Living Alone: Its Place in Household Formation and Change

by Joan Chandler, Malcolm Williams, Moira Maconachie, Tracey Collett and Brian Dodgeon

Sociological Research Online, Volume 9, Issue 3, <http://www.socresonline.org.uk/9/3/chandler.html>.

Received: 8 Jan 2004     Accepted: 26 Aug 2004    Published: 31 Aug 2004

Abstract

In recent decades there has been a significant rise in the numbers of people who live alone and it was predicted that by 2002 that a third of all households will be single-person households. The predicted increase has occurred with indications of continued growth in this type of living arrangement. Furthermore, although living alone remains common among older age groups, the largest growth has been within younger populations.

This demographic trend has attracted speculation about the numbers of people who will experience solo living, the stability of living alone in people's biography, and the impact of gender differences in the likelihood and stability of living alone. To answers these questions, this paper uses longitudinally linked Census data from England and Wales to explore the household origins and household destinations of working age people who live alone. This longitudinal data derives from the 1971, 1981 and 1991 Censuses.

The data from this analysis confirms other research demonstrating the increasingly numbers of non-retired people who live alone. Furthermore it demonstrates that once a person lives alone, they are more likely to continue to live in that household arrangement than any other and that the tendency to live alone and to continue to live alone is more likely amongst younger cohorts of people. It also demonstrates that the largest increase in living alone in amongst men, but that once women live alone they are more likely to continue to live alone.

These findings have an important bearing on current debates about 'individualisation', the contemporary experience of family life, life course trajectories and the emergent life styles of younger populations.

Keywords: Living Alone, Single Person, Household, Longitudinal Study, Cohort Study

Introduction

1.1 This paper uses longitudinally linked Census data from England and Wales to explore the household origins and destinations of working age people who live alone. The exploration provides an insight into the likelihood and stability of solo living within the early adult and middle years of the life course of men and women.

1.2 The past 30 years has seen a significant growth in the number of lone occupant households in the UK. One indication of the trend is that while the population of Great Britain has grown by 5% over the past three decades, the number of households with one occupant has grown by 31% (Social Trends, 2003:42). Turning from households to individuals as the basis of trend data, in the 2001 Census 15.6% ie. 16% of all adults under pensionable age were living alone. The trend towards living alone is also seen to be not just a feature of the UK but to characterise Europe more generally (Kaufmann, 1994).

1.3 While there is some knowledge of the characteristics of people who live alone, there has been little work to investigate what kinds of household they lived in prior to living alone and how living alone is located in the life course. Furthermore how might the position and longevity of living alone in the life course be changing? In other words, once a person lives alone do they continue to live alone, or do they eventually live in other household arrangements?

1.4 In this paper we begin by briefly reviewing the key characteristics of lone occupant households and the context of the growth of such living arrangements. We then describe the data set used - the Office of National Statistics (ONS) Longitudinal Study (LS), and its strengths and weaknesses in this context. The main part of the paper is a description of the household careers between 1971 and 1991 of a cohort of residents of England and Wales, aged 15 to 44 in 1971. The cohorts described all have, at some point, been enumerated in a lone occupant household. Finally we conclude with some modest speculations about the place of living alone in the life course and how this may be changing. Throughout, our emphasis is on living arrangements not on relationships, in the sense of intimate attachments (being partnered, in a marriage or other relationship). The focus of this study is on those who identified themselves as living alone in the Census in 1971, 1981 and 1991.

The Social Characteristics of Solo Livers in England and Wales

2.1 Who are the growing populations of solo-livers? Although changes in the elder population have contributed most to the demographic change, solo living is a growing trend in all age groups within the adult population. The largest increase, and also an accelerating increase, is in the proportion of those living alone who are below pensionable age. This has grown from 6% and 8% in 1971 and 1981 to 11% in 1991 and 15% in 2002 (Social Trends, 2003:42).

2.2 Solo living is an emergent feature of young adult lives, with 11% of men and 6.5% of women aged 25-29 living alone at the time of the 1991 Census. Furthermore, the pattern of solo living differs between men and women. One important reason for this is that women who are exiting partnerships are more likely to remain resident with children. Thus it is probably unsurprising that the biggest increase in the proportion of one-person households is found amongst men under 65, with the proportion having risen from 3% in 1971 to 10% in 2000. This trend is thought to be likely to continue, with the proportion of households with men living alone predicted to rise to 14% by 2021 (Social Trends, 2003:43).

