USA: +1(669)289-8683 
Sign in
I need this paper by tomorrow no later than 10 a.m PACIFIC TIME.
Writer must read all pages of the reading.
Writer must answer the summary questions step by step.
The summary section contains 4 parts that must be answered. Some of these questions are straight forward. Section (3 )is very important if theres 3 different types of data given each data must be answered with the questions (A-E) per data.
Im looking forward to working with one person throughout this course. Make me happy and ill provide you with the rest of the upcoming work. thanks
Eachsummary will consist of the following four sections. Include these as section headings.
(1) Author(s), publication date, and title of reading.
(2) Research questions or objectives that guide the paper; page number(s) on which you found this information.
In your own words, in a few sentences, explain what the authors are asking or trying to find out in this research study.
(3) Data and analysis for this study; page number(s) on which you found this information.
Identify the three most important types of data that the authors used to answer their research questions. (It is possible that the authors may use fewer than three types of data.)
For each type of data you identify, provide the following information:
(a) Content/description of the data. What are the data about?
(b) Form of the data. Is the data quantitative or qualitative? In the form of numbers, words, images, sounds…? Is it primary or secondary data?
(c) Data collection method. How did the researcher gather or find the data?
(d) Data source. From whom or what did the researcher collect the data?
(e) Analytic methods. How did the authors analyze the data?
(4) Major findings made by this study; page number(s) on which you found this information.
In your own words, what are the two or three most important findings the researchers made in this study? Summarize key findings that the researchers made in response to the questions/objectives you identified above in (2). One to three sentences per point is sufficient.
The Living, Moving and Travel Behaviour of
the Growing American Solo: Implication
s
for Cities
Devajyoti Deka
[Paper first received, November 2012; in final form, March 2013]
Abstract
Between 1930 and 2010 the share of single-person, or solo, households in the US
increased from 6 per cent to almost 28 per cent, whereas the share of married-
couple households decreased from 79 per cent to 49 per cent. Yet solo households
have received little attention in urban planning and transport research. Given th
e
significant increase of solo households in US cities, this study identifies the distinc-
tive dwelling, moving and travel characteristics of the American solo households,
and examines the reasons for their attraction to cities. It uses historical data from
census Public Use Microdata Samples and recent national data from the American
Housing Survey and the National Household Travel Survey. Descriptive statistics,
basic statistical tests, binary logit models and Heckman sample selection models are
used to examine various relationships. Some of the transport-related and environ-
mental implications of the findings are discussed.
Introduction
In his recent book, Going Solo: The
Extraordinary Rise and Surprising Appeal to
Living Alone, sociologist Eric Klinenberg
(2012) brings to the fore the growing ten-
dency among American adults to live in
single-person, or solo, households. For the
first time in centuries, Klinenberg men-
tions, the majority of American adults are
now single and approximately 31 million
adults, or one of every seven, live in a solo
household. From Klinenberg’s contention
that the US is simply following a pattern
experienced for a longer duration by other
advanced countries like Japan, Germany,
Canada, France and Australia, it can be
anticipated that the growth of solo house-
holds in the US will continue in the fore-
seeable future.
Devajyoti Deka is in the Alan M. Voorhees Transport Centre, Edward J. Bloustein School of
Planning and Public Policy, Rutgers, The State University of New Jersey, 33 Livingston Avenue,
New Brunswick, New Jersey 08901, USA. Email: ddeka@ejb.rutgers.edu.
Article
51(4) 634–654, March 2014
0042-0980 Print/1360-063X Online
! 2013 Urban Studies Journal Limited
DOI: 10.1177/004209801349223
3
at LOYOLA MARYMOUNT UNIV on January 8, 2016usj.sagepub.comDownloaded from
http://usj.sagepub.com/
Given the significant growth of solo
households in the US over the past decades
and its potential impact on urban areas in
the future, this paper seeks to provide a
better understanding about the distinctive
characteristics of the American solo house-
holds. It analyses the potential reasons for
their growth in cities, compares their travel
and moving patterns with adults from mar-
ried-couple households and discusses the
implications of their growth for urban
America. The paper is prepared with the pre-
mise that understanding the characteristics
of the solo households and the reasons for
their distinctive living, travel and moving
patterns is beneficial for urban planners,
transport planners and policy-makers.
Understanding the growth of solo
households in US cities is important for
several reasons. First, given the massive loss
of population in central cities over the past
decades, the growth of any type of house-
hold and jobs in cities could be perceived
as a positive sign. Between 1950 and 1990,
the proportion of metropolitan residents in
central cities decreased from 57 per cent to
37 per cent, whereas the proportion of jobs
decreased from 70 per cent to 45 per cent
(Mieszkowski and Mills, 1993). Therefore,
the apparent attractiveness of cities for solo
households, especially working adults from
solo households, provides a glimmer of
hope about potential recovery of US cities.
Secondly, as land is more compactly devel-
oped in central cities than suburban areas,
continued growth of solo households in
central cities can be expected to create a
more sustainable metropolitan land devel-
opment pattern in the future. Thirdly, solo
households exhibit travel patterns that are
more conducive to environmentally sus-
tainable cities than the travel patterns of
two-earner households. As shown in this
paper and other studies, solo households
own fewer cars and workers from solo
households commute shorter distances,
spend less time commuting and use auto-
mobiles less frequently. Considering that
the increase in the use of the automobile
has often been attributed to the growth of
two-earner households (Cervero, 1989;
Downs, 1992), the replacement of two-
earner households by solo households
could potentially help to reduce society’s
dependence on the automobile and pro-
mote the use of public transit.
This paper focuses on the US as a whole
instead of specific regions or cities of the
country. Hence, data from several sources
are used at the national level and city-spe-
cific comparisons are avoided. All compari-
sons in the paper are made between adult
persons from solo households and married-
couple households. By definition, married
couples are those with two spouses living
together, whereas solo households are com-
posed of only one adult. Single parents are
not included in any comparison because few
people potentially make a choice between
remaining solo and getting married with the
expectation of becoming a single parent.
Although many elderly persons live in solo
households, the paper places a greater
emphasis on working adults because of their
potential impact on the economic wellbeing
of cities today and in the future.
Data and Methods
This study uses data from the 2009 National
Household Travel Survey (NHTS), the 200
9
American Housing Survey (AHS) national
sample, and Public Use Microdata Sample
(PUMS) from several censuses and the 2010
American Community Survey (ACS). For
analytical purposes, it uses descriptive sta-
tistics, basic statistical tests, binary logit
models and the Heckman sample selection
model.
The following section, or the third sec-
tion of the paper, uses historical PUMS
SOLO HOUSEHOLDS IN THE US 635
at LOYOLA MARYMOUNT UNIV on January 8, 2016usj.sagepub.comDownloaded from
http://usj.sagepub.com/
data from 1930 to 2010 to demonstrate the
growth of solo households and the labour
force, and to compare the relative attrac-
tiveness of cities for adults from married-
couple households and solo households.
The characteristics of the solo households
in the US are explored by using binary logit
models with 2010 ACS PUMS data in the
section following the literature review on
the distinctive characteristics of solo house-
holds. In the subsequent section, compari-
sons are made between solo households
and adults from married-couple house-
holds regarding their living and travel char-
acteristics. For this purpose, descriptive
statistics from the 2009 NHTS and three
Heckman sample selection models with
data from the 2009 AHS national sample
are used.
The Heckman sample selection model
(Heckman, 1976, 1979) was used because it
has clear advantages over ordinary least
squares and binary logit or probit models. It
uses a system of equations to jointly predict
the probability of selection (for example, the
decision to be solo) and the outcome vari-
able (for example, commuting trip distance
or time). Since the model predicts an out-
come variable on the basis of the probability
of being selected into a group, the predicted
outcome can be interpreted as the result of
self-selection into the group. Thus the
model has a causal element that is absent in
ordinary least squares models. A number of
studies in transport have used the Heckman
model, including Cooke and Ross (1999),
Deka (2002) and Vance and Iovanna (2007).
Studies in other fields have used it to com-
pare SAT scores for coached and uncoached
students (Briggs, 2004), treatment duration
for individuals treated by two different
drugs (Leslie and Ghomrawi, 2008), profits
for family-controlled and non-family firms
(Maury, 2006) and child-abuse reporting
cases for Black and White populations (Ards
et al., 1998).
The analysis of travel characteristics of
solo households is followed by a comparison
of the moving behaviour of solo and mar-
ried-couple households. Three binary logit
models were used with 2009 AHS data to
make these comparisons. Additional infor-
mation about the specific models has been
provided in the sections where they have
been used.
The Growth of American Solo
Households and the Potential
Reasons for Their Growth
To verify Klinenberg’s observations about
the growth of solo households, especially
their growth in cities, historical census
PUMS data were obtained from the
Integrated Public Use Microdata Series
(Ruggles et al., 2010). The changing shares
of solo households as well as population
and labour force from solo households,
obtained from this source, are presented in
Figure 1. As evident, the share of solo
households increased from 6 per cent to
almost 28 per cent between 1930 and 2010.
