See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/33041237

Explaining Homelessness: a Critical Realist Perspective

Article in Housing Theory and Society · April 2005

DOI: 10.1080/14036090510034563 · Source: OAI

CITATIONS READS

141 2,721

1 author:

Suzanne Fitzpatrick
Heriot-Watt University
81 PUBLICATIONS 1,872 CITATIONS

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

ESRC Welfare Conditionality: Sanctions, Support and Behaviour Change View project

Severe Poverty, Destitution and Complex Needs View project

All content following this page was uploaded by Suzanne Fitzpatrick on 09 June 2015.

The user has requested enhancement of the downloaded file.

This article was downloaded by: [Heriot-Watt University]
On: 29 January 2015, At: 01:17
Publisher: Routledge
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered
office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Housing, Theory and Society

Publication details, including instructions for authors and
subscription information:
http://www.tandfonline.com/loi/shou20

Explaining Homelessness: a Critical

Realist Perspective
Suzanne Fitzpatrick
a
Centre for Housing Policy , University of York , UK
b
Centre for Housing Policy , University of York , Heslington,
York, YO10 5DD, UK E-mail:
Published online: 19 Aug 2006.

To cite this article: Suzanne Fitzpatrick (2005) Explaining Homelessness: a Critical Realist
Perspective, Housing, Theory and Society, 22:1, 1-17, DOI: 10.1080/14036090510034563

To link to this article: http://dx.doi.org/10.1080/14036090510034563

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the
“Content”) contained in the publications on our platform. However, Taylor & Francis,
our agents, and our licensors make no representations or warranties whatsoever as to
the accuracy, completeness, or suitability for any purpose of the Content. Any opinions
and views expressed in this publication are the opinions and views of the authors,
and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content
should not be relied upon and should be independently verified with primary sources
of information. Taylor and Francis shall not be liable for any losses, actions, claims,
proceedings, demands, costs, expenses, damages, and other liabilities whatsoever
or howsoever caused arising directly or indirectly in connection with, in relation to or
arising out of the use of the Content.

This article may be used for research, teaching, and private study purposes. Any
substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing,
systematic supply, or distribution in any form to anyone is expressly forbidden. Terms &
Conditions of access and use can be found at http://www.tandfonline.com/page/terms-
and-conditions
 Housing, Theory and Society,
Vol. 22, No. 1, 1–17, 2005

Explaining Homelessness: a Critical

Realist Perspective

SUZANNE FITZPATRICK
Downloaded by [Heriot-Watt University] at 01:17 29 January 2015

Centre for Housing Policy, University of York, UK

ABSTRACT This article attempts to engage with developing critical realist perspectives in
housing and urban policy to propose a more rigorous framework for analysing the causes of
homelessness. The article is framed mainly in the context of the extensive UK literature on this
topic, but the theoretical arguments it pursues are intended to have wider applicability. It
contends that the prevailing ‘‘new orthodoxy’’ in explanations of homelessness, which attempts to
integrate both ‘‘structural’’ and ‘‘individual’’ causes, is useful at a descriptive level, but is
unsatisfactory at a more profound conceptual level. Previous attempts to provide more
theoretically informed accounts of homelessness – including positivist, social constructionist,
feminist and postmodernist/poststructuralist – are also critiqued from a critical realist standpoint.
The complex, emergent and non-linear explanatory framework employed by realists is argued to
enable a coherent causal analysis to be maintained in the face of the diverse circumstances
associated with homelessness. Poverty, spatial concentrations of disadvantage and domestic
violence are used as illustrative examples of potential inter-related causes of homelessness to
sketch out a preliminary realist account of this persistent social problem.

KEY WORDS: Homelessness, Critical realism

Next to the US, the UK has probably the largest body of research on homelessness
in the developed world (Fitzpatrick & Christian, in press), with most of it rooted in
the housing and social policy research traditions. The ‘‘causes’’ of homelessness have
been much discussed within this literature, but the debate has suffered from a lack of
conceptual and theoretical clarity (Neale 1997). The disparate causal factors thought
to be associated with homelessness, such as unemployment, housing shortages,
mental illness and relationship breakdown, are often presented in an undifferentiated
list, with neither their relationship to each other nor to wider explanatory
frameworks rigorously investigated. The very concept of ‘‘causes’’ in relation to
homelessness is usually treated as either unproblematic (Greve 1991, Johnson,
Murie, Nauman & Yanetta 1991) or else is dismissed as misconceived (Pleace 1998,
Randall & Brown 1999). Furthermore, and despite the emphasis placed by most

Correspondence Address: Suzanne Fitzpatrick, Centre for Housing Policy, University of York, Heslington,
York, YO10 5DD, UK. Email: [email protected]

1403-6096 Print/1651-2278 Online/05/010001–17 # 2005 Taylor & Francis Ltd

DOI: 10.1080/14036090510034563
 2 S. Fitzpatrick

academic commentators on the ‘‘structural’’ causes of homelessness, the concept of a

‘‘social structure’’ remains ill-defined within this body of work.
This paper contends that ‘‘critical realism’’ provides a key theoretical frame-
work within which to advance our understanding of the causes of homelessness.
While the paper is framed mainly in the context of the extensive UK literature on this
topic, the conceptual arguments it pursues are intended to have much wider
applicability. ‘‘Realism’’ is not a substantive social theory, but rather a philosophical
doctrine which asserts that ‘‘…the world has an existence independent of our
perception of it’’ (Williams & May 1996:81), with ‘‘critical realism’’ focusing
specifically on the nature of explanations in the social (and natural) world (Bhaskar
1989, Sayer 1992, 2000, Archer 1995, Williams 2003). While ‘‘social construction-
ism’’ (within the broad ‘‘interpretivist’’ school of thought) continues to dominate
Downloaded by [Heriot-Watt University] at 01:17 29 January 2015

housing studies with a theoretical bent (Somerville & Bengtsson 2002a), critical
realism has come to prominence in sociology and elsewhere in social science
(Williams 2003) and is beginning to emerge in specific areas of housing and urban
research (Pawson & Tilley 1997, Allen 2000, Somerville & Bengtsson 2002a, b). Two
recent papers by Williams (2001, 2003) discussed in detail below, have used
homelessness as an illustrative example in elaborating on the methodological
possibilities and challenges of critical realism for empirical research in the social
sciences.
This paper begins by summarizing the critical realist approach to causation within
social science, before critiquing existing explanations of homelessness from this
perspective. It argues that most existing accounts of homelessness (implicitly or
explicitly) assume a positivistic notion of causality, or, alternatively, focus only on
the ‘‘social construction’’ of homelessness, bypassing the crucial issue of explanation
entirely. The arguments offered by feminists in the housing and homelessness field
are critiqued, but are preserved in modified form. The contribution of Giddens’
(1984) theory of ‘‘structuration’’ is highlighted, but is placed in the wider context of
realist research approaches. The paper concludes by considering the methodological
and ontological implications of a critical realist approach for explaining
homelessness in the UK and elsewhere.