2.3 Single men who live alone have different characteristics to women living alone - they are more likely to be younger and to have never married. Also, for men, living alone it is common at either end of the social class scale whereas for women it is closely associated with professional status. The fact that solo living is common for women across the life course is indicative of the greater financial independence of women today compared with earlier generations.

2.4 Solo living is also a feature of urban life, with over 40% of households in part of Central London, 39.1% of households in Manchester and 33.2% of households in Birmingham, containing people who live on their own (Census 2001). Furthermore, younger lone occupant households are found predominantly in urban areas whereas older lone occupant households have a more varied distribution, although rural and traditional retirement areas have significant concentrations. (Hall et al 1997).

The social significance of living alone

2.5 There are two important foci of analysis and debate that inform our understanding of the growth of living alone among younger populations: a discussion of the 'second demographic transition' (van de Kaa 2002, Odgen and Hall 2004) which highlights the structural shifts in populations and their living arrangements; and a discussion of contemporary family values and approaches to intimacy (Giddens 1992; Beck-Gernscheim, 2002; Bauman 2003). Here living alone may be seen as one of the structural consequences of the new sentimental order (Bawin-Legros, 2001; Bawin-Legros, 2004) and a context for individuals establishing intimate relations and develop reciprocities. However, before we discuss our study, we need to take a closer look at our theoretical start points.

Family diversity

2.6 For the past 20 years, much of the debate in the sociology of the family has attempted to capture and chart trends in family diversity. This has ranged from Gittins (1985) querying a notion of 'the family' to Morgan's (1996) more open analysis of 'family practices' and discussions of 'intimacy' by Giddens (1992) and Jamieson (1998, 1999). There have also been several empirical analyses of particular forms of 'new families' (Silva and Smart, 1999) including lone-parenting, step-families, cohabitation, and gay couples/parenting. However, as the dominant interest in the literature is in new forms of families, solo living is less frequently discussed, despite its growing incidence in people's lives.

Families and households

2.7 The analysis of lone occupant households provides insight into another long-standing area of interest for sociologists of family life, that is the relationship between families and households. Allan and Crow (2001) note the way in which contemporary analyses of family life demand greater precision in the use of the terms 'family' and 'household', but also how difficult this precision is, as their usage mesh together in everyday life and frequently within sociology. Relationships and living arrangements are often ambiguously reported and analysts frequently have to read between the lines. One source of ambiguity relates to the rise of cohabitation and joint parenting in the absence of marriage. Another dislocation between relationships and living arrangements is associated with the rise of solo living.

2.8 Within this schema solo living among younger age groups is variously seen as 'non-family living' (Goldscheider and Waite, 1991) or even to connote the rise in 'non-familial families' (Allan and Jones, 2003) or 'post-familial' families ( Beck-Gernsheim, 2002). However, such characterisation misses part of the point. Living alone does not necessarily mean lack of contact with kin or the absence of a relationship. Roseneil and Budgeon (2004) argue that intimacy and care increasingly takes place 'beyond the family' between partners who are not living together and within friendship networks. They continue their argument by suggesting that sociologists should focus on the 'cultures of intimacy' that are emerging as a result of changing ways of living. Similarly, Bertram (1994) (and more recently Levin, 2004) has discussed the lives of couples who inhabit separate households, who 'live apart together'. Furthermore, Jamieson's (2003) interviews with single people in their twenties found that over half described themselves as in a relationship.

2.9 While the emphasis of this paper is on living arrangements rather than on relationships, on living alone rather than being single/unpartnered, a fuller understanding of social patterns of living alone does provide useful background information for debates around personal life in the early 21st century.

Life course

2.10 Family diversity and greater dislocation between family and household has in recent decades led sociological discussion away from a life cycle model to a life course frame of analysis. Life course analysis links together family relationships and household forms within biographical time (Elder, 1985). It can be used to capture the distinctive pattern of relationships, and their dynamics over time, that different generations and age cohorts experience. Life course analysis emerged from naturalistic metaphors about family situations of individuals within their life cycles. In the past the emphasis was on the smooth and predictable movement through stages and living arrangements. Currently there is an emphasis on trajectories and journeys where, despite constraints, the pattern is at root a product of choice. Within this, solo living is not an end state, but an any-time potential for adult lives.

2.11 Solo living has a different social significance depending where it is located in the life course. Solo living among twenty-year-olds is likely to be associated with a different life style and set of meanings than solo living for those in their forties or those in their sixties. In the Jamieson et al study (2003) very few of the young people interviewed saw being single and living alone as long-term option. And it was seen as more transitional for young women than young men. It was something that characterised their stage of life, an interlude of choice and freedom, and an opportunity to establish their careers and have greater control over their resources before firming up on partnerships and more fully sharing lives and homes. Given the perceptions of the twenty-somethings, questions arise about the chances of the younger age cohorts living alone as they age? How transitory has solo living has proved for earlier cohorts? How does it relate to transitions to other types of household? The particular social contours of solo living are revealed by examining individual variables, but putting these factors together suggests the differing place of solo living in life styles and life courses.

