
| Kirkland | 6.6 | 3.5 | 7.1 | 3.6 |
| Inner | 8.2 | 3.5 | 7.1 | 4.5 |
| Outer | 7.8 | 3.3 | 6.5 | 4.2 |
As seen in the Table 20, the average number of links within each household type was similar for all locations. Across household types, those with children had the greatest number of stops per day, and households with one adult had the fewest.
Tables 20 and 21 suggest that respondents from both the mixed-use and King County had similar travel patterns in terms of the number of stops and the number of chained trips made per day. This is reasonable considering that travel demands on individuals in any type of area should also be similar. Individuals still need to travel to shop for groceries or buy clothes—regardless of where they live.
The average number of trip links per chain can be derived by combining Tables 20 and 21; the ratio is shown in Table 22.
Table 22 shows that the majority of all chains have one or two links or stops. Seniors have consistently more links per chain.
Table 22. Average Daily Trip Links per Chains per Household (Weekdays)
| Household Type | ||||
|---|---|---|---|---|
| With Child(ren) | 1 Adult | 2+ Adults | Senior | |
| Queen Anne | 1.66 | 1.47 | 1.52 | 1.96 |
| Wallingford | 1.72 | 1.57 | 1.57 | 1.87 |
| North Seattle | 1.41 | 1.42 | 1.41 | 1.61 |
| Kirkland | 1.72 | 1.49 | 1.64 | 2.09 |
| Inner | 1.47 | 1.37 | 1.36 | 1.52 |
| Outer | 1.45 | 1.24 | 1.43 | 1.82 |
The nature of the survey respondents’ chaining behavior can also be explored by analyzing the length of the trip chains as classified by the beginning or the ending link. For Table 23, the data from the three mixed-use neighborhoods are combined.
As seen in Table 23, for both the King County and mixed-use data, chains initiated or finishing at home are longer than those started elsewhere. Trips ending at work in the King County data were about as long as trips ending at home. However, in the mixed-use data, trips ending at work were notably shorter than trips ending at home, which suggests that mixed-use respondents made more stops coming from work than they did traveling to work.
Table 23. Average Chain Length in Miles by Initial or Terminating Purpose (Weekdays)
| Beginning | Ending | |||||||
|---|---|---|---|---|---|---|---|---|
| Mixed | (n) | King | (n) | Mixed | (n) | King | (n) | |
| Home | 7.8 | 5309 | 10.4 | 3658 | 7.5 | 5292 | 9.9 | 3593 |
| Work | 6.3 | 1810 | 10.2 | 1828 | 5.1 | 1847 | 9.4 | 1850 |
| Other | 5.5 | 1791 | 7.0 | 2250 | 5.3 | 1771 | 8.5 | 2291 |
| (n) = number of person chains | ||||||||
Further investigation of chain length can be completed by examining the starting and ending purpose of each chain as shown in Table 24.
Table 24. Average Chain Length by Initial and Terminating Purpose
| Home | Work | Other | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Mixed | (n) | King | (n) | Mixed | (n) | King | (n) | Mixed | (n) | King | (n) | |
| Home | 8.3 | 2795 | 13.0 | 873 | 7.1 | 1289 | 11.4 | 1194 | 6.1 | 1192 | 6.9 | 1509 |
| Work | 5.9 | 1121 | 10.8 | 1192 | 5.3 | 176 | 12.2 | 50 | 3.4 | 443 | 6.1 | 543 |
| Other | 5.5 | 1320 | 8.7 | 1525 | 3.4 | 303 | 7.4 | 573 | 7.1 | 148 | 10.3 | 193 |
| (n) = number of person chains | ||||||||||||
For the mixed-use respondents, the longest chains are those that (1) begin and end at home; (2) begin at home and end at work; and (3) begin and end at other locations. The shortest chains are those that (1) begin at work and end at another purpose; or (2) begin at another purpose and end at work. This situation indicated that non-discretionary, work-based trips tended to be longer than more flexible, discretionary trips for other purposes (e.g., shopping, personal reasons) The longest chains were those that both started and ended at home. This category includes the greatest number of trips, and it probably includes many shopping trips from home wherein the respondent stayed less than one hour at the trip destination. Stops of less than 90 minutes would not create a new chain under this project's definition.
As seen in both Tables 23 and 24, the chains completed by the King County inhabitants were generally longer than those of the mixed-use inhabitants, but they followed the same patterns between purposes. However, one difference is that discretionary trips by King County inhabitants from work to other destinations were relatively longer. This suggests that the King County inhabitants may be more likely to complete errands as they travel from work or that in the suburbs you need to travel farther.
Trip Stops
Given the neotraditional movement’s emphasis on trips to locations
near home, one factor of interest is how many trip destinations are within
a short distance from home. Table 25 addresses trip
ends that are less than two roadway miles from each respondent's
household.
Table 25. Percent of Trip Stops By Distance from Households (Weekdays)
| Distance of Stops from Household Location | |||
|---|---|---|---|
| 1.0 Mile | 1.5 Miles | 2.0 Miles | |
| Mixed Use | 17.4% | 25.4% | 38.7% |
| King County | 4.5% | 11.6% | 18.2% |
This table clearly shows that the respondents in the mixed-use neighborhoods made almost twice as many trips to stops within 2 miles of home than did the King County respondents. The difference between the data sets is especially evident for trips less than one mile from home.
Work Travel
A number of studies have indicated that understanding urban daily travel
behavior requires consideration of not only an individual’s household
location, but his or her workplace location as well. Hanson, for example,
using travel diary data from a Swedish city, concluded that many
households’ daily trips were tied to the journey to and from the work
place (1980). Hodge, using travel diary data collected in King County,
concluded that, “The journey to work remains a critical element of
urban trip making, both as organizer of discretionary travel and household
activities” (1991).
The following tables highlight the importance of the work trip in daily travel patterns and their role as part of multi-purpose trips. Table 26 shows the percentage of links (trips) that involve a work stop.
Table 26. Percent of All Trip Links Involving a Work Stop (Weekdays)
| All Day | AM1 | PM2 | |
|---|---|---|---|
| Queen Anne | 33.9% | 53.4% | 32.8% |
| Wallingford | 30.7% | 57.5% | 36.6% |
| Kirkland | 29.1% | 55.7% | 30.3% |
| King County | 31.6% | 50.8% | 35.5% |
| 1 Any trip link
that starts between 6 and 9 A.M. 2 Any link that starts between 3 and 6 P.M. |
|||
During the morning commute, more than one half of all trip links involved a work stop while about a third of all the evening commute trip links involved a work stop. The King County respondents’ distribution of links per day is not notably different from that of the mixed-use respondents.