The Critical Realist Challenge to Positivism and Interpretivism

Realism has been used to challenge both ‘‘positivist’’ and ‘‘interpretivist’’ approaches
to ‘‘causation’’ in the social world. Positivists traditionally adopt David Hume’s
‘‘constant conjunctions’’ theory of causality which, in modern terminology,
translates into the search for observable ‘‘empirical regularities’’, that is, statistically
significant correlations between ‘‘variables’’ (Hollis 1994). ‘‘Interpretivists’’, on the
other hand, are primarily concerned with the ‘‘meanings’’ people attach to social
situations, with some arguing that human actions are not governed by ‘‘cause’’ and
‘‘effect’’ at all, but rather by the rules that we use to interpret the social world
(Williams & May 1996). Realists concur with many aspects of the interpretivist
critique of positivism – and in particular acknowledge the centrality of human
perceptions and reasoning to the study of social science – but do not accept that
social science can be reduced wholly to the interpretation of meaning:
 Explaining Homelessness 3

… there is much more to the social world than agent’s understandings of it. [In
particular] real structures in the social world can impose themselves upon
agents both in a way they do not understand and without agents’ knowledge of
their existence. (Williams 2003)

While realists ‘‘rescue’’ causality from the dismissal of interpretivists, their

definition is entirely different to that of the positivists. They regard causal powers as
necessary tendencies of social objects and structures which may or may not be
activated (and produce ‘‘actual’’ effects) depending on conditions (Sayer 1992). The
presence of other (contingently related) causal mechanisms may often – or even
always – prevent correspondence between cause and effect, which is why the presence
(or absence) of empirical regularities is not a reliable guide to the (non-)existence of
Downloaded by [Heriot-Watt University] at 01:17 29 January 2015

‘‘real’’ causal powers. Realist explanations of actual social events and phenomena
are not ‘‘mono-causal’’ and deterministic, but rather ‘‘complex’’ (with intricate
feedback loops linking multiple causal mechanisms); ‘‘emergent’’ (from this
complexity new properties may emerge which cannot be deduced from the individual
components); and ‘‘non-linear’’ (small changes in these complex relationships can
bring about sudden and dramatic outcomes) (Byrne 1998, 1999, Williams 2001,
2003).
Social structures are central to the realist analysis of causation, and are defined by
Sayer (1992) as ‘‘sets of internally related objects or practices’’ existing at a range of
levels:

Contrary to common assumption, structures include not only big social objects
such as the international division of labour but small ones at the interpersonal
and personal levels (e.g. conceptual structures) and still smaller non-social ones
at the neurological level and beyond. (p. 92)

Unlike many natural structures, social structures are only relatively enduring
(Stones 2001); the most durable are argued to be those which ‘‘…lock their
occupants into situations which they cannot unilaterally change and yet in which it is
possible to change between existing positions’’ (Sayer 1992:95–96). As social
structures are dependent on human actors to reproduce them, people can also effect
their transformation. At the same time, these pre-existing structures constrain (and
enable) human actions, with some people having more options allowed them by
structures than others (the comparison between Giddens’ ‘‘structuration’’ theory and
the ‘‘analytical dualism’’ of realist ontology is discussed below).

Critiquing Existing Explanations of Homelessness from a Realist Perspective

This section of the paper uses the critical realist perspective to critique existing
explanations of homelessness in the UK and elsewhere. It begins by tracing the
developing understanding of homelessness in the UK over the post-war period, with
the predominant shift being from ‘‘individual’’ to ‘‘structural’’ explanations. It
then explores how a ‘‘new orthodoxy’’ has arisen which attempts to combine
these polarized positions, but argues that this is essentially pragmatic rather than
theoretically robust. Attempts to impose more theoretically rigorous understandings
 4 S. Fitzpatrick

of homelessness from positivist, interpretivist and feminist perspectives are then

critically reviewed, before the contribution of ‘‘structuration’’ theory, and its
compatibility with a realist perspective, is highlighted.

Emergence of the ‘‘New Orthodoxy’’– a Practically Adequate Response?

Explanations of homelessness in the UK and elsewhere have traditionally been
divided into two broad categories: ‘‘individual’’ and ‘‘structural’’ (Neale 1997).
Individual explanations focus on the personal characteristics and, especially,
behaviours of homeless people. Macro-structural explanations, on the other hand,
locate the causes of homelessness in broader social and economic structures, such as
adverse housing and labour market conditions, reduced social security protection,
Downloaded by [Heriot-Watt University] at 01:17 29 January 2015

rising levels of poverty and increasing family fragmentation. Up until the 1960s,
explanations of homelessness in the UK tended to emphasize individual pathology,
often focusing on the ill-health and/or substance dependencies of homeless people.
But the screening of a popular television drama about a homeless family, ‘‘Cathy
Come Home’’, in 1966 and the establishment of the housing pressure group Shelter in
the same year prompted a significant shift towards more structurally-orientated
accounts. This was reinforced by a series of academic studies which forcefully put the
case that homelessness was the result of malign social and economic forces (see
Robson & Poustie 1996 for a more detailed historical account). The most influential
of these studies was a major Government-funded report published in 1981, which
attributed homelessness primarily to an insufficient supply of affordable accom-
modation for those in weak economic positions (Drake, O’Brien, & Beiuyck 1981).
However, this housing market-based account of homelessness quickly ran into
trouble during the 1980s as research repeatedly demonstrated the non-housing
problems experienced by many single homeless people, particularly with regards to
mental health, drugs and alcohol (Pleace 1998). Most academic commentators
therefore now attempt to weave together consideration of both macro-structural and
individual factors in their explanations of homelessness (Dant & Deacon 1989,
Fitzpatrick 1998, Kennett & Marsh 1999, Fitzpatrick, Kemp & Klinker 2000), but at
the same time they usually continue to assert the overall primacy of structural causes.
This leads them to a position described by Pleace (2000) as the ‘‘new orthodoxy’’ (a
similar standpoint has also been adopted by some US authors, see, for example,
Metraux & Culhane 1999). The key assertions of this new orthodoxy are as follows:

N structural factors create the conditions within which homelessness will occur; and
N people with personal problems are more vulnerable to these adverse social and
economic trends than others; therefore
N the high concentration of people with personal problems in the homeless
population can be explained by their susceptibility to macro-structural forces,
rather than necessitating an individual explanation of homelessness.