'Individualization' and the life course

2.12 The rise of solo living is frequently seen as an indicator of 'individualization'. White (1994) suggests that there is a strong cultural preference in the USA for privacy and independence, and when individuals have sufficient resources one of the ways they are able to realise such values by living alone. Beck-Gernsheim (2002) makes a similar argument about the impact of 'individualization' on family life in Germany. Furthermore, van de Kaa (2002), Ogden and Hall (2004) and Bawin-Legros (2001, 2004) identify demographic and cultural change as Europe-wide. Beck-Gernsheim argues that the greater choice and the greater fragility of contemporary relationships means that lone occupant households are more likely to occur throughout an adult's life and have different implications and meanings at these different points. Solo living is likely to be more common before co-habitation, after de-coupling (especially for men) and as a response to labour market instability and the demands for career flexibility.

2.13 Bawin-Legros' analysis (2001, 2004) of the new sentimental order emphasised the re-formation of intimacy values and their implications for inter-dependencies. As peoples' reference points shift from collectivities to the individual, from their obligation to others to their freedom to choose and their right to privacy this interplays with living arrangements. Here an analysis of living alone provides an examination of a structural context resonant for these cultural shifts. This argument has been echoed by others who link solo living to lifestyle choice associated with the professionalisation of the labour market (Hamnett, 1994) and in the housing market (Hall et al, 1997; Lyons, 1996) with a consumption oriented lifestyle facilitated by not having children and having access to greater wealth.

2.14 Within a more individualised society, living alone may not only be a more frequent feature of people's lives but also a more permanent, more socially recognised and distinctive feature of the lives of present and future generations. Solo living amongst younger generations may be normalised within a cohort's experience. In the future, it may be an increasingly common experience for a larger proportion of the population and a more permanent feature of adult lives and household organisation. In this context, questions about the incidence and the stability of solo living in the lives of the non-pensionable population merit more systematic investigation.

Cohort analysis and longitudinal study

2.15 Sociological theorists and social commentators have linked the increased flux and diversity in family forms to 'individualization' as an emergent feature of social structure and contemporary culture. To follow this theme empirically, any analysis of the changing shape of the life course and the place of solo living in household formation and change necessitates cohort analysis. Such analysis assumes that 'each age cohort is to some degree embedded in sets of social, political and economic relationships which are different from those of previous cohorts and mark it off as distinct' (Allan and Jones, 2003: 3). Cohort experiences reflect broader historical shifts in such areas as labour markets, family obligations and cultural understandings of marriage and partnership. They also reflect trends in life course decision-making and the ways in which individuals understand and describe the decisions they have made.

2.16 Methodologically, cohort analyses of life courses require longitudinal study in order to follow each age cohort over time. The vehicle we will use to systematically examine living alone is the Census-derived ONS Longitudinal Study, which links individual records in 1971, 1981, 1991 and soon 2001. It permits us to ask questions about living alone in younger age cohorts at time points spanning a 20-year period. It is therefore able to provide a clear picture of the likelihood and the stability of living alone in England and Wales.

Using the LS to track Lone Occupant Households

3.1 The LS is a set of records of various events held by ONS relating to just over 1% (about 500,000 people) of the population of England and Wales. These can be linked in a variety of ways for analysis. Initially, all people born on each of four dates in any year were selected from information given in the 1971 Census. From 1971, as new births occur on these four dates each year and as immigrants with these birth dates register with the National Health Service, these people join the LS. Another sample of all those giving the selected birth dates was taken from the 1981 Census and their Census records were incorporated into the LS and individually linked to any existing 1971 record. This procedure was again repeated after the 1991 Census. Thus the LS represents a continuous sample of the population rather than a sample taken at any one time point only. Census information is also included for all people living in the same household as the LS member.

3.2 In the research reported here a particular age cohort has been selected: all of those LS members enumerated at home in each of the 1971, 1981 and 1991 Censuses who were of working age (15-44) in 1971. Thus this age cohort allows us to track adult household moves, but selects out the elderly. The first feature has the advantage of allowing us to know that no children were involved in certain moves, say from/into married couple households with dependent children[1]. A move from or to this arrangement usually represents a marriage or a divorce. The second feature allows us to discount the long-present tendency for the elderly, especially elderly women, to live alone.