Table 27 shows the percentage of chains that involve at least one work stop. If trip chains, involving a work stop are examined, as in Table 24 above, the predominance of the work trip is more apparent. Between 40 and 50 percent of all daily trip chains include a work stop. During both the morning and evening commute, this percentage increases to over 50 percent.
Table 27. Percentage of All Trip Chains Involving a Work Stop (Weekdays)
| All Day | AM1 | PM2 | |
|---|---|---|---|
| Queen Anne | 48.4% | 56.2% | 57.6% |
| Wallingford | 43.6% | 59.3% | 52.9% |
| Kirkland | 41.9% | 57.1% | 50.9% |
| King County | 44.8% | 55.0% | 51.9% |
| 1 Any trip chain
that starts between 6 and 9 A.M. 2 Any trip chain that starts between 3 and 6 P.M. |
|||
The contribution of the work trip to daily travel can also be explored by looking at average mileage for both work and non-work chains. Table 28 shows length for work chains, and Table 29 shows length for non-work chains. As seen in the table, except for the senior households category (which tends to include retired individuals with few work trips, and small survey sample sizes), King County work chains were slightly less than twice the length of the mixed-use chains.
Table 28. Average Daily Trip Mileage Per Work Chain
| Household Type | Mixed | King County | ||
|---|---|---|---|---|
| Mileage | (n) | Mileage | (n) | |
| With Children | 4.9 | 481 | 9.2 | 703 |
| 1 Adult | 4.8 | 212 | 7.1 | 226 |
| 2+ Adults | 4.9 | 706 | 9.1 | 912 |
| Senior | 5.1 | 61 | 5.1 | 45 |
| (n) = number of daily person chains | ||||
Table 29. Average Daily Trip Mileage Per Non-work Chain
| Household Type | Mixed | King County | ||
|---|---|---|---|---|
| Mileage | (n) | Mileage | (n) | |
| With Children | 6.0 | 1182 | 10.1 | 1512 |
| 1 Adult | 6.4 | 516 | 8.5 | 279 |
| 2+ Adults | 5.8 | 1303 | 10.1 | 1412 |
| Senior | 7.4 | 403 | 9.5 | 470 |
| (n) = number of daily person chains | ||||
As shown in Table 29, the mixed-use residents’ non-work chains had about 40 percent less mileage than those of King County. A comparison of Tables 28 and 29 reveals that work chains typically had slightly lower mileage than non-work chains.
Regional Work Trips
One concern when comparing the mixed-use and King County data was
confounding effects due to differential accessibility to Seattle’s
Central Business District (CBD). The CBD is a major employment center for
King County, as such it can be expected to attract a large number of work
trips. Both Queen Anne and Wallingford are close to the CBD; Queen Anne is
about two miles and Wallingford four miles away. This proximity raised
concerns that any average trip length for these two neighborhoods would be
shorter than other locations simply because work trips to the CBD would
reduce the average trip length. These shorter work trips potentially could
obscure some of the transportation effects related to mixed use.
As a means of investigating the CBD’s capture of work trips, the location of each respondent’s workplace was identified for both the mixed-use and King County data. Table 30 shows the percentage of work trips that remained in the same areas as the household location, and those that traveled to the Seattle CBD and to other zones. It is apparent from Table 30 that the Seattle CBD is indeed a significant generator of work travel for Queen Anne and Wallingford. The CBD also attracts the same level of work trips from the north Seattle zone. This finding is particularly relevant to this research because the north Seattle study area includes the Queen Anne and Wallingford neighborhoods. Because of the equal percentage of work trips traveling to the CBD from each of these areas, we conclude that differences in average trip lengths between these areas are probably not unduly influenced by travel to the CBD.
Table 30 indicates that Seattle’s CBD is a major location for work sites for King County’s inner and outer zones. This is reasonable given the large size of these areas. However, as expected, most of the work sites for these two zones remained internal to the areas. The majority of the work locations for the Kirkland residents remain within the inner King zone.
Table 30. Work Trip Destinations (%)
| Location | Within Location | To CBD | To North Seattle | To Inner King | To Outer King |
|---|---|---|---|---|---|
| Queen Anne | 10.5 | 30.9 | 41.6 | 11.5 | 4.5 |
| Wallingford | 10.4 | 24.8 | 46.4 | 11.4 | 5.2 |
| Kirkland | 14.3 | 11.6 | 6.4 | 52.9 | 16.5 |
| North Seattle | 42.0 | 31.0 | 42.0 | 8.4 | 6.1 |
| Inner King | 52.7 | 12.6 | 9.2 | 52.7 | 10.5 |
| Outer King | 44.5 | 6.8 | 4.1 | 31.0 | 44.5 |
Household Location and Commercial Establishments
Since each mixed-use household address was geocoded to a latitude and
longitude, it was possible to determine each household’s distance
from commercial streets. This information made it possible to relate
travel behavior of individuals to the accessibility to local goods and
services. Accessibility was measured by the straight line distance between
each household and the nearest commercial street. Commercial streets were
selected based on concentrations of establishments providing goods and
services used on a routine basis, including grocery stores, convenience
stores, restaurants, dry cleaners, and drug stores.
One tenet of the mixed-use movement is that nearby commercial establishments reduce the need to drive. One test of this idea is to compare levels of walking for mixed-use residents living at different distances from commercial areas. Figure 3 shows the percentage of shopping trips that were completed on foot by households at five different distances from the commercial streets. This analysis includes only shopping trips that have at least one trip end within a census tract that includes the mixed-use neighborhoods. As expected, the figure indicates that the farther mixed-use inhabitants live from a commercial street, the less likely their shopping trips will be on foot (and more likely in an automobile). This trend is particularly noticeable for the Queen Anne and Wallingford data. Over 65 percent of the residents from Queen Anne and 50 percent of those from Wallingford, who also lived within 0.1 mile of a commercial street, walked to shop. In contrast, fewer than 25 percent of those respondents who lived more than 0.2 mile from commercial establishments walked. (These trips could be going anywhere--not just to the local commercial street).
Figure 3. Total Shopping Trips by Walking Related to Household Distance from Commercial Streets
The Kirkland data showed a less obvious trend because of low numbers of walk trips and small survey sample sizes. Kirkland also had a more dispersed pattern of commercial activity than did the other two mixed-use neighborhoods, rendering any patterns less obvious.
The same analytical process was applied to recreation and personal trip purposes (Figure 4). Personal and recreational purposes include eating and drinking, pleasure trips, and family/personal business. As seen in the figure, the overall relationship between walking trips and distance is also noticeable for recreation/personal trips. Since many of these purposes involve commercial establishments, it is not surprising that this level of walking shows a similar trend to shopping purposes.