This new orthodoxy provides a far more ‘‘practically adequate’’ explanation of

homelessness than the individual and structural accounts that preceded it, but is
clearly unsatisfying from a theoretical point of view. As has been noted by other
authors, the individual/structural divide in homelessness explanations is unhelpfully
 Explaining Homelessness 5

crude (see Neale 1997, Pleace 1998, 2000), and mirrors the now largely discredited
notion of a strict agency/structure dichotomy in sociological theory (Stones 2001). It
also fails to accommodate convincingly a whole range of factors that might plausibly
contribute to homelessness, particularly where, as is often the case, structural factors
are restricted to macro-level social and economic forces, and individual factors are
assumed to be limited to personal behaviours (on the part of the homeless person).
For example, experience of poor parenting is not a macro-structural issue, and yet it
is hardly ‘‘behavioural’’ in the sense of being within the control of the homeless
person him- or her-self. There are a great many factors which could be interpreted as
operating at either a structural or individual level – should the breakdown in a
homeless person’s marriage be considered an individual problem or the result of a
structural trend towards growing family fragmentation? Furthermore, how can the
Downloaded by [Heriot-Watt University] at 01:17 29 January 2015

new orthodoxy account for those cases of homelessness arising from acute personal
crises where structural factors can seem virtually absent (as Crane 1999 has argued is
often the case with older homeless people)? The most fundamental weakness of the
new orthodoxy, however, is that it lacks any clear conceptualization of causation.
What is it about these structural and individual ‘‘factors’’ that generate home-
lessness?
It will be argued in the final section of this paper that realist theory can overcome
the weaknesses in this (essentially atheoretical) ‘‘new orthodoxy’’ through its
explicit conceptualizations of causation, structure and stratification in the social
(and material) worlds. First, however, it is necessary to consider existing more
theoretically-informed attempts to explain homelessness in order to highlight the
conceptual weaknesses of these accounts and to make the case that critical realism
can help to address these weaknesses.

The Positivists: Seeking Regularities

Most explanations of homelessness in UK have implicitly employed an ‘‘empirical
regularity’’ notion of causation in keeping with the positivist nature of the majority
of housing research (Jacobs, Kemeny & Manzi 1999). This positivist conception of
causation was made explicit in the work of Randall and Brown (1999) who comment
as follows:

Housing shortages, poverty, unemployment, personal difficulties such as

mental health, drug or alcohol problems are sometimes said to be the causes of
rough sleeping. However, there are continuing problems of rough sleeping in
areas with no housing shortage. Equally, the great majority of people in
poverty or with mental health, or substance abuse problems, do not sleep
rough. … It follows that housing shortages, poverty, mental health and
substance misuse problems cannot be said to cause rough sleeping. (p. 5)

There is a clear assumption here that the factors associated with homelessness can
only possess causal force if there is a perfect match between their presence and
homelessness resulting (Allen 2000). This is a very strong version of the empirical
regularities notion of causation – demanding not only statistically valid associations
between proposed causal factors and homelessness, but 100% correlations
 6 S. Fitzpatrick

(like Hume’s ‘‘constant conjunctions’’). Pleace’s (1998) analysis is subtler, but

appears also to be underpinned by a positivist conception of causation:

Single homelessness and rough sleeping are never one thing or another,
sometimes the structural factors seem all-important, sometimes it is relation-
ships breaking down, loss of a job or a host of other factors that seem almost to
be unique to each individual who experiences homelessness… Instead of being
confronted by patterns, clear relationships and shared characteristics, there is
the impression of variation above all else, rather than a central tendency…
none provides a satisfactory explanation of all forms of homelessness or what
is known about single homelessness on a case by case basis. (p. 56)
Downloaded by [Heriot-Watt University] at 01:17 29 January 2015

I would argue that the research evidence in fact demonstrates a recurring pattern
of life events and circumstances implicated in ‘‘pathways’’ into homelessness
(Fitzpatrick 2000), although, as Pleace identifies, the precise combination of these
factors differs significantly from person to person (Anderson & Tulloch 2000,
Fitzpatrick et al. 2000). Pleace, like Randall and Brown, reconceptualizes these
recurring factors as contributing towards ‘‘increased risks’’ (rather than causes) of
homelessness. Such an ‘‘increased risk’’ approach undoubtedly has value at the
descriptive level in that it appears to be empirically well-grounded, and thus is
practically useful in enabling ‘‘micro-level’’ prevention measures to be targeted
appropriately. But it is unsatisfactory at an explanatory level: if we have dispensed
with the concept of causation, how can we understand why these factors lead to an
increased risk of homelessness?
Realism is capable of providing a resolution to this problem. It reinstates these
‘‘risk factors’’ as ‘‘real’’ causes of homelessness if they can be shown to have a
tendency to bring about homelessness, even if they only bring about actual
homelessness on some occasions. For realists, the varying circumstances of each
homeless person highlighted above is to be expected in an open social system where a
multitude of structures are contingently (and unpredictably) related, and where there
is scope for human agency within the range of options that these structures enable.
Thus these mixed patterns do not negate the idea of ‘‘causes’’ of homelessness, but
merely the positivistic notion of causation as predictable, empirical regularities.

The Interpretivists: Seeking Meaning

The main challenge to positivist analyses of homelessness has come from ‘‘social
constructionism’’ within the interpretivist tradition. Hutson and Liddiard (1994)
were among the first authors in the homelessness field to take this approach. They
explored the diverse ways in which youth homelessness is interpreted and presented
by a range of key actors – the media, homelessness agencies, politicians and young
homeless people themselves – and argued that:

No longer should we simply aim to observe and measure aspects of society,

such as crime or homelessness, as if they were objective facts. Instead we should
concentrate on understanding the meanings and interpretations that people
 Explaining Homelessness 7

apply to the social world and to social phenomena, such as youth

homelessness. (p. 24)

The main question that this sort of statement begs is whether social construc-
tionists believe that there is an underlying (social and/or material) reality which is
being mediated through these social and cultural processes (the ‘‘weak’’ construc-
tionist approach), or whether all reality is simply the product of ‘‘ways of seeing’’
(the ‘‘strong’’ constructionist approach) (Lupton 1999:35). In other words, is an
exploration of the ‘‘meanings’’ attached to homelessness by the range of social actors
an alternative or additional exercise to investigating its ‘‘real’’ causes? This reflects a
broader debate about whether realists and constructionists fundamentally disagree
about the nature of social reality, or simply focus on different aspects of reality and
Downloaded by [Heriot-Watt University] at 01:17 29 January 2015

thus ‘‘talk past’’ each other (see Manzi’s 2002 argument that there are no ‘‘strong
constructionists’’ active in the housing studies field and Somerville & Bengtsson’s
2002b response).
A later, and more developed, contribution from the constructionist school of
thought has been made by Jacobs et al. (1999). They focus on the contested nature of
the definition of homelessness, arguing that:

… the struggle by different vested interests to impose a particular definition of

homelessness on the policy agenda is critical to the way in which homelessness
is treated as a social problem. (p. 11)