3.3 The overall cohort is divided into 10-year age bands, thus allowing a life course comparison within each age cohort and between different age cohorts. In the latter case, for example, we can compare cross-sectionally those who were 35-44 in 1971, with those in the same age group in 1981. It is important to note that whilst our analyses are movements in and out of households, our unit of analysis is the individual LS member. Furthermore living alone is defined here as an individual who lives in a separate household in which he or she does not share communal housekeeping (Celsius, 2004). In order to track lone occupant households we have used a household structure variable from the Census.

3.4 The household structure variable is based on the building blocks of Minimal Household Units (MHU), originally derived by Overton and Ermisch (1984) and developed for the LS by Williams and Dale (1991) and Dale et al (1996). All households can be divided into one or more MHUs. The variable used here is based on a typology which also distinguishes MHUs on the basis of age and on the combination of MHUs in the same household. A dependent child is defined as under 16 years (15 in 1971), or under 19 but still in full-time education. Married couples are those who classified themselves as married in the Census.

3.5 The full 12 household categories (Williams and Dale 1991; Dale et al 1996) are:

  1. Elderly Person living alone
  2. Elderly Married couple
  3. One person in household, less than 65
  4. Two or more non married adults in household, no elderly
  5. Non elderly couple, no children in the household
  6. Married Couple with dependent children
  7. Married Couple with dependent children and non-dependent children and others, or other children.
  8. Married Couple, no dependent children, with non-dependent children or others, no elderly person in household.
  9. Lone parent with dependent children.
  10. Lone parent with dependent child(ren) with non-dependent child or others
  11. Two or more families, most with dependent children - no elderly people
  12. Complex households with at least one elderly person

3.6 Household types 1 and 2 were eliminated automatically because of the age cohort selection. Categories 9 and 10 were combined because the numbers in category 10 in the current analyses were very small. Likewise category 11 was very small (and distinguished from 12 only by the absence of an elderly person) and was thus combined with 12.

3.7 For the purposes of this study it has been possible to simplify the presentation of the above household typology in the following way:

Strengths and Weaknesses of the LS in Household Research

3.8 The obvious strength of the LS is the ability to track individuals over long periods of time. Because it is based on the England and Wales Census, it has universal coverage, with a response rate of over 98% and is not skewed geographically. Even relatively small numbers will be statistically significant (Hattersley and Creeser 1995) and the sample here can be seen as a fair representation of the population of England and Wales. The LS is the only UK data set of this size able to track individuals and (to an extent) members of their households over a 20-year (soon to be 30-year) period.

3.9 LS data are essentially descriptive and cannot provide explanations of why people moved into or out of any particular household arrangement, though cross tabulations with other variables can provide valuable clues about lifestyle choices, or indeed constraints. In this paper we have made some modest speculations, but for the most part we are content to describe the kinds of moves people made. Although the LS can be linked to event data (Hattersley and Creeser 1995) for the most part only Census variables can be used.

3.10 A limitation of the LS, and indeed of most longitudinal data sets, is that it is a snapshot. It is based upon Census enumeration and the state recorded on Census day may not hold, or have held, for very long. A person enumerated as living in a lone occupant household in April 1981 may have moved there that same month, but conversely may have lived in this arrangement for most of the decade. A person enumerated in (say) a lone occupant household in two Censuses may have remained in that household arrangement for the whole period, or conversely may have lived in other arrangements in the intervening years. Nevertheless one assumes that within the sample such a diversity of states would cancel each other out.

3.11 Our approach to investigating the stability of living alone and its place in the life course involves following a cohort of individual LS members and examining the types of households in which they lived at each of the three Census points. The LS enables us to look back from 1991 to see the household origins those living in lone occupant households and forwards from 1971 to see the destinations of those then living alone A final point to emphasise is that the unit of analysis throughout the study is the individual LS member.

The Household Origins and Destinations of Individuals Living Alone

4.1 The social context of a lone occupant households changed considerably during the period 1971 to 1991, though as we have noted above, the overall number of people living in lone occupant households increased throughout this period. This is borne out by the data presented in Table 1. Looking at our whole sample of those aged 15-44 in 1971, we find that the proportional increase in those living alone is the largest, rising from 1.6% in 1971 to 3.8% in 1981 and 8.4% in 1991. In contrast there was a notable decrease in those living in couple-based households. The proportion of people living in such households fell by 6.5% between 1971 and 1991.