During this period, the share of married-
couple households decreased from 79 per
cent to 49 per cent. Because of the presence
of children in non-solo households, the
increase in the share of population in solo
households is less substantial than the
increase in the share of solo households.
However, the share of labour force from
solo households has been growing more
rapidly than population.
Consistent with Klinenberg’s assertion
about the growth of solo households in
cities, the figure shows that the share of
solo households, population and labour
force in central cities is significantly higher
than the country as a whole. While the
share of solo female labour force in central
cities has been higher than male labour
force since 1930, the share of solo male
636 DEVAJYOTI DEKA
at LOYOLA MARYMOUNT UNIV on January 8, 2016usj.sagepub.comDownloaded from
http://usj.sagepub.com/
labour force in central cities has increased
considerably since 1970. As of 2010, the
share of solo male and female labour force
in central cities was 15 per cent and 16 per
cent respectively.
Klinenberg identifies economic develop-
ment, the welfare state and cultural change
as the primary reasons for the growth of
solo households. He contends that the
rising status of women, new communica-
tions technologies, mass urbanisation and
increasing longevity contributed to the
trend. An additional reason for the growth
of solo households in the US could be the
diminishing attractiveness of marriage. As
Warren and Tyagi (2003) emphasised, mar-
riage for many working adults has become a
trap that leaves them with less savings for the
rainy day compared with solo households.
Since women have historically contribu-
ted more to the growth of solo households
in cities than men, Kleinenberg’s assertion
about the rising status of women as a
reason for the growth of solo households
requires further discussion. Available data
on women’s education and earnings seem
to provide some support to the claim that
women’s status has risen in recent decades.
In 1960, women received 35 per cent of the
bachelor’s degrees in the US, whereas they
received 58 per cent of the degrees in 2004
(Buchmann and DiPrete, 2006). The 2010
ACS PUMS data for persons in the labour
force 25 or older show that 51 per cent of
women from solo households have a bache-
lor’s degree or higher level of education,
whereas 47 per cent of women from mar-
ried-couple households and 45 per cent of
men from both types of household have
that distinction. Thus women from solo
households do have a certain level of advan-
tage in education over women from mar-
ried-couple households and men.
The increasing level of education for
women has been accompanied by an
increase in earnings. Although women from
Figure 1. Share of solo households, persons and labour force in the US, excluding group
quarters, 1930–2010.
Source: estimated from 1930–2000 censuses and 2010 ACS PUMS data.
SOLO HOUSEHOLDS IN THE US 637
at LOYOLA MARYMOUNT UNIV on January 8, 2016usj.sagepub.comDownloaded from
http://usj.sagepub.com/
both solo and married-couple households
still earn substantially less than male work-
ers, the increase in educational attainment
among women has reduced the gender gap
in earnings. That earned income for women
has increased in recent years is evident from
Figure 2, which shows the average earnings
of persons age 25 and older from solo and
married-couple households. Due to the
2007–09 recession, the earnings for all per-
sons in the labour force stagnated in 2010.
Yet a comparison between 1990 and 2010
shows that the gap between male and
female earnings decreased, in all places as
well as in central cities, for both married-
couple households and solo households.
Although the earnings for the female labour
force from married-couple households are
rapidly increasing, consistent with their
respective educational attainment, women
from solo households continue to earn
more.
Taking a cue from economists who
believe that marriage is an economic insti-
tution and that the decisions to marry and
divorce can be explained by utility theory
(for example, Becker, 1974; Becker et al.,
1977), one can argue that, due to increasing
earnings, women today feel less compelled
to marry for security than in the past.
Becker’s theory suggests that both husband
and wife benefited from production effi-
ciencies due to the division of labour in tra-
ditional households, but, as Isen and
Stevenson (2010) point out, these produc-
tion efficiencies have disappeared over time
due to technological advances. The decrea
se
in the production efficiencies can be
expected to reduce the motivation to marry
for both men and women, thereby leading
to the growth of solo households.
Another rationale for an individual to go
solo is the solo’s ability to move quickly
whenever desirable employment opportu-
nities arise. Their ability to move could
potentially allow them to earn a higher
salary in the job market. At an age of quick
job turnover, solo workers certainly have
an advantage.
What brings solos to cities? According
to Klinenberg, educated solo men and
women are increasingly flocking to large
cities of the US to live independent lives.
Klinenberg emphasised the liberal culture
as a reason for their concentration in
cities. However, as shown in Figure 2, an
earnings differential also makes cities
more attractive to solos than married-
couple households. While men from mar-
ried-couple households earn nearly the
same in central cities as elsewhere, both
men and women from solo households
earn far more by living in central cities.
Since men have continued to remain the
primary earners of married-couple house-
holds, the attractiveness of central cities
for these households is lower than solo
households.
The Distinctive Characteristics of
Solo Households in Past Studies
A number of studies that compared the resi-
dential location and travel patterns of differ-
ent types of household provide evidence of
the distinctive nature of solo households. In
a study for the San Francisco Bay Area, Bhat
and Guo (2007) found that individuals from
solo households are less likely to own cars
and more likely to live in areas with high
street density. In a study for the Louisville,
Kentucky–Indiana metropolitan area, Scott
and Horner (2008) found that persons from
solo households live in areas with greater
access to activities, especially commercial
activities within walking distance. Similarly,
Brownstone and Golob (2009) found in a
California study that, compared with others,
solo households locate more often in high-
density areas, drive fewer miles and con-
sume less fuel.
638 DEVAJYOTI DEKA
at LOYOLA MARYMOUNT UNIV on January 8, 2016usj.sagepub.comDownloaded from
http://usj.sagepub.com/
Studies from other countries provide
more extensive evidence about solo house-
holds. In a British study, Dargay and Hanly
(2007) found that solo workers are less
likely to commute by automobile than
others. Noting that workers from solo
households spend less time commuting,
Vandersmissen et al. (2003) concluded in a
study for Québec City, Canada, that devoid
of family constraints, solo workers have a
greater opportunity than others to locate
close to work. A study by Howley (2009
)
found that solo households are less likely
than others to move out of a central-city
area of Dublin. Eluru et al. (2009) found in
a study for Zurich, Switzerland, that solo
persons are more likely to move for educa-
tion and employment compared with per-
sons from multiple-person households.
Other studies have identified additional
travel and transport attributes that distin-
guish solo households from others. For
example, Roorda et al. (2010) found that
persons from solo households make more
trips per person compared with persons
from married-couple households in several
Canadian cities. Analysis conducted by
Arentze, et al. (2005) in the Netherlands
showed that solo individuals make more
shopping trips than others. However, as
noted by Newsome et al. (1998), solo indi-
viduals often have to make more trips than
individuals from multiple-person house-
holds because of their sole responsibility of
household activities.
To the detriment of solo households, sev-
eral studies have shown that married-couple
households can save significantly because of
economies of scale generated by the sharing
of goods and services within the household.
For example, Lazear and Michael (1980)
showed that married couples have an advan-
tage over solo households because they save
substantially on the consumption of shelter,
food and other goods because of sharing.
Similarly, Nelson (1988) found that larger
households generate more savings in the
consumption of shelter, food, household
Figure 2. Personal earned income in constant 1999 dollars for persons in the labour force
age 25 and over in the US, excluding group quarters, 1990–2010.
Source: estimated from 1990 and 2000 censuses and 2010 ACS PUMS data.
SOLO HOUSEHOLDS IN THE US 639
at LOYOLA MARYMOUNT UNIV on January 8, 2016usj.sagepub.comDownloaded from
http://usj.sagepub.com/
furnishings and transport compared with
smaller households. A number of other stud-
ies have provided additional insights about
the economies of scale due to sharing of
goods and services by household members
(Deaton and Paxson, 1998; Gan and Vernon,
2003; Fernández-Villaverde and Krueger,
2002; Browning et al. 2006). These studies
leave little doubt that solo households sacri-
fice the benefits of sharing housing space and
household goods and services. For the
growth of solo households, countervailing
factors must offset these sacrifices.
Who are the Solo in America?
Although Klinenberg cites a number of
potential reasons for the growth of solo
households, his objective was to describe a
phenomenon rather than to provide a sta-
tistical discourse. This study takes a more
methodological approach. To identify the
factors associated with the likelihood of
living in a solo household, three binary
logit models were used with 1 per cent
PUMS data for the US from the 2010 ACS.
The results are presented in Table 1. In
model 1, the likelihood of living in a solo
household is predicted by including all
individuals in the dataset excluding those
below age 25 and people living in group
quarters. In model 2, the data are further
restricted to workers in the 25–64 age
group to examine if their likelihood of
living in solo households differed from the
likelihood for all individuals. This model
was included especially in view of the large
proportion of persons in the labour force
from solo households in cities. Model 3 is
more elegant than model 1 and model 2 in
that it combines the first two models and it
can be used to show the relationships
between the dependent variable and the
independent variables separately for work-
ers and non-workers. In this model, all
variables are interacted with employment
status except for the dummy variable on
employment.