Figure 4. Total Personal/Recreation Trips by Walking Related to Household Distance from Commercial Streets
Travel Mileage
Travel distance information from the PSRC’s King County data was
compared to data from the mixed-use neighborhoods. During this stage of
analysis, an effort was made to control for sample bias, which was
achieved by comparing travel mileage between similar household types and
incomes. Because of small sample sizes, various categories were
aggregated, and different analysis zones were used.
The average daily mileage by mode for Queen Anne and Wallingford combined (the Seattle mixed-use neighborhoods), Kirkland; north Seattle; and the inner and outer areas (the King County suburban areas) is shown in Figure 5. For all modes the following progression was observed: the Seattle mixed-use neighborhoods had the lowest mileage per day, north Seattle the next lowest, followed by Kirkland. The King County suburban areas had the highest daily mileage. Across modes, automobile use had the highest mileage. For transit the difference in average mileage for the two mixed-use neighborhoods and the King County suburban areas was 14 miles per day. For automobile use, this difference was almost 16 miles a day.
Figure 5. Average Total Daily Mileage by Mode (Weekdays)
Figure 6 compares average daily travel mileage per trip by purpose. Again, there was a notable progression of trip mileage: trip length increased from Seattle mixed-use to north Seattle to Kirkland to suburban King County. In most cases, the Queen Anne + Wallingford mixed-use respondents traveled half the distance per trip than did those living in suburban King County. Across purposes, work trips had the highest average mileage, and shopping trips had the lowest trip mileage.
Figure 6. Average Trip Mileage per Person by Purpose (Weekdays)
This average daily travel information can be subdivided by income. Since it was shown previously that daily mileage varies with household income, daily average mileage was separated into higher and lower income categories. Figure 7 shows the travel mileage for individuals from households with high and low incomes. Again, the Seattle area mixed-use neighborhoods showed the lowest mileage, and the King County suburban areas showed the highest. Those from households with lower incomes consistently traveled fewer miles per day than those from higher income households.
Figure 7. Average Daily Person Mileage for Household by Location and Income (Weekdays)
The travel mileage data can be broken down in more detail by location. Figure 8 shows some of the same data as above, but disaggregated into the three mixed-use neighborhoods and the three King County zones. As with the previous figures, the Seattle mixed-use neighborhoods had the lowest daily person mileage, and the suburban King County areas had the highest. The Kirkland mixed-use neighborhood respondents had higher mileage than other mixed-use neighborhoods and north Seattle, but lower mileage than the King county suburban zones. This finding supports the idea that Kirkland is a transitional neighborhood between mixed land use and traditional suburban land use.
Figure 8. Figure 8-Average Daily Person Mileage by Household Location (Weekdays)
The results depicted in Figure 9 support some of the earlier findings in that individuals from households with children traveled the most and that those from households with seniors traveled the least. Those who lived in the Seattle area mixed-use neighborhoods consistently traveled fewer miles than the respondents from the King County data sets. In every case, the two Seattle mixed-use neighborhoods also had a lower average mileage than similar households in north Seattle.
Figure 9. Average Daily Person Mileage by Household Type (Weekdays)
Table 31 summarizes average daily travel mileage for several locations, household types, and two income levels. With the exception of categories with a small sample size, respondents from the Seattle mixed-use neighborhoods (Queen Anne and Wallingford together) had the lowest mileage for each household type and income category. North Seattle was the next lowest, followed by the inner King County cities, and then outer King County. Except for the senior households category (characterized by a small sample size), the higher income households had higher average daily mileage than their lower income counterparts.
Table 31. Average Daily Travel Mileage by Household type and Annual Household Income (Weekdays)
| < $35,000 | > $35,000 | |||
|---|---|---|---|---|
| miles | (n) | miles | (n) | |
| With Child(ren) | ||||
| Queen Anne + Wallingford | 13.13 | 96 | 15.26 | 663 |
| North Seattle | 25.88 | 44 | 26.93 | 96 |
| Inner | 30.17 | 79 | 34.20 | 189 |
| Outer | 36.73 | 133 | 41.45 | 242 |
| One Adult | ||||
| Queen Anne + Wallingford | 16.53 | 183 | 17.56 | 100 |
| North Seattle | 20.53 | 61 | 18.15 | 12 |
| Inner | 30.85 | 38 | 36.45 | 16 |
| Outer | 37.23 | 39 | 28.00 | 10 |
| Two Adults | ||||
| Queen Anne + Wallingford | 11.75 | 264 | 15.48 | 669 |
| North Seattle | 20.58 | 117 | 24.49 | 140 |
| Inner | 27.34 | 66 | 31.88 | 166 |
| Outer | 37.85 | 117 | 36.64 | 279 |
| Senior | ||||
| Queen Anne + Wallingford | 12.04 | 153 | 16.17 | 98 |
| North Seattle | 17.27 | 60 | 11.69 | 10 |
| Inner | 23.79 | 56 | 21.55 | 21 |
| Outer | 38.04 | 50 | 28.63 | 19 |
| (n) = number of daily person trips | ||||
Travel Time
As noted in the literature search, Gordon and Richardson (1994) pointed
out that while work trip distances have increased, so have travel speeds,
confirming a finding supported by this data analysis. Hupkes (1982)
summarized trip rates and travel times for the
U. S. and European countries and reported the average daily travel per person to range from 65 minutes to 84 minutes. The U. S. travel time in Hupkes' paper was 83 minutes for 1965/66 and was an average of 44 urban areas. Purvis (1994) calculated an average for the San Francisco Bay Area of 82.5 minutes per person in 1990. These observations seem to be confirmed the data analysis reported in Table 32, Table 33 and Table 34. These tables also show that substantial differences in daily travel distances among areas analyzed were not maintained when travel time was taken into account.
Table 32 indicates that for all ages all areas were clearly similar in the number of minutes spent traveling per day. Of all the age groups, the 18- to 24-year-olds in Queen Anne, Wallingford and outer King County tended to spend the most time traveling. For the remaining age groups, those in the middle ages categories had longer travel times. The Seattle area average of about 90 minutes compares fairly well with the Bay Area when you consider that the Seattle survey collected no travel data from those younger than age 16 and the Bay Area started with age 5. NPTS (1995) reports shorter and fewer trips for these younger people and leaving them out raises the average travel time for those remaining.