Table 1. Comparison of the living arrangements of the sample showing their household types at each Census point: 1971, 1981, 1991

It should be noted that the sampling method for the LS is based on the individual, rather than the household, care should be taken in interpreting comparative percentages of those living in households of varying sizes. Because a household with five residents has five times as much of a chance of containing a person born on one of the four selected LS dates as a household with one resident, large households tend to be over-represented in the LS compared to the full Census.



4.2 Having set the household context for our LS data set, we now move to examining our sub-sample of those who were living alone at any of the three Census points. In all of the tables that follow, men and women are compared in ten year age groups.

Origins 1971 and 1981

4.3 We begin by looking back in time to the household origins of those living alone in 1981 and 1991. Focusing only on those living alone in 1981, Table 2 shows the types of households in which those living alone in 1981 were living in 1971. Overall the most common 1971 household origin for those living alone 10 years later was that of couples with dependent children. In many cases, this movement would be accounted for by younger people leaving the parental home to live alone (the LS member could be any member of the household). The second most common origin was that of a lone occupant household, with 14.3% of all those living alone in 1981 also having lived alone in 1971. There was an overall sex difference, with 12.9 % of men and 16.1% of women remaining in lone occupant households.

4.4 Amongst those who lived alone in 1981 there was an age cohort gradient, with each older cohort more likely to have been living alone in both Census years than each of the younger cohorts. There were also differences between men and women. Only 3.1% of men and 3.9% of women in the youngest cohort had lived in lone occupant households in both Census years, but in the 25-34 cohort there was a much greater sex difference as 20.5% of women in this cohort lived alone in both years compared to 14.2% amongst men. The sex difference was in the opposite direction in the oldest cohort, with an increase of over 10% for men, but just over 1% for women.

4.5 For men, the greater number in the 35-44 cohort living alone in both Census years (as compared to the 25-34 cohort) may indicate a departure some time before from a couple household (due to divorce or separation). Indeed there is some evidence to suggest a large proportion of the men who had moved into lone occupant households by 1981 had lived in couple households (with children) in 1971. Many more men than women in the older cohorts were in households characterised as 'couples with dependent children' in 1971.

Table 2. 1971 Household Origins of Those Living Alone in 1981. Age 1971

Click here to see the main points illustrated using bar charts

4.6 Table 3 follows the same format as Table 2, but moves the comparison point 10 years on, and shows the 1981 household origins of those living alone in 1991. As with 1971-1981 we can examine movement over time, but additionally we can make cross-sectional comparisons between the two 10-year periods. It is however important to remember that because we were following all those LS members between 15 and 44 years of age in 1971 (who were also present in 1981 and 1991) there was an ageing effect in the cohorts. Thus, those people 15-24 in 1971 were 25-34 in 1981.

4.7 Overall there was a large increase between 1981 and 1991 in the number of single- person households. In 1981 5,965 members of the sub-sample moved into or continued to live in lone occupant households, but by the next Census this had increased to 13,027. The cohort analyses suggest not only the growing likelihood of living alone but a growth in the numbers who continued to live alone. Looking at the household origins of 1981's solo livers, 14.3% of all those living alone in 1981 had previously lived alone in 1971. However, in 1991 25.3% of those living alone had also lived alone in 1981. Furthermore, unlike the previous 10 years, the most common household origin (looking back to 1981) for both men and women was the lone occupant household.

4.8 Differences between men and women in many of the other categories are also apparent. More men than women left complex households (16.2% compared to 10%) and households with dependent children (15% compared to 9.4%), whilst more women than men left lone-parent households (12% compared to 4.6%) or households where there were couples with no children (13.9% compared to 9%). All suggest that the biographic points and the household contexts were different for men and women, reflecting and adding to the gendered experience of solo living.

4.9 Between 1981 and 1991 men were more likely to have continued to live alone than women. While cross-sectional comparisons of the 25-34 and 35-44 age cohorts (living alone in both Census years) showed similar patternings between the sexes, it is in the ageing of the cohorts where sex differences emerged, with 34.2% of men and 23.8% of women aged 45-54 in 1981 living alone in 1981 and 1991. Therefore although living in a lone occupant household 10 years before was the biggest predictor of being in that same arrangement10 years later, this became more the case for men as they got older.

Table 3. 1981 Household Origins of those living alone in 1991. Age 1981

Click here to see the main points illustrated using bar charts

Destinations 1981 and 1991

4.10 Tables 4 and 5 might be thought of as an opposite image of the previous two tables. Here we examine what kind of households those who previously lived in lone occupant households moved to. As might be expected from the foregoing, a large proportion of those in lone occupant households remained in such households.