Living in a solo household is the depen-
dent variable in all three models, which was
coded 1 for persons living in a solo house-
hold and 0 for others. The independent vari-
ables in the models are the demographic and
socioeconomic characteristics of the indi-
viduals in the dataset. In addition, a dummy
variable on central city was included to
examine the extent to which central-city
living is associated with being solo.
The results of the three models in Table
1 are consistent with each other in terms of
the signs of the coefficients and their statis-
tical significance. Although many other
characteristics of individuals are associated
with the likelihood of living in a solo
household, the three models show that
ageing contributes to the likelihood more
than other characteristics. This is more evi-
dent in model 1 and model 3 than in model
2 because in these two models the persons
older than age 65 are also included. Model
1 shows that a person who is 85 or older is
almost 900 per cent more likely, and a
person between age 75 and 84 is almost 400
per cent more likely, to live in a solo house-
hold compared with a 25–34-year-old
person (e2.265 = 9.63 and e1.609 = 4.997).
Model 1 also shows that being female and
being African American increases the likeli-
hood of living in a solo household, whereas
being Hispanic, a foreign-born immigrant,
or a non-English speaker decreases the like-
lihood. The latter results are consistent with
the notion that immigrants are more
family-oriented than non-immigrants.
Consistent with Klinenberg’s assertion
about the growth of solo households in
cities, both models show a positive associa-
tion between central-city living and being
solo. Persons with personal earned income
between $40,000 and $60,000 are the most
likely to live in solo households, whereas
persons with lower and higher earnings are
640 DEVAJYOTI DEKA
at LOYOLA MARYMOUNT UNIV on January 8, 2016usj.sagepub.comDownloaded from
http://usj.sagepub.com/
T
a
b
le
1
.
B
in
a
ry
l
o
g
it
m
o
d
e
l
re
su
lt
s
o
n
th
e
li
k
e
li
h
o
o
d
o
f
li
v
in
g
in
a
so
lo
h
o
u
se
h
o
ld
M
od
el
1
L
ik
el
ih
oo
d
of
li
vi
n
g
so
lo
:
a
ll
p
er
so
n
s
M
od
el
2
L
ik
el
ih
oo
d
of
li
vi
n
g
so
lo
:
w
or
ke
rs
on
ly
M
od
el
3
L
ik
el
ih
oo
d
of
li
vi
n
g
so
lo
:
a
ll
p
er
so
n
s
(i
n
te
ra
ct
io
n
m
od
el
)b
C
oe
ff
i
c
ie
n
t
(b
)
S
.E
.
C
oe
ff
ic
ie
n
t
(b
)
S
.E
.
C
oe
ff
ic
ie
n
t
(b
)
S
.E
.
C
o
n
st
a
n
t
–
2.
8
0
7
*
0
.0
1
5
–
2
.7
6
1
*
0
.0
1
8
–
1
.3
8
3
*
0
.0
0
3
E
m
p
lo
ye
d
p
er
so
n
a
–
1
.3
5
0
*
0
.0
1
8
L
es
s
th
an
h
ig
h
sc
h
o
o
l
–
0.
0
4
3
*
0
.0
0
7
–
0
.3
2
0
*
0
.0
1
3
–
0
.2
4
7
*
0
.0
1
3
H
ig
h
sc
h
o
o
l
d
ip
l
o
m
a
–
0.
0
8
8
*
0
.0
0
5
–
0
.1
6
5
*
0
.0
0
8
–
0
.1
3
5
*
0
.0
0
8
S
o
m
e
co
ll
eg
e
o
r
as
so
ci
at
e’
s
d
eg
re
e
(r
ef
er
en
t)
B
ac
h
el
o
r’
s
d
eg
re
e
0
.0
3
5*
0
.0
0
6
0
.1
8
9
*
0
.0
0
8
0
.1
5
3
*
0
.0
0
8
P
o
st
-b
ac
ca
la
u
r
e
at
e
ed
u
ca
ti
o
n
0
.
0
1
3*
0
.
0
0
1
0
.
0
4
5
*
0
.0
0
2
0
.0
3
4
*
0
.0
0
2
F
em
al
ea
0
.2
4
1*
0
.0
0
4
-0
.0
60
*
0
.0
0
6
0
.0
9
7
*
0
.0
0
6
A
fr
ic
an
A
m
er
ic
an
a
0
.2
8
6*
0
.0
0
7
0
.3
2
1
*
0
.0
0
9
0
.3
3
2
*
0
.0
0
9
H
is
p
an
ic
a
–
0.
2
3
8
*
0
.0
1
0
-0
.2
35
*
0
.0
1
3
–
0
.
2
2
1
*
0
.0
1
3
F
o
re
ig
n
-b
o
rn
i
m
m
ig
ra
n
ta
–
0.
3
3
4
*
0
.0
0
9
–
0
.3
0
6
*
0
.0
1
2
–
0
.2
7
9
*
0
.0
1
2
L
an
gu
ag
e
at
h
o
m
e
is
E
n
gl
is
h
a
0
.
2
6
8*
0
.0
0
9
0
.
2
5
7
*
0
.0
1
3
0
.2
4
8
*
0
.0
1
3
C
en
tr
al
-c
it
y
re
si
d
en
ta
0
.4
9
4*
0
.0
0
6
0
.5
9
1
*
0
.0
0
8
0
.5
7
7
*
0
.0
0
8
A
ge
2
5
–
3
4
(r
ef
er
en
t)
A
ge
3
5
–
4
4
–
0.
1
3
1
*
0
.0
0
8
–
0
.1
43
*
0
.0
0
9
–
0
.
1
9
6
*
0
.0
0
9
A
ge
4
5
–
5
4
0
.2
7
4*
0
.0
0
8
0
.
2
0
3
*
0
.0
0
8
0
.1
1
9
*
0
.0
0
9
A
ge
5
5
–
6
4
0
.6
7
4*
0
.0
0
7
0
.
5
6
6
*
0
.0
0
8
0
.4
8
9
*
0
.0
0
9
A
ge
6
5
–
7
4
1
.0
4
2*
0
.0
0
8
0
.8
7
3
*
0
.0
1
2
A
ge
7
5
–
8
4
1
.
6
0
9*
0
.0
0
9
1
.3
1
1
*
0
.0
2
2
A
ge
8
5
o
r
h
ig
h
er
2
.2
6
5*
0
.0
1
1
1
.9
0
2
*
0
.0
6
1
P
er
so
n
al
ea
rn
ed
in
co
m
e
le
ss
th
an
$
2
0
k
0
.0
5
9*
0
.0
1
1
0
.2
0
8
*
0
.0
1
2
0
.1
5
0
*
0
.0
1
2
P
er
so
n
al
ea
rn
ed
in
co
m
e
$
2
0
–
40
k
0
.1
6
4*
0
.0
1
1
0
.3
4
3
*
0
.0
1
2
0
.2
7
7
*
0
.0
1
2
P
er
so
n
al
ea
rn
ed
in
co
m
e
$
4
0
–
60
k
0
.2
7
8*
0
.0
1
1
0
.3
9
1
*
0
.0
1
2
0
.3
5
2
*
0
.0
1
2
P
er
so
n
al
ea
rn
ed
in
co
m
e
$
6
0
–
80
k
0
.2
6
1*
0
.0
1
2
0
.3
3
3
*
0
.0
1
3
0
.3
1
2
*
0
.0
1
3
P
er
so
n
al
ea
rn
ed
in
co
m
e
$
8
0
–
10
0
k
0
.2
1
4*
0
.0
1
5
0
.2
5
3
*
0
.0
1
5
0
.2
4
5
*
0
.0
1
5
P
er
so
n
al
ea
rn
ed
in
co
m
e
$
1
0
0
k
o
r
h
ig
h
er
(r
ef
er
en
t)
(c
on
ti
n
u
ed
)
SOLO HOUSEHOLDS IN THE US 641
at LOYOLA MARYMOUNT UNIV on January 8, 2016usj.sagepub.comDownloaded from
http://usj.sagepub.com/
T
a
b
le
1
.
(C
o
n
ti
n
u
e
d
)
M
od
el
1
L
ik
el
ih
oo
d
of
li
vi
n
g
so
lo
:
a
ll
p
er
so
n
s
M
od
el
2
L
ik
el
ih
oo
d
of
li
vi
n
g
so
lo
:
w
or
ke
rs
on
ly
M
od
el
3
L
ik
el
ih
oo
d
of
li
vi
n
g
so
lo
:
a
ll
p
er
so
n
s
(i
n
te
ra
ct
io
n
m
od
el
)b
C
oe
ff
ic
ie
n
t
(b
)
S
.E
.
C
oe
ff
ic
ie
n
t
(b
)
S
.E
.
C
oe
ff
ic
ie
n
t
(b
)
S
.E
.