Jacobs et al. conceptualize the key ideological battle as between those who treat
homelessness as a structural problem requiring broad welfare measures, and those
with a minimalist definition who view homelessness as resulting from individual
fecklessness (a conceptualization which mirrors the structure versus agency debate
discussed above). As compared with Hutson and Liddiard, they take a more
explicitly ‘‘critical’’ (yet orthodox) approach, arguing that those in positions of
power use their resources to establish a dominant discourse through which social
‘‘problems’’ are defined and dealt with. The authors imply at several points that they
(unlike strong constructionists) recognize an underlying reality beneath these
competing social constructions of homelessness, commenting, for example, that
‘‘… the structure of a society does provide the basis for the emergence of certain
kinds of ‘‘problems’’ (p. 13). However, they do not explicitly acknowledge the
limitations of the constructionist approach in providing explanations of phenomena
such as homelessness, but rather make a number of assertions relating to the
underlying causes of homelessness without offering substantiating evidence (e.g.
‘‘…the fewer resources that are committed to an area like housing, the more
homelessness there is likely to be’’ (p. 13)).
As noted earlier, realists are as critical as interpretivists of positivist approaches to
social science, and agree that the social construction of meaning is of central
importance in social research. Realists would entirely accept Jacobs et al’s
contention that the definition of a social problem like homelessness has ‘‘real’’
impacts on policy-making, which in turn is likely to contribute to the underlying and
proximate causes of homelessness (Kennett & Marsh 1999). However, where realist
accounts differ from those of strong constructionists at least is in insisting on the
 8 S. Fitzpatrick

social and material reality underlying these competing discursive accounts of social
phenomena. Therefore while they would accept that definitions of homelessness are
‘‘socially constructed’’ rather than ‘‘objectively given’’, they would maintain that the
housing and social conditions they refer to have a ‘‘real’’ existence whether or not
they are defined as constituting homelessness (whilst also accepting that such
definitions may affect the ‘‘lived experience’’ of these conditions through ‘‘negative
labelling’’ and ‘‘internalization’’ or ‘‘resistance’’ to these labels, see Jenkins 1996,
Wardhaugh 1999). They would further argue that, for any definition of homelessness
to have ‘‘causal weight’’, it must constitute a ‘‘realistic’’ category with internal
conceptual coherence – this point is discussed in detail in the last section of the
paper.
Downloaded by [Heriot-Watt University] at 01:17 29 January 2015

The Feminists: Finding the Roots in Patriarchy

Feminist academics have launched a sustained critique of mainstream analyses of
housing and homelessness in recent years (Watson & Austerberry 1986, Tomas &
Dittmar 1995, Watson 1999). Women, they argue, are often powerless to define their
own housing needs or to house themselves independently from a man because of
their weak economic position and the patriarchal assumptions embedded in housing
policy and practice. These factors, together with women’s vulnerability to domestic
abuse and violence (Buck, B. Exploring the relationship between repeat homelessness
and domestic abuse. An interim report to the Homelessness Task Force.
Unpublished) (Pawson, Third & Tate 2001), have led numerous mainstream as well
as feminist authors to identify being female as a particular risk factor predisposing a
person to homelessness (Greve 1991, Johnson et al. 1991, Robson & Poustie 1996).
However, these attempts to establish the special vulnerability of women to
homelessness do not stand up to empirical scrutiny. Neale (1997) points out, for
example, that in the UK women (as mothers) have frequently been given a special
claim to local authority housing, and access to Housing Benefit has to a significant
degree broken the link between earning power and access to accommodation
(although the quality of the available accommodation will often be very poor). Even
more inconvenient to feminist analyses of homelessness is the repeated evidence from
across the UK (and overseas) that single homeless men far outnumber single
homeless women (Kemp 1997, Fitzpatrick et al. 2000). Many commentators have
dealt with these awkward statistics by suggesting that female homelessness tends to
take ‘‘hidden’’ forms, with women more likely than men to stay with friends and
relatives rather than approach homelessness agencies for help (Watson &
Austerberry 1986, Greve 1991, Webb 1994, Edgar & Doherty 2001). However, the
(very limited) UK evidence on gender and homelessness (as opposed to women and
homelessness) indicates that women are more likely than men to approach local
authorities and housing associations when they find themselves homeless, and to be
treated more sympathetically by these agencies than their male counterparts,
suggesting that female rather than male homelessness is most likely to be ‘‘visible’’ in
the official statistics (Fitzpatrick 2000, Cramer & Carter 2001).
As it happens, the gender imbalance in the single homeless population is probably
more than compensated for by the very high vulnerability of lone parent families
to homelessness, almost all of whom are headed by women (Smith et al. 1996,
 Explaining Homelessness 9

Fitzpatrick 2000, Pleace & Fitzpatrick 2004). However, even if this were not the case,
realism could provide a means of defending the feminist association of homelessness
with patriarchal social structures. This is because realist conceptions of causation are
not dependent on empirical regularities. So male oppression of women could still be
one of a number of social structures with a ‘‘tendency’’ to cause homelessness, even
if men predominated in the homeless population. The production of actual
homelessness would be dependent on the interaction of patriarchy with other
(contingently related) social structures (Walby 1990). An important point, seldom
acknowledged in the homelessness literature, is that some gendered factors
associated with homelessness almost certainly disproportionately affect men. For
example, qualitative evidence indicates that lack of social support is a key factor
which can precipitate homelessness and we know that women tend to build up
Downloaded by [Heriot-Watt University] at 01:17 29 January 2015

stronger kinship and other social relationships than men (Finch 1989). It may also be
that domestic training from an early age enables most women to manage
independent living and domestic crises more effectively than men (Fitzpatrick
2000, Cramer & Carter 2001). Thus gendered (including patriarchal) social
relationships can disadvantage men as well as women in relation to homelessness,
and are in any case only one set of social structures with the power to cause
homelessness.

The Contribution of ‘‘Structuration’’ Theory: Compatible with Realism?

