Content uploaded by Erik Elldér
Author content
All content in this area was uploaded by Erik Elldér on Feb 03, 2016
Content may be subject to copyright.
The changing role and importance of
the built environment for daily travel
in Sweden
Erik Elldér
Göteborg 2015
Avdelningen för Kulturgeografi
Human Geography Unit
Institutionen för Ekonomi och samhälle
Department of Economy and Society
Handelshögskolan vid Göteborgs Universitet
School of Business, Economics and Law
Viktoriagatan 13
University of Gothenburg
405 30 Göteborg
Viktoriagatan 13
SE 405 30 Gothenburg
ISSN 0343-6663
ISBN 91-86472-75-5
© Erik Elldér
Printed by Kompendiet
Göteborg 2015
Cover design Ulf Hammarkärr
www.ulfhammarkärr.com
ABSTRACT
Elldér, Erik. 2015. The changing role and importance of the built environment for daily travel in Sweden.
Publications edited by the Departments of Geography, University of Gothenburg, Series B,
no. 126. Department of Economy and Society, University of Gothenburg, Gothenburg. ISBN
91-86472-75-5.
Geography, in terms of the built environment and location patterns, was traditionally, and still
is, emphasized by many scholars, policymakers, and planners as greatly influencing people’s
daily travel behaviour. However, taking recent decades of rapidly increasing mobility
capabilities (physical as well as virtual) into account, and the related increase in individual
choice opportunities, others argue that the importance of geographic factors has gradually
dissolved. Starting from this discussion, the overall aim of this thesis is to examine the current
role and relative significance of the built environment for the geographical extension of
individuals’ daily travel in Sweden. The thesis is based on three empirical studies in which
particular attention is paid to detailing the impact of geographic factors on various daily travel
activities (paper I); exploring possible changes over time in the importance of the built
environment for home–work distances (paper II); and the potential relaxing of the
relationships between locational structures and travel behaviour when people regularly use
ICTs and telework (paper III). All three papers apply multivariate quantitative approaches to a
unique combination of detailed, high spatially resolved micro-data, including the national travel
surveys and register data of the total population.
An overall conclusion of the thesis is that the proximity of various aspects of the built
environment to home still plays an important role in how far people in Sweden travel daily.
However, the analyses, informed by theory emphasizing everyday spatiotemporal constraints,
reveal that these relationships have become relaxed in several important respects. First, the
specific time–spatial constraints associated with different daily activities that motivate trips and
travel are key and also differentiating factors. When considering trips taken during holidays and
for everyday leisure purposes, the built environment is less important for the observed daily
travelled distance. Whereas service trips to a greater extent is associated with the built
environment surrounding home, and work trips even more. Second, important changes occur
over time, here examined in the case of work trips. Workers living in the same neighbourhood
increasingly travel divergent distances between home and work. This suggests a continued
decrease in the influence of the built environment on work related travel. Third, in terms of
time-spatial relaxation, a rapid increase of telework lately is an important case. The built
environment influences teleworkers’ daily travel to a lesser extent than it does regular workers’
daily travel since telework allows for the freer scheduling of daily activities in time and space.
Conclusively, the results confirm the importance of considering spatiotemporal constraints
related to daily activities when exploring the role of the built environment and its importance
for daily travel. More generally, the thesis also remind us that the importance of the built
environment changes as an integral part of larger societal transformations connected with
development of mobility technologies and profound socio-economic and demographic
changes.
Keywords: built environment, travel, distance, daily, commuting, telework, residential location,
Sweden, activities, mobility, spatiotemporal constraints
ISSN 0343-6663
ISBN 91-86472-75-5
© Erik Elldér
Printed by Kompendiet
Göteborg 2015
Distribution:
Department of Economy and Society
P.O. Box 625
SE 405 30 Gothenburg
Acknowledgements
As with so much else the writing of a doctoral thesis is truly a collective effort.
Many people have contributed to this work. First and foremost I want to thank
my main supervisor, Bertil Vilhelmson, for countless comments, discussions,
encouragement, advice and inspiration, which made this thesis far better than it
otherwise would have been. Anders Larsson has been a brilliant co-supervisor
in general and fellow traveller in particular. Urban Fransson was also part of
the supervising team in the early stages giving valuable guidance. Many thanks
to Katarina Haugen who gave important comments on my half time
manuscript, and to John Östh for a stimulating discussion and many good
advices during my final seminar. Jerry Olsson and Andrew Byerley were
internal discussants during the final seminar and also had many able
comments. At the Human Geography unit, I have had the great fortune of
being part of the mobility research group. This has been an important and
fruitful forum for me, where, besides Bertil, Anders and Jerry, Eva Thulin,
Lotta Frändberg, Ana Gil Solá and Erik Hysing all have been active and
discussed my research. I also want to thank the rest of my colleagues at the
unit for lots and lots of coffee drinking, discussions and for generally enriching
the working environment. Thanks to Ulf Ernstson for being a fellow GIS
nerd, Robin Biddulph for various language advice, and all the PhD students for
navigating the academic maze with me. Moreover, I want to acknowledge all
the feedback I have received from the academic community – from editors,
anonymous referees, conference and workshop participants, etc. – that greatly
assisted this research in many ways. Finally, thanks to my parents, Tomas and
Marie, and my brother Johan for invaluable support and for always being there,
and to my dearest Josefin och världens mest fantastiska unge Noah för att ni
finns och gör livet så mycket bättre!
Göteborg, 3 May 2015
Contents
1 Introduction
1
1.1 Background
1
1.2 Aim
4
1.3 Outline of the thesis
5
1.4 Central concepts and delimitations
5
1.5 The Swedish case and context
7
2 Theoretical framework
10
2.1 Introduction
10
2.2 An activity-based approach
11
2.2.1 Choice in the context of constraints
11
2.2.2 The spatiotemporal constraints of daily activities
12
2.2.3 The role of individual needs, wishes, and capabilities
15
2.3 The role of the built environment in daily travel
17
2.3.1 Facility constraints
17
2.3.2 Relationships between the built environment and daily travel
19
2.3.3 Controversies
23
2.4 The changing role and importance of the built environment
26
3 Data and methods
31
3.1 Point of departure
31
3.2 Data
31
3.2.1 Complementary data sources
31
3.2.2 GILDA
32
3.2.3 National travel survey
33
3.2.4 Accessibility calculations
34
3.3 Methods
36
3.3.1 Empirical considerations
36
3.3.2 Statistical methods
39
3.4 Limitations
41
4 Paper summaries
43
5 Concluding discussion
46
5.1 Main conclusions
46
5.2 Implications for future research
49
5.3 A practical view
52
6 Sammanfattning (Swedish summary)
54
7 References
57
Appendices
Paper I: Elldér, E. 2014. “Residential Location and Daily Travel Distances: The
Influence of Trip Purpose”. Journal of Transport Geography 34, 121–130.
Paper II: Elldér, E. 2014. “Commuting Choices and Residential Built Environments
in Sweden, 1990–2010: A Multilevel Analysis”. Urban Geography 35(5), 715–734.
Paper III: Elldér, E. 2015. “Does Telework Weaken Urban Structure–Travel
Relationships?”. Accepted for publication in Journal of Transport and Land Use.
1
1 Introduction
1.1 Background
This thesis concerns how land use in terms of the built environment influences
people’s daily travel – a “hot topic” in current debates on how to manage
mobility through spatial planning strategies, for example, limiting urban sprawl
by building denser cities. It also contributes to a long-standing issue in
geographical theorizing and research concerning the relationships between
location patterns, individual capabilities, and human spatial interaction. The
specific concern of my thesis is the geographical extension of people’s daily
travel activities, and how these relate to the various built environments of
people’s homes, locally as well as regionally. I make a case for this by drawing
attention to some of the drastic changes that increasing mobility has generated
over the last century.
In Sweden, average daily travel has increased from a few kilometres a
day in the early 20th century to about 45 kilometres one hundred years later
(Frändberg and Vilhelmson, 2011). A series of innovations in transport
technology has increased the geographical reach of people’s daily activity
spaces twenty-fold. All developed countries have experienced similar trends,
and developing countries are largely following the same path (Banister, 2012).
This development continues to have profound geographical consequences,
including suburbanization, increased labour market ranges, and the spatial
extension of leisure activities. While individual reach and freedom of choice
have increased greatly, high levels of mobility also bring many drawbacks and
costs: car dependence and urban sprawl, congestion and pollution, path
dependencies, and lock-in situations. The currently most prominent example
of a negative impact is greenhouse gas (GHG) emissions from transport,
which are increasing at a faster rate than emissions from any other sector and
have more than doubled globally since 1970 (Sims et al., 2014). Eighty per cent
of this increase comes from road vehicles. In 2010, transport accounted for
27% of total energy use globally and is expected to make up an even larger
share in the future. In the absence of effective policy interventions, transport-
related GHG emissions are expected to more than double by 2050 (Sims et al.,
2014). In Sweden, transport-related emissions continue to grow despite the
2
rapidly increasing energy efficiency of new cars (Johansson, 2011). This calls
for careful attention to reducing traffic to more sustainable levels and to the
potential of various planning measures to do so.
From a geographer’s perspective, the contested effects of increased
mobility evoke the concern of the relationships between the locations of
human activities and the needs for daily travel – and to what extent these could
be influenced by urban and regional planning. For example, in the newly
released Fifth Assessment Report – Mitigation of Climate Change, the
Intergovernmental Panel on Climate Change (IPCC) emphasizes improved
land use and transport planning as major mitigation strategies (Seto et al., 2014;
Sims et al., 2014). The IPCC proceeds from the argument that the built
environments of cities and regions establish path dependencies that have long-
term effects on travel. Infrastructure and spatial planning efforts therefore play
a key mitigation role in providing templates for future low-carbon travel
behaviour. The IPCC draws heavily on a vast, longstanding, and still growing
academic literature that interrogates the associations between transport
demand and the built environment (e.g., Boarnet, 2011; Ewing and Cervero,
2001, 2010; Newman and Kenworthy, 1999; Næss, 2012; Salon et al., 2012).
This literature provides one starting point for this thesis. Strong links have
been found in a wide range of geographical contexts; for example, people
living in densely built environments close to a mix of daily amenities and
served by good public transport and cycling connections generally travel
shorter distances and less by car every day than do people living in suburban
and sparsely populated areas. Based on these results, many influential scholars
(e.g., Newman and Kenworthy, 1999) and policymakers (e.g., the IPCC) imply
that there is more or less a straightforward causal connection between denser
location patterns – i.e. increased proximity – and reduced travel.
This view is, however, challenged by another important theoretical
concern in human geography and related academic fields, namely, whether
travel and destination choices are becoming less constrained by location and
proximity over time (see, e.g., Giuliano, 1995; Kwan and Weber, 2003; Miller,
2007). Geographers have long since highlighted the problems of using
geographic location patterns as the single determinant of human spatial
behaviour (e.g., Cox and Colledge, 1969; Olsson, 1965). More nuanced views
have been established that stress the importance of taking the analytical point
of departure in individuals, their needs and capacities to perform activities at
3
various locations, and the role of everyday spatiotemporal constraints in doing
so (see, e.g., Hägerstrand, 1970; Jones et al., 1983). Accordingly, when access
to rapid mobility resources, such as cars, increases, people’s opportunities to
more freely choose where to perform daily activities also widen and become
more flexible, and travel and location patterns follow suit. Simply stated,
people might no longer be as dependent on the built environment and what is
in geographic proximity, and can perform everyday activities in a number of
different places in the same day.
In addition, not only have changing capacities for physical travel
between locations possibly weakened built environment–travel relationships,
but the virtual mobility enabled by information and communication
technologies (ICTs) could also have similar implications. ICTs have the
potential to further relax the spatiotemporal constraints of daily life and make
traditional temporal and spatial patterns of activity participation less dependent
on geographical location and proximity (see, e.g., Dal Fiore et al., 2014; Kwan,
Dijst and Schwanen, 2007; Lenz and Nobis, 2007; Lyons, 2009; Van Wee,
Geurs and Chorus, 2013).
These discussions generate many important questions deserving further
exploration, some of them highlighted in this study. Theoretical and
methodological developments focusing on individual needs and wants have
produced a series of somewhat mixed and contradictory results as regards the
importance of the built environment in daily travel (Ewing, Deanna and Li,
1996; Handy, 1996; Kitamura, Mokhtarian and Laidet, 1997; van de Coevering
and Schwanen, 2006; Van Wee, 2013). Some of these studies even find travel
to be relatively independent of what they define as the built environment while
also accounting for various individual aspects, such as car access and use,
household situation, and socioeconomic characteristics. In addition, from a
dynamic perspective, a key question is the extent to which the influence of
location patterns is actually decreasing over time when it comes to daily travel,
bearing in mind the possible effect of improved research that better capture
individual explanations. In the background lurks the worry that the “individual-
based turn” and approaches – i.e., the search for explanations at an individual
level in transport research – have somewhat neglected the challenge of
properly defining and measuring geographical contexts and built environments.
For example, as increasing mobility has allowed daily travel to destinations
farther from the local neighbourhood, recent research has emphasized the
4
importance of regional built environment structures (Boarnet, 2011; Næss,
2011). Dealing satisfactorily with these issues of dynamics and scale, however,
makes specific demands of data and methods. This includes a persistent need
to properly describe the individual (e.g., as regards socio-demographics), her
actual daily travel (e.g., as regards distances, modes, and purposes), and other
potential modes of access (e.g., via ICTs). In a society where the ability to
move physically as well as virtually is increasing, the factors underlying the
geographical extent of our daily activities are changing and multifaceted. This
includes social issues (e.g., changing intergroup mobility divides; see Frändberg
and Vilhelmson, 2014; Östh and Lindgren, 2012; Solá, 2013) and virtual access
and labour market changes (e.g., recent rapid increases in telework; see Green,
2004; Vilhelmson and Thulin, 2015).
Several considerations need further investigation in order to settle to
what extent, and in what dimensions, for whom, and at what levels, the spatial
opportunities manifested in the built environment are actually playing a role
and changing in importance. This improved knowledge is relevant for
evaluating the efficient and effective use of the spatial policy instruments in
which, for example, the IPCC and many urban and regional authorities set
great store. My thesis contributes to these discussions by scrutinizing the
importance of built environments for people’s daily travel for various
purposes, considering whether its role is changing over time and whether ICTs
play a role in this process. The constituent empirical studies of the thesis
contribute by using and combining rich sources of geocoded micro-level data
capturing people’s daily travel and ICT use, the spatial opportunities provided
by transport systems, and the location patterns of the built environment in
Sweden.
1.2 Aim
Due to increased access to faster means of transport and ICTs, people’s travel
and activity patterns can be expected to become increasingly heterogeneous,
flexible, and less dependent on proximity and the location of activities. In this
study, I explore whether this relaxation is occurring. Specific attention is paid
to detailing the impact of the built environment on different travel purposes,
exploring possible changes over time, and, finally, the potential role of ICT use
in relaxing spatial constraints. I believe that these issues are central not only to
5
many theoretical discussions in geography but also to many key challenges
facing planners and policymakers now and in the future, particularly in relation
to the mitigation of transport-related GHG emissions.
The overall aim of this thesis is to examine the role and relative significance of
the built environment for the geographical extension of individuals’ daily travel in Sweden.
This aim is pursued through three empirical studies in which the specific
research questions are:
Does the built environment affect daily distance travelled
differently when individuals travel for different purposes? (Paper I)
Is there a trend whereby the built environment is of decreasing
importance when it comes to home–work distance? (Paper II)
Does the built environment influence daily travel behaviour
differently when people regularly use ICTs and telework? (Paper
III)
1.3 Outline of the thesis
The heart of this thesis is the three papers in the appendices, referred to in the
text by their roman numerals. The papers are theoretically framed and their
results are discussed and concluded in the five chapters of this thesis summary.
This introductory chapter continues by broadly defining some of the central
concepts and setting the scene in the geography of Swedish daily mobility
trends. The theoretical framework of the thesis and a discussion of issues
explored in the papers are presented in the second chapter. The methods and
data are presented in the third chapter. The fourth chapter presents a summary
of each paper. Finally, the thesis summary is rounded off with a concluding
discussion on the joint theoretical and practical contributions and
recommendations derived from the empirical studies.
1.4 Central concepts and delimitations
Two groups of key concepts that play a central role in the thesis can be derived
from the aim and research questions. These concern (i) people’s daily travel and
mobility and (ii) the built environment. It is important that the reader be attentive to
how these concepts are broadly defined before reading the rest of the text
6
since they take on different meanings in academic literature and in everyday
life. Basically, these meanings concern the dimensions and scale levels at which
mobility and space are observed.
The terms travel and mobility are used mainly to broadly describe people’s
actual movements between locations and activities in geographical space. People’s
movements between places can be studied within different time periods and at
different levels of spatial scale, ranging from long-distance infrequent
international travel, on the one extreme, to short-distance movements. Against
that range, this thesis is concerned mainly with daily movements, including everyday
travel to work, services, and leisure. These daily movements are mainly local and
regional, not that there are no important interactions with other temporal
mobility scales. Trips and travel can be measured in three interrelated
dimensions – i.e., travel time, distance, and frequency – that are more or less
relevant to the problem studied. This study focuses on the geographical
dimension of daily travel, mainly in terms of travel distances. Travel distance is,
however, closely related to and a foundation for other aspects of daily mobility,
such as speed, mode choice, and time pressure (Banister, 2011). These and
other delimitations and operationalizations are discussed in the theoretical
framework following this introductory chapter. How daily travel is further
operationalized empirically is elaborated on in chapter 3, which presents the
data and methods used in the constituent papers of the thesis.
The built environment broadly designates the location patterns of
potential destinations that in various ways might affect people’s daily travel,
i.e., the locations of various facilities (e.g., stores, workplaces, and schools)
people potentially use to perform certain daily activities in relation to where
they live. The literature on built environment–travel relationships is vast and
various concepts describing the built environment are in use, including land
use, urban structure, urban form, accessibility, density, and proximity. One
important point to bear in mind is that the meanings of the concepts differ if
approached from a theoretical or an empirical angle. From an empirical point
of view, the meanings appear straightforward and much a question of what
built-environment features are being measured. For example, Seto et al. (2014)
defines the four main metrics of the built environment as density, land use
mix, connectivity, and accessibility. To be measurable, these metrics are
operationalized in often crude and simplistic ways, for example, density in
terms of population per neighbourhood and land use mix in terms of the ratio
7
of jobs to residents. In this dissertation, and in the Swedish case, access to
unique micro-level register data for the total populations of individuals and
firms, geo-referenced at a high spatial resolution, makes it possible to design
and test a wide range of metrics. This process is further described and
discussed in chapter 3. From a theoretical perspective, questions such as why
and under what conditions various features of the built environment can be
expected to be important to people’s daily travel are emphasized. The
theoretical framework presented in chapter 2 elaborates on these aspects and
inform the empirical operationalization and analysis presented later.