Table 32. Average Daily Minutes of Travel per Person by Age Group (Weekdays)
| 15-17 | 18-24 | 25-34 | 35-44 | 45-54 | 55-64 | 65-98 | All Ages | Total (n) | |
|---|---|---|---|---|---|---|---|---|---|
| Queen Anne | 89 | 111 | 90 | 92 | 99 | 81 | 81 | 91.8 | 670 |
| Wallingford | 84 | 96 | 92 | 93 | 90 | 88 | 83 | 91.1 | 594 |
| N. Seattle | 89 | 85 | 93 | 88 | 89 | 78 | 75 | 86.2 | 596 |
| Kirkland | 99 | 81 | 86 | 98 | 100 | 96 | 80 | 90.1 | 589 |
| Inner King | 61 | 96 | 88 | 99 | 88 | 89 | 82 | 89.6 | 665 |
| Outer King | 57 | 109 | 95 | 94 | 92 | 96 | 100 | 93.8 | 925 |
| n = number of person days, italics/emphasis = n less than 25 | |||||||||
Table 33 is interesting in the variability, as well as the similarity apparent among household trips in the mixed-use and King County areas. In a number of age categories, individuals from the outer suburbs had the longest time travel (one adult 65+, children under six) but for another types of household this area had among the shortest travel times (two adults < 35). The Wallingford neighborhood had the longest travel times for the several household types (one adult < 35, one adult 35-64) but among the shortest for households 65+. The Queen Anne neighborhood had the longest travel time for households with two adults 35-65 but the shortest for one adult 65+.
Table 33. Average Daily Minutes of Travel per Household Category (Weekdays)
| child <6 | child 6-17 | one adult <35 | one adult 35-64 | one adult 65+ | two adults <35 | two adults 35-64 | two adults 65+ | Total (n) | |
|---|---|---|---|---|---|---|---|---|---|
| Queen Anne | 88 | 97 | 75 | 90 | 66 | 103 | 90 | 91 | 671 |
| Wallingford | 81 | 85 | 100 | 100 | 87 | 93 | 102 | 79 | 595 |
| North Seattle | 88 | 81 | 89 | 86 | 54 | 93 | 88 | 77 | 596 |
| Kirkland | 83 | 88 | 76 | 91 | 83 | 90 | 105 | 81 | 591 |
| Inner | 102 | 81 | 83 | 82 | 82 | 88 | 107 | 94 | 665 |
| Outer | 95 | 88 | 86 | 87 | 145 | 73 | 97 | 90 | 925 |
| n = number of person days, italics/emphasis = n less than 25 | |||||||||
Table 34 indicates that the great difference in travel mileage between the mixed-use neighborhoods and the King County area is not nearly as apparent as the difference in travel times. The average speed for each area shows significantly lower travel speeds for the Queen Anne and Wallingford respondents compared to other areas. Given that these areas had a higher use of the slower transit, bike and walk modes, and higher levels of congestion, this finding is reasonable.
Table 34. Average Daily Time vs. Average Daily Travel Miles (Weekdays)
| Average Daily Travel Minutes | Average Daily Travel Mileage | Average Travel Speed (MPH) | |
|---|---|---|---|
| Queen Anne | 92 | 18.2 | 11.9 |
| Wallingford | 91 | 16.9 | 11.1 |
| North Seattle | 86 | 22.4 | 15.6 |
| Kirkland | 90 | 27.1 | 18.1 |
| Inner | 90 | 30.3 | 20.2 |
| Outer | 93 | 38.5 | 24.8 |
Data Analysis — Weekend Travel
Overview
The previous section detailed weekday travel characteristics of respondents in both the mixed-use neighborhoods and in greater Seattle. While the journey to work still dominates transportation research, travel for shopping, as well as family and personal business, is the fastest growing element of household vehicle miles traveled (Comsis 1994). Weekend travel primarily consists of these categories, and the potential transportation benefits for mixed-use residents who can shop nearby are intuitive.
This section looks at weekend travel from the mixed-use survey. While no comparisons could be made with the PSRC data (wherein no weekend data were collected), descriptive statistics regarding weekend travel for the three mixed-use neighborhoods are presented. General travel characteristics, variations between time of day, and Saturday versus Sunday, as well as a separate look at short walking trips, provide insight into the weekend travel patterns of mixed-use respondents.
Because only the mixed-use neighborhood data are used for this section, the analysis includes all trips made by survey respondents, including those trips under five minutes in length. Again all survey respondents are over age 15 years.
Households
There were 775 people living in almost 450 households providing weekend
travel data for this study (see Table 35 ). In
general, the demographics of these households are comparable with the
mixed-use households described earlier; the weekend data are merely a
subset of the overall data set. As before, these data do not evaluate
children under 15 years of age because they did not fill out the travel
diary surveys.
Table 35. Number of Households and People with Weekend Trips
| Number of Households | Number of Participants | |
|---|---|---|
| Queen Anne | 146 | 257 |
| Wallingford | 156 | 283 |
| Kirkland | 144 | 235 |
General Travel Characteristics
Basic Trip Information
A total of 5,699 weekend trips were taken by respondents in the three
neighborhoods. Table 36 depicts the trip distribution
by neighborhood and day.
Table 36. Weekend Trip Information
| Total number of trips | Number of Saturday trips | Number of Sunday trips | |
|---|---|---|---|
| Queen Anne | 2036 | 1119 | 917 |
| Wallingford | 1946 | 1062 | 884 |
| Kirkland | 1717 | 948 | 769 |
Trip Purpose
Distribution by trip purpose is shown in Figure 10.
Unlike weekday travel, trips for school and work accounted for less than
seven percent of all trips. Thus, they are not considered to be a factor
on weekends. Trips for shopping, personal, and "home" accounted
for more than 90 percent of all trips. Trip purposes by percentage were
generally similar across the three neighborhoods.
Figure 10. Trip Purpose Distribution (Weekends)
Trip Mode
The predominant modes of choice in all three neighborhoods for weekend
travel were either car or walking (Figure 11). Since
less than five percent of all trips utilized a bus, bike or "other"
mode, later analysis involving travel modes will include only car and
walking trips.
Figure 11. Trip Mode Distribution (Weekends)
The corresponding percentages for auto and walk travel are displayed in Table 37. While all three neighborhoods chose the auto predominantly as a travel mode, Queen Anne and Wallingford saw high percentages of walk trips, while Kirkland had only half the percentage walk trips of the other two neighborhoods.
Table
37. Weekend Auto and Walk Percentages
(% of trips)
| Auto | Walk | |
|---|---|---|
| Queen Anne | 76.2 % | 18.9 % |
| Wallingford | 75.5 % | 18.8 % |
| Kirkland | 89.1 % | 9.1 % |
Trip Length
Trip length is among the most frequently used measures of travel. The
overall average trip length in miles for each of the three neighborhoods
is shown in Table 38. Queen Anne and Wallingford had
similar numbers, while Kirkland's average length was somewhat longer.
Because the average lengths for the two predominant modes (walking and
auto) was so different, individual averages are provided as well.