N
2
,0
5
8
,2
8
4
c
1
,2
5
2
,4
3
7
d
2
,0
5
8
,2
8
4
c
-2
lo
g
li
k
el
ih
o
o
d
1
,6
4
2
,3
5
0
8
6
8
,5
6
4
1
,7
1
5
,0
8
6
N
ag
el
k
er
k
e
R
2
0
.1
0
3
0
.0
4
2
0
.0
4
5
M
o
d
el
ch
i-
sq
u
ar
ed
1
2
6
,
5
9
5
(p
\
0
.0
0
0
1
)
2
7
,1
0
8
(p
\
0
.0
0
0
1
)
5
3
,8
5
8
(p
\
0
.0
0
0
1
)
a
D
u
m
m
y
va
ri
ab
le
,
co
d
ed
1
fo
r
sh
o
w
n
ca
te
go
ry
an
d
0
fo
r
o
th
er
s.
b
N
o
te
th
at
al
l
in
d
ep
en
d
en
t
va
ri
ab
le
s
ex
ce
p
t
‘E
m
p
lo
ye
d
p
er
so
n
’
ar
e
in
t
e
ra
ct
ed
va
ri
ab
le
s
in
m
o
d
el
3
.
c I
n
cl
u
d
es
al
l
p
er
so
n
s,
ag
e
2
5
an
d
o
ve
r,
ex
cl
u
d
i
n
g
gr
o
u
p
q
u
ar
te
rs
.
d
In
cl
u
d
es
al
l
w
o
r
k
er
s,
ag
e
2
5
–
6
4
,
ex
cl
u
d
in
g
gr
o
u
p
q
u
ar
te
rs
.
N
ot
e:
*
si
gn
if
ic
an
t
at
th
e
1
p
er
ce
n
t
le
ve
l.
S
ou
rc
e:
es
ti
m
at
ed
fr
o
m
th
e
2
0
1
0
A
m
er
ic
an
C
o
m
m
u
n
it
y
S
u
rv
ey
P
u
b
li
c
U
se
M
ic
ro
S
a
m
p
le
d
at
a.
642 DEVAJYOTI DEKA
at LOYOLA MARYMOUNT UNIV on January 8, 2016usj.sagepub.comDownloaded from
http://usj.sagepub.com/
less likely. Higher education increases the
likelihood of living in a solo household,
although the influence of higher education
is higher for workers than the general pop-
ulation. According to model 2, a worker
with a bachelor’s degree is about 21 per
cent more likely to live in a solo household
compared with a person with an associate
degree or some college education. For the
general population, being female increases
the likelihood of living in a solo household
by 27 per cent, but for workers aged 25–64,
women have a 6 per cent lower likelihood
of living in a solo household. Further exam-
ination of the data reveals that a reason for
this discrepancy is that a far larger number
of elderly widows live in solo households
compared with widowers. In male solo
households, 34 per cent are divorced, 42
per cent have never married and 14 per
cent are widowed, whereas in female solo
households, 29 per cent are divorced, 25
per cent have never married and an over-
whelming 40 per cent are widowed (poten-
tially due to women’s greater longevity than
men). In sum, although Klinenberg’s asser-
tions are mostly validated by the analysis,
the significant likelihood of living in a solo
household by the elderly suggests that these
individuals are being solo because of their
circumstances instead of going solo on their
own volition. Yet the analysis shows that
higher education, a middle-class income
and central-city residence are positively
associated with the likelihood of living in a
solo household.
The Distinctive Travel and Living
Characteristics of the American
Solo
The characteristics of solo households
found in the literature pertaining to other
developed countries can also be observed in
the US. Basic statistics from the 2009
NHTS, presented in Table 2, show that
men from solo households, on average,
travel a significantly shorter distance and
spend less time commuting to work than
men from married-couple households. Solo
women commute a shorter distance than
women from married-couple households,
but spend an almost identical amount of
time commuting. Solo persons’ higher pro-
pensity to use public transit and walk, and
a lower propensity to use personal vehicles,
is evident for both men and women. As
expected, a far smaller proportion of solo
households own vehicles compared with
married-couple households and the average
number of vehicles owned by solo house-
holds is also smaller than married-couple
households. Solo women have a lower pro-
pensity to own vehicles than solo men as
well as men and women from married-
couple households. Consistent with other
studies, American solo men and women
make a greater proportion of trips for shop-
ping/errands and personal business/obliga-
tions compared with men and women from
married-couple households.
The travel patterns of American solo
households are potentially associated with
their dwelling characteristics. As evident in
Table 2, less than half of the solo men and
women live in detached single houses,
whereas around 75 per cent of married
couples live in such dwellings. Almost 54
per cent of solo men and 45 per cent of the
solo women live in rented dwellings, com-
pared with about a quarter of the men and
women from married-couple households.
By living in rented homes, solo households
can minimise their commuting distance
and time, although they spend more on
shelter than married-couple households on
a per-earner basis. The 2009 US Consumer
Expenditure Survey data analysed by the
author shows that the average per-earner
shelter expenditure for an adult in a mar-
ried-couple household is $1394 for a
SOLO HOUSEHOLDS IN THE US 643
at LOYOLA MARYMOUNT UNIV on January 8, 2016usj.sagepub.comDownloaded from
http://usj.sagepub.com/
Table 2. Comparison of travel and dwelling characteristics of persons aged 25–64 from solo
households and married-couple households in the US
Couple Solo
Male Female Male Female
Mean trip distance (miles)a
All trip purposes 13.08 9.02 10.09 7.18
Work trips only 14.87 10.22 10.93 8.64
Mean trip duration (minutes)a
All trip purposes 22.14 18.17 20.57 18.05
Work trips only 27.09 21.78 22.69 21.55
Mode used for all trip purposes
(percentages)b
Car, truck, van and SUV 87.2 87.2 77.4 80.6
Public transit (excluding
taxi, ferry, school bus and
inter-regional transit)
1.5 1.6 4.2 4.1
Walk 8.7 10.0 14.0 13.5
Bicycle 1.0 0.3 2.2 0.2
Other 1.6 1.0 2.2 1.6
Total 100 100 100 100
Mode used for work trip purposes
(percentages)b
Car, truck, van and SUV 91.5 93.9 85.8 88.5
Public transit (excluding
taxi, ferry, school bus and
inter-regional transit)
3.1 2.8 6.1 7.0
Walk 2.3 2.4 3.9 3.7
Bicycle 1.2 0.2 1.5 0.2
Other 1.9 0.6 2.7 0.7
Total 100 100 100 100
Trip purpose (percentages)b
Home 33.8 33.6 32.7 31.1
Work 20.4 12.1 18.8 17.0
Shopping/errands 15.9 19.2 20.6 22.4
Social/recreational 10.1 10.3 11.6 10.0
Family personal business/
obligations
2.9 3.4 3.4 4.7
Meals 7.2 6.5 7.3 6.9
Other purposes 9.6 14.9 5.6 7.9
Total 100 100 100 100
Mean number of vehicles in
householda
2.40c 2.37c 1.28c 1.02c
Percentage of households
with at least one vehiclea
97.6 97.0 86.7 85.0
(continued)
644 DEVAJYOTI DEKA
at LOYOLA MARYMOUNT UNIV on January 8, 2016usj.sagepub.comDownloaded from
http://usj.sagepub.com/
quarter, whereas the expenditure for a solo
male is $1660 and for a solo female is $1561
(US Bureau of Labor Statistics, 2010).
Although the data presented in Table 2
suggest that the travel characteristics of the
solo are environmentally more sustainable
than individuals from married-couple
households, they do not show that their
distinctive travel characteristics are linked
to their decision to be solo. A model that
takes into account this link is the Heckman
sample selection model.
The Heckman model was used to com-
pare the commuting time and distance of
workers with the 2009 AHS national sample
data. The dataset was restricted to persons
aged 25–64 belonging to solo households or
married-couple households in urbanised
areas. Persons from other types of house-
hold, such as single-parent households,
were omitted so that comparisons could be
made directly between solo households and
married-couple households. The results of
the two models are presented in Table 3
(models 1 and 2). The models were esti-
mated by using both STATA and SAS to
ensure that the results are identical.
The Heckman model consists of two
components: a selection component and an
outcome component. In addition to the
coefficients and standard errors of the vari-
ables, the model generates two critical para-
meter estimates: sigma (s) and rho (r).
While sigma is the adjusted standard error
of the outcome model, rho represents the
correlation between the errors of the two
models and indicates whether the factors
affecting selection are associated with the
outcome. For example, if the selection group
is solo individuals, and rho is statistically sig-
nificant and negative, it would indicate that
the outcome variable, commuting time or
distance, for solos is shorter than individuals
from married-couple households because of
the selection effect.
In addition to the models on commuting
distance and time, a Heckman sample selec-
tion probit model estimating the likelihood
of commuting by automobile is presented in
Table 2 (model 3). Similar to the typical
Heckman model with a continuous outcome
variable, the binary outcome variable of this
model is predicted on the basis of a selection
component. The outcome component of the
Table 2. (Continued)
Couple Solo
Male Female Male Female
Percentage living in rented
homea
24.4 26.9 53.6 45.1
Dwelling type (percentages)b
Detached single house 76.4 74.8 45.3 48.3
Row house or townhouse 12.2 13.5 40.5 40.3
Other housing types 11.3 11.7 14.2 11.4
Total 100 100 100 100
aSolo male and female are different from married-couple men at the 5 per cent significance level
on t-test.
bDifferences are significant at the 5 per cent level on chi-squared test.
cAccording to 2010 ACS PUMS, these figures are 2.33, 2.28, 1.15 and 0.97 respectively.