1.5 The Swedish case and context
Empirically, this dissertation analyses travel/built environment-related
patterns, processes, and developments in Sweden. Historically, the interlinked
trends of built environment, infrastructure, and travel in Sweden have been
similar to those of many other countries of the global north (Banister, 2012;
Frändberg and Vilhelmson, 2014; Metz, 2010; Millard-Ball and Schipper, 2010;
Vilhelmson, 2007). The beginning of mass motoring in the early 1950s
generated exponential growth in daily travel that continued for several decades
and allowed many people to move from cities to their rural hinterlands in the
1970s, resulting in increased urban sprawl. The average daily distance travelled
in Sweden increased from about 1 kilometre in the early 19th century to about
10 kilometres in the 1950s; from there, it increased rapidly, finally peaking a
few years into the 21st century at around 45 kilometres (Frändberg and
Vilhelmson, 2011). From 1978 to 2006, the average distances travelled for
work, leisure, and shopping increased the most (>50%), while school, child
care, social, and health care trip lengths increased only slightly (10–20%).
Sweden is now witnessing small reductions in total daily travel distances, the
stagnation of car use, and densification in the largest cities (Frändberg and
Vilhelmson, 2014). The car is the dominant mode of transport (accounting for
59% of all daily trips in 2006). Car use has peaked in recent years, but
continues to increase for some cohorts and trip purposes. For example, the
average distance commuted to work by car increased by 27% for men and 31%
for women between 1995 and 2011 (Elldér, 2014a). Sweden does not
distinguish itself from the rest of Europe in its modal split of passenger
transport and its motorization rates are similar to that of the rest of Europe
8
(Eurostat, 2014); in 2006 there were 464 cars per 1000 inhabitants in Sweden
compared with the average of 455 in EU 27.
In the case of ICT and virtual mobility use and access, there have been
drastic changes in recent decades (Vilhelmson and Thulin, 2008): for example,
between 1990 and 2000, Swedes spent 80% more time on ICT use, and only
30% had home Internet access in 1998, versus over 70% in 2005. Telework has
also increased rapidly in recent years: 17% of Swedish workers reported
teleworking regularly in 2011, versus only 10% in 2005–2006 (Vilhelmson and
Thulin, 2015).
Furthermore, the geography of the built environment and the
population differ from those of many other countries. In 2013, 9.6 million
people lived in Sweden at an average population density of 23.7 inhabitants per
square kilometre. In the European Union, only Finland has a lower population
density than Sweden (Eurostat, 2014). The geography of the Swedish
population, however, is highly diverse and includes metropolitan cities, small
towns, and very sparsely populated areas. The population is concentrated
mainly in the southern and coastal areas, as shown in Figure 1. In 2010, 85.1%
of Swedes lived in built-up areas
1
and 35.6% lived in the three largest urban
regions of Stockholm, Gothenburg, and Malmö (SCB, 2010). Urbanization is
slowly continuing into the 21st century, but at a faster rate in Sweden’s largest
urban regions. Furthermore, partly due to its geographically dispersed
population structure, Sweden has more transport infrastructure per capita than
do many other countries; there are, for example, 1.2 kilometres of railway and
0.2 kilometres of motorway per 1000 inhabitants in Sweden versus 0.6
kilometres of railway and 0.15 kilometres of motorway in the EU (Eurostat,
2014).
In summary, Sweden is a country with a highly heterogeneous built
environment, and – like many other countries of the global north – is
characterized by high degrees of physical and virtual mobility. Although recent
years have seen a slight decrease in daily travel distances, travel among certain
groups and for specific purposes continues to increase, and ICT use continues
to climb. The papers delve deeper into these developments because they are
1
The Swedish official definition of a built-up area (termed ‘locality’ by Statistics Sweden)
is any area with at least 200 inhabitants that also meet the criterion that houses are not
farther than 200 metres apart.
9
central to the discussion of changes in the relationships between the built
environment and daily travel.
Figure 1. Population distribution in Sweden, 2008 (source: GILDA).
10
2 Theoretical framework
2.1 Introduction
The theoretical understanding of daily travel has changed profoundly in the last
fifty years. In general, the research has shifted from a technical focus on
infrastructure development, network characteristics, and sheer flows of
vehicles, to a social-science orientation emphasizing human travel behaviour
and the role of transport in society and policymaking. If the mid-20th-century
focus was estimating aggregated transport flows in networks and between
zones in order to “predict and provide” for infrastructure construction, travel
is now more often approached as a behavioural or activity-based phenomenon
in the context of people’s use of time and space. In geography, this was
associated with a change of focus in several important ways: viewing travel as
an outcome of individual decision processes rather than derived directly from
locational patterns; disaggregating the level of analysis from aggregated flows
of traffic to individual (and household) trip making; and deriving travel from
individual needs to perform various activities distributed in time and space
rather than analysing trips isolated from their socio-spatial context. It is also in
individual capabilities, needs, and wishes that I take my theoretical starting
point, adopting a time-geographical activity-based approach, as outlined in
section 2.2. This approach considers people’s everyday life as a sequence of
activities, such as working, eating, exercising and sleeping, and notes that travel
stems from a need to schedule these activities in different places. Central to
this approach are the constraints to which this process is subject, for example,
the geographical accessibility of locations. As my thesis sets out to investigate
how and to what extent people’s daily travel is related to and determined by
the spatial opportunities provided by the built environment, I delve into these
spatial constraints in section 2.3. Finally, section 2.4 elaborates on certain
processes of socioeconomic and mobility change that may have important
implications for built environment–travel relationships, and therefore are
explored empirically in the papers.
11
2.2 An activity-based approach
2.2.1 Choice in the context of constraints
My main theoretical point of departure is in the human activity approach
(Axhausen and Gärling, 1992; Fox, 1995; Jones et al., 1983; Vilhelmson, 2007).
This approach emerged from mounting criticism of “positivistic” theories of
spatial interaction, network analysis, and gravity models (and associated
traditional trip-based forecasting methodology, four-step-planning models,
etc.) central to transport research and planning at the time (and often still in
use in practical planning). Heavily inspired by Hägerstrand (1970) and
Lenntorp (1976) – and the famous claim to pay attention to “people in regional
science” – Jones et al. (1983) developed the “human activity approach” to
better understand daily travel. Unlike the standard transport models of the
time, which mainly constituted statistical descriptions of trends and provided
no behavioural content, Jones et al. built on a growing contemporary literature
of human activity studies, especially a branch that studied activity patterns.
Building on Chapin’s (1974, as cited by Jones et al., 1983) work, human
activity patterns can be said to derive from individuals’ various physical and
physiological needs. The needs and wants of everyday life are translated into a
set of activities that are spontaneous, planned, or routinely scheduled into a
pattern (or a sequence of activities) in time and space each day. Some of
people’s basic daily activities, such as sleeping and cooking, are normally
performed at home. Many activities, however, demand specialized facilities
outside the dwelling in order to be carried out. To shop for groceries, we often
need to visit the grocery store, to attend a medical appointment we need to go
to the hospital, and so on. Daily travel behaviour can therefore be seen as the
result of a process whereby people match the demand for activities against the
supply of facilities (in both time and space) in preferred sequences. Travelling
is therefore understood as a demand derived from the need to perform
activities at geographically separated facilities. However, drawing on
Hägerstrand’s (1970) time-geographic approach, Jones et al. (1983, p. 266) also
emphasized that the process of meeting this demand is subject to several
space-time constraints levied by “physiological, economic and cultural factors”
and, as is central here, “the nature of space itself”. Hägerstrand (1970) initially
identified three main types of constraints: capability constraints refer to the
12
capabilities individuals have to perform activities, including allocating time for
biological needs and the mobility resources possessed; coupling constraints require
that individuals and objects come together in time (at certain points in time, in
a specific sequence, and for certain periods) and space (at certain physical
locations) for activities to be performed; and authority constraints regulate access
to facilities at particular times, for example, during working hours.
By synthesizing these two contrasting approaches – i.e., choice and
constraint thinking – into a basis for understanding daily activity patterns and
travel behaviour, Jones et al. added a behavioural aspect to time-geography (cf.
Vilhelmson, 2007), summarizing it as “choice in the context of constraints”
(Jones et al., 1983, p. 266). The core idea of “choice in the context of
constraints” also sums up the general theoretical impetus for the empirical
investigations reported in the appended papers. The implications of various
spatial constraints constituted by the location of facilities and inherent in the
built environment for the shaping of daily travel constitute the focus of all
three papers.
In sum, in this thesis, daily travel is theoretically understood as a demand
derived from people’s needs and wishes to perform daily activities that are geographically
dispersed; various types of constraints are crucial for understanding the spatial outcome (e.g.,
in terms of daily travel distance) of this process. A fundamental question, then, is the
extent to which travel relates to, or is determined by, the location patterns of
the built environment. The answer is not as obvious as it may appear at first
glance, and the human activity approach clearly emphasizes that other factors
contribute. It is therefore appropriate to conduct a deeper review of the central
constraints of daily activities, spatial ones in particular.
2.2.2 The spatiotemporal constraints of daily activities
The needed and wanted activities of daily life per se are not only an important
starting point for understanding the origin of travel, as their associated (and
varying) spatiotemporal fixity levels can be expected to have further key
consequences. First, an individual’s daily activity pattern – seen through a time-
geographical lens – results from solving an allocation problem in which both
time and space are limited resources (Hägerstrand, 1970; Jones et al., 1983).
Within time–spatial boundaries, people have to make trade-offs between
activities. The fact that activities are derived from various types of need – and
13
are more or less fixed at certain locations – has been an important feature from
the very start of activity-based approaches.
Efforts to classify needs and the daily use of time for various mobility-
related activities are central in this context. A basic categorization of activity
types in relation to need is typically made by roughly differentiating between
mandatory and discretionary activities. Chapin (1974, as cited by Jones et al.,
1983), for example, early on identified two main activity groups: activities that
satisfy subsistence needs (including sleeping, eating, and working) and activities
that fulfil culturally, socially, and individually defined needs. For better or
worse, such basic categorizations still underlie much activity-based analysis of
travel behaviour (Doherty, 2006). Ås (1978) provides a general framework for
placing activities in priority order based on freedom of choice and constraints,
as follows: 1. necessary time, 2. contracted time, 3. committed time, and 4. free
time. Note, however, that Ås does not attend to the location of activities and
time use – i.e., to the extent to which activities are fixed or flexible in space (cf.
Vilhelmson, 1999). However, the framework can be used as a basis for further
theoretical elaboration concerning spatial constraints on travel as follows (cf.
paper I).
People satisfy their basic biological needs during necessary time, which
includes activities such as sleeping, eating, and personal hygiene. These
activities are generally characterized by little flexibility and are often spatially
tied to the home (Ellegård and Vilhelmson, 2004). The second group of
activities are performed during contracted time, which mainly refers to paid work
and participating in education. These activities are often associated with
relatively little individual choice once they are decided upon (i.e., once a
“contract” is signed). The time–spatial premises (e.g., working hours and
location) are largely not individually determined. Activities performed during
committed time are predominantly linked to household work such as grocery
shopping. These are activities that also have to be carried out on a regular
basis, but could be expected to be associated with more individual flexibility
concerning when or where to perform them than are activities performed
during contracted time. People have greater abilities both to postpone such
activities and to decide where to perform them. Most committed activities,
however, must be performed during a limited amount of time, many during a
regular week, for example. The time people then have left can be considered
free time, which can be used for various leisure activities. Activities executed
14
during this time can theoretically also be expected to be the most flexible in
time and space.
Furthermore, the linkages between daily time use, prioritization of daily
activities, and spatiotemporal constraints are important when it comes to
analysing the geographical extension of daily travel for various purposes (i.e.,
activities). For example, mandatory activities such as sleeping and working are
often defined as fixed in time and space and are used as a basis when
modelling how other activities are distributed in time and space (Schwanen and
Dijst, 2003). Cullen and Godson (1975) early on demonstrated that these
activities and the places where they are performed are used as anchors around
which other activities are ordered. In later studies, work has been treated as an
activity that often tends to bind other activities in time and space (Schwanen,
Kwan and Ren, 2008). A person who is employed and must perform wage
labour eight hours a day at a specific location has limited opportunities to
engage in leisure activities far from their home and work locations. If this
person also has many obligations during committed time, his or her spatial
opportunities for daily activities are still further constrained.
Note, however, that conditions in society have changed since the early
conceptualizations of time-geography and activity-based approaches, possibly
influencing the need for activities, together with their prioritization and
associated spatiotemporal fixity for many people. For example, such conditions
include an increasingly knowledge-intensive and “flexible” labour market
(Green, 2004) and the associated rapid development and spread of ICT use
(Mokhtarian and Tal, 2013). These developments have also laid the
groundwork for criticism of the traditional categorizations of activities
discussed above and established a need to re-examine the original time-
geographic constraints as regards capability, coupling, and authority (e.g.,
Kwan and Schwanen, 2008). Some scholars have questioned the traditional
activity categorizations and sought more salient attributes of activities (see, e.g.,
Akar, Clifton and Doherty, 2012; Doherty, 2006). Using data from an in-depth
week-long activity scheduling survey, Doherty (2006) found significant
variations in spatiotemporal attributes between and within the traditional
categorizations. Another emerging literature studies the fragmentation of
activities in time and space enabled by ICT (see, e.g., Alexander et al., 2011;
Couclelis, 2000; Lenz and Nobis, 2007). The fragmentation hypothesis posits
that ICT weakens the relationships between activities, space, and time by
15
enabling many activities to be fragmented into smaller components distributed
in time and space.
These processes and their theoretical implications for the spatial
constraints of various activities and for the role of the built environment in
shaping everyday travel for various purposes are further discussed in section
2.4.
2.2.3 The role of individual needs, wishes, and capabilities
Though activities performed during necessary time, especially sleep, are very
fixed and must be executed by all individuals every day, the scheduling of most
daily activities in time and space is subject to individual needs, wishes, and
capabilities. There are important inter-individual variations both in what
activities are performed daily and in their associated spatiotemporal
constraints. In the activity-based approach, scheduling constraints at the
individual and household levels are therefore central to understanding travel
behaviour (Axhausen and Gärling, 1992; Fox, 1995; Jones et al., 1983). Jones et
al. (1983) paid considerable attention to the links and interactions among
household members, individual lifecycle stages, and daily trip patterns. The
other members of the household to which a person belongs contribute
significantly to shaping individual activity patterns. Individuals have to share
the available mobility resources within the household in terms of, for example,
car access or available time outside paid and unpaid labour. Many individuals
must consider other household members’ needs and constraints when planning
their own and joint activities. For example, if there are children in the
household, their needs must also be met and integrated into the scheduling of
the parents’ activities.
Much empirical work has shed light on the importance of these factors.
A common approach is to investigate how socio-demographic factors relate to
travel. Important factors found in most contexts include gender, household
composition and life course stage, income, education, car access, and daily
pursuits (in the Swedish context, see, e.g., Elldér, Solá and Larsson, 2012;
Frändberg and Vilhelmson, 2011; Öhman and Lindgren, 2003; Östh, 2007;
Sandow, 2011). All these factors mediate the capabilities that enable and
constrain daily activity spaces. Gaining access to a car, for example, greatly
widens the opportunities for individuals to choose where to perform activities
16
(Vilhelmson, 2007). Likewise, a higher salary can be invested in mobility
resources, giving opportunities for extended activity spaces (Swärdh, 2009).
Gender also has important implications for many reasons, including the
unequal division of unpaid work in the home limiting the capabilities for many
women to participate in certain activities (Solá, 2013; Solá and Vilhelmson,
2012). These factors are also closely related to the duties included among the
daily pursuits in which individuals are engaged (e.g., wage labour, education,
and household work), duties that are highly decisive for daily activity
participation and thus travel. Berg et al. (2014), for example, illustrated how the
transition to retirement fundamentally changes experienced time–spatial
constraints and mobility.
The present study further gauges the relative significance of the built
environment for daily travel in relation to individuals’ capabilities, needs, and
wishes – mainly operationalized as the socio-demographic factors mentioned.
These factors are closely related to other intervening factors highlighted in
recent literature as influential in shaping daily travel, for example, various
socio-cultural factors and related individual interests, attitudes, and desires (see,
e.g., Næss, 2013). As discussed below, personal desires and wishes can
challenge the importance of the built environment, for example, when distant
destinations are favoured over nearby ones. Similarly, also influencing
individual interests as well as preferred activities and destinations, social
networks and their influence on travel have recently been stressed (see, e.g.,
Dugundji et al., 2012; Tilahun and Levinson, 2011). Furthermore, personal and
socio-demographic factors are modified by the structural conditions of society
(cf. Næss, 2006), as there is a range of socially conditioned activities in which
many people must participate (e.g., wage labour and education). In addition,
lifestyles and attitudes are structurally influenced: individuals are situated in a
range of social contexts and networks that bring people together and enable
information sharing, affecting what activities are needed and wanted, and what
travel choices are made. Altogether, various factors operating at different scales
influence people’s capabilities, needs, and wishes for activity participation and
travel.
So far, I have built a theoretical understanding of the geographical
extension of daily travel as a consequence of our need and want to perform
activities that are separated in space. I have also emphasized that activities are
associated with various time–spatial constraints. These constraints vary
17
depending on the individual needs that activities fulfil, on how these daily
activities are scheduled in time and space, and on individual capabilities, needs,
and wishes. Such individual factors have received much attention in the
literature in recent decades as people (i.e., “actors” or “decision makers”) have
moved to the centre of transport studies, concurrent with a paradigm shift
(e.g., the behavioural turn, the time-geographic turn, and activity-based
thinking) in social science in general and in human geography in particular.