4.11 Table 4 shows that between 1971 and 1981 34.4% of the sample remained in lone occupant households and there was little sex difference, with 30.6% of men and 39.8% of women remaining in a lone occupant household. Living alone appeared to be a more stable experience for women from their mid-twenties onwards; once they lived alone they were more likely than men to continue to live alone.

Table 4. 1981 Household Destination of those living alone in 1971. Age 1971

Click here to see the main points illustrated using bar charts

4.12 Table 5 shows an increase in the number of individuals who continued to live alone, with percentages rising from a third (34.4%) between 1971 and 1981 and over half (55%) between 1981 and 1991. This increase may be due to age, with older age cohorts more likely to continue to live alone. Overall, more women than men continued to live alone.

4.13 When we compared the same age groups and the likelihood of people living alone, there appeared to be little change for women aged 25-34 or for both men and women aged 35-44. However there was a significant increase in men aged 25-34 continuing to live alone in that 10 year interval. Between 1971 and 1981, 22.0% of men in that age group continued to live alone, whereas between 1981 and 1991 the proportion had risen to 32.2%. This is indicative of the increasing numbers of men who became solo-livers.

4.14 These tables provide some clues about the longevity of living alone by showing who moved out from such arrangements. Turning to those who moved from living alone to living in a couple-based household, the data is indicative of the rising age of forming initial partnership/age at first marriage. Focussing on the 25-34 age group of men and women, over the 10-year time interval, there was a small rise in the proportion moving from living alone to a household with no children. There was little sex difference for this age group in their movement into households with dependent children.

Table 5. 1991 Household Destination of those living alone in 1981. Age 1981

Click here to see the main points illustrated using bar charts

4.15 As the LS is giving a snapshot in time, men may have lived, in the intervening years, in couples and with young children. However, in the ten year time lapse, such relationships may have come and gone in their lives. The fleeting nature of such relationships is suggested by another aspect of the data set. Throughout the time period the proportion of the 'ever-married'[2] in our sample steadily rose from 67.6% in 1971 to 89.3% in 1981 and 92.5% in 1991. When we look at our youngest cohort (15-24) 69.7% of them were never-married in 1971, only 17.4% remained never-married in 1981.

4.16 However when we turned to the next age group, those 35-44, we found a growing proportion of men and women continuing to live alone and the proportion was very similar when we compared them over the two time intervals; just over 50% of men and 60% of women remained living alone. Those who 10 years later lived in couple-based households, for men these households were more likely to contain dependent children than those of women. Also for this older age group, a much higher proportion of women than men now living alone were previously in lone-parent households.

4.17 Table 6 is a little different to the previous two tables in so far as it follows individuals through all three Censuses, but selecting all those persons who lived alone in both 1971 and 1981. The 851 persons in this table represent 0.5% of the total sample. Seventy eight percent of this sub-sample had been enumerated in lone occupant households at all three Censuses and can therefore be reasonably assumed to be the core of people who had lived alone through much of their working lives. Such longevity of living alone was especially prevalent in the oldest age cohorts in both sexes. Women in the youngest cohort (35-44) were more likely than men to be living alone at all three Census points (73.3% compared to 54.5%). Secondly the table provides some indication of the kinds of exits from living alone for people in early to late middle age. The youngest person would be 35 in 1991 and the oldest 64. Nevertheless 20.5% of men in the youngest cohort became part of couples with dependent children. This was much less likely for women (6.7%) and is explained by most moving to or beyond normal child- bearing age. However across both sexes there were proportionally sizeable numbers of moves into couples with no children, possibly indicating the formation of new partnerships, but without children.

Table 6. 1991 Household Destination of those living alone in 1971 and 1981. Age 1991

Click here to see the main points illustrated using bar charts

Changing Patterns of Solo Living in the Different Age Cohorts

5.1 Because our sample consists of three different age cohorts in 1971 it is possible to compare patterns of solo living between the cohorts. Not only were these cohorts at a different stage in their life course in 1971, but would almost certainly hold different values and follow different social practices. A comparison of the cohorts therefore allows us to compare intergenerational differences in living patterns. Specifically we can compare those who were between 25 and 34 in 1971 with those who were between 25 and 34 10 years later. Similarly we can compare those who were 35-44 in 1971 with those who were 35-44 ten years later.

5.2 25 to34 age group. Of men who were aged between 25 and 34 in 1971 a total of 929 were living alone by 1981. This compares to 2,040 of men of the same age in 1981 living alone by 1991 (120% increase). Thus those in the younger cohort of men were twice as likely to live alone at the same stage in their lives. The pattern was very similar for women with 556 of the 1971 25-34 year olds living alone by 1981 compared to 1,175 in the same age group between 1981 and 1991 (111% increase).