Table 38. Average Weekend Mileage Per Trip
| Ave. Trip Length | Ave. Auto. Trip Length | Ave. Walk Trip Length | |
|---|---|---|---|
| Queen Anne | 3.9 | 4.5 | 0.3 |
| Wallingford | 4.0 | 4.9 | 0.5 |
| Kirkland | 5.1 | 5.6 | 0.4 |
Beyond average trip length, the distribution of trip lengths in the three neighborhoods is also of interest insofar as it is distinctly different. Figure 12, Figure 13, and Figure 14 are trip length histograms for each neighborhood. Queen Anne had many trips under one-half mile, with very few longer trips. Wallingford's histogram was less angled, but still indicated an emphasis on shorter trips. Kirkland however, beyond the very short trips, and saw a resurgence of trips at the 4-mile mark, and again at the 10-mile mark.
Figure 12. Trip Length Histogram - Queen Anne (Weekends)
Figure 13. Trip Length Histogram - Wallingford
Figure 14. Trip Length Histogram - Kirkland (Weekends)
Trip Duration
Average weekend trip duration varied from 16.7 minutes for Queen Anne to
19.3 minutes for Wallingford. The Kirkland average duration was between
the two at 18 minutes. As with the weekday data, the time duration of
trips in these areas, due to increase speeds, show less variability than
the distance in miles. Average auto trip duration is listed in
Table 39 (walking trips are addressed later). Queen
Anne respondents spent the least time traveling by car for each trip. Yet
as indicated in the second part of the table, Queen Anne residents spend
the most amount of time traveling on a daily basis. This is consistent
with the average daily travel minutes for weekday travel, and concurs with
later findings that Queen Anne residents travel more frequently than those
living in the other mixed use neighborhoods.
Table 39. Average Minutes of Travel
| Per trip (autos only) | Per day (all modes) | ||||
|---|---|---|---|---|---|
| Weekend | Saturday | Sunday | Weekend | Weekday comparison | |
| Queen Anne | 16.3 | 16.3 | 16.2 | 94 | 92 |
| Wallingford | 19.5 | 19.4 | 19.7 | 89 | 91 |
| Kirkland | 17.9 | 17.8 | 17.9 | 90 | 86 |
Total Distances Per Day
The average total distance per person per day is shown in Table
40. Kirkland respondents had 20 percent longer distances than did
Queen Anne and Wallingford respondents. This may correspond to the longer
individual trip distances seen in Table 38. Yet
because the average duration was longer, the general travel speeds are
somewhat higher in Kirkland, as noted above. This confirms patterns seen
in the weekday analysis where Kirkland had the highest travel speeds of
the three mixed use neighborhoods.
Table 40. Average Daily Mileage per Person (Weekend)
| Weekend | Saturday | Sunday | |
|---|---|---|---|
| Queen Anne | 21.5 | 24.7 | 18.1 |
| Wallingford | 18.3 | 19.5 | 17.0 |
| Kirkland | 24.5 | 27.0 | 21.7 |
The average total distance per household per day is shown in Table 41. Wallingford households in general traveled the least, while residents in all three neighborhoods traveled much less on Sunday than on Saturday.
Table 41. Average Daily Mileage Per Household (Weekends)
| Weekend | Saturday | Sunday | |
|---|---|---|---|
| Queen Anne | 36.8 | 42.6 | 30.8 |
| Wallingford | 31.3 | 33.5 | 28.9 |
| Kirkland | 37.7 | 42.4 | 31.5 |
Frequency
This measure is indicated by trips per person per day or by trips per
household per day. Table 42 shows that the overall
frequency is not significantly different among the three neighborhoods,
although Queen Anne residents traveled most often. All residents tended to
stay home more often on Sunday.
Table 42. Average Weekend Trip Frequency
| number of trips per hh | number of trips per person | |||
|---|---|---|---|---|
| Saturday | Sunday | Saturday | Sunday | |
| Queen Anne | 10.5 | 8.9 | 6.1 | 5.2 |
| Wallingford | 8.9 | 7.3 | 5.0 | 4.2 |
| Kirkland | 8.4 | 7.2 | 5.2 | 4.7 |
Because mixed-use neighborhoods may encourage more walk trips, it is important to study frequency by mode (Table 43). Auto trip frequencies were very similar, but a difference is clearly apparent in the walk trips per person per day. Kirkland saw only half the frequency of the two mixed-use neighborhoods.
Table 43. Average Daily Trips per Person by Mode (Weekend)
| Automobile Trips | Walking Trips | |||||
|---|---|---|---|---|---|---|
| All Wknd | Sat. | Sun. | All Wknd | Sat. | Sun. | |
| Queen Anne | 4.31 | 4.70 | 3.90 | 1.07 | 1.04 | 1.10 |
| Wallingford | 3.48 | 3.79 | 3.17 | 0.86 | 0.91 | 0.82 |
| Kirkland | 4.43 | 4.70 | 4.14 | 0.45 | 0.44 | 0.46 |
Number of people in party
The average number of people in a party for each trip varied from 1.60
for Wallingford, to 1.70 for Kirkland to 1.73 for Queen Anne. The
distribution shows very similar behaviors among the three neighborhoods,
with more than 80 percent of all trips taking place either alone or with
one other person.
Chaining Information
As people link more of their travel together, traditional travel
measures such as number of trips may no longer reflect the amount of
travel accurately. Calculating the number of links per chain provides a
better measure of the efficiency of a resident's travel. Tabular results
are shown in Table 44, while the distribution for
links per chain is shown in Figure 15. For all three
neighborhoods, about half of all trips had more than one link.
Table 44. Number of Links per Chain (Weekend)
| Queen Anne | 1.88 |
|---|---|
| Wallingford | 1.68 |
| Kirkland | 1.79 |
Figure 15. Links per Chain Distribution (Weekends)
Data Comparison
Day of Week
The travel patterns of mixed-use residents were not necessarily the same
as opposed to Saturday and Sunday. The top part of Table
45 shows a summary of several travel measures for the two weekend
days. In general, travel distance decreased on Sunday (except for the
average distance per trip in Wallingford). Travel frequency decreased in
all three neighborhoods on Sunday. Travel duration and efficiency (links
per chain) remained relatively constant over the weekend.
The bottom portion of Table 45 lists compatible travel measures for the same residents on an average weekday. Trip distance (average trip length) and duration were similar from weekdays to weekends, while the number of trips and distance per person per day appeared to increase on the weekends.