Arguably, the roles of geographical context and space have been downplayed
somewhat, at least concerning the potential relaxation of their role over time
and concerning the impacts of ICTs. Accordingly, the shift has resulted in
controversies concerning our understanding of the role of the built
environment in shaping daily travel patterns. These controversies and the
general role of spatial constraints and opportunities in daily activity
participation – the group of constraints in focus here – are discussed in the
next section.
2.3 The role of the built environment in daily travel
2.3.1 Facility constraints
In initial conceptualizations of the human activity approach, Jones et al. (1983)
identified “facility constraints” manifested by and related to the built
environment. Many of the needed and wanted activities of daily life require
specialized facilities adapted to the activities to be performed. In other words,
facilities such as schools, workplaces, public offices, shops, and parks shape
people’s activity possibilities as people visit them to perform daily activities.
Usually, several spatiotemporal constraints are encountered when matching
facilities and activities. Some facilities are only available during specific hours
and at a limited number of locations. Therefore, when individuals seek to
satisfy their personal demand for daily activities, there is a need (to various
extents) to match this demand with the supply of facilities. The fact that this
supply (manifested in the built environment) is unequally distributed in space is
a central departure point of this thesis and constitutes the basis of each of its
constituent papers. The main research questions share a concern with the role
of the built environment in people’s daily travel in various dimensions: when
travelling for various purposes (activities) (paper I); how this influential role
18
changes over time (paper II) and when certain activities (in this case,
teleworking) are performed in virtual instead of physical space (paper III). The
answers to these questions are not only of theoretical interest in understanding
the mobility dependencies of society; they also help us ponder the effectiveness
of spatial planning and policy measures intended to promote sustainable travel
(e.g., reduce distance travelled via energy-consuming modes of travel) by
changing the built environment (e.g., densifying cities and applying urban
planning principles based on compact cities, containment, and proximity).
This raises the question of how properly to observe and measure the
built environment and its supply of opportunities for human activity. A wide
range of operationalizations is presented in the literature. Various
characteristics – for example, proximity to destinations (including
agglomerations in city centres, jobs, and various service facilities), density (e.g.,
population or jobs per hectare), and land use mixes (e.g., job-to-worker ratios
and entropy measures) – can all be seen as proxies and measures of the varying
spatial supply of facilities mediating the geographical extension of daily travel. I
will elaborate further on this in the next section.
When it comes to theoretical explanations of spatial behaviour and
organization, the friction of distance and geographical proximity have
traditionally been central factors in geography (Couclelis, 1996; Miller, 2007).
As physical relocation costs resources (e.g., time, energy, and money), actors
are more likely to use geographically easily accessible facilities. As Couclelis
(1996) highlights, this is manifested in the widely cited Tobler’s (1970, p. 236)
“first law of Geography”, i.e., “everything is related to everything else, but near
things are more related than distant things”, and in many classical theories
addressing, for example, the size and geographical distribution of cities
(Christaller, 1933) and household localization (Alonso, 1964). The same line of
reasoning can be applied to daily travel in relation to the proximity to and
geographical accessibility of various facilities in relation to where people live
(Næss, 2006). Some people live in city centres with a high concentration of
activity possibilities, such as workplaces, stores, and various recreational
facilities, while others live peripherally, far from urban areas and
agglomerations of the facilities needed for daily activity participation. These
spatial constraints are often emphasized as highly influential on daily travel
(Ewing and Cervero, 2010; Newman and Kenworthy, 1999), while others note
that the friction of distance is increasingly being challenged by the spread of
19
rapid means of transport that relax the capability constraints on individuals’
daily reach (e.g., Kwan and Weber, 2003).
Before further elaborating on these relationships and discussions, it is
important to point out that the spatial variation in the use of facilities is closely
related to the individual needs and wants of daily life and the prioritization of
various activities. There are qualitative differences, and proximity is often
challenged by taste and preferences (Haugen, 2012; Næss, 2006, 2013). The
nearest grocery store might, for example, not be well-enough stocked to supply
the ingredients for a certain preferred dish. Other facilities are essentially
equally suited for the purpose of a certain activity; for example, all mailboxes
are equally suited for posting a letter. However, even for such facilities,
proximity to home is not always the most influential factor, as activities may be
performed at nearby facilities linked to other activities (e.g., shopping for
groceries on the way home from work) (McGuckin, Zmud and Nakamoto,
2005; Næss, 2006). The coupling constraints arising from coordinating an
individual’s sequencing of activities in time and space, that is, matching daily
needs with the supply of facilities, are important in this context. Following the
basic time-geographical postulate, it is necessary that individuals and facilities
be coordinated in the same place for a certain time. The grocery store, for
example, not only must be located near the workplace, but also must have
opening hours that coincide with the individual’s working hours. Such links
between activities, individual constraints, and spatial constraints exemplify how
these are not mutually exclusive, but closely interrelated factors. Accordingly,
the next section will discuss previously measured relationships between the
spatial supply of facilities manifested in the built environment and daily travel.
2.3.2 Relationships between the built environment and daily travel
The relationships between the built environment and travel demand have
grown to become one of the most researched subjects in the urban planning
literature (Ewing and Cervero, 2010). Boarnet (2011) describes the background
to the “explosion” of built environment–travel studies in the mid 1990s in
terms similar to those of the activity-based approaches. The rise and increasing
problems of congestion and emissions moved transport planning beyond the
traditional “predict and provide” approach to a more behavioural focus on
how to break habits, curb demand, and promote shifts from car-based to
20
transit-based mobility. Many scholars and policymakers then realized that the
geographical locations of trip origins (e.g., housing) and destinations (e.g., jobs)
could also be used to influence travel by using spatial planning as a tool to, for
example, place them closer together.
Research into daily travel and the built environment is typically
empirically focused (Boarnet, 2011; Ewing and Cervero, 2010). Many of the
studies are based on travel diaries and use statistical methods to test potential
relationships between various measures of the built environment and travel
behaviour. A general trend within this literature in recent decades has been an
increasing level of detail as regards spatial context and related characteristics,
while early studies often compared travel and built environment measures
aggregated to the city level (e.g., Newman and Kenworthy, 1999) or
neighbourhood level (e.g., Cervero, 1989). As regards spatial resolution, the
development of GIS and of computer power have allowed for more micro-
level analysis, which permits detailed measures of the built environment at
specific locations (e.g., travel time by different modes from home to the closest
grocery store) and of individual travel behaviour. More recent studies are also
micro-level in that they typically control for various individual variables (cf.
section 2.2.3). The crucial behavioural dimension – travel – is commonly
measured in terms of trip rates, travel distance and time, and mode choice.
The extensive literature on built environment–travel relationships
includes several reviews
2
and even reviews of reviews.
3
I will therefore provide
only a concise overview, and then focus on certain aspects relevant to the
research questions posed here. Overall conclusions often highlighted in the
literature are that people living in denser inner-city areas with diversified land use
travel shorter distances and use more sustainable transportation modes (e.g.,
public transport, cycling, and walking) than do others. Note, however, that
many scholars question such simplified generalizations, which will be further
elaborated on in the following sections. Each constituent paper of my thesis
also provides a detailed overview of the literature directly relating to its theme.
In an influential paper, Cervero and Kockelman (1997) summarize the
main features of the built environment associated with daily travel as the “Ds
2
See, e.g., Boarnet (2011) and Ewing and Cervero (2001, 2010); see Næss (2012) for a
review of studies performed in a Nordic context.
3
See, e.g., Gebel, Bauman and Petticrew (2007).
21
of the built environment” – now commonly cited in built environment–travel
studies – namely, density, diversity, and design. These features of the built
environment could all be seen as proxies for the varying spatial supply of
facilities constituting constraints on the geographical extension of daily activity
participation, as discussed above. Density typically measures the number of land
uses (e.g., population and jobs) within each area of analysis or the number of
land uses relative to the geographical size of the area (e.g., inhabitants per
square kilometre). Density is, to date, the most studied variable, being relatively
easy to determine. Diversity typically measures the mix of land use within the
area of analysis. A classical and simple measure is the job-housing balance
(Cervero, 1989), which captures the spatial mix of jobs and housing within
each area. Entropy measures of diversity taking into account many different
land uses have recently become more common (Boarnet, 2011). Design often
refers to the characteristics and design of the street network (Marshall and
Banister, 2000). Examples of characteristics are street density, cul-de-sac
occurrence, proportions of four-way intersections, presence of separated
bicycle lanes, etc.
Complementing these three Ds, more recent studies often use more
detailed proxies of the built environment by directly taking account of the
actual travel opportunities provided by transport systems. For example, as a
more detailed proxy of built-environment features, another D has been added,
namely, destination accessibility (Ewing and Cervero, 2001, 2010). Destination
accessibility takes into account the ease (often by various transport modes)
with which various desired activities can be reached from certain locations.
One concern is that many of these features are statistically often highly
correlated with each other; for example, densely populated areas often also
allow for more mixed land use and better destination accessibility. It could
therefore be difficult to evaluate whether or to what extent certain features of
the built environment exert a causal influence on daily travel. There is, for
example, recent discussion in the literature of local versus regional built
environment–travel relationships (Boarnet, 2011; Næss, 2011). Local
neighbourhood-scale variables assessing, for example, density or diversity
within a census tract have been found to be less important than regional
variables. Such regional variables as the distance to the closest city centre are to
some extent proxies for a range of other built-environment features (e.g., a
range of destinations, including jobs, education, and services, often clustered in
22
city centres). In the case of travel distance, a key aspect is the importance of
longer trips. Longer trips, outside the neighbourhood, constitute large shares
of the total distance travelled during a day, making regional-level variables
more important. Likewise, in light of increasing mobility, some authors speak
in terms of the regionalization of activities when daily activity spaces extend
from the local to the regional (Haugen and Vilhelmson, 2013), having
consequences for the relationships between the built environment and daily
travel. Haugen and Vilhelmson (2013), for example, found that while better
local-level spatial access is associated with shorter travel distances, a larger
supply of service amenities on a regional level leads to longer travel distances.
These issues of scale and causality place high demands on methods and data.
The traditional quasi-experimental design of neighbourhood comparison
studies (Saelens, Sallis and Frank, 2003) might miss such important dynamics.
Data capturing the built environment therefore need to extend geographically
far outside the residential neighbourhoods of the sampled individuals. The
analysis techniques must simultaneously be able to take into account various
scales, including the individual, neighbourhood, and regional scales. These
issues have important implications for the empirical approach of this thesis and
are further discussed in chapter 3 outlining methods and data.
The results of qualitative studies confirm the importance of regional
built-environment features in various contexts, for example in Hangzhou
Metropolitan Area, China (Næss, 2013) and Copenhagen, Denmark (Næss,
2006). Næss notes a strong tendency for interviewees to emphasize, rather than
the proximity of the closest facility, the importance of having the opportunity
to choose from a range of facilities. Daily travel demand could therefore be
expected to be more influenced by how the residence is located relative to
concentrations of facilities (e.g., city centres), in line with the arguments
presented above. Another key conclusion from Næss’s studies is the
significance of various individual factors, as introduced in section 2.2.3,
emphasizing the relevance of activity-based approaches. Travelling for
activities in free time can be expected to result mostly from personal interests
and preferences. Such discussions of the relative importance of the built
environment and individual characteristics and other controversies in the
literature are discussed in the following sections.
23
2.3.3 Controversies
Many scholars highlight the contradictory results and controversies within the
literature on built environment–travel relationships (see, e.g., Ewing, Deanna
and Li, 1996; Hamidi et al., 2015; Kitamura, Mokhtarian and Laidet, 1997;
Pontes de Aquino and Timmermans, 2010; Stead, 2001; Van Acker and
Witlox, 2011; van de Coevering and Schwanen, 2006; Van Wee, 2013). A key
aspect of these debates is the theoretical and methodological advances and
discussions in the research field as discussed above; in simplified (and static)
theoretical terms, these concern whether the approach assumes that geography
explains the process (i.e., location patterns shape individual travel-activity
decisions and choices) or that individual processes explain spatial behaviour
and outcomes. In an empirical sense, this is most evident in the increasing
inclusion of various individual factors in the analysis. When controlling for
various individual characteristics, similar built environmental proxies were
found to have no or contradictory effects on travel (Ewing, Deanna and Li,
1996). Giuliano and Small (1993, p. 1485), for example, argued that “evidence
is accumulating that in modern cities the effects of commuting cost are
swamped by variations in household characteristics, preferences and locational
amenities”. Handy (1996) concluded early on that more advanced research
methods analysing travel behaviour at the individual micro-level generally find
smaller built environment effects. To date, many studies in various ways
explore the relative importance of locational and individual characteristics for
daily travel. These studies reach divergent conclusions, some finding a strong
built environment effect (e.g., Shuttleworth and Gould, 2010), other find little
or no such effect (e.g., Bagley and Mokhtarian, 2002; Kitamura et al., 1997;
Weber and Kwan, 2003).
Van Wee (2013) lists several reasons why the results differ, including
research methods, geographical scale and scope, time horizon, factors
included, and indirect effects. Underlying many of these differences, there are
often various theoretical, methodological, and contextual explanations. A most
obvious explanation is often found in how travel behaviour is operationalized.
For example, Meurs and Haaijer (2001), when studying various trip purposes
defined by the number of trips during a week, found that travelling to work
was almost exclusively explained by individual factors. This comes as no
surprise, as the number of work trips can be expected to be a function mainly
of employment type and daily pursuit, while travel distance and time are not.
24
The geographical scope can also be expected to have consequences. For
example, large metropolitan areas give people many choices as to where to
perform activities and might relax built environment–travel relationships
compared with more sparsely populated areas. However, some authors go
further and argue that contemporary spatial behaviour is becoming more
complex due to technological and socioeconomic developments, a behaviour
that is increasingly disentangled from the built environment (Kwan and Weber,
2003; Miller, 2007). These discussions raise many important questions to which
I will return in the next section.
The idea of a simple and “direct causality” between the built
environment and travel is further challenged from other theoretical
perspectives. For example, based on economic utility theory, some scholars
view the features of the built environment mainly as factors that influence the
cost of travel by determining travel times (Boarnet and Crane, 2001, as cited by
Boarnet, 2011; van Wee, 2011). This view is part of a long tradition, notably
among urban economists, in which individuals are seen as homogeneous
rational actors weighing travel costs, wages, land rent, etc., to reach various
spatial outcomes (e.g., Alonso, 1964; Brueckner, 2000). Such perspectives,
however, are increasingly challenged by the complexities underlying
contemporary spatial behaviour derived from the need to perform activities, as
outlined throughout this thesis (cf. Giuliano, 1989; Kwan and Weber, 2003;
Mokhtarian, 2005). In this context, it is appropriate to comment on the
criticism of the view of travel as a derived demand (Mokhtarian, 2005;
Mokhtarian and Salomon, 2001). Ory and Mokhtarian (2005), for example,
highlight several occasions when travel is an end in itself, including for physical
exercise, variety-seeking, adventure seeking, and escape. This view obviously
challenges a narrow economic–rational approach treating travel as a disutility
to be minimized and also casts some doubts on activity-based approaches.
However, I see no direct contradictions between the theoretical understanding
constructed here and a focus on choice in the context of constraints and the
perception that travel is sometimes wholly or partly an end in itself. In fact, as
further discussed in the next section, this constitutes an important aspect of
my hypotheses. When more trips are not made in connection with activities
during contracted time (paper I), and when mobility technologies evolve and
become more available (papers II and III), the individual meaning of travel as a
25
(leisure) activity in itself might be more important than, for example, the
location patterns of the built environment.
Also challenging the direct causality is the discussion of to what extent
people weigh future travel abilities, needs, and preferences when they choose
their residential location, i.e., the residential self-selection hypothesis
(Mokhtarian and Cao, 2008). The self-selection hypothesis taken to its
theoretical extreme claims that the built environment has no effect on daily
travel because individuals freely choose their place of residence according to
their travel needs. However, this hypothesis is questioned due to, for example,
the fact that even though people largely weigh future travel needs when
changing residential location, the underlying location patterns of the built
environment still have an effect by allowing households to make that choice
(Næss, 2009, 2014). Cao, Mokhtarian and Handy’s (2009) review of 38 self-
selection studies also found no studies that did not establish a significant
travel–built environment relationship after controlling for self-selection. A rich
set of socio-demographic variables appears to account for most self-selection
(Bhat and Guo, 2007; Brownstone and Golob, 2009, as cited by Guerra, 2014).
In summary, a main lesson from the extensive literature on the
relationship between the built environment and travel is the importance of
placing studies in theoretical and geographical context, and pay attention to
individual constraints and opportunities. Daily travel is fundamentally a human
behaviour and is in that sense very complex and multifaceted. The
determinants work on several scales – individual, household, local, regional,
societal, etc. – and therefore need to be understood within these various
contexts. As touched upon in the following section, some of the
“controversies” might unravel if seen through an activity-based lens focusing
on spatiotemporal constraints. Given the activities that individuals want and
need in their everyday life and given the various time–spatial constraints and
opportunities to which these activities are connected, the built environment
can be expected to have a range of meanings and roles. In view of the
complexity underlying contemporary society, various processes and factors
could be expected to contribute to major differences in the geographical
extension of daily activities for individuals living in similar built environments.
It is therefore a key task for geographers to explore these processes and
factors, and specifically ask for whom, when, where, and how the location
patterns of the built environment influence people’s everyday geographies.
26
However, before embarking on such an empirical endeavour, it is
important to take a step back, and relate the built environment’s potential role
in and importance for personal travel to a broader, long-term view of socio-
spatial change associated with processes of time–space convergence. In the
next section, I therefore cite such conceptualizations to justify closer analysis
of the specific relationships between the built environment and daily travel
explored in the papers.
2.4 The changing role and importance of the built environment
The well-established concept of time–space convergence denotes the
fundamental process of how transport and communication technology affect
people’s use of geographical space. The concept was introduced in the 1960s
by Janelle (1969),
4
who demonstrated that the diffusion of transport
innovations brought places closer together in time and thus also in relative and
perceived space; i.e., when the time required to move between locations
decreases, the significance of geographical distance subsides. Typical of
geographical research at the time, Janelle developed a quantitative model
illustrating how increasing time–space convergence led to major changes in the
location of supply, in this case food stores in the western United States. The
concept of time–space convergence and its implications are now widely
acknowledged theoretically and empirically (Dodgshon, 1999; Harvey, 1989;
Knowles, 2006). The concept has become even more prominent because of
the rapid development of ICT and its potential impact on spatial processes and
behaviour (Kwan, 2002) – ultimately expressed in terms of the “death of
distance”. The convergence discussion is central to my thesis focusing on
everyday life and the role of the built environment in shaping people’s daily
activity patterns. The rapid time–space convergence occurring in Sweden in the
20th century was discussed in section 1.5 and is indicated by the access to and
use of physical space-bridging technologies that have caused the actual daily
distance moved to increase dramatically.