Source: estimated from the 2009 National Household Transportation Survey.
SOLO HOUSEHOLDS IN THE US 645
at LOYOLA MARYMOUNT UNIV on January 8, 2016usj.sagepub.comDownloaded from
http://usj.sagepub.com/
T
a
b
le
3
.
H
e
c
k
m
a
n
m
o
d
e
l
re
su
lt
s
o
n
c
o
m
m
u
ti
n
g
d
is
ta
n
c
e
,
ti
m
e
a
n
d
th
e
li
k
e
li
h
o
o
d
o
f
u
si
n
g
a
u
to
m
o
b
il
e
fo
r
c
o
m
m
u
ti
n
g
M
od
el
1
C
om
m
u
te
d
is
ta
n
ce
M
od
el
2
C
om
m
u
te
ti
m
e
M
od
el
3
A
u
to
m
ob
il
e
u
se
C
oe
ff
ic
ie
n
t
(b
)
t-
va
lu
e
C
oe
ff
ic
ie
n
t
(b
)
t-
va
lu
e
C
oe
ff
ic
ie
n
t
(b
)
t-
va
lu
e
O
u
tc
om
e
co
m
p
on
en
t
C
o
n
st
an
t
5
.8
9
2
*
*
1
.8
3
1
5
.8
1
0
*
3
.3
5
0
.9
8
1
*
3
.3
2
A
g
e
o
f
p
er
so
n
0
.
4
6
2
*
3
.1
3
0
.5
3
4
*
2
.4
6
0
.0
0
7
0
.5
2
A
ge
sq
u
ar
ed
–
0
.0
0
5
*
–
3
.2
1
–
0
.0
0
6
*
–
2
.3
7
–
0
.0
0
0
1
–
0
.6
0
F
em
al
ea
–
1
.
3
9
2
*
–
3
.7
0
–
0
.9
1
6
*
*
–
1
.6
6
0
.0
7
8
*
2
.4
9
A
fr
ic
an
A
m
er
ic
an
a
1
.0
0
7
*
*
1
.8
2
2
.5
7
2
*
3
.1
5
–
0
.1
9
2
*
–
4
.0
0
H
is
p
an
ic
a
2
.4
7
2
*
3
.5
7
3
.
2
9
4
*
3
.2
3
0
.2
7
3
*
4
.9
4
R
en
te
ra
–
1
.4
8
1
*
–
3
.0
9
–
1
.
5
1
3
*
–
2
.1
5
–
0
.0
3
2
–
0
.8
0
N
u
m
b
er
o
f
b
ed
ro
o
m
s
in
d
w
el
li
n
g
0
.7
2
4
*
3
.2
0
0
.1
5
1
0
.4
6
0
.0
9
6
*
4
.7
3
N
u
m
b
er
o
f
ve
h
ic
le
s
in
h
o
u
se
h
o
ld
0
.5
5
7
*
2
.2
2
–
1
.0
6
0
*
–
2
.8
9
0
.5
0
2
*
1
7
.6
7
L
iv
es
in
ce
n
tr
al
ci
ty
a
–
3
.5
0
5
*
–
7
.
9
9
–
1
.3
1
0
*
–
1
.9
9
–
0
.5
6
7
*
–
1
6
.3
4
M
o
ve
d
in
th
e
p
as
t
5
ye
ar
sa
0
.4
0
2
0
.9
4
–
0
.8
4
0
–
1
.3
4
D
is
ta
n
ce
to
w
o
rk
in
m
il
es
0
.0
2
3
*
1
1
.8
1
S
ig
m
a
1
2
.7
0
3
*
8
1
.1
0
1
8
.5
2
2
*
8
5
.8
1
S
el
ec
ti
on
co
m
p
on
en
t
C
o
n
st
an
t
–
1
.0
9
4
*
–
2
6
.9
1
–
1
.0
9
7
*
–
2
6
.9
3
–
1
.0
9
3
*
–
2
8
.6
5
A
ge
2
5
–
3
4
0
.1
0
3
*
3
.5
6
0
.1
0
3
*
3
.5
6
0
.0
9
8
*
3
.4
5
A
ge
3
5
–
4
4
–
0
.
1
6
0
*
–
5
.
7
6
–
0
.1
6
0
*
–
5
.7
6
–
0
.1
3
5
*
–
5
.0
6
A
ge
4
5
–
5
4
-0
.0
6
2
*
–
2
.3
6
–
0
.0
6
2
*
–
2
.3
5
–
0
.0
6
4
*
–
2
.
5
4
A
ge
5
5
–
6
4
(r
ef
er
en
t)
A
fr
ic
an
A
m
er
ic
an
a
0
.3
4
7
*
1
0
.9
1
0
.3
4
8
*
1
0
.9
4
0
.3
3
8
*
1
0
.7
5
H
is
p
an
ic
a
–
0
.1
7
6
*
–
4
.9
7
–
0
.1
7
6
*
–
4
.9
7
–
0
.2
0
5
*
–
5
.
8
8
F
o
re
ig
n
b
o
rn
a
–
0
.3
5
4
*
–
1
1
.2
4
–
0
.3
5
2
*
–
1
1
.1
3
–
0
.2
9
6
*
–
1
0
.9
9
(c
on
ti
n
u
ed
)
646 DEVAJYOTI DEKA
at LOYOLA MARYMOUNT UNIV on January 8, 2016usj.sagepub.comDownloaded from
http://usj.sagepub.com/
T
a
b
le
3
.
(C
o
n
ti
n
u
e
d
)
M
od
el
1
C
om
m
u
te
d
is
ta
n
ce
M
od
el
2
C
om
m
u
te
ti
m
e
M
od
el
3
A
u
to
m
ob
il
e
u
se
C
oe
ff
ic
ie
n
t
(b
)
t-
va
lu
e
C
oe
ff
ic
ie
n
t
(b
)
t-
va
lu
e
C
oe
ff
ic
ie
n
t
(b
)
t-
va
lu
e
L
es
s
th
an
b
ac
h
el
o
r’
s
d
eg
re
e
(r
ef
er
an
t)
B
ac
h
el
o
r’
s
d
eg
re
e
0
.0
8
9
*
3
.6
5
0
.0
9
4
*
3
.8
8
0
.0
4
8
*
2
.1
1
P
o
st
-b
ac
ca
la
u
re
at
e
ed
u
ca
ti
o
n
0
.1
4
7
*
4
.9
7
0
.1
5
2
*
5
.1
5
0
.0
9
8
*
3
.5
3
S
al
ar
y
in
co
m
e
le
ss
th
an
$
2
0
k
–
0
.0
6
6
*
*
–
1
.6
6
–
0
.0
6
1
–
1
.5
4
–
0
.0
1
9
–
0
.5
2
S
al
ar
y
in
co
m
e
$
2
0
–
4
0
k
0
.4
1
6
*
1
0
.4
1
0
.4
1
6
*
1
0
.3
4
0
.3
9
9
*
1
0
.7
3
S
al
ar
y
in
co
m
e
$
4
0
–
6
0
k
0
.3
6
2
*
9
.2
0
0
.3
6
1
*
9
.1
2
0
.3
7
0
*
1
0
.0
5
S
al
ar
y
in
co
m
e
$
6
0
–
8
0
k
0
.2
1
9
*
5
.1
8
0
.2
1
7
*
5
.1
1
0
.2
2
0
*
5
.5
5
S
al
ar
y
in
co
m
e
$
8
0
–
1
0
0
k
0
.2
4
5
*
4
.9
9
0
.2
4
8
*
5
.0
6
0
.2
4
0
*
5
.2
3
S
al
ar
y
in
co
m
e
$
1
0
0
k
o
r
h
ig
h
er
(r
ef
er
en
t)
L
iv
es
in
ce
n
tr
al
ci
ty
a
0
.4
0
4
*
1
8
.6
9
0
.4
0
4
*
1
8
.6
9
0
.4
0
1
*
1
8
.6
9
R
h
o
–
0
.1
8
0
*
–
3
.7
0
–
0
.1
3
0
*
–
2
.3
4
–
0
.9
9
2
*
–
1
9
6
.0
1
L
am
b
d
a
–
2
.2
8
9
–
2
.4
1
2
N
2
2
,6
1
6
2
2
,6
1
6
2
2
,6
1
6
L
o
g
li
k
el
ih
o
o
d
–
2
9
,7
9
8
–
3
1
,6
2
5
–
1
2
,6
1
7
A
IC
5
9
,6
5
2
6
3
,3
0
6
2
5
,2
8
8
S
ch
w
ar
z
cr
it
er
io
n
5
9
,8
7
7
6
3
,5
3
0
2
5
,5
0
5
W
a
l
d
ch
i-
sq
u
ar
ed
1
7
9
(p
\
0
.0
0
0
1
)
4
9
(p
\
0
.0
0
0
1
)
1
0
4
2
(p
\
0
.0
0
0
1
)
a
D
u
m
m
y
va
ri
ab
le
,
co
d
ed
1
fo
r
sh
o
w
n
ca
te
go
ry
an
d
0
fo
r
o
th
er
s.