Postmodernists and poststructuralists have mounted a profound attack on
traditional social science in recent years (Hollis 1994). These ‘‘post-critiques’’
challenge assumptions about causality because they do not accept that there is a
single oppressive force impinging on people’s lives (such as capitalism or patriarchy),
nor that we can design rational solutions to social problems based on sure
foundations of knowledge (see Watson 1999 for discussion of the poststructuralist
contribution to understanding homelessness). Neale (1997) argues, however, that the
inherent tendency of postmodernists and poststructuralists to subjectivity, relativity
and irrationality limits their potential for practical application through social policy
measures, and the focus on individual agency can lead to the power of (macro) social
structures being neglected. She therefore advocates Giddens’ (1984) theory of
structuration as an alternative, and more promising, way out of the structure-versus-
agency debate than the ‘‘post-critiques’’. Giddens’ theory rests on the idea of
‘‘duality’’ of structure whereby social structures are constituted by human agency,
but at the same time enable and constrain that social action. Human agents make
conscious, reflective choices within the range of options allowed them by existing
power structures, but their actions also have unintended consequences; one of which
is to reproduce or, on occasion, to transform society. Using this perspective, Neale
concludes that:

…there are forces in operation which make it likely that some people, and not
others, will become homeless in any given set of circumstances. Nevertheless,
because personal circumstances are not predetermined and because power
structures operate at different levels, there will be various ways of effecting
changes to human lives. Homeless people cannot, therefore, be defined as
 10 S. Fitzpatrick

either deserving or undeserving, entirely responsible for their problems, or

victims of circumstances beyond their control. (p. 47)

This analysis is entirely consistent with the realist approach being proposed in this
paper, and in fact Giddens’ structuration theory has been proposed as a key example
of realist social theory (Williams & May 1996). The principal concern of Stones
(2001) was to demonstrate the compatibility of structuration theory with realism,
against Archer’s (1995) argument that the Giddens’ duality of structure is at odds
with the ‘‘analytical dualism’’ which lies at the heart of the realist approach.
‘‘Analytical dualism’’ postulates that structural conditions can be separated from
action through an emphasis on temporality in which ‘‘…structure precedes action
which, in turn, leads to a more or less attenuated structural outcome or elaboration
Downloaded by [Heriot-Watt University] at 01:17 29 January 2015

which, in turn, provides the preconditions for action, and so on.’’ (Stones 2001:180).
Thus, social structures have a ‘‘real’’ existence which provides ‘‘…‘‘slots’’ that pre-
exist the particular human agents that subsequently inhabit, reproduce or transform
these position-practices’’ (Stones 2001:183, see also Bengtsson’s & Somerville’s
2002b:150 reference to ‘‘frozen past action’’). Stones contends that, while Giddens
does not elaborate the sequencing that Archer draws our attention to, his account of
constraints upon agency make clear that social structures both pre-exist agency and
can have a causal influence on agents’ practices. Stones goes on to argue that, not
only are structuration and realism compatible theoretical frameworks, but they have
complementary strengths and weaknesses. For example, realism provides a more
convincing account of ‘‘meso’’ and ‘‘macro’’ structural levels than structuration
theory, but can be enhanced by the more ‘‘hermeneutically informed’’ concept of
duality employed by Giddens. Thus Neale’s advocacy of structuration theory as an
appropriate framework for understanding the causes of homelessness can be located,
I would argue, in the broader context of realist social theory.

Implications of a Critical Realist Approach for the Study of Homelessness

This paper has critiqued previous theoretical understandings of the causes of
homelessness from a realist perspective, and has attempted to demonstrate the
potential value of the realist framework in overcoming the weaknesses identified in
these earlier approaches. The final section of the paper reviews the methodological
and theoretical implications of taking a realist perspective in the study of
homelessness.

Methodological Implications: What Type of Research Should be Undertaken?

Despite critical realism’s theoretical elegance, it is not a philosophical programme
that translates easily into empirical research (Williams & May 1996). Realist social
science is a demanding undertaking because it requires not only a description of
social relations and events, but also a commitment to explaining them by means of
uncovering ‘‘hidden’’ dimensions of social reality. Theory-building is needed to
postulate ‘‘abstracted’’ aspects of reality in order to define necessary (‘‘internal’’)
tendencies which may constitute causal mechanisms. Attempts must then be made to
demonstrate the existence and mode of operation of these hypothetical mechanisms
 Explaining Homelessness 11

through empirical test and the elimination of alternative plausible explanations.

Bhaskar (1989) calls this process ‘‘retroduction’’. The overall aim is to achieve a
correspondence between the ‘‘transitive’’ objects of science (the hypothetical
mechanisms) and the ‘‘intransitive’’ objects of reality (the ‘‘real’’ mechanisms).
It might reasonably be argued that critical realist research, focused as it is on
hypothesis development and empirical testing, is little different from more
traditional methodologies. However, it is at the level of theorization and
interpretation that realism differs profoundly from more traditionally ‘‘empiricist’’
approaches (Burrows 1989). Pawson and Tilley (1997), for example, in calling for
‘‘realistic evaluations’’ in urban policy, advocate familiar (qualitative) research
methods, but through proposing a focus on ‘‘what works, for whom, in what
circumstances’’ they offer a rigorously argued alternative to experimental designs
Downloaded by [Heriot-Watt University] at 01:17 29 January 2015

intended to quantify the ‘‘effects’’ of a programme in controlled circumstances.

Perhaps the most important distinguishing feature of critical realist approaches
to social science is the close attention paid to conceptual validity and clarity,
particularly with regards to the ‘‘qualitative nature’’ of the social object under
scrutiny. In this spirit Williams (2001, 2003) has mounted an important challenge to
the assumption that homelessness is a ‘‘realistic category’’ amenable to causal
analysis; his arguments are given detailed consideration in the next section.

Ontological Implications: Does ‘‘Homelessness’’ Exist?

Williams (2001) has postulated that:

…there is no such thing as homelessness, but instead a range of heterogeneous

characteristics that give rise to a wide range of symptoms that we term
‘‘homelessness’’. (p. 1)

The crux of his argument is, to use Sayer’s terminology, that homelessness is a
‘‘bad abstraction’’ (or ‘‘chaotic conception’’) meaning that it ‘‘…arbitrarily divides
the indivisible and/or lumps together the unrelated and the inessential’’ (1992:138).
However, it is unclear from Williams’ papers whether he views homelessness as a
chaotic conception because it arises from distinctive causal processes (he uses the
term ‘‘antecedent conditions’’) or because, as an emergent condition, it lacks
conceptual ‘‘unity and autonomous force’’ (Sayer 1992:138)1.
To take the point of antecedent conditions to begin with, I would argue that
consistency in these is not central to the ‘‘realistic’’ meaningfulness of a social object
or property. In other words, a realist analysis of causation can be ‘‘complicated’’
(comprising a range of separate causal processes) as well as ‘‘complex’’ (a result of
interacting feedback loops between inter-related causal components). The challenge
for a realist analysis of homelessness may then be to identify the range of separate
(and complex) causal routes into this experience (as I attempted to do in examining
the diversity of young people’s ‘‘pathways’’ into and through homelessness, though
without using an explicit realist conceptual framework – see Fitzpatrick 2000).
The conceptual coherence of the emergent condition of homelessness is,
in contrast, crucial to its categorization as a ‘‘real’’ property appropriate for causal
investigation. Williams seems to question the conceptual coherence of homelessness
 12 S. Fitzpatrick

primarily on intersubjective grounds: he argues that it is not consistently defined

across time and place, and in particular it is not a ‘‘meaningful’’ category to those
subject to it. However, intersubjective identification is not central to realist
conceptions of social objects and properties – the agents ‘‘in focus’’ do not have to
recognize an object or mechanism affecting them for its presence to be ‘‘real’’
(Stones 2001, see also Williams 2003 own comments quoted above). Williams’
analogy of ‘‘Greekness’’ is perhaps not apposite here: as a category concerned
primarily with cultural identity, it would be difficult to argue for its existence in the
face of a lack of recognition of those subject to it. Homelessness, on the other hand,
is not a cultural phenomenon but rather a signifier of objective material and social
conditions, and as such intersubjective recognition is, I would argue, less central to
its existence2.
Downloaded by [Heriot-Watt University] at 01:17 29 January 2015