4
The concept of time–space convergence is closely related to the concepts of “time–
space compression” (Harvey, 1989) and “time–space distanciation” (Giddens, 1984)
that were introduced later but describes generally similar processes.
27
Time–space convergence is not only associated with increased activity
spaces, but also constitutes a process that increases the complexity of the
relationships between individuals, built environments, and travel, all in all
indicating weakening built environment–travel relationships (Brotchie, 1984;
Giuliano, 1995; Kwan and Weber, 2003; Miller, 2007; Mokhtarian and Tal,
2013). Spatial constraints generally relax when individuals gain more control
over where to perform daily activities, making travel patterns more
heterogeneous and varied. This relaxation of constraints has also had
repercussions for location decisions (e.g., where to locate shopping centres,
housing, and workplaces), producing more dispersed location patterns and
regional structures (cf. Brotchie, 1984) commonly referred to in terms of urban
sprawl and regional enlargement. In theory, time–space convergence presents a
radical contrast to the simplified theoretical hypothesis underlying some of the
literature on built environment–travel relationships, and to many of its
planning implications. This hypothesis is based partly on the idea that
individuals who share spatial constraints and opportunities (e.g., living in the
same neighbourhood and sharing identical built environments) largely share
similar daily activity spaces and patterns. However, apart from that, there
seldom exists such direct mono causality (cf. section 2.3.3): increasing
opportunities to overcome geographical distances probably increase the
importance of capabilities and constraints directly linked to individuals (and
their households), as the capacity to travel rapidly increases individual variation
in travel in terms of destination and distance. This raises concerns regarding
the traditional approach, which views individuals as more or less homogenous
actors who largely shape similar daily activity patterns according to the nearby
general opportunities that the built environment offers (cf. Miller, 2007).
From the perspective of this thesis, it is important to stress that few
empirical studies of daily travel and the built environment test the hypothesis
that spatial factors are declining in importance over time, i.e., time–space
convergence in ongoing, an issue also important in relation to discussions of
peak travel (Frändberg and Vilhelmson, 2011, 2014; Metz, 2010). Some studies
are exceptions, having presented varying results, but without finding evidence
of major changes over time in recent decades (e.g., Grunfelder and Nielsen,
2012; Guerra, 2014; Susilo and Maat, 2007; Zegras and Hannan, 2012). Paper II
contributes directly to this discussion by examining whether the relative
28
significance of residential location and the built environment has decreased in
Sweden in recent decades when it comes to commuting distances.
Another important aspect of time–space convergence concerns the
rapid development of ICTs. ICTs augment the space-transcending capabilities
of individuals and relax the coupling constraints of human interaction.
Research into the links between ICT and the demand for daily travel has a
fairly long tradition (Andreev, Salomon and Pliskin, 2010; Mokhtarian, 1990;
Mokhtarian and Tal, 2013; Salomon, 1986). This literature has traditionally
been interested in whether ICT substitutes, generates, or/and complements
travel, where substitution would clearly reinforce convergence in terms of
“distanciation” and complexity. More recently, the possible effects of the
modification and fragmentation of daily activities, time use, and mobility have
been examined (Lenz and Nobis, 2007). A general conclusion from this
literature is that ICT access and use increase the complexity underlying daily
travel choices (Dal Fiore et al., 2014; Mokhtarian and Tal, 2013). However,
there is little empirical scrutiny of how ICT influences built environment–
travel relationships (see paper III for a detailed review). The potential of ICTs
to ease various spatial constraints is often theoretically emphasized, for
example, by possibly replacing traditionally location-based activities (e.g., work
and banking) with virtual ones, and by creating opportunities for individuals to
choose more freely where to perform daily activities (Van Wee, Geurs and
Chorus, 2013). These discussions of the lessening of spatial constraints due to
virtual mobility provide a key justification for paper III’s focus on one
prominent ICT-based activity, telework. The ability of a person to telework
(often enabled by using ICTs) might relax daily spatiotemporal constraints and
erode the strong relationships between the built environment and daily travel
behaviour on workdays – an important issue to explore given the strong focus
on commuting trips in traditional distance- and location-based models used for
predicting and planning transport.
Processes of time–space convergence also relate to several other broad
social changes that potentially further complicate and erode relationships
between the built environment and daily mobility. Many of these changes are
linked to the rapid labour market changes of recent decades, changes
characterized by, for example, increased flexibilization and uncertainty, less
manual and more knowledge-intensive employment, increasing double-career
households, a smaller proportion of the total population gainfully employed,
29
and a larger proportion of the population in higher education and retirement
(Borggren, 2011; Green, 2004; Green, Hogarth and Shackleton, 1999; Östh
and Lindgren, 2012; Standing, 2011; Van Ham, 2002; Weber and Kwan, 2003).
Important topics in the context of these processes concern how people choose
between daily activities and travel purposes, what daily activities are actually
being carried out, and how the time–spatial constraints associated with certain
everyday activities have changed (cf. section 2.2.2). For example, fewer people
are permanently employed with distinct time–spatial boundaries, instead being
engaged in flexible working practices and non-manual jobs (Green, Hogarth
and Shackleton, 1999). This increases commuting tolerance and opportunities
to plan where and when work is performed (e.g., by teleworking), relaxing
work-related spatiotemporal constraints for some workers.
5
In turn, this may
also change the priority order of everyday activities. For example, certain
activities conducted during committed time (e.g., picking up and dropping off
children at school) could become more constrained in time and space than
contracted-time activities. Furthermore, as larger proportions of the population
retire, while others are engaged in higher education, larger shares of these
groups’ daily activities will be associated with committed and free time,
possibly accompanied by greater spatiotemporal autonomy (especially in free
time). For these reasons, it is important to study whether the relative
importance of the built environment and individual choice to travelling
behaviour in contemporary society is conditional upon trip purpose (i.e.,
activity) – a main justification for paper I.
In summary, the described developments raise many important
questions. When opportunities to travel (physically as well as virtually) increase,
the importance of the (local) localization patterns of the built environment for
daily mobility can be hypothesized to decrease, while individual-related factors
might play an increasingly prominent role. This hypothesis needs to be
examined empirically in various respects. Is it valid for all travel purposes? Is
the built environment really decreasing in importance over time? What it is the
role of ICTs in this context? Answering these questions requires a theoretical
understanding of the determinants of travel, of the importance time, scale, and
5
This “flexibilization” also has many downsides for individual workers (Standing, 2011).
Under precarious working conditions, for example, it is more difficult to make longer-
term mobility decisions about, for example, residential relocations, and people can be
forced to endure long commuting distances.
30
the varying spatiotemporal constraints of daily activities, of individuals’ needs
and wants in daily life, and, not least, of how to properly operationalize the
built environment. These factors were all important when designing the
empirical studies, as described in more detail in the following chapter.
31
3 Data and methods
3.1 Point of departure
When outlining the theoretical framework, I argued that engaging with built
environment–travel relationships in contemporary society makes specific
demands in terms of the data needed and methods to be used. Besides focal
information concerning individuals’ daily travel behaviour, it is also crucial to
characterize their capabilities, obligations, as well as daily activities that involve
travel. Furthermore, there is a need for detailed information on spatial context,
i.e., the facilities inherent in the built environment on different spatial scales,
and on the physical opportunities to reach these. Conducting geographic
research within the Swedish context therefore has its advantages, because of
good access to detailed, high spatially resolved micro-data covering the total
population and spanning comparatively long periods of time. My thesis seeks
to benefit from this comparative advantage and advances the research field by
applying a multivariate quantitative approach (e.g., multi-level statistical
models) to a unique combination of such rich data sources. In other words, it
is the quantification of actual behaviour, distances, built environments,
background factors, etc., that is in focus. This is associated with a number of
ontological possibilities and limitations (Næss, 2004; Sayer, 1992). For example,
the thesis does not extend to personal attitudes, meanings, and desires,
excluding individual perceptions and valuations of geographical proximity in
various respects.
This chapter provides a general overview of data and methodological
considerations, while further details can be found in the constituent papers of
the thesis. The sources of data are presented in the next section. This is
followed by an account of the empirical approach, including empirical
operationalizations of key concepts, statistical methods, and a discussion of
important limitations.
3.2 Data
3.2.1 Complementary data sources
Three basic complementary data sources are used (in different combinations)
32
in the thesis’ papers. These data enable a joint description and analysis of how
individual daily travel activities, given varying individual mobility capabilities,
are associated with spatial opportunity structures, i.e., the underlying built
environment patterns of transport infrastructure, housing, workplaces,
services, and other facilities. Information about spatial opportunities at the
local and regional levels, as well as certain information about individual
mobility capabilities, are retrieved from the Geographical Individual
Longitudinal Database for Analysis (GILDA). GILDA comprises geocoded
micro-data on every individual and organization in Sweden. Information
concerning actual travel activities is retrieved from the Swedish National Travel
Survey – RES. RES includes survey data on all everyday trips of large
randomized samples of the Swedish population. Furthermore, a GIS-based
accessibility tool adds information about potential access to opportunities
facilitated by transport networks. GILDA is used in all three papers to define
various characteristics of the built environment, and in paper II also to define
certain individual-level variables. RES is the main source of data used in papers
I and III to define both individual characteristics and travel behavioural
variables. The highest possible geographical resolution of residential
neighbourhoods is then used to integrate information from RES and GILDA.
In addition, paper III takes into account actual opportunities to reach
destinations provided by the public transport and car-road networks. The rest
of section 3.2 further describes these data, followed by an account in section
3.3 of how the data are used in the empirical analysis (see also Table 1).
3.2.2 GILDA
The GILDA database is administrated by the Human Geography unit at the
University of Gothenburg, Sweden. It comprises official register data provided
by Statistics Sweden from the database Longitudinell Integrationsdatabas för
Sjukförsäkrings- och Arbetsmarknadsstudier (LISA) (SCB, 2011). Annually
updated data, starting from 1990, from educational, income, employment,
health insurance, and population registers are integrated in this database. It
holds data on every individual, sixteen years or older, registered in Sweden as
of the 31st of December each year. A wide range of variables regarding
demographics, education, employment and unemployment, income, and social
security are included. Gainfully employed individuals are linked to their
33
workplaces, which are characterized by several variables, for example, industry
sector and number of employees. From a geographical perspective, a key focus
in my thesis is the inclusion of geographical coordinates for the places of
residence and work of each individual (with a 100-metre resolution). This
information is fundamental to the research designs applied in the papers, as it
allows the definition of variables that describe the spatial characteristics and
locational patterns of the built environment (see section 3.3.1).
3.2.3 National travel survey
RES comprises data on the everyday travel of the Swedish population aged 6–
84 years and is conducted by Transport Analysis, a governmental transport
policy agency. Paper I is based on RES 2005-06 (SIKA, 2007) and paper III on
RES 2011 (Trafikanalys, 2012). The initial sample of RES 2005-06 included
40,928 individuals and the response rate was 68%. RES 2011 initially included
39,596 individuals and the response rate was 43%. The survey is carried out by
phone interviews referring to prepared individual travel diaries. Every
individual is randomly assigned a survey day during the studied period and the
diary is sent by mail a few days beforehand. Information about two main types
of trips are included in RES: all trips taken on the survey day and any trips
longer than 100 kilometres taken during the month before the survey day.
Since the focus here is on daily travel, only trips taken on the survey day are
used. Furthermore, the respondents are also asked questions about various
background conditions (e.g., income, household situation, and car access) that
may affect their travelling behaviour.
All relocations outside the respondent’s dwelling are surveyed,
including short walks. These are defined and arranged in a hierarchy of main
trips, partial trips, and trip elements. A trip is defined as a main trip if it starts
and ends at a facility defined as a main trip location. A main trip location is the
respondent’s main dwelling, holiday house, workplace, school, or temporary
overnight abode. The main trip consists of one or more partial trips, which
emerge when the respondent performs an errand outside of a main trip
location. The partial trip is divided into trip elements when the respondent
changes travel mode. Furthermore, each main trip, partial trip, and trip element
is characterized by a range of variables. For example, respondents are asked to
34
report the main purpose of main and part trips, while for trip elements they are
asked to state, for example, travel mode, distance, and time.
3.2.4 Accessibility calculations
For paper III, which is delimited to the greater Gothenburg area, GILDA was
complemented with a GIS-based accessibility tool
6
for the purpose of more
precisely capturing the potential given by transport systems to reach built
environment facilities (e.g., workplaces and stores). This tool computes travel
times by car and public transport between 500 × 500-metre cells using a
detailed “door-to-door” approach. Because of this comparatively high
geographical resolution, it was impossible to run such an analysis nationwide.
The tool uses several input data sources. The computing of car travel times is
relatively straightforward. First, the centre point of each cell is linked with a
straight line to its closest road segment. The road network is compiled from
the official Swedish road database (NVDB) provided by the Swedish Transport
Administration. Second, travel time is computed by means of simple shortest-
path analyses based on the speed limit and length of the road links. Travel
times by public transport are drawn mainly from the timetable database of the
regional public transport authority (Västtrafik). Based on this data, an origin–
destination matrix of travel times between all public transport stops within the
study area was compiled. The selected travel times describe the fastest
connections when arriving at each stop within a timeframe set to 7–8.30 am on
a regular weekday. Finally, the estimated times it takes to walk via the road
network to the closest public transport stop from the starting and destination
cells are also added to the travel times. These combinations of data make it
possible to define detailed built-environment variables by measuring travel
times by car and public transportation between the residential location of the
respondents in RES and key destinations defined using GILDA. One such
variable is how many jobs can be reached within a certain travel time by car or
public transport from the respondent’s home location.
6
This tool was initially developed by Dr. Svante Berglund, KTH Royal Institute of
Technology, for Skåne, Sweden. I developed it further as part of a research project
about accessibility planning for Västra Götaland, Sweden. For further description of the
tool and its underlying data, see Larsson et al. (2011) or Larsson and Elldér (2014) in
English.
Table 1. Basic information regarding the data and methods used in the papers.
Paper I
Paper II
Paper III
Aim
Examine whether residential
location relative to individual
attributes affects daily distance
travelled when individuals travel
for various purposes.
Analyse how home–work
distance varied among workers
and across residential areas in
Sweden, 1990–2010.
Examine how urban structure
relates to the daily travel
behaviour of teleworkers
compared with workers who do
not telework.
Data source
RES and GILDA
GILDA
RES, GILDA, and accessibility
instrument
Method
Multilevel modelling
Multilevel modelling
Pearson correlations and OLS
regressions
Study area
Sweden
Sweden
Gothenburg, Sweden
Time period
2005/2006
1990, 1995, 2000, 2005, and 2010
2011
Number of
observations
12,370
140,531–171,859 depending on
the year studied
2500
Definition
of travel
Total daily distance travelled for
all purposes, including work,
services, and leisure, on both
weekdays and weekends
Euclidian home–work distance
Total daily travel distance and
time on workdays
35
36
3.3 Methods
3.3.1 Empirical considerations
This section presents and discusses overall empirical considerations such as
concept definitions and the treatment and merging of data. More detailed
considerations regarding, for example, variable definitions and data selection
procedures are described in each paper.
Travel-activity locations of origin
In measuring travel behaviour, I take operational departure in people’s
residential location. This plays a central role in shaping the regularity of daily
activity patterns as most daily trips start from and return to people’s homes
(Ellegård and Vilhelmson, 2004). The home is also the only place that can
reasonably be considered of relatively equal importance across various
population groups and activities. This increases the possibility of making
comparisons between the three studies. Besides, the residential location is of
practical importance as it serves as a data merging key in the empirical designs.
For ethical reasons, it is impossible to identify and match information on the
same person in RES and GILDA. However, geo-referenced information
makes it possible to build detailed spatial variables using GILDA and to link
these to the actual travel behaviour of individuals defined by RES.
However, locations other than the home might also be important
anchoring points for the organization of a person’s daily travel activities in
specific cases, and have when possible been considered in the research.
Arguably, this is particularly relevant to workplaces, as work often acts as an
anchor around which other activities are scheduled (Saxena and Mokhtarian,
1997; Schwanen, Kwan and Ren, 2008). As GILDA provides detailed
information on the workplace location, this was explored when building the
statistical models for paper II. The results suggest that the workplace location
also constitutes a source of variation in commuting distance, but that the
residential location is much more important. Other daily activity nodes,
especially children’s schools and day-care centres, could sometimes be
expected to play a part but are not covered in the data sources.
37
The issue of spatial resolution
Once the location on which to base the empirical analyses has been chosen,
new challenges are encountered. A first operational dilemma, also highlighted
by previous research, concerns the definition and spatial delimitation of
residential areas (Horner and Murray, 2002; Kwan and Weber, 2008). In many
studies, this is a main limitation because data are often based on various
administrative areas that are more or less suitable in terms of both travel
behaviour and the built environment. Built environment–travel relationships
have traditionally largely been studied at an aggregate level, sometimes that of
whole cities (e.g., Newman and Kenworthy, 1999) and often at the
neighbourhood level (Saelens, Sallis and Frank, 2003). The many pitfalls
associated with using different area definitions and aggregations – i.e., the
spatial resolution – have long been highlighted (Openshaw, 1984; Robinson,
1950). When data are grouped into spatial units, important individual variation
might be missed, and different area definitions can also create different results.
The highest spatial resolution in RES is delimited to Small Area Market
Statistics (SAMS) areas and is therefore used as a basis for capturing the built
environment in all papers (the second highest spatial resolution is that of
municipalities, which are considerably less detailed). SAMS areas are statistical
areas defined by Statistics Sweden for the purpose of representing residential
neighbourhoods (Amcoff, 2012; SCB, 2005); there are about 9200 SAMS in
Sweden, with an average of around 1000 people living in each. Like most area
definitions, the validity of SAMS is open to question. In the geographical
literature, most prominently as regards neighbourhood effects, SAMS are
sometimes pointed out as highly problematic since the number of individuals
included in each area varies between zero and tens of thousands (Amcoff,
2012). The most important aspect as regards area definitions in my thesis,
however, is that these are homogeneous in terms of the built-environment
features theoretically expected to influence daily travel distances.