Table 45. Travel Measures by Day of Week
| Distance per person per day | Distance per hh per day | Distance per trip | Minutes per trip | Trips per person per day | Links per chain | |
|---|---|---|---|---|---|---|
| Saturday | ||||||
| Queen Anne | 24.7 | 42.6 | 4.2 | 16.8 | 6.1 | 1.9 |
| Wallingford | 19.5 | 33.5 | 3.9 | 19.2 | 5.0 | 1.7 |
| Kirkland | 27.0 | 42.2 | 5.4 | 17.9 | 5.2 | 1.8 |
| Sunday | ||||||
| Queen Anne | 18.1 | 30.8 | 3.5 | 16.5 | 5.2 | 1.9 |
| Wallingford | 17.0 | 28.9 | 4.1 | 19.4 | 4.2 | 1.6 |
| Kirkland | 21.7 | 31.5 | 4.7 | 18.1 | 4.7 | 1.8 |
| Average Weekday | ||||||
| Queen Anne | 18.7 | 31.9 | 4.0 | 17.0 | 7.2 | 1.7 |
| Wallingford | 17.1 | 28.1 | 3.8 | 17.4 | 6.7 | 1.7 |
| Kirkland | 28.0 | 45.3 | 4.9 | 17.3 | 10.0 | 1.8 |
There was very little variation in trip purpose between Saturday and Sunday. As indicated in Table 46, respondents from all three neighborhoods did a significant amount of travel for shopping and personal purposes. Kirkland saw a 50 percent drop in work trips between Saturday and Sunday.
Table 46. Trip Purpose by Day of Week (% of Trips on that Day)
| Work | Shop | Personal | Home | |
|---|---|---|---|---|
| Saturday | ||||
| Queen Anne | 6.2 % | 20.9 % | 39.9 % | 32.2 % |
| Wallingford | 7.0 % | 19.8 % | 35.7 % | 36.8 % |
| Kirkland | 6.3 % | 22.6 % | 37.7 % | 32.7 % |
| Sunday | ||||
| Queen Anne | 5.9 % | 19.7 % | 38.5 % | 35.7 % |
| Wallingford | 7.1 % | 17.5 % | 36.1 % | 38.6% |
| Kirkland | 3.0 % | 22.9 % | 39.0 % | 35.1 % |
| Italics = n less than 25 | ||||
The trip mode analysis is again restricted to auto and walking insofar as those are the only modes that are factors in these neighborhoods. Table 47 shows the variation in trip mode throughout the weekend. Walking percentages increased in all the study neighborhoods on Sunday.
Table 47. Trip Mode by Day of Week (Percent of Trips on that Day)
| Auto | Walking | |
|---|---|---|
| Saturday | ||
| Queen Anne | 77.3 % | 17.1 % |
| Wallingford | 76.0 % | 18.3 % |
| Kirkland | 89.8 % | 8.4 % |
| Sunday | ||
| Queen Anne | 74.8 % | 21.0 % |
| Wallingford | 75.0 % | 19.3 % |
| Kirkland | 88.3 % | 9.9 % |
Hourly Distributions
This analysis investigated when residents travel during the day.
Figure 16 shows the hourly distribution for Saturday
and Sunday. Typically, people travel later on Sundays than Saturdays. This
is apparent in the differing peak locations in the distribution table.
Figure 16. Hourly Distribution by Time of Day
Other Weekend Research
Two studies discussed in the literature search explored weekend travel
in suburban areas. Table 48 reflects Murakami's (1996)
finding of average weekend trip length of 7.9 miles for suburban areas
throughout the country. The three mixed use neighborhoods have, to varying
degrees, a shorter average trip length than Murakami reported. The second
column of the table shows the weekday average trip length comparison
between the mixed use neighborhoods and the PSRC study areas. It is
interesting to note that the mixed use neighborhood trip lengths are very
similar between weekends and weekdays, and the Murakami number closely
resembles the average trip length for outer ring suburbs.
Table 48. Murakami - Average Trip Length Comparison
| Weekend | Weekday | |
|---|---|---|
| Murakami (suburban) | 7.9 | |
| North Seattle | - - | 4.7 |
| Inner | - - | 6.1 |
| Outer | - - | 8.0 |
| Queen Anne | 3.9 | 4.0 |
| Wallingford | 4.0 | 3.8 |
| Kirkland | 5.1 | 4.9 |
The second study (Hu 1996) cataloged travel behavior by day of week for suburban areas. The author detailed multitudes of variables and travel characteristics. Table 49 shows a few select measures that can be compared to measures in this data set and to average weekday numbers from the PSRC data. Such a comparison shows that travel frequencies are similar regardless of household location while the average trip lengths clearly increase in more suburban areas. In addition, women tend to travel more often than men. These findings concur with the ratio between mixed use trip length and suburban trip seen in Murakami’s comparison above.
Table 49. Hu - Weekend Travel Comparison
| Freq. by household income 40k+ | Freq. by gender (male/female) | Ave. trip length for shopping (miles) | |
|---|---|---|---|
| Hu (suburban) | approx. 4.8 | approx. 2.9 / 2.9 | 5.9 |
| N. Seattle* | approx. 5.0 | 4.7 / 5.3 | 5.1 |
| Inner* | approx. 5.1 | 4.7 / 5.3 | 8.7 |
| Outer* | approx. 5.0 | 4.5 / 5.3 | 12.3 |
| Queen Anne | approx. 5.8 | 5.6 / 5.7 | 3.1 |
| Wallingford | approx. 4.7 | 4.6 / 4.7 | 3.2 |
| Kirkland | approx. 5.1 | 4.9 / 5.1 | 3.5 |
| * = Weekday data | |||
Walking Trips
Much has been written about the possibility of walking trips
substituting for auto trips in mixed-use neighborhoods. A number of
studies reviewed in the literature search indicated that people will use
their cars less in neighborhoods where goods and services are nearby.
The mixed-use data for this study are more complete than for most because the database includes short walking trips. This section takes a specific look at these weekend walking trips. There were a total of 749 pedestrian trips in the data base completed by 293 individuals. See Table 50 for the neighborhood distributions of these trips.
Table 50. Weekend Walk Trip Distribution
| Number of Walking Trips | Number of Individuals With Walking Trips | |
|---|---|---|
| Queen Anne | 384 | 113 |
| Wallingford | 365 | 118 |
| Kirkland | 156 | 62 |
Age
The average age of people who undertook walking trips is shown in
Table 51. It does not appear different from the
average age of all the study respondents.
Table 51. Average Age of People Who Walk on Weekends
| Average Age (Years) | |
|---|---|
| People Who Walk | |
| Queen Anne | 42.5 |
| Wallingford | 37.4 |
| Kirkland | 49.3 |
| All Participants | |
| Queen Anne | 41.6 |
| Wallingford | 39.6 |
| Kirkland | 48.7 |
Household Type
An interesting finding is that people who walk appear to come from
larger households on average than that of the respondents as a whole.
There are more adults and children in households with walking trips (Table
52).