Williams further objects to homelessness as a realistic category because it does not

form a ‘‘self-organizing’’ system (with the large homeless ‘‘community’’ in the
‘‘Bullring’’ in London given as an example of where such an interacting, self-
organizing system can emerge at a more differentiated level than the ‘‘universal’’
category of homelessness). However, there need be no actual interaction between
members of a category of people for that category to be meaningful from a realist
perspective; what is required is conceptual rather than empirical relations of
connection. Purely formal relations – of similarity and dissimilarity – do not suffice
(Sayer 1992), but overlapping, shared (not necessarily identical) experiences which
give rise to similar impacts on individuals (that is, similar emergent attributes and
causal tendencies) would be good evidence of the existence of a ‘‘real’’ category of
homelessness (here Williams’ comments on ‘‘Greekness’’ are apposite).
Is there such conceptual coherence in the ‘‘universal’’ category we currently
define as homelessness? Perhaps not. It may well be that a focus on more internally
homogeneous subgroups within the ‘‘homeless’’ population would yield better
explanations of the social problems associated with ‘‘suboptimal’’ housing
circumstances. So if, for example, single people sleeping rough and living in
hostels shared many similar experiences (especially if there was extensive movement
between these ‘‘types’’ of homelessness), and exhibited similar health and other
impacts of their experience of homelessness, they may constitute a ‘‘real’’ category
which could, for instance, be entirely separate from that of families living in bed
and breakfast hotels3. Systematic investigation of this point would require, as
Williams has argued, empirically-informed theorizing about meaningful categor-
izations within the homeless experience, and a re-analysis of existing empirical
material with this critical realist framework in mind; new data collection might also
be required. That said, and as I have argued above, I think there is already
sufficient evidence of similar life histories and recurring characteristics amongst at
least some groups of homeless people to conclude that such a realist conceptual
interrogation would be unlikely to find that the reference category had to be
reduced to the single case to be meaningful (but see Williams 2001)4. So it is with
this question mark over the appropriate scope of the ‘‘real’’ conceptualization of
homelessness, that the last section of the paper sketches out what a preliminary
realist analysis of the causes of homelessness (appropriately defined) could look
like.
 Explaining Homelessness 13

A Realist Analysis of the Causation of Homelessness

A central ontological assumption of realists is that the world is structured,
differentiated and stratified, and that it is the business of science to uncover these
structures. Allen (2000), for example, in his critical realist analysis of housing and
illness, has proposed an ontologically stratified conception of causation at the
physiological (body), psychological (mind) and sociological (housing/home) levels,
with each strata interacting with the others and none treated as logically prior. Thus,
according to Allen, the causes of ill-health cannot simply be reduced to mechanisms
on the physiological strata as what happens there is co-determined by mechanisms
at the social and psychological levels – with the contingent (‘‘external’’) relations
between these strata producing actual health effects.
Downloaded by [Heriot-Watt University] at 01:17 29 January 2015

Taking a similar realist approach to theorizing the causes of homelessness, one

could hypothesize that causal mechanisms may exist on four levels:

N economic structures – social class interacts with other stratification processes and
welfare policies to generate poverty and to determine poor individuals
‘‘and households’’ (non-) access to material resources such as housing, income,
employment and household goods.
N housing structures – inadequate housing supply and a deterioration in
affordability can squeeze out those on lower incomes; tenure and allocation
policies, coupled with the collective impacts of private choices, can lead to
residential segregation and spatial concentration of the least advantaged groups.
N patriarchal and interpersonal structures – can lead to the emergence of domestic
violence, child neglect or abuse, weak social support, relationship breakdown, etc.
N individual attributes – personal resilience can be undermined by mental health
problems, substance misuse, lack of self-esteem and/or confidence.

Within the chaos/complexity perspective that is increasingly influential in realist

social theory (Byrne 1998), each of these strata of potential causal mechanisms
would be as considered part of a ‘‘multiple set of nested systems’’, with no hierarchy
of influence assumed between them (Byrne 1999). From this complex system of
nested probabilities, ‘‘…the clustering of a particular set of antecedents in any
individual [increases] the probability of particular outcomes.’’ (Williams 2001:3).
Complicating matters of prediction, however, are ‘‘non-linear’’ dynamics that mean
that a small change in one of these complex relationships can lead to dramatically
different consequences (positive or negative) (Byrne 1999). The ‘‘triggers’’, or
‘‘immediate causes’’, often associated with homelessness, for example, may be
understood as an instance of such non-linear dynamics (Fitzpatrick et al. 2000).
Taking the role of ‘‘poverty’’ as my starting point, I now attempt to offer a
preliminary sketch of a realist analysis of homelessness.
Research has repeatedly indicated that most but not all homeless people come
from circumstances of poverty; at the same time, it is clear that most poor people do
not experience homelessness (Fitzpatrick 2000). For a critical realist, the fact that
homelessness arises amongst non-poor people indicates only that poverty is not a
‘‘necessary condition’’ of homelessness; it does not remove the possibility of its being
 14 S. Fitzpatrick

one of a range of causal factors. Similarly, the lack of universality of the homeless
experience amongst poor people is not the central concern of critical realist
approaches. The key question for a realist is not what proportion of poor people are
homeless, but rather what is it about poverty that could cause homelessness.
The obvious response is that homelessness can arise from poor people’s inability
to compete in a ‘‘tight’’ housing market where pricing is the key rationing
mechanism: and indeed affordability problems are strongly correlated with use of
temporary accommodation for homeless households amongst local authorities in
England (Pleace & Fitzpatrick 2004). However, homelessness also exists in areas of
the UK where access to affordable housing (subsidized by Housing Benefit) is not
problematic, and amongst groups for whom priority is given in access to social
housing (where bureaucratic rather than pricing mechanisms are therefore the key
Downloaded by [Heriot-Watt University] at 01:17 29 January 2015