Geographically, the SAMS are quite detailed within cities, but can be very large
in the sparsest-populated parts of Sweden. This makes it more problematic to
portray the built environment within and near SAMS far away from urban
agglomerations and to define exact geographical points for defining variables.
One measure taken to minimize potential biases has been to use the
geographical potential of GILDA and define a demographic centre of gravity
based on all inhabitants registered in each SAMS. Furthermore, when fitting
38
models for paper II, the high spatial resolution of GILDA was also used to
evaluate different area definitions, and both administrative areas (i.e.,
municipalities, postcode areas, parishes, and SAMS areas) as well as 1- and 5-
kilometre cells were used. If much larger areas are not used (e.g.,
municipalities, which give smaller built environment effects), the overall results
are similar.
Spatial characteristics of the built environment
A wide range of variables that capture various aspects of the built environment
on different scales – local (e.g., within residential areas, i.e., SAMS) as well as
regional – defined from GILDA has been evaluated. A typical local built
environment variable is the population density within SAMS. The more
regional variables can, for example, describe the distance from the population
centre of each SAMS to its closest city centre. Papers I and II use Euclidian
distances, but paper III uses accessibility calculations to account for the
opportunities provided by cars or public transport. It is thus possible to
compare the results of variables describing the proximity to similar facilities
but with different distance constructs (Paper III), and to evaluate potential
biases in more crudely defined variables (Papers I and II). A great many
variables that capture different aspects of the built environment are defined
initially for each paper, for example, the shortest distance to various facilities
(e.g., grocery stores and day-care centres), density measures (e.g., population,
jobs, and streets), and regional location (e.g., distance to closest city or
suburban centre). The papers further describe the selection and exact
definitions of the built-environment variables used. Paper III also includes a
more detailed list and evaluation of different variables.
Measuring trips
The main dependent variable in all three papers is travel distance. Distance is
chosen as it underlies other travel dimensions such as speed, time, and
frequency (cf. section 1.4). Focusing on distance (in contrast to time and
speed) also more explicitly addresses sustainable mobility, including
environmental impact (e.g., energy use and GHG emissions), social
differentiation (e.g., equal access), and wellbeing and health (e.g., physical
activity) (Banister, 2008, 2011). However, though measuring travel distances
may appear straightforward it can be problematic. Various methods can be
used, including self-reporting, Euclidian distances (calculated from trip origins
39
and destinations), GPS tracking, and vehicle miles travelled, each having
various benefits and drawbacks (see e.g., Bricka and Bhat, 2006; Sparks, Bania
and Leete, 2011; Wolf, Oliveira and Thompson, 2003). The use and
comparison of multiple data sources, in which distances are measured
differently, again provides opportunities to assess potential biases. In papers I
and III, based on RES, the respondents report the distances travelled in a
survey. The dependent variable in paper I is defined as the total daily distance
travelled for different purposes, including work, service, and leisure on both
weekdays and weekends. Paper III uses the total distance travelled on the
survey day for all purposes. In both papers, every trip (via all transport modes)
is included. Trips are observed regardless of how long (or short) they are, but
do not include travel abroad. For paper II, there is no information available on
the actual travel distance. The geographical information in GILDA is instead
used to calculate Euclidean distances between home and work. To reduce
possible biases (e.g., weekly commuters residing in temporary housing on
weekdays), workers registered as living more than 200 kilometres from work
are excluded.
To sum up, an overall intention has been to take advantage of the
particular qualities of different data sources to strengthen the validity and
reliability of the empirical operationalizations and analyses. The geographic
dimension has been central here, not only in practically integrating the data,
but also to increase comparability between studies. The rich datasets provide
good opportunities to operationalize and analyse the relationships between the
built environment and travel. The next section briefly discusses the analytical
techniques used.
3.3.2 Statistical methods
The complexity of the relationships between the built environment and daily
travel, combined with the high resolution of the data, places specific demands
on the statistical analysis techniques. In light of the many previous studies that
have found strong effects of various individual factors, an initial basic
requirement is that the approach should allow for multivariate analysis in
which such potential covariates are controlled for. Both GILDA and RES
contain a number of variables describing individual and household-related
characteristics previously found important. Each paper includes a set of such
40
variables comprising, for example, car access, income, education, and family
situation. This is to ensure that the variation in travel behaviour potentially
explained by the built environment is not due to, for example, certain groups
of people being clustered in similar geographical contexts. Both GILDA and
RES generally provide good opportunities for this, the main difference being
that GILDA does not include any information about driving licensing or car
access. As expected, in the empirical analysis in papers I and III based on RES,
these variables proved important. Part of this variation is probably captured by
other socioeconomic variables available in GILDA, for example, income and
life course.
Two main multivariate statistical analysis techniques are used:
multilevel linear modelling using MLwiN software (papers I and II) and
ordinary least squares (OLS) regression modelling using SPSS software (paper
III). Multi-level models are an umbrella term for various statistical models that
all address the hierarchical or clustered structure present in most behavioural
data (Goldstein, 2011). There are natural hierarchies in many datasets,
especially those that involve spatial information. For example, people living
within the same residential neighbourhood tend to be more like each other
than are people randomly chosen from the rest of the population (e.g., the self-
selection hypothesis discussed in section 2.3.3). This data clustering is ignored
in standard regression models, which assume independence between
observations. Multilevel models recognize that such hierarchical data levels
exist by allowing effects to work at each defined data level. Several facets of
the importance of such scale issues have been discussed. The determinants of
daily travel distances could theoretically be expected to work at various scales.
Individuals are situated in and share different contexts, such as residential
neighbourhoods, regions, workplaces, and households, that need to be
considered analytically. In addition, the effects of various characteristics of the
built environment have also been found to differ across scales. Multilevel
models are specifically designed to address these matters. The hierarchical
structure of the models also provides additional analytical opportunities that
have been important in answering the research questions in papers I and II. By
dividing the analysis into different hierarchical levels (i.e., individuals clustered
in residential areas), for example, it is possible to derive the similarity of the
distances travelled by individuals living in the same neighbourhood.
41
3.4 Limitations
This section highlights some key limitations as regards data and methods. First,
the analysis is mainly limited to measuring the impact of the built environment
on the distance that individuals travel daily. Although distance is fundamental
in many respects, people’s daily travel-activity patterns are obviously
multifaceted, incorporating other important dimensions for which the built
environment can be expected to differ in role and importance from that of the
distance context. Travel time is one example, and has to some extent been
taken into consideration by being included as the dependent variable in paper
III. In this case, the built environment proved to be less significant for daily
travel time than it was for distance, but displayed similar inter-group
differences. However, the relationships between different travel dimensions
(e.g., distance, time, and speed) might vary depending on trip purpose. The
distance to work, for example, can be expected to be covered in a shorter time
(i.e., at higher speed) than leisure trips, which are often slower. This is only
partly captured in travel distance. In relation to this, it may also be noted that
the empirical models focus on estimating changes in the mean of the
dependent variable travel distance. Like many other variables that describe
spatial interaction, travel distance is highly positively skewed and, again, the
distribution can differ among various travel activities. Taking discretionary
travel as an example, some trips might be short distance (e.g., walks in the
neighbourhood) while others are considerably longer (e.g., road trips on the
weekend). Consequently, models of the mean might miss explanations of parts
of the distribution. Fitting models with other assumptions, such as quantile
regressions based on the median, might change some of the inferences. This
particularly applies when large outliers, i.e., very long distances, are present in
the analysis. However, since the focus is on daily travel (and different
delimitations have been applied to this end), few outliers are present in the
various datasets. If less frequent long-distance travel had been included, the
risk of biases would have been higher.
Furthermore, the population of the empirical studies does not include
children 16 years old or younger or people older than 84 years. The daily travel
patterns of these excluded groups are likely to differ than from those included
in the analysis of the thesis. Finally, it is important to acknowledge that the
cross-sectional nature of the empirical designs makes it difficult to definitely
infer the direct causal effects of the built environment (Cao, Moktharian and
42
Handy, 2009; van de Coevering, Maat and van Wee, 2015). In particular, the
scheduling of daily travel is in various ways interrelated with longer-term
mobility decisions and habits, such as the decision to buy a car or move one’s
place of residence. Experimental designs and longitudinal approaches could
allow more robust conclusions regarding causal relationships and effects;
however, such conclusions would still be difficult to reach considering the
complexity of factors contributing to individual travel behaviour and that daily
travel occurs within relatively “open systems” (Næss, 2004, 2015). As
elaborated on in the theoretical framework description (e.g., section 2.2), the
built environment does not in itself create travel, but is one of many
contributory factors. To discover the incentives and mechanisms underlying
the measured relationships between the built environment and daily travel,
qualitative methods are also needed. Note finally that each paper further
discusses its specific limitations. Related recommendations for future studies
are given in section 5.2.
43
4 Paper summaries
Paper I
This paper takes its point of departure in the often cited controversies
regarding the extent to which proximity and the built environment relative to
individual choice explain daily travel demand. It contributes by investigating
whether the relative importance of these factors to travel distance changes
when different travel purposes are considered. Trip purpose is examined from
an activity-based perspective focusing on the varying spatiotemporal
constraints of daily activities. For this purpose, multilevel statistical models are
applied combining two rich sources of geocoded micro-level data: Swedish
National Travel Survey data from 2005–2006 captures individual travel
behaviour while Swedish register data capture the built environment. Separate
models are fitted to examine the extent to which everyday travel distances for
various purposes vary among individuals who share residential locations. Three
main categories of trip purposes are modelled: trips to activities performed
during contracted time (work trips), committed time (e.g., service trips to the
grocery store and for health care), and free time (e.g., various leisure trips).
The results indicate that the influence of residential location and the
built environment on daily distance travelled is conditional on trip purpose in a
nationwide Swedish context. Statistically significant proportions of the
variation in daily distance travelled to work and, to a lesser extent, on service
errands are dependent on proximity to various features of the built
environment. Daily distances travelled for leisure activities vary greatly among
people living in the same neighbourhood, however, and in this context,
variables capturing the built environment were found to have no significant
effects. These results confirm the importance of considering the time–spatial
constraints of activities when examining the relationships between location,
individual characteristics, and travel behaviour. Also, from a policy perspective,
these results suggest that measures intended to alter the built environment to
reduce the volume of travel will be most effective in addressing work trips,
while trips taken during free time will likely be little affected. In addition, the
multilevel models applied reveal several important interactions between
variation in travel distances across residential locations and individual
characteristics of which researchers should be aware, especially when
examining service trips.
44
Paper II
Some scholars emphasize that everyday travel behaviour is becoming less
location based and increasingly dependent on individual choice, mainly due to
increasing mobility capabilities. Few studies, however, apply a dynamic
approach and attempt to quantify how or in what direction the relative
significance of built environment factors for everyday travelling has developed
in recent decades. Doing so calls for further empirical examination using
appropriate data and statistical models. The empirical analysis presented in this
paper focuses on work travel: as it is generally more closely related to the built
environment than are other types of travel (cf. paper I), it is often emphasized
in traditional distance- and location-based models and policies used in
predicting and planning transportation. Trends in the relative significance of
residential location for home–work distance in Sweden, 1990–2010, are
explored. The analysis has two aims: first, to evaluate to what extent workers
who live in the same residential area travelled similar home–work distances
during the studied period and, second, to evaluate whether there has been any
change in the effect of individual and spatial characteristics on the home–work
distance. The paper uses official register data from GILDA covering the home
and work locations of the total Swedish working population.
Two results from the analysis that are important in relation to previous
research can be highlighted. First, the explained variance in commuting
distance attributed to the individual level is low compared with that found
elsewhere. Most previous studies were based on single metropolitan areas that
are comparatively geographically homogeneous and generally include better
choice opportunities as regards possible job locations. The nationwide Swedish
case, however, includes a wide variety of built environments ranging from large
urban agglomerations to very sparsely populated areas. The most important
result, however, concerns changes over time. The results indicate growing
variation in home–work distance for workers living in the same residential
neighbourhoods, indicating that the significance of residential location for
home–work distance decreased throughout the studied period. This leads to a
conclusion that relationships between daily travel and the built environment
could indeed change over a few years. This possibility calls for studies
exploring whether changes are occurring in other dimensions as well, for
example, as regards additional travel activities or in other geographical
contexts.
45
Paper III
The third paper studies work-related travel in detail, considering how it relates
to the built environment by focusing on a key ICT-based activity, i.e., telework,
that has increased rapidly in recent years in Sweden. It is often argued that, due
to its space-transcending abilities, ICT use reduces the importance of proximity
as an organizing principle of travel. However, as telework has so far not been
that widespread, most previous studies use small samples, unrepresentative of
entire workforces. More precisely, this paper investigates whether the built
environment influences daily travel behaviour differently when people
telework in urban contexts. Regression models are applied to address whether
and to what extent travel is associated with various measures of the built
environment and key destination accessibility relative to the home location in
Gothenburg, Sweden. The analysis treats groups of workers defined by
teleworking practices. Micro-level data from the Swedish National Travel
Survey, 2011, capture individual travel behaviour, and Swedish register data on
the location of all firms and individuals combined with a GIS-based tool
measuring travel times by car or public transport capture the built
environment.
The results indicate that telework weakens the relationship between the
built environment and daily travel in various dimensions. First, workers are
differentiated in terms of those who did and did not report teleworking
regularly. The total daily travel distance and time are significantly related to the
built environment for both groups, but the built-environment variables explain
considerably more of the variation in travel for non-teleworkers. Second, only
regular teleworkers were analysed, and differentiated based on whether or not
they teleworked on the survey day. For those respondents, the built-
environment variables explain little of the variation in travel time and distance,
and retain few significant effects when fitting the models. These results lead to
the conclusion that telework allows various mobility strategies that together
foster more spatially heterogeneous daily travel behaviour, more dependent on
personal attributes than on the home location relative to various built-
environment features. Planners and policymakers should monitor whether the
number of teleworkers continues increasing. If so, traditional distance- and
location-based models and policies for predicting and planning transport may
prove less accurate and effective than currently assumed.
46
5 Concluding discussion
5.1 Main conclusions
Geography, in terms of the built environment and location patterns, was
traditionally, and still is, emphasized by many scholars, policymakers, and
planners as greatly influencing people’s daily travel behaviour. However, taking
recent decades of rapidly increasing mobility capabilities (physical as well as
virtual) into account, and the related increase in individual choice
opportunities, others argue that the importance of geographic factors has
gradually dissolved. Starting from this discussion (and controversy), the overall
aim of this thesis is to examine the current role and relative significance of the
built environment for the geographical extension of individuals’ daily travel in
Sweden. Particular attention is paid to detailing the impact of geographic
factors on various daily travel activities, including the potential relaxing of the
relationships between locational structures and travel behaviour enabled by
space-transcending ICT use, and to detecting actual changes over time. This
final chapter proceeds from these concerns and outlines the main findings
from my three empirical studies and relates them to an overall framework and
understanding of the issues at hand. The main conclusions are first
summarized in this section, which is followed by discussions of implications
for further research (section 5.2) and of practical lessons (section 5.3).
An overall conclusion of my thesis is that the proximity of various aspects of
the built environment to home still plays an important role in how far people in Sweden
travel daily. However, my analyses, informed by theory emphasizing everyday
spatiotemporal constraints and using rich sources of geocoded micro-data,
reveal that these relationships have become relaxed in several important
respects. First, the specific time–spatial constraints associated with different
daily activities that motivate trips and travel are key and also differentiating factors when it
comes to this relaxation. When considering trips taken during holidays and for
everyday leisure purposes, the built environment is less important for the
observed daily travelled distance. Activities carried out during daily free time
are related mainly to the individual’s own capabilities and wants rather than to
his or her physical environment. This means that, compared with the firmly
established routines and space–time fixities of work-related travel, school trips,
etc., personal preferences and interests as regards daily free time are translated
47
into activity patterns less dependent on the built environment surrounding the
place of residence. Apparently, individuals can schedule these activities more
freely in time and space. Other travel activities are, however, more dependent
on proximity and the local environment. It has been demonstrated that the
distance Swedes cover to perform activities in relation to committed time (e.g.,
various service errands) is more associated with proximity and the built
environment surrounding the home. Commuting to work, being the prime
activity in relation to contracted time use, still overall displays the strongest
relationship with various land use measures compared with other trip
purposes.
Important urban and regional changes over time are, however,
observed as regards work trips. The study indicates that the influence of the built
environment on work-related travel has weakened in recent decades. This means that
workers living in the same neighbourhood increasingly travel divergent
distances between home and work and presumably also to more dispersed job
locations. This suggests a continued decrease in the influence of distance on
labour market choices and, hence, on commuting destination; this change
coincides with a general increase in commuting, urban sprawl, and “regional
enlargement” occurring in contemporary Sweden. This supports the
hypothesis that increased personal mobility capabilities reduce the importance
of the built environment for daily travel distances. At the same time, it is
important to recognize that several interrelated processes at work potentially
also influence the spatial constraints on travel. Complexity makes it difficult to
entirely disentangle the exact roles of key contributing factors, as elaborated on
in the next section.
Other key spatiotemporal constraints traditionally associated with work
and daily commuting are also being relaxed and new opportunities have
opened up. The development of virtual mobility opportunities and other
labour market changes (e.g., the growth in non-manual employment) have
resulted in the increased decoupling of work-related activities from regular
workplace locations – at least for some groups. The rapid recent increase in
telework is evidently an important case of this. My study demonstrates that the
built environment influences teleworkers’ daily travel to a lesser extent than it does regular
workers’ daily travel. Telework allows for the freer scheduling of daily
activities, and is used to make travel less dependent on location structures and
relatively more based on individual needs and wishes.
48
Furthermore, increasing mobility capabilities are presumably also a key
factor when it comes to what aspects of the built environment are most
prominent in explaining variation in daily travel distances. This study
demonstrates that the regional built-environment variables are generally more important
than are the characteristics of local neighbourhoods in explaining the daily travel distance.
When they have the opportunity, many people extend their daily activity
spaces, which makes regional location structures more important. This
observation remains robust when running models on the different sets of data.