Table 52. Household Characteristics of People Who Walk on Weekends
| Household Size | Number of Adults per Household | Number of Children per Household | |
|---|---|---|---|
| People who Walk | |||
| Queen Anne | 2.50 | 1.94 | .52 |
| Wallingford | 2.47 | 2.07 | .43 |
| Kirkland | 2.02 | 1.81 | .21 |
| All Participants | |||
| Queen Anne | 2.16 | 1.69 | .47 |
| Wallingford | 2.15 | 1.81 | .34 |
| Kirkland | 2.04 | 1.72 | .32 |
Tables 53 and 54 list the walking rates by day of week for various household types. The first set of rates includes only households with walking trips, while the second set of numbers displays the average walking rates based on all households. Individual numbers for “Households with walking trips” are not included because the number count of households is low (between 2 and 25).
Table 53. Average Daily Walking Trip Per Household by Household Type (Saturday)
| Household Type | Total (walking trips per household) | |||||
|---|---|---|---|---|---|---|
| With Child(ren) | 1 Adult (no children) | 2 Adults (no children) | Seniors | |||
| Households with Walking Trips | ||||||
| Queen Anne | 3.67 | |||||
| Wallingford | 3.40 | |||||
| Kirkland | 2.67 | |||||
| All Households | ||||||
| Queen Anne | 2.23 | 1.46 | 2.28 | .55 | 1.79 | |
| Wallingford | 1.59 | 1.63 | 1.81 | 1.17 | 1.75 | |
| Kirkland | .76 | .25 | .86 | 1.05 | 0.71 | |
| Italics = (n) households less than 25 | ||||||
Table 54. Average Daily Walking Trip Per Household by Household Type (Sunday)
| Household Type | Total (walking trips per household) | |||||
|---|---|---|---|---|---|---|
| With Child(ren) | 1 Adult (no children) | 2 Adults (no children) | Seniors | |||
| Households with Walking Trips | ||||||
| Queen Anne | 3.94 | |||||
| Wallingford | 3.05 | |||||
| Kirkland | 2.81 | |||||
| All Households | ||||||
| Queen Anne | 1.80 | 1.59 | 2.20 | 2.00 | 1.87 | |
| Wallingford | 2.00 | 1.27 | 1.49 | 0.71 | 1.46 | |
| Kirkland | 0.19 | 1.00 | .73 | 0.80 | 0.72 | |
| Italics = (n) households less than 25 | ||||||
Two-adult households without children walk most often. Interestingly, households with children do not walk less than other types of households. There does not appear to be a large difference between Saturday and Sunday walking rates within the three neighborhoods.
Number of Vehicles
An interesting finding (Table 55) is that people
who walked tended to have similar or slightly more vehicles than other
survey respondents.
Table 55. Average Number of Vehicles of Those Who Walk
| Average Number of Vehicles | |
|---|---|
| People Who Walk | |
| Queen Anne | 1.96 |
| Wallingford | 1.64 |
| Kirkland | 1.65 |
| All Participants | |
| Queen Anne | 1.74 |
| Wallingford | 1.62 |
| Kirkland | 1.89 |
Annual Income
Table 56 shows walking rates by annual income for
the three neighborhoods. Lower income residents walk more often than those
in higher income households. Queen Anne residents walk most often, while
as shown earlier, Kirkland's walking rates are lower than those in the
other two neighborhoods.
Table 56. Average Daily Walking Rates by Annual Income
| Annual Household Income | ||
|---|---|---|
| Less than $35,000 a year | More than $35,000 a year | |
| Number of trips per person per day | ||
| Queen Anne | 1.43 | 0.89 |
| Wallingford | 0.82 | 0.77 |
| Kirkland | 0.51 | 0.43 |
| Number of trips per household per day | ||
| Queen Anne | 1.88 | 1.71 |
| Wallingford | 1.24 | 1.53 |
| Kirkland | 0.73 | 0.71 |
Trip Length
Figure 17 illustrates the trip length histogram for
walking trips. Seventy-three percent of trips were less than one-half
mile, and 40 percent of the trips were less than one-quarter mile. Very
few people will undertake walking trips of more than one mile.
Figure 17. Trip Length Histogram for Walk Trips (Weekends)
Conclusion
The large body of literature reviewed for this paper generally supports the notion that mixed-use or neotraditional neighborhoods can reduce the amount of travel for most households, as measured by the number of miles traveled. The research underlying this paper generally found support for these notions, although we concur with others that the linkage is very complex. Residents of the two mixed-use neighborhoods in Seattle traveled 27 percent fewer miles than the remainder of North Seattle, 72 percent fewer than the inner suburbs and 119 percent fewer than the outer suburbs. If one of these mixed-use neighborhood were somehow relocated to the outer suburbs would it travel characteristics remain the same? It's doubtful, but indications from this research based on looking at various breakdowns of trip and household types make it clear that substantial reductions in travel distances can be accomplished with appropriate urban design.
The paper also looked at weekend travel for the mixed-use neighborhoods. This analysis showed that travel miles on Saturday were about 25 percent greater than Sunday, and Saturday travel was 12 percent greater than the average weekday. Distance per trip for weekend travel was essentially the same as weekday. Comparison of the mixed-use neighborhood weekend data to NTPS weekend travel for suburban sites showed a similar ratio of travel distances as found for comparisons of weekday travel in mixed-use sites and King County suburbs. There is some evidence that mixed land uses has the same effect on weekend trips as weekday trips.
This paper also gives credence to the few researchers who have looked at travel time rather than distance as a principal measure. The large differences among the areas reported for travel distance are not seen when considering travel time. The travel time was about 90 minutes per person regardless of where that person lived. Variation by age and family life cycle stage was also remarkably small. This "travel time budget" of about 90 minutes is an interesting finding and compares favorably to previously cited studies.
This research has several implications for travel demand modeling. First, in order to model new (old) neighborhood forms, short trips must be handled much better than in the past. The sheer number of short trips and the fact that they are substituting for longer trips that would be made in more modern suburbs dictates they be modeled more faithfully. Transportation zone boundaries swallow entire neighborhoods, making consideration of pedestrian and many bicycle trips very difficult. Second, if travel time budgets are as uniform as found in this work and shown in others, perhaps they could be used more in the calibration and validation process to assure that models operate within time constraints by various parameters. Third, the travel time budget issues and close ties between land use and short trips reinforce the notion that feedback loops are an increasingly important part of the travel forecasting process.
Bibliography
Adler, Thomas and Moshe Ben-Akiva (1979) "A Theoretical and Empirical Model of Trip Chaining Behavior," Transportation Research B, Vol. 13b, 243-257.
Bookout, Lloyd W (1992a) "Neotraditional Town Planning: A New Vision for the Suburbs?" Urban Land 51, no. 1 January 1992: 20-32.
Bookout, Lloyd W (1992b) "Neotraditional Town Planning: Cars Pedestrians and Transit," Urban Land 51, no. 2 February 1992: 10-15.