5.3 35 to 44 age group. The pattern in this age group is similar but less pronounced with a 36% increase from 1,105 (1971 to 1981) to 1,715 (1981 to 1991) amongst men. Amongst women there was also a 36% increase from 1,188 (1971 to 1981) to 1,844 (1981-1991). From this we can conclude that younger age groups are more likely to live alone at the same stage of their lives than their forebears 10 years earlier.

5.4 A further question is how likely were different age cohorts to continue to live alone? A comparison of those who were 25-34 in 1971 with those who were 25-34 in 1981 shows that more of the younger age group continued to live alone (Tables 1 and 2). Amongst men aged 25-34 14.2% were living alone in 1971 and continued to live alone in 1981, whereas 28.0% of men in this older group ten years later lived alone in 1981 and in 1991. For women the increase was less: from 20.5% (1971-81) to 22.8% (1981-1991).

5.5 This study therefore shows that the number of people living alone in the study period increased, especially among younger age groups, and suggests that within that period individuals were more likely to live alone for long periods of time. It also suggests major sex differences in the pattern of living alone: from their mid-20's onwards, women were much more likely to remain living alone than men. But young men appeared to be catching up fast with the greatest change between the 2 sets of data in Tables 4 and 5 being the growing proportion of men aged 25-34 who continued to living alone in 1991 as compared with 1981.

Conclusions

6.1 The foregoing analyses confirm other research showing the increasing number of non- retired people who are living alone for part of their adult lives. Additionally it demonstrates that once a person lives alone they are more likely to continue to live in that arrangement than any other. Furthermore the chances of continuing to live alone rose significantly with age. Indeed the research provides some evidence to refute suggestions that living alone is a temporary phase in one's adult life. Moreover the tendency to live alone and to continue to live alone appears to be growing amongst younger cohorts of people. The twenty-somethings interviewed by Jamieson and her colleagues (2003) may envisage living alone as a time-limited circumstance, but it may prove to be more permanent that they think it is going to be.

6.2 The analyses also show differences in the pattern of living alone between the sexes and suggest that the transition to solo-living was distinctively gendered. The largest increase in living alone was among men. Also men moved to living alone following relationship breakdown whereas for women living alone was more likely to be contingent not only on relationship breakdown but also on children leaving home. However, once women were living alone in their 30's, they were more likely to continue to live alone than men. The different place of living alone in the biographies of men and women will enhance the differences in the ways in which men and women exchange care and develop intimacy.

6.3 Cross-sectional Census data from 2001 shows that the number of non-retired adults living alone continues to rise. It therefore seems likely that there will be a continuation of the trends reported here when 2001 LS data are linked to the current analyses. We expect not only more adults to be living alone but also that higher proportions of those living alone in their 20's and 30's will continue to live alone.

6.4 Finally, the data illustrate the structural fall-out of 'individualization' and the new sentimental order. In whatever ways individuals may conceptualise intimacy and approach personal and family relationships, these are increasingly thought about and acted upon from a household base of individuals living alone. Furthermore the shape of the life course appears to be altering. In the past living alone was seen as a brief interlude or end-state in adult lives and, outside gerontology, attracted little research interest. With significant numbers of younger people living alone and their growing propensity to continue to live alone, new questions are raised about the form and exigencies of 'non-family living' and about how those living alone engage in family practices and partnerships. We believe the answers to these questions will provide a new dimension to 'family' studies.

Notes

1 Some of the youngest cohort (15-24 in 1971) who move would have been moving from the parental home. Thus, a move in this cohort between 1971 and 1981 from a household characterised as couples with dependent children and others or non dependent children or others could be either a move from the parental household, or from a marriage.

2'Ever-married' includes all those who are or have been married. At any Census point it includes those who are married, divorced and widowed.

Acknowledgements

This article draws on research funded by the ESRC (RES-000-22-0081) under the Small Grants Scheme. The authors gratefully acknowledge the ONS for access to the Longitudinal Study.

References

ALLAN, G. AND CROW, G. (2001) Families, Households and Society. Basingstoke: Palgrave.

ALLAN, G. AND JONES, G. (2003) 'Introduction', Social Relations and the Life Course ed'd by G Allan and G Jones. Basingstoke: Palgrave.

BAUMAN, Z. (2003) Liquid Love: Cambridge: Polity Press.