“The proximity of home to larger agglomerations of destinations in city
centres” is a key determinant in all cases in which built-environment variables
have significant effects. This testifies to the importance of longer trips outside
the neighbourhood, as such trips often make up most of the daily distance
travelled. It also indicates that the relaxation of spatial constraints works up to
a certain range, beyond which distance again becomes truly crucial from the
perspective of many daily activities.
Taken together, the empirical results of my study confirm the
importance of considering spatiotemporal constraints related to daily activities
when exploring the role of the built environment and its importance for daily
travel. Some travel purposes (e.g., commuting to a regular workplace) leave less
room for free individual spatial scheduling and therefore are more closely
related to the built environment. While other travelling (e.g., for leisure
purposes, broadly defined) is open for a range of mobility strategies
conditional on individual needs and wants, and to a relatively smaller extent on
distance to locations and opportunities. In this context, the results also remind
us that the importance of the built environment changes as an integral part of
larger societal transformations connected with processes of both time–spatial
convergence and profound socioeconomic change. For example, regular work
has traditionally been an important anchor, spatially and temporally, around
which people’s daily activity patterns have been planned and formed. This
situation has changed for some workers as labour markets have become more
“flexible”, for example, by enabling telework. Nevertheless, considering the
weaker influence of the built environment on non-work travel activities, an
ongoing decrease in the proportion of the population engaged in (permanent)
work is likely to play a more important role overall. These concerns and the
findings summarized in this section have important implications for future
research and practice, as discussed in the two remaining sections.
49
5.2 Implications for future research
Many authors have drawn attention to controversies and discrepancies in
results concerning the built environment’s role in shaping daily travel, and
some have argued that geography is of little importance (e.g., Bagley and
Moktharian, 2002; Snellen, Borgers and Timmermans, 2002), while others
strongly emphasize the role of urban structure and location (e.g., Newman and
Kenworthy, 1999; Seto et al., 2014)
7
. My results conclusively contend that any
of these positions risks conveying too simplistic a view of the situation (also
with implications for policy and planning; see next section). From a research
perspective, this points to the importance of situating empirically grounded built
environment–travel studies within a basic activity-based understanding of daily travel and
associated spatiotemporal constraints. In this context, the study conclusively leads to
several suggestions for future research, most prominently as regards the scope
and scale of future studies.
A first suggestion concerns the increased importance of considering
regional structures when seeking to explain daily travel. If only the built-
environment features of the local neighbourhood are taken into account,
important locational drivers (e.g., regional points of supply and
agglomerations) will be overlooked. Studies ignoring this fact may find no or
very weak built-environment effects for the wrong reasons. This underscores
the importance of situating built environment–travel studies within contextual
knowledge of the area of study as regards access on the local and regional scales and
associated spatial constraints. The relationships between location patterns and
travel distances are likely to differ across regions and countries (cf. Cervero,
2013). Results from my study indicate that built-environment variables still
contribute greatly to explained variance in commuting distance – more than
the results of many similar studies in other geographical contexts. This is partly
contingent on scale, that is, because a nationwide Swedish case is used that
includes a large variety of geographical contexts ranging from very sparsely
populated areas to large metropolitan areas, resulting in more apparent spatial
differences. Paradoxically, the direct link between the built environment and
travel could be expected to be of lesser magnitude if studies were delimited to
larger urban agglomerations with good overall accessibility to daily activities
7
See also section 2.3.3
50
(where individuals have more choice opportunities). An interesting topic for
future research would therefore be to design comparative studies evaluating
such dynamics. For the purpose of further scrutiny of the joint influence of
various scales of the built environment – both across and within urban
agglomerations – on travel, multilevel models similar to those applied here are
important tools (cf. the concluding sections of papers I and II).
A challenging task for future research concerns the relative importance of
individual and spatial factors in supplementing dimensions as regards capabilities, needs, and
wants. Overall, the activity-based conceptualization will continue to provide a
basic appraisal. Many people still follow relatively similar life courses that entail
more or less mandatory activities fairly fixed in time and space, activities such
as work and care for children. The built environment obviously plays a decisive
role in shaping the total daily trip length in these cases. However, for some
travel activities in contemporary society, the importance of geographic context
and proximity to the home location is diminishing in relative terms (i.e., in
relation to individual capabilities, needs, and preferences). Leisure travel and
travel during teleworking days are examples of such more flexible options
highlighted here. In these cases, various individual experiences, desires, and
practices are expected to be central determinants. In addition, though all
physical mobility is place bound and corporeal by definition, the theoretical
understanding of travel as a demand derived directly from the need to perform
geographically separated activities needs to be nuanced in many cases. The
demand is “less derived” when individuals have increasing opportunities to
engage in substitute virtual activities that are not fixed to certain locations and
times. A thorough understanding of these processes cannot rest solely on the
quantitative research designs that currently dominate this area of research (cf.
Næss, 2015).
Another aspect as regards the relative importance of the built
environment and individual choice calling for further study is how and to what
extent the spatiotemporal constraints associated with certain activities change
over time, concerning work in particular. The role of work is evidently shifting
and in some cases becoming less fixed. This is crucial, because home–work
relationships (and associated locations) are often regarded as having a key
organizing impact on other parts of the daily activity pattern. When workplaces
are used as a basis for understanding travel activities and how location patterns
impact travel, it is essential to account for individual workers’ spatiotemporal
51
constraints in various ways (e.g., regarding telework opportunities). However,
other (non-work) activities are also changing rapidly and merit further study as
regards their dependence on proximity and built-environment location
patterns. For example, education and service activities can increasingly be
accessed virtually. In my study, a rather crudely defined group of committed
time activities was examined. The time–spatial constraints associated with these
activities may vary greatly (Doherty, 2006). Some of these activities (e.g., taking
care of children) are usually very fixed in time and space, and proximity to
appropriate facilities (e.g., day-care centres) is crucial, while many other
activities conducted during committed time are subject to increasing individual
choice opportunities (e.g., buying groceries via the Internet). Likewise, the role
and significance of location also differ among what are labelled free-time
activities. An important subject in this case, not explored here, concerns the
role of non-built environments, including the physical geographies of nature
and vegetation giving outdoor recreation possibilities.
Additional important interrelationships between individual capabilities,
daily activities, and the built environment merit further study and theorizing
(on empirical grounds). In particular, there proved to be interactions between
individual and spatial scales as regards travel undertaken to carry out activities
in relation to committed time. In this case “access to a car” plays a key role
giving significantly more options for when and where these activities can be
performed. This is, however, not the case when it comes to commuting, as
many people do not have similar choice opportunities for work activities. In
this context, my study also raises important questions regarding how the effect
of individual mobility capabilities varies in different built environments. A
reasonable hypothesis is, for example, that the effect of car access is stronger
in sparser location structures with accordingly fewer choice opportunities.
Similar reasoning can also be applied in many other dimensions, opening up
important avenues for future research, for example, how different stages in the
life course (e.g., parenthood) and associated spatiotemporal constraints affect
the importance and roles of the built environment when people schedule
various travel activities. This further connects to a research focus on processes
of social differentiation and exclusion in relation to local and regional access;
for example, for what activities and population groups is local versus regional
access important.
52
Furthermore, my study points to the need for continued empirical
exploration of changes over time in the effects of the built environment. This is an
under-researched topic in the current literature that is only partly covered in
my thesis as regards work trips. The analysis should be enlarged to encompass
other travel activities, segments of the population, and geographical
delimitations. Repeated cross-sectional data over periods of time, as used here,
are important for the purpose of identifying change, but other approaches are
also needed to disentangle possible underlying causal factors. Besides
longitudinal approaches, quasi-experimental before-and-after studies could be
used to compare the daily travel of “treatment groups” with that of control
groups to isolate various effects of individual capabilities, spatiotemporal fixity
of activities, and changes to the built environment (e.g., residents in
neighbourhoods subject to densification strategies).
All in all, my thesis clearly testifies to the importance of a theoretical
understanding of the spatiotemporal constraints of individuals studied in terms of space
bridging, accessible temporal capabilities, and daily activities needed and wanted. The
activity-based approach is therefore an important theoretical and empirical
departure when studying the travel choices and mobility dependencies of daily
life – not least when exploring under what circumstances and to what extent
built structures are likely to influence daily travel in the future. Needless to say,
this will continue to merit study in the future. Like most phenomena involving
human behaviour, the underlying factors continuously change in time and
space. When it comes to built environment–travel relationships, a range of
possibly related processes has been set in motion. These include ongoing
urbanization (e.g., larger urban agglomerations, increasing choice
opportunities, and increased segregation), urban structure adaptations to high
mobility levels (e.g., urban sprawl and lock-in effects), further labour market-
related developments (e.g., policies of regional enlargement, flexibilization,
insecurity, and increasing commuting tolerance), and the development of ICT
practices (e.g., virtual substitution of more activities and telework practices).
5.3 A practical view
From a policy and planning perspective, my study demonstrates the
importance of viewing cities and regions as relational spaces. The built
environment of the region and the proximity to regional agglomerations greatly
53
impact daily travel distances. This highlights the importance of focusing on the
location and concentration of city centres within regions in urban and regional
planning. From this perspective, typical “new-urbanism” ideas, often with a
strong focus on the local neighbourhood, might be less effective than expected
(cf. Næss, 2011). Furthermore, policymakers and planners should heed the
increasing complexity of travel and destination choices, a complexity that
makes it problematic to use spatial planning and urban restructuring measures
with the sole aim of reducing people’s travel distances and associated general
costs (e.g., congestion and pollution). Urban planning principles based on
compact cities and the containment of urban areas are probably necessary, yet
are not sufficient conditions given the interaction of a complex web of factors.
Trips that concern free-time activities, where individuals have considerable
spatiotemporal flexibility, are very difficult to address by means of changes in
the built environment alone. In addition, commuting, which has traditionally
been the target of location-based policies and prediction models, have become
increasingly difficult to influence. This confirms the need to combine different
measures when planning for sustainable mobility and cities. Much daily travel
in Sweden today may be more responsive to instruments directly targeting the
individual level of travel decision making, such as fuel taxes, congestion
charges, and awareness campaigns. However, as discussed throughout my
thesis, this situation could change if individuals’ capabilities to travel quickly
should decline, affected, for example, by rising fuel prices.
Still, the built environment constitutes one of the foundations of
sustainable mobility. To make individually targeted measures acceptable and
effectual, alternatives enabling changes in daily travel, for example,
densification strategies, are essential. Without increased proximity, it is
infeasible for many people to reduce their car use and accustom themselves to
more sustainable options, such as improved public transportation and cycling.
This also raises questions of mobility divides and social welfare issues that
policymakers should take into account. Mobility resources, like many other
societal resources, are quite unevenly distributed. Car access is a crucial factor
in this context. The opportunities for daily activity participation that the local
built environment offers can play an important role not only in fostering
ecological sustainability but also in creating improved social justice in cities.
54
6 Sammanfattning (summary in Swedish)
Avhandlingen studerar den byggda miljöns betydelse för människors
vardagsresande. Inom urban och regional planering knyts numera ofta stora
förhoppningar till att den geografiska närhet som täta och funktionsintegrerade
bebyggelsemönster kan erbjuda skall bidra till ett mer hållbart vardagsresande.
Samtidigt medverkar människors växande möjligheter att överbrygga
geografiska avstånd, genom ökande bilinnehav, förbättrad kollektivtrafik etc.,
till att minska betydelsen av närhet och den byggda miljöns utformning. Den
tilltagande rumsliga flexibiliteten återspeglas bland annat i den exponentiella
ökningen av individers dagliga reslängd som skedde under 1900-talet, och i de
senaste decenniernas snabba utveckling i användning av informations och
kommunikationsteknologi (IKT) för att överbrygga avstånd.
Avhandlingen syftar därför till att undersöka den byggda miljöns roll
och relativa betydelse för den geografiska utsträckningen av individers
vardagsresande i dagens Sverige. Detta genomförs i tre empiriska
forskningsartiklar där särskild uppmärksamhet riktas mot tre viktiga aspekter:
att undersöka skillnader i den byggda miljöns betydelse för olika resesyften
(paper I); att utforska hur den byggda miljöns betydelse förändras över tid när
det gäller en viktig relation i vardagen, nämligen för arbetsresor (paper II); och
slutligen att undersöka den potentiellt ökade rumsliga flexibilitet och vad den
innebär för resandet när individer använder IKT regelbundet och
distansarbetar (paper III).
Artiklarna utgår från ett teoretiskt ramverk som betonar samspelet
mellan individens resurser, roller och omgivning som betydelsefulla när de
vardagliga aktivitets- och resmönstren formas med hänsyn till olika
tidsrumsliga restriktioner. I artiklarna analyseras sedan med multivariata
kvantitativa metoder unika kombinationer av innehållsrika geokodade
mikrodata inbegripet nationella resvaneundersökningar och registerdata
gällande hela befolkningen och samtliga arbetsplatser i Sverige. Detta möjliggör
en integrerad beskrivning och analys av hur individers dagliga reseaktiviteter,
givet olika resurser och restriktioner (t.ex. bilinnehav, familjeförhållanden),
samvarierar med den byggda miljöns lokaliseringar av transportinfrastruktur,
bostäder, arbetsplatser, service och andra viktiga samhällsfunktioner.
Avhandlingens övergripande slutsats är att den geografiska närheten till
olika samhällsfunktioner och verksamheter viktiga i vardagslivet i relation till
55
bostaden fortfarande spelar en viktig roll för hur långt individer i Sverige reser
dagligen. Dock visar analyserna att betydelsen av närhet luckras upp i flera
viktiga avseenden. Tidsrumsliga restriktioner förknippade med de aktiviteter
som motiverar resor i vardagen är viktiga och differentierande faktorer. När
endast helg- och fritidsresor beaktas är den byggda miljön av relativt liten
betydelse när det kommer till daglig reslängd. Personliga preferenser avseende
aktiviteter formar då dagliga aktivitetsmönster relativt oberoende av närhet.
Andra aktiviteter är däremot mer beroende av närhet och den byggda miljöns
lokaliseringsmönster och täthet. Hur långt individer färdas för att nå service
och utföra hushållsrelaterade aktiviteter samvarierar i viss utsträckning med
den byggda miljön som omger bostaden. Arbetsresor uppvisar det starkaste
sambandet jämfört med andra resesyften.
Dock observeras viktiga urbana och regionala förändringar över tid
gällande arbetsresor. Avhandlingen visar att den byggda miljöns påverkan på
arbetsresor har försvagats i Sverige under de senaste decennierna. Det innebär
t ex att förvärvsarbetande som bor i samma område (dvs med samma
omgivningsförhållanden) uppvisar alltmer divergerande avstånd mellan
hemmet och arbetet. Detta indikerar en avtagande betydelse av avstånd för
geografiska val på arbetsmarknaden; en förändring som också sammanfaller
med en generell ökning av pendling och sk regionförstoring i dagens Sverige.
Detta stödjer hypotesen att ökade individuella mobilitetsresurser minskar
betydelsen av den byggda miljöns lokaliseringsmönster för vardagsresandet.
Men även andra tidsrumsliga restriktioner kopplade till arbete och
daglig pendling luckras upp. Utvecklingen av virtuella mobilitetsmöjligheter
och en generell flexibilisering av arbetsförhållanden i stort bör ha resulterat i en
större individuell frikoppling från arbetsplatsen. Distansarbetets snabba
utbredning under de senaste åren är en central indikator. Avhandlingens
resultat tyder på att distansarbete ytterligare försvagar sambandet mellan den
byggda miljön och dagliga resor i olika dimensioner. Variationen i
distansarbetares dagliga restid och reseavstånd förklaras i mycket liten
utsträckning av den geografiska tillgängligheten till arbetsplatser och andra
viktiga samhällsfunktioner i relation till bostaden. Distansarbete möjliggör ett
bredare spektrum av dagliga mobilitetsstrategier (t ex att arbeta hemifrån och
avstå från resor alternativt resa mer på fritiden) eftersom det dagliga
aktivitetsmönstret då inte är knutet till en specifik arbetsplats.
56
Vidare visar avhandlingen att regionala lokaliseringsmönster generellt
är viktigare än egenskaper i det lokala bostadsområdet för att förklara dagliga
reseavstånd. När möjligheten ges utvidgar många sina dagliga
aktivitetsutrymmen bortom närområdet, vilket gör regionala
lokaliseringsstrukturer viktigare. En särskilt viktig faktor är bostadens läge i
förhållande till den regionala huvudortens centrum, där en mer nära
lokalisering generellt ger mindre resande. Detta vittnar om vikten av längre
resor utanför grannskapet då sådana resor ofta utgör en majoritet av den
dagliga reslängden.
Avhandlingens resultat påminner om att den byggda miljöns betydelse
för individers rörlighetshandlingar förändras som en integrerad del av större
samhällsomvandlingar: både gällande transport- och kommunikationsteknikens
utveckling och djupgående socioekonomiska förändringar. Till exempel har
fast arbete traditionellt varit en viktig nod i tid och rum kring vilka individers
dagliga aktivitetsmönster har formats. Men denna situation förändras för vissa
förvärvsarbetare när arbetsmarknaden blir mer flexibel i tid och rum, till
exempel genom distansarbete och olika former av mobilt arbete. Med tanke på
det betydligt svagare inflytande som den byggda miljön har på fritidsresandet,
kommer sannolikt också en fortsatt minskning av andelen av befolkningen
som deltar i fast arbete att spela en viktig roll totalt sett.
Slutligen vittnar avhandling om vikten av att beakta individers
tidsrumsliga restriktioner i vardagen vad gäller deltagande i aktiviteter och
tillgång till rumsöverbryggande resurser och tid. Detta är centralt för att förstå
resval och mobilitetsberoenden i vardagen, inte minst när det gäller att
fastställa under vilka omständigheter och i vilken utsträckning den byggda
miljön påverkar vardagsresandet i framtiden.
57
7 References
Akar, G., Clifton, K. J., & Doherty S. T. 2012. Redefining activity types: Who
participates in which leisure activity? Transportation Research Part A:
Policy and Practice, 46(8), 1194–1204.
Alexander, B., Hubers, C., Schwanen, T., Dijst, M., & Ettema, D. 2011.
Anything, anywhere, anytime? Developing indicators to assess the
spatial and temporal fragmentation of activities. Environment and
Planning B: Planning and Design, 38(4), 678–705.