Cervero, Robert and John Landis (1995) "The Transportation -- Land Use Connection Still Matters," University of California Transportation Center. Access No. 7 Fall.
City of Los Angeles (1978) Weekend Travel Characteristics On Santa Monica Mountain Surface Street Routes, Program Development Division, Department of Traffic, City of Los Angeles Staff Report. April 24.
Comsis Corporation (1994) "NPTS Urban Travel Patterns, 1990 Nationwide Personal Transportation Survey (NPTS)," Federal Highway Administration, FHWA-PL-94-018, June.
Crane, Randall (1996) "Cars and Drivers in the New Suburbs -- Linking Access to Travel in Neotraditional Planning" Journal of the American Planning Association Winter.
Downs, Anthony (1992) Stuck in Traffic: Coping with Peak-Hour Traffic Congestion The Brookings Institute, Washington D.C.
Ewing, Reid, Padma Haliyur, and G. William Page (1994) "Getting Around a Traditional City, A Suburban PUD, and Everything In Between," Resource Papers for the 1994 ITE International Conference. [Also in Transportation Research Record 1466. 1993.]
Frank, Lawrence, and Gary Pivo (1994) Relationships Between Land Use and Travel Behavior in the Puget Sound Region, Washington State Transportation Center (TRAC).
Friedman, Bruce, Stephen P. Gordon and John B. Peers (1992) The Effects of Neotraditional Neighborhood Design on Travel Characteristics, Transit Oriented Development Impacts on Travel Behavior, edited by Calthorpe Associates.
Giuliano, Genevieve (1995) University of California Transportation Center. Access Spring.
Gordon, Peter, and Harry W. Richardson (1994) Geographic Factors Explaining Worktrip Length Changes, Nationwide Personal Transportation Survey Implications of Emerging Travel Trends. Edited by Walcoff and Associates. July.
Handy, Susan (1991) "Neo-Traditional Development: The Debate" The Berkeley Planning Journal, Vol. 6, 135-144.
Hanson, Susan (1980) "The Importance of the Multi-Purpose Journey to Work in Urban Travel Behavior", Transportation 9, 229-248.
Hodge, David (1991) Development of Method of Analysis for Planning Transit Systems Components in and Around Major Activity Centers. Part 1: Trip Chaining. The Behavioral Basis for the Design of Circulation Systems for Major Activity Centers, Transportation Northwest, Seattle, University of Washington, September.
Holtzclaw, John (1991) Explaining Urban Density and Transit Impact on Auto Use. Transit Oriented Development Impacts on Travel Behavior. Edited by Calthorpe Associates August 21, 1992. [also prepared for the Nature Resources Defense Council and The Sierra Club, presented to the State of California Energy Resources Conservation and Development Commission, January]
Hu, Juayang (1996) Travel Behavior By Day Of Week In The United States Using The 1990 Nationwide Personal Travel Survey Department of Civil Engineering, University of Maryland, College Park, Maryland. June 6.
Hu, Patricia S. and Jennifer Young. (1993) National Personal Transportation Survey -- 1990 NPTS Databook Volume 1, Oak Ridge National Laboratory for Federal Highway Administration, FHWA-PL-94-010A, November.
Hupkes, Geurt. (1982) "The Law of Constant Travel Time and Trip-Rates" Futures, February.
Kitamura, Ryuichi, Patricia L. Mokhtarian, and Laura Laidet (1994) A Micro-Analysis of the Land-use and Travel In Five Neighborhoods in the San Francisco Bay Area Institute of Transportation Studies, University of California, Davis November.
Kulash, Walter, et. al. (1990) Traditional Neighborhood Development: Will the Traffic Work? Prepared for the American Society of Civil Engineers, Successful Land Development: Quality and Profits Conference, March.
Langdon, Philip (1994) A Better Place to Live: Reshaping the American Suburb The University of Massachusetts Press, Amherst.
McNally, Michael G. and Sherry Ryan (1993) A Comparative Assessment of Travel Characteristics for Neotraditional Developments. Transportation Research Record 1400. [also in 72nd Annual Meeting of the Transportation Research Board, Washington DC, January 1993.]
Murakami, Elaine (1996) Weekend Travel Tables Using the 1990 Nationwide Personal Transportation Survey. Via E-mail and Fax. Federal Highway Administration and Comsis Corporation, May.
Murakami, Elaine, and W.T. Watterson (1992) "The Puget Sound Transportation Panel After Two Waves," Transportation 19, 141-158.
Pisarski, Alan E (1992) Nationwide Personal Transportation Survey -- Travel Behavior Issues in the 90s. U.S. Department of Transportation Federal Highway Administration, July.
Pivo, Gary, Paul Hess and Abhay Thatte (1995) Land Use Trends Affecting Auto Dependence in Washington's Metropolitan Areas, 1979-1990. Washington State Transportation Center (TRAC) for the Washington State Department of Transportation, July.
Prevedouros, Panos D, and Joseph Schofer (1991) Trip Characteristics and Travel Patterns of Suburban Residents, Transportation Research Record 1328.
Puget Sound Council of Governments (1990) Household Travel Surveys, 1985-1988 Puget Sound Region June 1990.
Puget Sound Governmental Conference (1971) Weekend Recreational Travel -- Development of a Concept, Prepared for Washington State Highway Commission, Seattle, Washington, September.
Purvis, Charles L. (1994) "Changes in Regional Travel Characteristics and Travel Time Expenditures in San Francisco Bay Area: 1960-1990," Transportation Research Record 1466.
Richardson, A.J. and W. Young (1982) "A Measure of Linked-Trip Accessibility," Transportation Planning and Technology Vol. 7, 73-82.
Southworth, Frank (1985) Multi-Destination, Multi-purpose Trip Chaining and its Implications for Locational Accessibility: A simulation Approach, Paper of the Regional Science Association, Vol. 57, 107-123.
Steiner, Ruth L (1994) Residential Density and Travel Patterns: Review of the Literature, Transportation Research Record 1466.
US Travel Data Center (1990) Weekend Travel: America's Growing Trend, Affiliate of the Travel Industry Association of America Washington, DC.
Voorhees and Associates, Inc. (1974) Statewide Travel Forecasting Procedures Including Activity Allocation and Weekend Travel -- Phase II -- Weekend Travel Model, Prepared for Federal Highway Administration Washington, DC, August.
Yai, Tetsuo, Harutoshi Yamada, and Naohisa Okamoto (1995) Nationwide Recreation Travel Survey in Japan: Outline and Modeling Applicability, Transportation Research Record 1493.
Zemotel, Linda (1993) Travel Patterns in Mixed-Use Neighborhoods, Phase 1 Draft, Washington State Transportation Commission Innovations Unit, November.