N
ot
es
:
*
si
gn
if
ic
an
t
at
th
e
5
p
er
ce
n
t
le
ve
l;
*
*
si
gn
if
ic
an
t
at
th
e
1
0
p
er
ce
n
t
le
ve
l.
S
ou
rc
e:
es
ti
m
at
ed
fr
o
m
th
e
2
0
0
9
A
m
er
ic
an
H
o
u
si
n
g
S
u
rv
ey
n
at
io
n
al
sa
m
p
le
.
SOLO HOUSEHOLDS IN THE US 647
at LOYOLA MARYMOUNT UNIV on January 8, 2016usj.sagepub.comDownloaded from
http://usj.sagepub.com/
three models, measuring commuting dis-
tance, commuting time and the likelihood of
automobile use, is shown in the top part of
Table 3, whereas the selection component,
measuring the likelihood of being solo, is
shown in the bottom part. In the selection
component of all three models, the depen-
dent variable is a dummy variable (solo = 1,
else 0), whereas the dependent variables in
the outcome component of model 1 and
model 2 are continuous (distance and time)
and binary in model 3 (automobile use = 1,
else 0).
The selection component of the three
models, which predicts the likelihood of
being solo, is consistent with the model
results in Table 1, where the likelihood of
being solo was predicted with 2010 ACS
PUMS data. Similar to the models in Table
1, the selection components of the models
in Table 3 show that the likelihood of being
solo is higher for central-city residents,
African Americans, highly educated persons
and persons with moderate earnings, while
it is lower for Hispanic persons, immi-
grants and persons in mid-life. The out-
come component of the three models also
shows results consistent with expectation.
According to the distance model (model 1),
women, renters and central-city residents
commute a shorter distance, whereas
African Americans, Hispanic persons, per-
sons living in large dwellings and persons
from households with a larger number of
vehicles commute a longer distance. These
results are consistent with the results in
Crane (2007), where distance was estimated
with a random-effects generalised least
squares model using past AHS data.
The commute time model (model 2) is
similar to the distance model for all vari-
ables except the number of vehicles in the
household. Although the number of vehicles
has a positive association with distance and
a negative association with commute time,
both results are consistent with expectation
because vehicles can reduce travel time. The
only variable that was not statistically signif-
icant in either of the models was the variable
on movers.
The negative sign of a statistically signifi-
cant rho in the commuting distance and
time models indicates that the selection into
the solo group decreases commute distance
and time. By using the parameter estimates
of the two models, it can be estimated that
an average solo person commutes 3.2 miles
and 3.4 minutes less than individuals from
married-couple households.
An additional variable, travel distance,
was included in the probit model on auto-
mobile use for commuting (model 3 in
Table 3) because mode choice widely varies
by distance. The variable on moving was
excluded because it could not be theorised
why or how moving would affect mode
choice. Although fewer variables are statisti-
cally significant in this model compared
with the models on commuting distance
and time, the parameter estimates are gener-
ally consistent with expectation, as they
show that central-city residents and African
Americans are less likely to commute by
automobile, whereas persons from house-
holds with a large number of vehicles and
persons living in larger dwellings are more
likely to do so. As expected, individuals
making longer trips are more likely to use
an automobile for commuting. Most impor-
tantly, the highly significant negative rho
indicates that persons from solo households
are less likely to use an automobile than per-
sons from married-couple households
because of a selection effect. According to
the model results, the automobile usage rate
for commuting for an average solo person is
4 per cent lower than that for an average
person from a married-couple household.
In sum, the models on commuting distance,
commuting time and automobile use for
commuting, estimated with AHS data, are
consistent with the statistics provided in
648 DEVAJYOTI DEKA
at LOYOLA MARYMOUNT UNIV on January 8, 2016usj.sagepub.comDownloaded from
http://usj.sagepub.com/
Table 2 from the 2009 NHTS. The models
provide further evidence that the travel pat-
terns of solo households are environmen-
tally more sustainable than the travel
patterns of persons from married-couple
households.
The structure of the Heckman model also
provides some insights about the relation-
ship between being solo, living in central
cities and having more sustainable commu-
tes. The selection component of the model
in Table 3 shows that the solos are more likely
to live in central cities and the outcome com-
ponent shows that, by living in cities, they
commute a shorter distance and use the auto-
mobile less often for commuting. The selection
and outcome components together indicate
that the higher propensity of the solo house-
holds to live in central cities adds to their pro-
pensity to commute a shorter distance and use
an automobile less often. However, living in
central cities is one of many reasons for their
observed travel patterns.
The Moving behaviour of the
American Solo
It can be hypothesised that solo persons dis-
play more sustainable travel characteristics
because: they have a greater ability to move;
and, they have a greater concern about job
proximity than acquiring larger housing
space compared with married-couple house-
holds. Solo households can be expected to
have a greater ability and propensity to
move because they have no constraints
related to other persons in the household,
such as spouse’s employment or children’s
school. Because they do not benefit from
sharing housing space, solo households can
also be expected to have a lower propensity
to move in order to acquire larger dwelling
units. It can be further hypothesised that
their lower attraction for housing space is
accompanied by a greater attraction for
proximity to work because they are the sole
earners in household with greater time con-
straints than an average adult from a mar-
ried-couple household. To test three specific
hypotheses—that solo households move
more often, move less to acquire larger
housing units and move more for greater
proximity to work—statistical models were
used with the 2009 AHS data.
The AHS inquires whether a respondent’s
household moved within the past 12
months. The responses to this question were
used to examine whether solo households
move more than married-couple house-
holds. Those who moved are subsequently
asked by the AHS the reasons for their
move. The response categories ‘‘needed a
larger house or apartment’’ and ‘‘to be
closer to work/school/other’’ were used to
examine the attraction of solo households
for housing space and proximity to work.
The restriction of the dataset to age 25–64 is
expected to eliminate most of those who
moved for greater proximity to school.
Among the movers from married couple-
households, 15 per cent men and 13 per
cent women aged 25–64 reported moving
for a larger dwelling, but only 3 per cent
men and 4 per cent women from solo
households mentioned doing so. In con-
trast, 7 per cent men and 8 per cent women
from married-couple households men-
tioned moving to be closer to work/school/
other, compared with 10 per cent men and 9
per cent women from solo households.
Since these differences may be because of
other characteristics of the individuals, the
relationship between household type and
moving was examined by using binary logit
models to control for variations in these
characteristics. Sample-selection probit
models, which would estimate the probabil-
ity of being solo first and use the probabil-
ities to predict moving, were considered
inappropriate because moving, the outcome
SOLO HOUSEHOLDS IN THE US 649
at LOYOLA MARYMOUNT UNIV on January 8, 2016usj.sagepub.comDownloaded from
http://usj.sagepub.com/
variable, refers to the past. Being associative,
binary logit models do not involve this con-
ceptual problem, although their results
cannot necessarily be interpreted as causal.
Results from three binary logit models
are presented in Table 4. In model 1, the
dependent variable is a dummy variable on
moving, which was coded 1 for movers and
0 for non-movers. In model 2, where the
coefficients were obtained with data from
movers only, the dummy dependent variable
indicates whether a household moved to
acquire a larger dwelling. Those who moved
to acquire larger dwelling were coded 1 and
the other movers were coded 0. Similarly, in
model 3, the dummy dependent variable
was coded 1 for those who moved for greater
proximity to work/school/other, and other
movers were coded 0.
As expected, model 1 shows that solo
men and women are more likely to move
than adult men and women from married-
couple households. According to the model
results, men and women from solo house-
holds are 30 per cent and 29 per cent more
likely to move in a 12-month period com-
pared with adult men and women from
married-couple households (e0.262 = 1.30
and e0.257 = 1.29). The model also shows
that younger adults and persons with lower
incomes are more likely to move compared
with older adults and persons with higher
incomes. Level of education does not seem
to influence the likelihood of moving to a
great extent, except that individuals with
the lowest level of education are less likely
to move than others. Consistent with expec-
tation, having children reduces the likeli-
hood of moving. The negative association
between the number of household vehicles
and moving can be interpreted as a trade-
off between transport cost and moving
because greater mobility using vehicles can
reduce the need for moving.
Fewer independent variables are statisti-
cally significant in model 2 and model 3
than in model 1, but both provide the
expected results. The coefficients of model 2
show that solo men and solo women respec-
tively, are 75 per cent and 67 per cent less
likely to move to acquire a larger dwelling
unit compared with adults from married-
couple households (1-e-1.399 = 0.75 and
1-e-1.123 = 0.67). Model 3 shows that both
solo men and solo women are more likely
to move for greater proximity to work com-
pared with adults from married-couple
households. However, only the variable on
solo men has a confidence level of 95 per
cent, whereas the variable on solo women
has a confidence level of 87 per cent. The
model shows that men and women from
solo households are 34 per cent and 27 per
cent more likely respectively to move for
job proximity compared with adults from
married-couple households. Model 3 also
shows that younger adults and individuals
with a high level of education are far more
likely to move for job proximity compared
with others. Overall, the model results are
consistent with the notion that adults from
married-couple households are more inter-
ested in acquiring larger dwelling space,
whereas solo households are more con-
cerned about job proximity. The lower con-
cern for larger housing space and a greater
concern for job proximity seemingly lead to
a more sustainable living and commuting
behaviour by the solo.