Alonso, W. 1964. Location and Land Use: Toward a General Theory of Land Rent.
Cambridge: Harvard University Press.
Amcoff, J. 2012. Hur bra fungerar SAMS-områdena i studier av
grannskapseffekter? En studie av SAMS-områdenas homogenitet
[How well do SAMS work in studies of neighbourhood effects?].
Socialvetenskaplig tidskrift, 19(2), 93–115.
Andreev, P., Salomon, I., & Pliskin, N. 2010. Review: State of teleactivities.
Transportation Research Part C: Emerging Technologies, 18(1), 3–20.
Ås, D. 1978. Studies of Time-Use: Problems and Prospects. Acta Sociologica,
21(2), 125–141.
Axhausen, K. W., & Gärling, T. 1992. Activity‐based approaches to travel
analysis: conceptual frameworks, models, and research problems.
Transport Reviews, 12(4), 323–341.
Bagley, M., & Mokhtarian, P. L. 2002. The impact of residential neighborhood
type on travel behavior: A structural equations modeling approach. The
Annals of Regional Science, 36(2), 279–297.
Banister, D. 2008. The sustainable mobility paradigm. Transport Policy, 15(2),
73–80.
Banister, D. 2011. The trilogy of distance, speed and time. Journal of Transport
Geography, 19(4), 950–959.
Banister, D. 2012. Viewpoint: Assessing the reality—Transport and land use
planning to achieve sustainability. Journal of Transport and Land Use, 5(3),
1–14.
Berg, J., Levin, J., Abramsson, M., & Hagberg, J-E. 2014. Mobility in the
transition to retirement – the intertwining of transportation and
everyday projects. Journal of Transport Geography, 38, 48–54.
58
Bhat, C. R., & Guo, J. Y. 2007. A comprehensive analysis of built environment
characteristics on household residential choice and auto ownership
levels. Transportation Research Part B: Methodological, 41(5), 506–526.
Boarnet, M. G. 2011. A Broader Context for Land Use and Travel Behavior,
and a Research Agenda. Journal of the American Planning Association,
77(3), 197–213.
Boarnet, M. G., & Crane, R. 2001. Travel by design: The influence of urban form on
travel. New York: Oxford University Press.
Borggren, J. 2011. Kreativa individers bostadsområden och arbetsställen: belysta mot
bakgrund av näringslivets omvandling och förändringar i bebyggelsestrukturen i
Göteborg [Creative individuals’ residential areas and places of work. In
light of economic transformation and changes in urban structure in
Gothenburg] (PhD thesis). Gothenburg: Departments of Geography,
University of Gothenburg.
Bricka, S., & Bhat, C. R. 2006. Comparative Analysis of Global Positioning
System-Based and Travel Survey-Based Data. Transportation Research
Record: Journal of the Transportation Research Board, 1972, 9–20.
Brotchie, J. F. 1984. Technological change and urban form. Environment and
Planning A, 16(5), 583–596.
Brownstone, D., & Golob, T. F. 2009. The impact of residential density on
vehicle usage and energy consumption. Journal of Urban Economics,
65(1), 91–98.
Brueckner, J. K. 2000. Urban Sprawl: Diagnosis and Remedies. International
Regional Science Review, 23(2), 160–171.
Cao, X., Mokhtarian, P. L., & Handy, S. L. 2009. Examining the Impacts of
Residential Self‐Selection on Travel Behaviour: A Focus on Empirical
Findings. Transport Reviews, 29(3), 359–395.
Cervero, R. 1989. Jobs-Housing Balancing and Regional Mobility. Journal of the
American Planning Association, 55(2), 136–150.
Cervero, R. 2013. Linking urban transport and land use in developing
countries. Journal of Transport and Land Use, 6(1), 7–24.
Cervero, R., & Kockelman, K. 1997. Travel demand and the 3Ds: Density,
diversity, and design. Transportation Research Part D: Transport and
Environment, 2(3), 199–219.
Chapin, F.S. 1974. Human Activity Patterns in the City: Things People Do in Time and
Space. New York: Wiley.
59
Christaller, W. 1933. Die Zentralen Orte in Süddeutschland. Jena: Gustav Fischer.
Couclelis, H. 1996. The death of distance. Environment and Planning B: Planning
and Design, 23(4), 387–389.
Couclelis, H. 2000. From sustainable transportation to sustainable accessibility:
can we avoid a new tragedy of the commons? In Information, Place and
Cyberspace: Issues in Accessibility, eds. Janelle, D. G. & Hodge, D. C.
Berlin: Springer.
Cox, K. R., & Golledge, R. G. (eds). 1969. Behavioral Problems in Geography: A
Symposium. Evanston: Northwestern University Press.
Cullen, I., & Godson, V. 1975. Urban networks: The structure of activity
patterns. Progress in Planning, 4(1), 1–96.
Dal Fiore, F., Mokhtarian, P. L., Salomon, I., & Singer, M. E. 2014. “Nomads
at last”? A set of perspectives on how mobile technology may affect
travel. Journal of Transport Geography, 41, 97–106.
Dodgshon, R. A. 1999. Human geography at the end of time? Some thoughts
on the notion of time - space compression. Environment and Planning D:
Society and Space, 17(5), 607–620.
Doherty, S. T. 2006. Should we abandon activity type analysis? Redefining
activities by their salient attributes. Transportation, 33(6), 517–536.
Dugundji, E., Scott, D. M., Carrasco, J. A., & Paez, A. 2012. Urban mobility
and social–spatial contact—introduction. Environment and Planning A,
44(5), 1011–1015.
Elldér, E. 2014a. Commuting choices and residential built environments in
Sweden, 1990–2010: a multilevel analysis. Urban Geography, 35(5), 715–
734.
Elldér, E. 2014b. Residential location and daily travel distances: the influence
of trip purpose. Journal of Transport Geography, 34, 121–130.
Elldér, E. 2015. Does telework weaken urban structure-travel relationships?.
Journal of Transport and Land Use. Manuscript accepted for publication.
Elldér, E., Solá A. G., & Larsson, A. 2012. Featured graphic: Spatial inequality
and workplace accessibility: the case of a major hospital in Göteborg,
Sweden. Environment and Planning A, 44(10), 2295–2297.
Ellegård, K., & Vilhelmson, B. 2004. Home as a Pocket of Local Order:
Everyday Activities and The Friction of Distance. Geografiska Annaler:
Series B, Human Geography, 86(4), 281–296.
60
Eurostat. 2014. Eurostat statistics [web page], accessed November 25, 2014.
http://epp.eurostat.ec.europa.eu/portal/page/portal/statistics/themes
Ewing, R., & Cervero, R. 2001. Travel and the built environment : a synthesis.
Transportation research record: Journal of the Transportation Research Board,
1780, 87–114.
Ewing, R., & Cervero, R. 2010. Travel and the Built Environment. Journal of the
American Planning Association, 76(3), 265–294.
Ewing, R., Deanna, M., & Li, S-C. 1996. Land use impacts on trip generation
rates. Transportation Research Record: Journal of the Transportation Research
Board, 1518, 1–6.
Fox, M. 1995. Transport planning and the human activity approach. Journal of
Transport Geography, 3(2), 105–116.
Frändberg, L., & Vilhelmson, B. 2011. More or less travel: personal mobility
trends in the Swedish population focusing gender and cohort. Journal of
Transport Geography, 19(6), 1235–1244.
Frändberg, L., & Vilhelmson, B. 2014. Spatial, generational and gendered
trends and trend-breaks in mobility. In Handbook of sustainable travel,
eds. Gärling, T., Ettema, D., & Friman, M. Dordrecht: Springer.
Gebel, K., Bauman, A. E., & Petticrew, M. 2007. The Physical Environment
and Physical Activity: A Critical Appraisal of Review Articles. American
Journal of Preventive Medicine, 32(5), 361–369.
Giddens, A. 1984. The constitution of society. Outline of the Theory of Structuration.
Polity press: Cambridge.
Giuliano, G. 1989. Research policy and review 27. New directions for
understanding transportation and land use. Environment and Planning A,
21(2), 145–159.
Giuliano, G. 1995. The weakening transportation-land use connection.
ACCESS Magazine, 6, 3-11.
Giuliano, G., & Small, K. A. 1993. Is the Journey to Work Explained by
Urban Structure? Urban Studies, 30(9), 1485–1500.
Goldstein, H. 2011. Multilevel statistical models. Hoboken: Wiley.
Green, A. E. 2004. Is Relocation Redundant? Observations on the Changing
Nature and Impacts of Employment-related Geographical Mobility in
the UK. Regional Studies, 38(6), 629–641.
Green, A. E., Hogarth, T., & Shackleton, R. E. 1999. Longer distance
commuting as a substitute for migration in Britain: a review of trends,
61
issues and implications. International Journal of Population Geography, 5(1),
49–67.
Grunfelder, J., & Nielsen, T. S. 2012. Commuting behaviour and urban form:
a longitudinal study of a polycentric urban region in Denmark.
Geografisk Tidsskrift-Danish Journal of Geography, 112(1), 2–14.
Guerra, E. 2014. The Built Environment and Car Use in Mexico City: Is the
Relationship Changing over Time?. Journal of Planning Education and
Research, 34(4), 394–408.
Hägerstrand, T. 1970. What about people in Regional Science?. Papers in
Regional Science, 24(1), 7–24.
Hamidi, S., Ewing, R., Preuss, I., & Dodds, A. 2015. Measuring Sprawl and Its
Impacts: An Update. Journal of Planning Education and Research 35(1): 35–
50.
Handy, S. 1996. Methodologies for exploring the link between urban form and
travel behavior. Transportation Research Part D: Transport and Environment,
1(2), 151–165.
Harvey, D. 1989. The condition of postmodernity. Oxford: Blackwell.
Haugen , K. 2012. The accessibility paradox. Everyday geographies of proximity, distance
and mobility (Phd thesis). Umeå: Department of Geography and
Economic History, Umeå University.
Haugen, K., & Vilhelmson, B. 2013. The divergent role of spatial access: The
changing supply and location of service amenities and service travel
distance in Sweden. Transportation Research Part A: Policy and Practice, 49,
10–20.
Horner, M. W., & Murray, A. T. 2002. Excess Commuting and the Modifiable
Areal Unit Problem. Urban Studies, 39(1), 131–139.
Janelle, D. G. 1969. Spatial reorganization: a model and concept. Annals of the
Association of American Geographers, 59(2), 348–364.
Johansson, H. 2011. Ökade utsläpp från vägtrafiken trots rekordartad
energieffektivisering av nya bilar [Increased emissions from road transport
despite record high energy efficiency of new cars]. Borlänge:
Trafikverket.
Jones, P. M., Dix, M. C., Clarke, M. I., & Heggie, I.G. 1983. Understanding travel
behaviour. Aldershot: Gower.
62
Kitamura, R., Mokhtarian, P. L., & Laidet, L. 1997. A micro-analysis of land
use and travel in five neighborhoods in the San Francisco Bay Area.
Transportation, 24(2), 125–158.
Knowles, R. D. 2006. Transport shaping space: differential collapse in time–
space. Journal of Transport Geography, 14(6), 407–425.
Kwan, M-P. 2002. Time, Information Technologies, and the Geographies of
Everyday Life. Urban Geography, 23(5), 471–482.
Kwan, M-P., Dijst, M., & Schwanen, T. 2007. The interaction between ICT
and human activity-travel behavior. Transportation Research Part A: Policy
and Practice, 41(2), 121–124.
Kwan, M-P., & Schwanen, T. 2008. The Internet, mobile phone and space-
time constraints. Geoforum, 39(3), 1362–1377.
Kwan, M-P., & Weber, J. 2003. Individual accessibility revisited: Implications
for geographical analysis in the twenty-first century. Geographical
Analysis, 35(4), 341–353.
Kwan, M-P., & Weber, J. 2008. Scale and accessibility: Implications for the
analysis of land use–travel interaction. Applied Geography, 28(2), 110–
123.
Larsson, A., Elldér, E., Ernstson, U., & Fransson, U. 2011. Analysverktyg för
tillgänglighetsberäkning med bil och kollektivtrafik i Västra Götaland.
Slutrapport [Analysis tool for calculation of accessibility by car and
public transport in Västra Götaland. Final report]. Gothenburg:
Department of Human and Economic Geography, University of
Gothenburg.
Larsson, A., & Elldér, E. 2014. Accessibility Tool for Road and Public
Transport Travel Time Analysis in Västra Götaland. In COST Action
TU 1002: Collecting accessibility instruments and improving their usability for
planning practices, http://www.accessibilityplanning.eu/wp-content/
uploads/2013/10/Accessibility-Tool-for-Road-and-Public-Transport-
Travel-Time-Analysis-in-V%C3%A4stra-G%C3%B6taland.pdf.
Lenntorp, B. 1976. Paths in space-time environments : a time-geographic study of
movement possibilities of individuals. Lund: Department of Geography,
Lund University.
Lenz, B., & Nobis, C. 2007. The changing allocation of activities in space and
time by the use of ICT—“Fragmentation” as a new concept and
63
empirical results. Transportation Research Part A: Policy and Practice, 41(2),
190–204.
Lyons, G. 2009. The reshaping of activities and mobility through new
technologies. Journal of Transport Geography, 17(2), 81–82.
Marshall, S., & Banister, D. 2000. Travel reduction strategies: intentions and
outcomes. Transportation Research Part A: Policy and Practice, 34(5), 321–
338.
McGuckin, N., Zmud, J. P., & Nakamoto, Y. 2005. Trip-Chaining Trends in
the United States: Understanding Travel Behavior for Policy Making.
Transportation Research Record: Journal of the Transportation Research Board,
1917, 199–204.
Metz, D. 2010. Saturation of Demand for Daily Travel. Transport Reviews, 30(5),
659–674.
Meurs, H., & Haaijer, R. 2001. Spatial structure and mobility. Transportation
Research Part D: Transport and Environment, 6(6), 429–446.
Millard‐Ball, A., & Schipper, L. 2010. Are We Reaching Peak Travel? Trends
in Passenger Transport in Eight Industrialized Countries. Transport
Reviews, 31(3), 357–378.
Miller, H. 2007. Place-Based versus People-Based Geographic Information
Science. Geography Compass, 1(3), 503–535.
Mokhtarian, P. L. 1990. A typology of relationships between
telecommunications and transportation. Transportation Research Part A:
General, 24(3), 231–242.
Mokhtarian, P. L. 2005. Travel as a desired end, not just a means. Transportation
Research Part A: Policy and Practice, 39(2–3), 93–96.
Mokhtarian, P. L., & Cao, X. 2008. Examining the impacts of residential self-
selection on travel behavior: A focus on methodologies. Transportation
Research Part B: Methodological, 42(3), 204–228.
Mokhtarian, P. L., & Salomon, I. 2001. How derived is the demand for travel?
Some conceptual and measurement considerations. Transportation
Research Part A: Policy and Practice, 35(8), 695–719.
Mokhtarian, P. L., & Tal, G. 2013. Impacts of ICT on travel behavior: a
tapestry of relationships. In The SAGE Handbook of Transport Studies,
eds. Rodrigue, J.-P., Notteboom, T., & Shaw S-L. London: SAGE.
Næss, P. 2004. Prediction, Regressions and Critical Realism. Journal of Critical
Realism, 3(1), 133–164.
64
Næss, P. 2006. Urban Structure Matters: Residential Location, Car Dependence and
Travel Behaviour. London: Routledge.
Næss, P. 2009. Residential Self‐Selection and Appropriate Control Variables in
Land Use: Travel Studies. Transport Reviews, 29(3), 293–324.
Næss, P. 2011. ‘New urbanism’ or metropolitan-level centralization? A
comparison of the influences of metropolitan-level and neighborhood-
level urban form characteristics on travel behavior. Journal of Transport
and Land Use, 4(1), 25–44.
Næss, P. 2012. Urban form and travel behavior: experience from a Nordic
context. Journal of Transport and Land Use, 5(2), 21–45.
Næss, P. 2013. Residential location, transport rationales and daily-life travel
behaviour: The case of Hangzhou Metropolitan Area, China. Progress in
Planning, 79, 1–50.
Næss, P. 2014. Tempest in a teapot: The exaggerated problem of transport-
related residential self-selection as a source of error in empirical
studies. Journal of transport and land use, 7(3), 57–79.
Næss, P. 2015. Built Environment, Causality and Travel. Transport reviews,
http://dx.doi.org/10.1080/01441647.2015.1017751.
Newman, P., & Kenworthy, J. R. 1999. Sustainability and Cities: overcoming
automobile dependence. Washington: Island Press.
Öhman, M., & Lindgren, U. 2003. Who is the long-distance commuter?
Patterns and driving forces in Sweden. CyberGEO: European Journal of
Geography, article 243.
Olsson, G. 1965. Distance and human interaction : A review and bibliography.
Philadelphia: Regional Science Research Institute.
Openshaw, S. 1984. The Modifiable Areal Unit Problem. Norwich: Geo Books.
Ory, D. T., & Mokhtarian P. L. 2005. When is getting there half the fun?
Modeling the liking for travel. Transportation Research Part A: Policy and
Practice, 39(2–3), 97–123.
Östh, J. 2007. Homogeneous regions and heterogeneous populations - creating
and analysing labor market regions for subgroups in Sweden. In Home
job and space. Mapping and modeling the labor market (Phd thesis). Uppsala:
Department of Social and Economic Geography,Uppsala University.
Östh, J., & Lindgren, U. 2012. Do changes in gdp influence commuting
distances? A study of swedish commuting patterns between 1990 and
2006. Tijdschrift voor Economische en Sociale Geografie, 103(4), 443–456.
65
Pontes de Aquino, A., & Timmermans, H. 2010. The Built Environment as a
Décor of Unfolding Housing Careers and Activity Travel Patterns:
Reflection and Research Agenda. Paper presented at The 12th World
Conference on Transport Research, Lisbon, Portugal.
Robinson, W. S. 1950. Ecological Correlations and the Behavior of
Individuals. American Sociological Review, 15(3), 351–357.
Saelens, B. E., Sallis, J. F., & Frank, L. D. 2003. Environmental correlates of
walking and cycling: Findings from the transportation, urban design,
and planning literatures. Annals of Behavioral Medicine, 25(2), 80–91.