BAWIN-LEGROS, B. 2001. Families in Europe: A private and political stake - intimacy and solidarity. Current Sociology, Vol 49, pp. 49-65.

BAWIN-LEGROS, B. 2004. Intimacy and the new sentimental order. Current Sociology, Vol 52, pp. 241-250.

BECK-GERNSHEIM, E. (2002) Reinventing the Family: in Search of New Lifestyles. Cambridge: Polity.

BERTRAM, H (1994) 'Die Stadt, das Individuum und das Verschwinden der Familie', Aus Politik und Zeitgeschicte, 29-30, pp.15-35

DALE, A, WILLIAMS, M. AND DODGEON, B. (1996) Housing Deprivation and Social Change. London: HMSO

ELDER, G (1985) 'Perspectives on the Life Course', Life Course Dynamics in G. Elder (Editor). Ithaca, NY: Cornell University Press. pp. 23-49.

GIDDENS, A. 1992. The transformation of intimacy. Cambridge: Polity Press,

GITTINS, D (1985). The Family in Question. Basingstoke: Macmillan,

GOLDSCHEIDER, F. K. AND WAITE, L. J. (1991) New Families, No Families. Berkley: University of California

HALL, R. OGDEN, P. E. AND HILL, C. (1997) 'The Pattern and Structure of One-person Households in England and Wales and France' International Journal of Population Geography Vol. 3, pp. 161 - 181.

HALL, R. AND OGDEN, P. (2003) 'The Rise of Living Alone in Inner London: Trends Among The Population of Working Age' Environment and Planning, Vol. 35, pp. 871 - 888.

HALL, R., OGDEN, P. E. AND HILL, C. (1999) 'Living Alone: evidence from England and Wales and France for the last two decades' in S McRae (editor) Changing Britain: Families and Households in the 1990s. Oxford: Oxford University Press

HAMNETT, C. (1994) 'Socio-economic Change in London: Professionalisation not Polarisation', Built Environment Vol. 20, pp. 192 - 203.

HATTERSLEY, L. AND CREESER, R. (1995) LS Series 7: Longitudinal Study 1971-1991. History, organisation and quality of data (London: HMSO).

LEVIN, I. (2004) 'Living Apart Together: A New Family Form', Current Sociology, Vol. 52, pp. 223 - 240.

LYONS, M. (1996), 'Employment, Feminisation and Gentrification in London, 1981 - 93', Environment and Planning A Vol. 28, pp. 341 - 356.

JAMIESON, L. 1998. Intimacy: Personal relationships in modern societies. Cambridge: Polity.

JAMIESON, L. 1999. Intimacy transformed? A critical look at the 'pure relationship'. Sociology, Vol 33, 477-494.

JAMIESON, L., STEWART, L.Y., ANDERSON, M., BECHOFER, F., MCCRONE, D. (2003) 'Single, 20-something and seeking ?' in Social Relations and the Life Course in G. Allan and G. Jones (editors). Basingstoke: Palgrave.

KAUFMANN, J.C. (1994) One person households in Europe. Population Vol. 49, Nos 4-5, pp.935-958.

MORGAN, D. (1996) Family Connections. Cambridge: Polity.

OGDEN, P. E. AND HALL, R. (2004) The second demographic transition, new household forms and the urban population of France during the 1990s. Transactions of the Institute of British Geographers, Vol 29, 88-105.

OFFICE FOR NATIONAL STATISTICS (2003) Social Trends 33. http://www.statistics.gov.uk/socialtrends

OFFICE FOR NATIONAL STATISTICS (2003) Census 2001. http://www.statistics.gov.uk/census2001

OVERTON, E. AND ERMISCH, J. (1984) 'Minimal Household Units', Population Trends No 35. London: OPCS

ROSENEIL, S AND BUDGEON, S (2004) 'Cultures of Intimacy and Care Beyond 'the Family': Personal Life and Social change in the Early 21st Century, Current Sociology Vol 52, pp 135-159.

SILVA, E. B. AND SMART, C. (eds) (1999) The New Family ? London: Sage.

VAN DE KAA, D. J. 2002. The idea of a second demographic transition in industrialized countries. Sixth Welfare Policy Seminar of the National Institute of Population and Social Security. Tokyo, Japan. 29 January 2002.

WHITE, L. (1994) 'Co-residence and Leaving Home: Young Adults and their Parents' in J Hagan and K Cook (editors) Annual Review of Sociology, Vol. 20, pp.81- 102.

WILLIAMS, M. AND DALE, A. (1991) 'Measuring Housing Deprivation Using the OPCS Longitudinal Study' LS Working Paper 72. London: SSRU.