Summary of the Findings and
Implications
The growth of solo households and workers
can be perceived as a sign of recovery for
American central cities. A lower propensity
to live in detached single homes and a
greater propensity to live close to work,
walk and use public transport make them
more attractive to cities than two-earner
households. To the benefit of themselves
650 DEVAJYOTI DEKA
at LOYOLA MARYMOUNT UNIV on January 8, 2016usj.sagepub.comDownloaded from
http://usj.sagepub.com/
T
a
b
le
4
.
B
in
a
ry
lo
g
it
m
o
d
e
l
re
su
lt
s
o
n
th
e
li
k
e
li
h
o
o
d
o
f
m
o
v
in
g
,
m
o
v
in
g
fo
r
la
rg
e
r
d
w
e
ll
in
g
u
n
it
,
a
n
d
m
o
v
in
g
fo
r
jo
b
p
ro
x
im
it
y
M
od
el
1
L
ik
el
ih
oo
d
of
m
ov
in
g
M
od
el
2
L
ik
el
ih
oo
d
of
m
ov
in
g
to
a
la
rg
er
d
w
el
li
n
g
M
od
el
3
L
ik
el
ih
oo
d
of
m
ov
in
g
fo
r
jo
b
p
ro
xi
m
it
y
C
oe
ff
ic
ie
n
t
(b
)
S
.E
.
C
oe
ff
ic
ie
n
t
(b
)
S
.E
.
C
oe
ff
ic
ie
n
t
(b
)
S
.E
.
C
o
n
st
an
t
–
1
.7
88
*
0
.1
0
2
–
2
.3
6
8
*
0
.2
9
7
–
3
.7
2
8
*
0
.3
4
9
S
o
lo
m
al
e
0
.2
6
2
*
0
.0
5
2
–
1
.3
9
9
*
0
.2
0
0
0
.2
9
3
*
0
.1
4
6
S
o
lo
fe
m
al
e
0
.2
5
7
*
0
.0
5
5
–
1
.1
2
3
*
0
.1
9
9
0
.2
4
2
0
.1
5
8
M
ar
ri
ed
-c
o
u
p
le
m
al
e
an
d
fe
m
al
e
(r
ef
er
en
t)
S
al
ar
y
in
co
m
e
le
ss
th
an
$
2
0
k
0
.3
7
4
*
0
.0
7
3
–
0
.2
7
8
0
.1
9
9
0
.3
2
1
0
.2
3
3
S
al
ar
y
in
co
m
e
$
2
0
–
40
k
0
.2
5
1
*
0
.0
7
7
–
0
.
1
4
8
0
.2
0
8
0
.3
4
7
0
.2
4
0
S
al
ar
y
in
co
m
e
$
4
0
–
60
k
0
.1
1
9
0
.0
7
7
–
0
.0
1
7
0
.2
0
8
–
0
.1
0
1
0
.2
5
0
S
al
ar
y
in
co
m
e
$
6
0
–
80
k
0
.0
6
0
0
.0
8
3
–
0
.0
7
0
0
.2
2
3
0
.1
1
5
0
.2
6
0
S
al
ar
y
in
co
m
e
$
8
0
–
10
0
k
0
.1
7
4
*
0
.0
9
6
–
0
.0
1
2
0
.2
5
3
–
0
.0
1
8
0
.3
0
0
S
al
ar
y
in
co
m
e
$
1
0
0
k
o
r
h
ig
h
er
(r
ef
er
en
t)
A
ge
2
5–
3
4
2
.0
5
1
*
0
.0
5
5
0
.3
4
4
0
.1
8
4
1
.2
0
7
*
0
.2
2
2
A
ge
3
5–
4
4
1
.0
5
9
*
0
.0
5
6
0
.3
0
8
0
.1
9
3
1
.0
3
6
*
0
.2
3
3
A
ge
4
5–
5
4
0
.5
1
3
*
0
.0
5
4
0
.0
1
4
0
.2
0
6
0
.7
4
7
*
0
.2
4
0
A
ge
5
5–
6
4
(r
ef
er
en
t)
W
h
it
e
ra
ce
a
–
0
.2
66
*
0
.0
4
4
0
.0
1
0
0
.1
2
3
–
0
.2
3
9
*
0
.1
2
1
H
is
p
an
ic
a
0
.2
0
0
*
0
.0
5
2
0
.0
8
4
0
.1
3
4
0
.2
1
5
0
.1
5
1
L
es
s
th
an
h
ig
h
sc
h
o
o
l
–
0
.5
50
*
0
.1
5
2
0
.5
6
8
0
.3
8
6
–
0
.4
9
7
0
.6
0
3
H
ig
h
sc
h
o
o
l
d
ip
lo
m
a
–
0
.0
73
0
.0
4
5
0
.1
5
6
0
.1
3
0
–
0
.1
1
5
0
.1
4
9
S
o
m
e
co
ll
eg
e
o
r
as
so
ci
at
e
d
eg
re
e
(r
ef
er
en
t)
B
ac
h
el
o
r’
s
d
eg
re
e
–
0
.0
73
0
.0
4
7
0
.2
6
4
*
0
.1
3
3
0
.3
9
5
*
0
.1
3
9
P
o
st
-b
ac
ca
la
u
re
at
e
ed
u
ca
ti
o
n
0
.0
5
7
0
.0
5
6
0
.3
0
8
0
.1
5
8
0
.6
2
2
*
0
.1
6
0
C
h
il
d
u
n
d
er
1
8
p
re
se
n
t
in
h
o
u
se
h
o
ld
a
–
0
.3
02
*
0
.0
4
7
0
.6
1
9
0
.1
1
1
–
0
.2
2
5
0
.1
4
9
N
u
m
b
er
o
f
h
o
u
se
h
o
ld
ve
h
ic
le
s
–
0
.2
90
*
0
.0
2
0
0
.0
4
3
0
.0
5
5
0
.0
7
3
0
.0
6
0
N
2
4
,7
0
6
b
4
8
8
8
c
4
88
8
c
–
2
lo
g
li
k
el
ih
o
o
d
2
1
,7
5
6
3
0
1
0
2
78
6
N
ag
el
k
er
k
e
R
2
0
.1
7
4
0
.1
0
6
0
.0
4
0
M
o
d
el
ch
i-
sq
u
ar
ed
2
82
2
(p
\
0
.0
0
0
1
)
2
6
0
(p
\
0
.0
0
0
1
)
8
8
(p
\
0
.0
0
0
1
)
a
D
u
m
m
y
va
ri
ab
le
,
co
d
ed
1
fo
r
sh
o
w
n
ca
te
go
ry
an
d
0
fo
r
o
th
er
s.
b
In
cl
u
d
es
al
l
p
er
so
n
s
in
ag
e
2
5
–
6
4
fr
o
m
so
lo
an
d
m
ar
ri
ed
co
u
p
le
h
o
u
se
h
o
ld
s
c I
n
cl
u
d
es
al
l
m
o
ve
rs
in
ag
e
2
5
–
6
4
fr
o
m
m
ar
ri
ed
-c
o
u
p
le
an
d
so
lo
h
o
u
se
h
o
ld
s
N
ot
e:
*
S
ig
n
if
ic
an
t
at
th
e
5
p
er
ce
n
t
le
ve
l.
S
ou
rc
e:
es
ti
m
at
ed
fr
o
m
th
e
2
0
0
9
A
m
er
ic
an
H
o
u
si
n
g
S
u
rv
ey
n
at
io
n
al
sa
m
p
le
.
SOLO HOUSEHOLDS IN THE US 651
at LOYOLA MARYMOUNT UNIV on January 8, 2016usj.sagepub.comDownloaded from
http://usj.sagepub.com/
and their employers, they can move more
freely than married couples to take advan-
tage of employment opportunities.
This research showed that, in addition to
ageing, the increase in education and earn-
ings of women could be a reason for the
growth of solo households. By living in cen-
tral cities, solo workers benefit more than
the traditional bread-winners from married-
couple households. This research further
showed that solo households are more likely
to move for job proximity and less likely to
move to acquire large dwelling units com-
pared with married-couple households. Solo
households have been able to bear the excess
expenses on shelter and transport so far
because of an earnings advantage, but if
their costs increase relative to their earnings,
the growth of solo households may slow
down. For the continued growth of solo
households in cities, there will have to be a
continued increase in earnings, stabilisation
of housing and transport costs, or both.
Their earnings will increase if jobs increase
because of higher demand for labour.