Salomon, I. 1986. Telecommunications and travel relationships: a review.
Transportation Research Part A: General, 20(3), 223–238.
Salon, D., Boarnet, M. G., Handy, S., Spears, S., & Tal, G. 2012. How do local
actions affect VMT? A critical review of the empirical evidence.
Transportation Research Part D: Transport and Environment, 17(7), 495–508.
Sandow, E. 2011. On the road. Social aspects of commuting long distances to work (Phd
thesis). Umeå: Department of Social and Economic Geography, Umeå
University.
Saxena, S., & Moktharian, P. L. 1997. The Impact of Telecommuting on the
Activity Spaces of Participants. Geographical Analysis, 29(2): 124–144.
Sayer, A. 1992. Method in social science: a realist approach. London: Routledge.
SCB. 2005. Geografin i statistiken : regionala indelningar i Sverige [Geography in
statistics : regional divisions in Sweden]. Stockholm: Statistiska
centralbyrån.
SCB. 2010. Tätorter 2010 [Built up areas 2010]. Stockhom: Statistiska
centralbyrån.
SCB. 2011. Longitudinell integrationsdatabas för sjukförsäkrings- och
arbetsmarknadsstudier (LISA) 1990-2009 [Integrated database for labour
market research]. Stockholm: Statistiska centralbyrån.
Schwanen, T., & Dijst, M. 2003. Time windows in workers' activity patterns:
Empirical evidence from the Netherlands. Transportation, 30(3), 261–
283.
Schwanen, T., Kwan, M-P., & Ren, F. 2008. How fixed is fixed? Gendered
rigidity of space–time constraints and geographies of everyday
activities. Geoforum, 39(6): 2109–2121.
Seto K. C., Dhakal, S., Bigio, A., Blanco, H., Delgado, G. C., Dewar, D.,
Huang, L., Inaba, A., Kansal, A., Lwasa, S., McMahon, J. E., Müller, D.
66
B., Murakami, J., Nagendra, H., & Ramaswami, A. 2014. Human
Settlements, Infrastructure and Spatial Planning. In Climate Change
2014: Mitigation of Climate Change. Contribution of Working Group III to the
Fifth Assessment Report of the Intergovernmental Panel on Climate Change, eds.
Edenhofer, O., Pichs-Madruga, R., Sokona, Y., Farahani, E., Kadner,
S., Seyboth, K., Adler, A., Baum, I., Brunner, S., Eickemeier, P.,
Kriemann, B., Savolainen, J., Schlömer, S., von Stechow, C., Zwickel,
T., & Minx, J.C. Cambridge: Cambridge University Press.
Shuttleworth, I., & Gould, M. 2010. Distance between home and work: a
multilevel analysis of individual workers, neighbourhoods, and
employment sites in Northern Ireland. Environment and Planning A,
42(5), 1221–1238.
SIKA. 2007. RES 2005-2006: Den nationella resvaneundersökningen [RES 2005–
2006: the national travel survey]. Stockholm: Statistiska centralbyrån.
Sims R., Schaeffer, R., Creutzig, F., Cruz-Núñez, X., D’Agosto, M., Dimitriu,
D., Figueroa Meza, M. J., Fulton, L., Kobayashi, S., Lah, O.,
McKinnon, A., Newman, P., Ouyang, M., Schauer, J. J., Sperling, D., &
Tiwari, G. 2014. Transport. In Climate Change 2014: Mitigation of Climate
Change. Contribution of Working Group III to the Fifth Assessment Report of
the Intergovernmental Panel on Climate Change, eds. Edenhofer, O., Pichs-
Madruga, R., Sokona, Y., Farahani, E., Kadner, S., Seyboth, K., Adler,
A., Baum, I., Brunner, S., Eickemeier, P., Kriemann, B., Savolainen, J.
Schlömer, S., von Stechow, C., Zwickel, T., & Minx, J.C. Cambridge:
Cambridge University Press.
Snellen, D., Borgers, A., & Timmermans, H. 2002. Urban form, road network
type, and mode choice for frequently conducted activities: a multilevel
analysis using quasi-experimental design data. Environment and Planning
A, 34(7), 1207–1220.
Solá, A. G. 2013. På väg mot jämställda arbetsresor: vardagens mobilitet i förändring och
förhandling [Towards gender equality? Women’s and men’s commuting
under transformation and negotiation] (PhD thesis). Gothenburg:
Departments of Geography, University of Gothenburg.
Solá, A. G., & Vilhelmson, B. 2012. Convergence or divergence? Changing
gender differences in commuting in two Swedish urban regions.
Cybergeo: European Journal of Geography, article 591.
67
Sparks, A. L., Bania, N., & Leete, L. 2011. Comparative Approaches to
Measuring Food Access in Urban Areas: The Case of Portland,
Oregon. Urban Studies, 48(8), 1715–1737.
Standing, G. 2011. The precariat: the new dangerous class. London: Bloomsbury
Academic.
Stead, D. 2001. Relationships between land use, socioeconomic factors, and
travel patterns in Britain. Environment and Planning B: Planning and Design,
28(4), 499–528.
Susilo, Y. O., & Maat, K. 2007. The influence of built environment to the
trends in commuting journeys in the Netherlands. Transportation, 34(5),
589–609.
Swärdh, J-E. 2009. Commuting time choice and the value of travel time (Phd thesis).
Örebro: Örebro University.
Tilahun, N., & Levinson, D. 2011. Work and home location: Possible role of
social networks. Transportation Research Part A: Policy and Practice, 45(4),
323–331.
Tobler, W. R. 1970. A Computer Movie Simulating Urban Growth in the
Detroit Region. Economic Geography, 46, 234–240.
Trafikanalys. 2012. Metodrapport RVU Sverige 2011 [Method report RVU
Sweden 2011]. Stockholm: Trafikanalys.
Van Acker, V., & Witlox, F. 2011. Commuting trips within tours: how is
commuting related to land use?. Transportation, 38(3), 465–486.
Van de Coevering, P., Maat, K., & van Wee, B. 2015. Multi-period Research
Designs for Identifying Causal Effects of Built Environment
Characteristics on Travel Behaviour. Transport Reviews,
http://dx.doi.org/10.1080/01441647.2015.1025455.
Van de Coevering, P., & Schwanen, T. 2006. Re-evaluating the impact of
urban form on travel patternsin Europe and North-America. Transport
Policy, 13(3), 229–239.
Van Ham, M. 2002. Job access, workplace mobility, and occupational achievement (Phd
thesis). Delft: Eburon.
Van Wee, B. 2011. Evaluating the impact of land use on travel behaviour: the
environment versus accessibility. Journal of Transport Geography, 19(6),
1530–1533.
68
Van Wee, B. 2013. Land use and transport. In The transport system and transport
policy, eds. Van Wee, B., Annema, J. A., & Banister, B. Cheltenham:
Edward Elgar.
Van Wee, B., Geurs, K. T., & Chorus, C. 2013. Information, communication,
travel behavior and accessibility. Journal of Transport and Land Use, 6(3),
1–16.
Vilhelmson, B. 1999. Daily mobility and the use of time for different activities.
The case of Sweden. GeoJournal, 48(3), 177–185.
Vilhelmson, B. 2007. The use of the car: Mobility dependencies of urban
everyday life. In Threats from car traffic to the quality of urban life : problems,
causes, and solutions, eds. Gärling, T., & Steg, L. Amsterdam: Elsevier.
Vilhelmson, B., & Thulin, E. 2008. Virtual mobility, time use and the place of
the home. Tijdschrift voor economische en sociale geografie, 99(5), 602–618.
Vilhelmson, B., & Thulin, E. 2015. Who and where are the flexible workers?
Exploring the current diffusion of telework in Sweden. Submitted
manuscript.
Weber, J., & Kwan, M-P. 2003. Evaluating the effects of geographic contexts
on individual accessibility: a multilevel approach. Urban Geography,
24(8), 647–671.
Wolf, J., Oliveira, M. S., & Thompson, M. 2003. Impact of Underreporting on
VMT and Travel Time Estimates: Preliminary Findings from the
California Statewide Household Travel Survey GPS Study.
Transportation Research Record: Journal of the Transportation Research Board,
1854, 189–198.
Zegras, P. C., & Hannan, V. A. 2012. Dynamics of Automobile Ownership
Under Rapid Growth. Transportation Research Record: Journal of the
Transportation Research Board, 2323(2), 80–89.
MEDDELANDEN FRÅN GÖTEBORGS UNIVERSITETS
GEOGRAFISKA INSTITUTIONER, SERIE B.
Doktorsavhandlingar i kulturgeografi och ekonomisk geografi,
Göteborgs universitet
Nr 1 Olof Wärneryd: Interdependence in urban systems. 1968
Nr 17 Lennart Andersson: Rumsliga effekter av organisationsförändringar.
Studier i lokalisering med exempel från skolväsendet. 1970
Nr 23 Lars Nordström: Rumsliga förändringar och ekonomisk utveckling. 1971
Nr 43 Kenneth Asp: Interregionala godstransporter i ett rumsligt system. 1975
Nr 44 Jan Lundqvist: Local and central impulses for change and development. A
case study of Morogoro District, Tanzania. 1975
Nr 56 Staffan Öhrling: Rural change and spatial reorganization in Sri Lanka.
Barriers against development of traditional Sinhalese local communities.
1977
Nr 57 Ulf Halloff: Inköpsresor i ett rumsligt system. Metodstudier på grundval
av empiriskt material från några stadsdelar i Göteborg. 1977
Nr 59 Lage Wahlström: Naturvården i regional och lokal planering. Geografiska
studier med exempel från Göteborgsregionen och övriga delar av de
västsvenska länen. 1977
Nr 60 Kent Persson: Sysselsättningen i centrum. Sysselsättningsförändringar i
stadscentrum, deras orsaker och verkan - med exempel från
Göteborg.1977.
Nr 63 Claes Göran Alvstam: Utrikeshandel och rumslig dynamik. En studie av
den väst-europeiska interna utrikeshandelns ländersammansättning
1955-1975. 1979
Nr 64 Sten Lorentzon: Ortsstruktur, arbetsresor och energiförbrukning.
Förändringar i bebyggelsestrukturen och energikonsumtionen vid
arbetsresor belysta med exempel från västra Sverige.1979
Nr 65 Bengt Holmgren: Transportförändringar och rumslig utveckling.
Geografiska studier av järnvägsnedläggningars effekter med exempel
från två västsvenska kommuner.1980
Nr 66 Rolf Pettersson: Omlokalisering av statlig verksamhet. Effekter på
arbetsmarknaden i mottagande orter.1980.
Nr 69 Åke Forsström: Commuting accidents. A study of commuting accidents
and casualties in some swedish regions during 1971. 1982.
Nr 71 Christina Nordin: Marchés, commerçants, clientèle. le commerce non
sédentaire de la région Parisienne - Etude de géographie humaine.1983
Nr 72 Kajsa Ellegård: Människa - produktion. Tidsbilder av ett
produktionssystem. 1983
Nr 73 Kjell Gustafsson: Tekoindustrin och förändringarna i den internationella
arbetsfördelningen. Konsekvenser för lokalisering och sysselsättning i
Sverige. 1983
Nr 74 Ingrid Johansson: Arbetsplatslokalisering i staden: dåtid-nutid-framtid.
Exempel från några stadsdelar i Göteborg. 1984.
Nr 75 Magnus Torell: Fisheries in Thailand. Geographical studies about the
utilization of resources in semi-enclosed seas. 1984
Nr 77 Bertil Vilhelmson: Resurser och resor. Äldres aktivitet och handikapp i
trafiken. 1985
Nr 78 Gerhard Gustafsson: Etik och lokala utvecklingsstrategier. Bevaring eller
förändring av människans livsvillkor. 1986.
Nr 79 Lars Aronsson: Turism och lokal utveckling. En turism-geografisk studie.
1989
Nr 80 Peter de Souza: Territorial production complexes in the Soviet Union -
with special focus on Siberia. 1989
Nr 81 Bertil Lundberg: Industriella beroenden. Rumslig och strukturell
förändring i ett värmlandsperspektiv. 1991
Nr 82 Thomas Jordan: Flows of pumps: Structure and Change in the
International Division of Labour. 1992
Nr 84 Joel Yrlid: Mission och kommunikation. Den kristna missionen och
transportnätets utveckling i Belgiska Kongo/Zaire 1878-1991. 1993
Nr 85 Martin Gren: Earth writing: Exploring Representation and Social
Geography In-Between Meaning/Matter. 1994
Nr 86 Sören Eriksson: Global shift in the aircraft industry. A study of airframe
manufacturing with special reference to the Asian NIEs.1995
Nr 87 Gabriel Bladh: Finnskogens landskap och människor under fyra sekler.
En studie av natur och samhälle i förändring. 1995
Nr 88 Anders Närman: Education and nation building in Kenya. Perspectives
on modernization, global dependency and local development
alternatives. 1995
Nr 89 Thomas Blom: Perspektiv på kunskap och utveckling. Om
högskoleutbildningens betydelse i perifera regioner. 1996
Nr 90 Inge Ivarsson: Integrated international production. A study of foreign
transnational corporations in Sweden. 1996.
Nr 91 Sang Chul Park: The technopolis plan in Japanese industrial policy.1997
Nr 92 Johan Dahl: A cry for water. Perceptions of development in Binga
district, Zimbabwe. 1997
Nr 94 Margareta Espling: Women's livelihood strategies in processes of change.
Cases from urban Mozambique. 1999
Nr 95 Lars-Gunnar Krantz: Rörlighetens mångfald och förändring.
Befolkningens dagliga resande i Sverige 1978 och 1996.1999
Nr 96 Per Assmo: Livelihood strategies and land degradation. Perceptions
among small-scale farmers in Ng'iresi Village, Tanzania. 1999
Nr 97 Anders Larsson: Proximity matters? Geographical aspects of changing
strategies in automotive subcontracting relationships: the case of
domestic suppliers to Volvo Torslanda assembly plant. 1999
Nr 98 Mikael Jonasson: The creation of places in traffic through performative
action. 2000
Nr 99 Matilde Mordt: Livelihoods and sustainability at the agrarian frontier. The
evolution of the frontier in Southeastern Nicaragua. 2001
Nr 101 Kersti Nordell: Kvinnors hälsa - en fråga om medvetenhet, möjlighet och
makt. Att öka förståelsen för människors livssammanhang genom
tidsgeografisk analys. 2002
Nr 102 Åsa Westermark: Informal livelihoods: Woman's biographies and
reflections about everyday life. A time-geographic analysis in urban
Colombia. 2003
Nr 103 Bodil Jansund och Ulrika Blom-Mondlane: Geografi-didaktik-praktik.
Interaktiva studier av förloppslandskapet. 2003
Nr 104 Alf Brodin: Baltic Sea ports and Russian foreign trade. Studies in the
economic and political geography of transition. 2003
Nr 105 Eva Thulin: Ungdomars virtuella rörlighet. Användningen av dator,
internet och mobiltelefon i ett geografiskt perspektiv. 2004
Nr 106 Daniel Rylander: Nätverkan och regional utveckling. Om
gränsöverskridande samarbete i södra Östersjöregionen och
nätverkstjänsternas roll i samhällsutvecklingen. 2004
Nr 107 Patrik Ström: The 'Lagged' Internationalization of Japanese Professional
Business Service Firms: Experiences from the UK and Singapore. 2004
Nr 108 Jonas Lindberg: Education for all in times of global transformations:
Aspirations and opportunities of poor families in marginal areas of Sri
Lanka. 2005
Nr 109 Ulf Ernstson: Kontrakt med naturen. Om spridning och implementering
av miljöledningssystem. 2006
Nr 110 Jerry Olsson: Responses to change in accessibility. Socio-economic
impacts of road investment: the distributive outcomes in two rural
peripheral Philippine municipalities. 2006
Nr 111 Iraê Baptista Lundin: Negotiating Transformation: Urban livelihoods in
Maputo adapting to thirty years of political and economic changes. 2007
Nr 112 Curt Nestor: Foreign Direct Investment in the Socialist Republic of
Vietnam 1988-2000: Geographical Perspectives. 2007
Nr 113 Lena Lindberg: The Regionalisation Process in Southeast Asia and the
Economic Integration of Cambodia and Laos into ASEAN. 2007
Nr 114 Kristina Thorell: Naturvårdsplanering med förankring i det lokala. Villkor
för delaktighet och underifrånperspektiv i vården av värden i landskapet.
2008
Nr 115 Jean Paul Dushimumuremyi: Spatial Distribution of Water Resources and
Accessibility to Water. The Case of Bugesera District in Rwanda. 2009
Nr 116 Théophile Niyonzima: Land Use Dynamics in the face of Population
Increase. A study in the Distrcits of Gatsibo and Nyagatare, Eastern
Province, Rwanda. 2009
Nr 117 Robin Biddulph: Geographies of Evasion. The Development Industry and
Property Rights Interventions in Early 21st Century Cambodia. 2010
Nr 118 Pelle Amberntsson: The Past of Present Livelihoods. Historical
Perspectives on Modernisation, Rural Policy Regimes and Smallholder
Poverty. A Case from Eastern Zambia. 2011
Nr 119 Kristina N. Lindström: Den massmediala (re)produktionen av turismens
platser. Geografiska perspektiv på journalistikens uttryck och
produktionsförhållanden. 2011
Nr 120 Jonathan Borggren: Kreativa individers bostadsområden och arbetsställen.
Belysta mot bakgrund av näringslivets omvandling och förändringar i
bebyggelsestrukturen i Göteborg. 2011
Nr 121 Anja K. Franck: From formal employment to street vending. Women´s
room to maneuver and labor market decisions under conditions of
export-orientation – the case of Penang, Malaysia. 2012
Nr 122 Mattias Sandberg: ”De är inte ute så mycket”. Den bostadsnära
naturkontaktens betydelse och utrymme i storstadsbarns vardagsliv.
2012
Nr 123 Ana Gil Sola: På väg mot jämställda arbetsresor? Vardagens mobilitet i
förändring och förhandling. 2013
Nr 124 Maja Essebo: Lock-in as make-believe. Exploring the role of myth in the
lock-in of high mobility systems. 2013
Nr 125 Elin Slätmo: Jordbruksmark i förändring. Drivkrafter bakom och
förutsättningar för offentlig styrning i Sverige och Norge. 2014