rationing device). Yet poverty also seems implicated in the causation of home-
lessness in these places and amongst these groups (Fitzpatrick et al. 2000). For a
critical realist, this is explicable because the connection between poverty and
homelessness is likely to be more complex than simply generating an inability to
‘‘purchase’’ housing. Rather, feedback loops between poverty and various of the
other potential causal mechanisms outlined above – operating through an array of
necessary (‘‘internal’’) and contingent (‘‘external’’) relationships – can be interpreted
as increasing the ‘‘weight of the weighted possibility’’ of homelessness amongst
certain poor people.
To take one such relationship, there is an ‘‘internal’’ (but asymmetric) relationship
between poverty and spatial concentrations of disadvantage – in other words,
one (spatial concentrations) could not exist without the other (poverty),
but the reverse is not true (the emergence of spatial concentration is contingent
upon interaction of poverty with the housing system). There has been much debate
about the emergence of ‘‘area effects’’ – including stigma, restricted social
networks/horizons, conflict, insecurity and criminal activity – in places experiencing
such concentrated disadvantage (Andersen 2002). There is also now some evidence
that, for poor people living in these areas, local social conflict can escalate into
external violence and threats, which may result in the abandonment of their home
and subsequent homelessness. In a recent study in Glasgow, for example,
Fitzpatrick, Jones and Pleace (in press), found strong (qualitative) evidence of
links between neighbourhood-based violence and family homelessness in deprived
areas of the city.
Domestic violence, in contrast, is ‘‘externally’’ related to poverty in that one can
exist without the other5, and either could be hypothesized to result in homelessness
independently of the other. However, poverty may also (contingently) impact on
domestic violence (making it more likely), and reverse causation is also possible
(with domestic violence making poverty more likely). Where they are found
in combination, poverty and domestic violence (regardless of their own causal
interrelationship) may increase the probability of homelessness, with particular
violent incidents, for example, providing the ‘‘trigger’’ for homelessness in a non-
linear dynamic fashion. Clearly, if an individual is also resident in an area of
concentrated disadvantage, or experiences one of the other causal factors identified
above, such as mental health problems or substance misuse, then the ‘‘weight of the
weighted possibility’’ of homelessness starts to increase substantially.
 Explaining Homelessness 15

Conclusion
This paper has attempted to engage with developing critical realist frameworks in
housing and urban policy to propose a more rigorous framework for analysing the
causes of homelessness in the UK and elsewhere. It has argued that the prevailing
‘‘new orthodoxy’’ in explanations of homelessness is useful at a descriptive level, but
unsatisfactory at a more profound conceptual level. Previous attempts to provide
more theoretically-informed explanations of homelessness – including positivist,
interpretivist, feminist and postmodernist/poststructuralist approaches – were also
examined and found wanting in various respects. Giddens’ theory of structuration
was argued to be helpful in overcoming the structure versus agency dichotomy, but
would, it was contended, be enhanced if integrated within the broader realist
Downloaded by [Heriot-Watt University] at 01:17 29 January 2015

conception of a layered social reality.

The paper also attempted to illustrate how a (complex) critical realist approach
could enable account to be taken of the full range of potential causal factors in
homelessness – and their necessary and contingent inter-relationships – while
avoiding making any one level ‘‘logically prior’’ to all others. In contrast to the ‘‘new
orthodoxy’’, therefore, it is unnecessary for those adopting a critical realist stance to
‘‘smuggle’’ individual factors into the analysis as merely making individuals
susceptible to the more fundamental structural causes – these personal factors can
be causes of homelessness in their own right without undermining the importance of
structural conditions in other cases. This realist framework also allows us to take full
account of ‘‘protective factors’’, such as strong family support or personal resilience,
which can operate to prevent homelessness: just because such benign counteracting
tendencies stop most people who are poor or experience domestic violence becoming
homeless does not, for a critical realist, stop poverty or domestic abuse being causes
of homelessness. Finally, and following Williams’ (2001, 2003), it was argued that
empirically-informed theorizing is now required to assess whether the identification
of (internally homogeneous) subgroups in the homeless population would assist with
the development of more ‘‘realistic’’ explanations of this persistent social problem
than the focus on the universal category of ‘‘homelessness’’ has hitherto allowed.

Acknowledgements
The author thanks Professor Chris Allen, Keith Kintrea, Professor Mark Stephens
and Professor Roger Burrows for their helpful comments on an earlier draft of this
paper. The author would also like to thank Bo Bengtsson and an anonymous referee
for their helpful suggestions. The usual disclaimers apply.

Notes
1. In fact, sometimes the two points seem to be conflated: in other words, it is implied that different
types of homelessness (diverse ‘‘suboptimal housing situations’’) arise from distinctive life
circumstances (see the vignettes in Williams 2001). However, no such direct mapping from prior
experience to type of homelessness is possible. For example, in a study of people begging, rough
sleeping and selling The Big Issue street paper in Glasgow and Edinburgh we encountered some
individuals who had became long-term rough sleepers via very different routes (some through the
impacts of poverty and multiple deprivation on their life chances from an early age, others through
the inability to cope with personal catastrophes later in life, such as bereavement) (Fitzpatrick &
 16 S. Fitzpatrick

Kennedy 2000). Conversely, others who had very similar life histories – involving trauma, drug use
and poverty – found themselves in distinctive homeless (and non-homeless) situations.
2. A related point made by Williams, that homelessness is often viewed as a ‘‘symptom’’ of other
‘‘principal’’ problems by many of those experiencing it, is something with which I would concur
(see for example, Fitzpatrick & Kennedy 2000). However, this (sometime) secondary status of
homelessness does not prevent it being a ‘‘real’’ emergent outcome appropriate for causal
analysis – but it does mean, from a social policy perspective, that it may not always be the most
important problem to focus on.
3. The antecedents to both categories of ‘‘homeless’’ may also be distinct, but that is not necessary to
the argument being pursued here.
4. At the same time, I certainly agree with Williams that the tendency to throw the ‘‘definitional net’’ of
homelessness ever wider in covering ever more sub-optimal housing conditions is unhelpful; in fact, I
would argue that it undermines not just the explanatory but also the descriptive value of the term.
5. This is the view that Sayer (1992) takes and I sympathize with, but see Walby (1990) for a
Downloaded by [Heriot-Watt University] at 01:17 29 January 2015

comprehensive discussion of the diverse theoretical perspectives on the interaction of class and
patriarchial social structures.