Although solo households have so far
exhibited living and travel patterns that are
conducive to environmentally sustainable
cities, there is no guarantee that they
will continue to exhibit these patterns. For
example, despite their higher propensity to
use public transit, according to PUMS
data, the share of transit commuting trips
by 25–64-year-old female solo workers
decreased from 9.6 per cent to 7.8 per cent
between 1990 and 2010, while the share for
solo male workers decreased from 6.5 per
cent to 6.2 per cent. During this period,
average commuting time by transit for solo
female workers increased from 40 minutes
to almost 46 minutes, while the time for
solo male workers increased from 39 min-
utes to 45 minutes. The decrease in transit
usage and increase in transit commuting
time were accompanied by an increase in
the average number of vehicles owned by
both solo females (from 0.90 to 0.97) and
solo men (from 1.13 to 1.15).
To provide better incentives to solo house-
holds, planners should address local public
transport needs. Although solo workers are still
far more likely to commute by transit and
spend less time commuting than workers from
married-couple households, if the quality of
transit service deteriorates, the commuting pat-
terns of solo workers could potentially become
less environmentally friendly.
Planners should also be concerned about
the long-term consequences of the growth of
solo households in cities. Although the living
and commuting patterns of solo households
are more sustainable than those of married-
couple households, many older adults are
solo because of their spouse’s death, divorce,
or separation. In the absence of other persons
in the household to take care of their travel
needs, the older adults from solo households
require special attention from society. Many
elderly solos need special transport services.
When today’s young solos become elderly in
future decades, the need for these services
will only increase.
Funding
This research received no specific grant from any
funding agency in the public, commercial or not-
for-profit sectors.
References
Ards, S., Chung, C. and Myers, S. L. Jr (1998) The
effects of sample selection bias on racial differ-
ences in child abuse reporting, Child Abuse
and Neglect, 23(12), pp. 103–115.
Arentze, T. A., Oppewal, H. and Timmermans,
H. J. P. (2005) A multipurpose shopping trip
model to assess retail agglomeration effects,
652 DEVAJYOTI DEKA
at LOYOLA MARYMOUNT UNIV on January 8, 2016usj.sagepub.comDownloaded from
http://usj.sagepub.com/
Journal of Marketing Research, 42, pp. 109–
115.
Becker, G. (1974) A theory of marriage, in: T. W.
Schultz (Ed.) Economics of the Family, pp. 293–
344. Chicago, IL: University of Chicago Press.
Becker, G. S., Landes, E. M. and Michael, R. T.
(1977) An economic analysis of marital
instability, The Journal of Political Economy,
85(6), pp. 1141–1187.
Bhat, C. R. and Guo, J. Y. (2007) A comprehen-
sive analysis of built environment characteris-
tics on household residential choice and auto
ownership levels, Transportation Research
Part B, 41, pp. 506–526.
Briggs, D. C. (2004) Casual inference and the
Heckman model, Journal of Educational and
Behavioral Statistics, 29(4), pp. 397–420.
Browning, M., Chiappori, P. and Lewbel, A.
(2006) Estimating consumption economies of
scale, adult equivalence scales, and household
bargaining power. Discussion Paper No. 289,
Department of Economics, Oxford University
(www.economics.ox.ac.uk/Research/wp/pdf/
paper289 ; accessed 11 June 2012).
Brownstone, D. and Golob, T. F. (2009) The
impact of residential density on vehicle usage
and energy consumption, Journal of Urban
Economics, 65, pp. 91–98.
Buchmann, C. and DiPrete, T. A. (2006) The
growing female advantage in college comple-
tion: the role of family background and aca-
demic achievement, American Sociological
Review, 71, pp. 515–541.
Cervero, R. (1989) Jobs–housing balancing and
regional mobility, Journal of the American
Planning Association, 55(2), pp. 136–150.
Cooke, T. J. and Ross, S. L. (1999) Sample selec-
tion bias in models of commuting time,
Urban Studies, 36(9), pp. 1597–1611.
Crane, R. (2007) Is there a quiet revolution in
women’s travel? Revisiting the gender gap in
commuting, Journal of the American Planning
Association, 73(3), pp. 298–316.
Dargay, J. and Hanly, M. (2007) Volatility of car
ownership, commuting mode and time in the
UK, Transportation Research Part A, 41, pp.
934–948.
Deaton, A. and Paxson, C. (1998) Economies
of scale, household size, and demand for
food, Journal of Political Economy, 106(5),
pp. 897–930.
Deka, D. (2002) Predicting commute time of
non-workers in the context of welfare reform,
Journal of Urban Affairs, 24(3), pp. 333–352.
Downs, A. (1992) Stuck in Traffic: Coping with
Peak-hour Traffic Congestion. Washington,
DC: The Brookings Institution Press.
Eluru, N., Sener, I. P., Bhat, C. R., Pendyala, R.
M. and Axhausen, K. W. (2009) Understand-
ing residential mobility: joint model of the
reason for residential relocation and stay
duration, Transportation Research Record,
2133, pp. 64–74.
Fernández-Villaverde, J. and Krueger, D. (2002)
Consumption over the life cycle: some facts from
consumer expenditure survey data. Working
Paper No. 9382, National Bureau of Eco-
nomic Research, Cambridge, MA (www.nber.
org/papers/w9382; accessed 11 June 2012).
Gan, L. and Vernon, V. (2003) Testing the Barten
model of economies of scale in household
consumption: toward resolving a paradox of
Deaton and Paxson, Journal of Political Econ-
omy, 11(6), pp. 1361–1377.
Heckman, J. (1976) The common structure of
statistical models of truncation, sample selec-
tion and limited dependent variables and a
simple estimator for such models, Annals of
Economic and Social Measurement, 5(4), pp.
475–492.
Heckman, J. (1979) Sample selection bias as a
specification error, Econometrica, 47(1), pp.
153–161.
Howley, P. (2009) Attitudes towards compact
city living: towards a greater understanding
of residential behavior, Land Use Policy, 26,
pp. 792–798.
Isen, A. and Stevenson, B. (2010) Women’s edu-
cation and family behavior: trends in mar-
riage, divorce and fertility. Working Paper
No. 15725, National Bureau of Economic
Research, Cambridge, MA (www.nber.org/
papers/w15725; accessed 12 June 2012).
Klinenberg, E. (2012) Going Solo: The Extraordi-
nary Rise and Surprising Appeal to Living
Alone. New York: The Penguin Press.
Lazear, E. P. and Michael, R. T. (1980) Family
size and the distribution of real per capita
income, American Economic Review, 70(1),
pp. 91–107.
Leslie, R. S. and Ghomrawi, H. (2008) The use
of propensity scores and instrumental variable
SOLO HOUSEHOLDS IN THE US 653
at LOYOLA MARYMOUNT UNIV on January 8, 2016usj.sagepub.comDownloaded from
http://usj.sagepub.com/
methods to adjust for treatment selection bias.
Paper No. 366-2008, SAS Global Forum
(www2.sas.com/proceedings/forum2008/366-
2008 ; accessed 20 July 2012).
Maury, B. (2006) Family ownership and firm
performance: empirical evidence from west-
ern European corporations, Journal of Corpo-
rate Finance, 12, pp. 321–341.
Mieszkowski, P. and Mills, E. S. (1993) The
causes of metropolitan suburbanization,
Journal of Economic Perspectives, 7(3), pp.
135–147.
Nelson, J. A. (1988) Household economies of
scale in consumption: theory and evidence.
Econometrica, 56(6), pp. 1301–1314.
Newsome, T. H., Walcott, W. A. and Smith, P.
D. (1998) Urban activity spaces: illustrations
and application of a conceptual model for
integrating the time and space dimensions,
Transportation, 25, pp. 357–377.
Roorda, M. J., Páez, A., Morency, C.,
Mercado, R. and Farber, S. (2010) Trip gen-
eration of vulnerable populations in three
Canadian cities: a spatial ordered probit
approach, Transportation, 37, pp. 525–548.
Ruggles, S. J., Alexander, T., Genadek, T.,
Goeken, R. et al. (2010) Integrated public use
microdata series: version 5.0 [machine-read-
able database]. University of Minnesota,
Minneapolis, MN.
Scott, D. M. and Horner, M. W. (2008) The role
of urban form in shaping access to opportuni-
ties: an exploratory spatial data analysis, Journal
of Transport and Land Use, 1(2), pp. 89–119.
US Bureau of Labor Statistics (2010) 2009 con-
sumer expenditure survey. CD-ROM. US
Bureau of Labor Statistics, Washington, DC.
Vance, C. and Iovanna, R. (2007) Gender and
the automobile: analysis of nonwork service
trips, Transportation Research Record, 2013,
pp. 54–61.
Vandersmissen, M., Villeneuve, P. and
Thériault, M. (2003) Analyzing changes in
urban form and commuting time, The Pro-
fessional Geographer, 55(4), pp. 446–463.
Warren, E. and Tyagi, A. W. (2003) The Two-
income Trap: Why Middle-class Parents Are
Going Broke. New York: Basic Books.
654 DEVAJYOTI DEKA
at LOYOLA MARYMOUNT UNIV on January 8, 2016usj.sagepub.comDownloaded from
www2.sas.com/proceedings/forum2008/366-2008
http://usj.sagepub.com/