References
Allen, C. (2000) On the ‘‘physiological dope’’ problematic in housing and illness research: towards a
critical realism of home and health Housing, theory & society, 17, pp. 49–67.
Andersen, H. S. (2002) Excluded places: the interaction between segregation, urban decay and deprived
neighbourhoods Housing, theory & society, 19, pp. 153–169.
Anderson, I. & Tulloch, D. (2000) Pathways through homelessness: a review of the research evidence
(Edinburgh: Scottish Homes).
Archer, M. (1995) Realist social theory: the morphogenetic approach (Cambridge: CUP).
Bhaskar, R. (1989) Reclaiming reality (London: Verso).
Burrows, R. (1989) Some notes towards a realistic realism: the practical implications of realist philosophies
of science for social research methods International journal of sociology and social policy, 9(4),
pp. 47–64.
Byrne, D. S. (1998) Complexity theory and the social sciences (London: Routledge).
Byrne, D. (1999) Social exclusion (Buckingham: Open University Press).
Cramer, H. & Carter, M. (2001) Homelessness: what’s gender got to do with it? (London: Shelter).
Crane, M. (1999) Understanding older homeless people (Buckingham: Open University Press).
Dant, T. & Deacon, A. (1989) Hostels to homes? The rehousing of homeless single people (Aldershot: Gower
Publishing).
Drake, M., O’Brien, M. & Beiuyck, T. (1981) Single and homeless (London: HMSO).
Edgar, B. & Doherty, J. (2001) Women & homelessness in Europe (Bristol: The Polity Press).
Finch, J. (1989) Family obligations and social change (Cambridge: Polity Press and Basil Blackwell).
Fitzpatrick, S. (1998) Homelessness in the European Union, in: M.Kleinman, W.Matznetter &
M.Stephens (Eds), European integration and housing policy (London: Routledge).
Fitzpatrick, S. (2000) Young homeless people (Basingstoke: Macmillan).
Fitzpatrick, S. & Christian, J. (in press) Comparing research on homelessness in the United Kingdom and
United States: what lessons can be learned? Journal of social issues.
Fitzpatrick, S., Jones, A. & Pleace, N. (in press) The support needs of homeless families (Edinburgh: NHS
Health Scotland).
Fitzpatrick, S., Kemp, P. A. & Klinker, S. (2000) Single homelessness: an overview of research in Britain
(Bristol: The Policy Press).
Fitzpatrick, S. & Kennedy, C. (2000) Begging, rough sleeping & The Big Issue in Glasgow and Edinburgh
(Bristol: The Policy Press).
Giddens, A. (1984) The constitution of society: outline of the theory of structuration (Cambridge: Polity
Press).
Greve, J. (1991) Homelessness in Britain (York: Joseph Rowntree Foundation).
Hollis, M. (1994) The philosophy of social science: an introduction (Cambridge: Cambridge University
Press).
 Explaining Homelessness 17

Hutson, S. & Liddiard, M. (1994) Youth homelessness: the construction of a social issue (Basingstoke:
Macmillan).
Jacobs, K., Kemeny, J. & Manzi, T. (1999) The struggle to define homelessness: a constructivist approach,
in: S.Hutson & D. (Eds), Clapham homelessness: public policies and private troubles (London: Cassell).
Jenkins, R. (1996) Identity (London: Routledge).
Johnson, B., Murie, A., Naumann, L. & Yanetta, A. (1991) A typology of homelessness (Edinburgh:
Scottish Homes).
Kemp, P. (1997) The characteristics of single homeless people, in: R.Burrows, N.Pleace & D.Quilgars
(Eds), Homelessness and social policy (London: Routledge).
Kennett, P. & Marsh, A. (Eds) (1999) Homelessness: exploring the new terrain (Bristol: The Policy Press).
Lupton, D. (1999) Risk (London: Routledge).
Metraux, S. & Culhane, D. P. (1999) Family dynamics, housing, and recurring homelessness among
women in New York City homeless shelters Journal of family issues, 20, pp. 371–396.
Manzi, T. (2002) Constructionism, realism and housing theory Housing, theory & society, 19, pp. 144–145.
Downloaded by [Heriot-Watt University] at 01:17 29 January 2015

Neale, J. (1997) Theorising homelessness: contemporary sociological and feminist perspectives,

in: R.Burrows, N.Pleace & D.Quilgars (Eds), Homelessness and social policy (London: Routledge).
Pawson, R. & Tilley, N. (1997) Realistic evaluation (London: Sage).
Pawson, H., Third, H. & Tate, J. (2001) Repeat homelessness in Scotland (Edinburgh: Scottish Homes).
Pleace, N. (1998) Single homelessness as social exclusion: the unique and the extreme Social policy and
administration, 32(1), pp. 46–59.
Pleace, N. (2000) The new consensus, the old consensus and the provision of services for people sleeping
rough Housing studies, 15(4), pp. 581–594.
Pleace, N. & Fitzpatrick, S. (2004) Centrepoint Youth Homelessness Index: an estimate of youth
homelessness for England (London: Centrepoint).
Randall, G. & Brown, S. (1999) Prevention is better than cure (London: Crisis).
Robson, P. & Poustie, M. (1996) Homelessness and the law in Britain (3rd edn) (London: Butterworths/
Planning Exchange).
Sayer, A. (1992) Method in social science: a realist approach (London: Routledge).
Sayer, A. (2000) Realism and social science (London: Sage).
Smith, J., Gilford, S., Kirby, P., O’Reilly, A. & Ing, P. (1996) Bright lights and homelessness: family and
single homelessness among young people in our cities (London: YMCA).
Somerville, P. & Bengtsson, B. (2002a) Constructionism, realism and housing theory Housing, theory &
society, 19, pp. 121–136.
Somerville, P. & Bengtsson, B. (2002b) Understanding contextualised rational action – authors response
Housing, theory & society, 19, pp. 148–152.
Stones, R. (2001) Refusing the realism-structuration divide European journal of social theory, 4(2),
pp. 177–197.
Tomas, A. & Dittmar, H. (1995) The experience of homeless women: an exploration of housing histories
and the meaning of home Housing studies, 10(4), pp. 471–491.
Walby, S. (1990) Theorizing patriarchy (Oxford: Blackwell).
Wardhaugh, J. (1999) The unaccommodated woman: home, homelessness and identity The sociological
review, 47(1), pp. 91–109.
Watson, S. (1999) A home is where the heart is: engendering notions of homelessness, in: P.Kennett &
A.Marsh (Eds), Homelessness: exploring the new terrain (Bristol: The Policy Press).
Watson, S. & Austerberry, H. (1986) Housing and homelessness: a feminist perspective (London: Routledge
and Kegan Paul).
Webb, S. (1994) My address is not my home: hidden homelessness and single women in Scotland (Edinburgh:
SCSH and GCSH).
Williams, M. & May, T. (1996) Introduction to the philosophy of social research (London: UCL Press).
Williams, M. (2001) Complexity, probability and causation: implications for homelessness research
Journal of social issues. Available at: http://www.whb.co.uk/socialissues/mv.htm.
Williams, M. (2003) The problem of representation: realism and operationalism in survey research
Sociological research online, 8(1). Available at http://www.socresonline.org.uk/8/1/williams.html.

View publication stats