‘ He wasn ’ t in that chair ’ : what loneliness means to widowed older people By

We have little robust empirical evidence that articulates what being lonely means to older people and even less knowledge about what loneliness means to older widows and widowers; this article addresses that deficit. We undertook a re-analysis of 125 interviews with older people (aged 55 98) that explored their experiences of widowhood. In this article, we focus on those interviews in which participants described themselves as experiencing loneliness by the spontaneous use of terms such as ‘‘lonely’’, ‘‘loneliness’’ or ‘‘lonesome’’. Almost half of the participants (42%) described themselves in that way without any prompting from the interviewer. In terms of understanding and describing the meaning of loneliness, 50% explained loneliness in terms of absence of either their spouse, a physical presence in the house or people. One-third (34%) discussed loneliness in relation to time and place: night, weekends and home, and 4% described the emotional impact of loneliness. Fifteen per cent just said they were lonely without elaboration, assuming a common understanding of what loneliness means. Our findings suggest that widowed people’s understanding and experience of loneliness resonates 1 Kate Mary Bennett, Institute of Psychology, Health and Society, University of Liverpool, Liverpool, UK 2 Christina Victor, School of Health Sciences and Social Care, Brunel University, London, UK International Journal of Ageing and Later Life, 2012 7(1): 33 52. # The Authors


Introduction
There is extensive quantitative literature in the United Kingdom, Europe, North America and Australasia documenting the prevalence of loneliness and the key risk factors thereof (Dykstra 2009;Fokkema et al. 2012;Victor et al. 2009;Yang & Victor 2011).Sheldon (1948), in his landmark study of old age in Wolverhampton, reported that 7.7% of those aged 60/65 and over were often or always lonely, compared with 13% of widowed participants, and that 78% of those who described themselves as lonely were widowed.Contemporary research has confirmed the statistical association between loneliness and widowhood in old age.Victor et al. (2005a), for example, report that in Britain loneliness amongst widowed individuals aged 65 and above was 20% compared with 1% among married, 7% among divorced and 9% among separated individuals.This is typical of the excess levels of loneliness found amongst widowed people across a diverse range of countries and is not a consequence of confounds, such as age and gender (e.g.Dykstra & Fokkema 2007 for the Netherlands, Kaasa 1998 for Norway, Steed et al. 2007 for Australia, Theeke 2009 for the United States of America and Wang et al. 2011 for China).
Two key conceptualisations of loneliness, cognitive deficit and loss, underpin the quantitative measurement of loneliness.The cognitive deficit perspective proposes that loneliness is a deficit between actual and desired quantity and/or quality of social engagement (Peplau et al. 1982).Within this framework, the Theory of Mental Incongruity (Dykstra & de Jong Gierveld 1994) suggests that loneliness is formed by a mismatch in the interrelationship of cognition, experience and expectations of social relationships (see also van Baarsen 2002).The loss perspective focuses on the absence of specific attachments and differentiates emotional loneliness (the loss of a key attachment figure such as a spouse) from social loneliness (a reduced social network) (see Weiss 1973).In his extended Bowlby theory, Weiss (2007) argues that whether or not a relationship is functionally substitutable is a key feature of the distress that manifests itself as loneliness.Both of these perspectives have relevance to the widowed population who experience reductions in their social networks (Bennett et al. 2005a; Utz et al. 2002) and the loss of a primary attachment relationship that is both functionally highly specific and ''non-substitutable'' (Weiss 1973).
What do studies of widowhood in old age have to say about loneliness?Loneliness has been reported as a feature of widowhood from the earliest studies (Atchley 1975;Bowling & Cartwright 1982;Lopata 1980;Parkes 1972).Lund et al. (1993) reported that it was the greatest difficulty faced by older widowed people.A number of recent studies have examined whether loneliness among widowed people can be explained more effectively by the loss perspective or by the Theory of Mental Incongruity (van Baarsen 2002).There is evidence both for the Theory of Mental Incongruity (Dykstra 1995) and for social loneliness (Utz et al. 2002).Dykstra & De Jong Gierveld (2004) found evidence for both emotional and social loneliness, when marital history was considered alongside gender.Being socially embedded provided protection from emotional loneliness for women but not men, for whom the marital bond was more important both socially and emotionally.Van Baarsen (1999) tested hypotheses from both Mental Incongruity Theory and emotional/social loneliness.She found that whilst there was much support for emotional loneliness, there was only some support for the theory of mental incongruity.What was most interesting was that social support, whilst not providing protection from loneliness in the short term, did so in the long term.Thus, it is important to consider the temporal dimension in widowhood and loneliness research as well as age, gender and factors linked to the quality of the marital relationship.To date, the strongest quantitative evidence supports an emotional loneliness explanation with loneliness caused by the loss of the spousal attachment figure (Stroebe et al. 1996).
As the evidence we have presented so far has shown, the focus in most studies of loneliness has been on enumerating the extent of the ''problem'' of loneliness and establishing the characteristics of those groups most at risk of experiencing this condition with the implicit objective of being able to ''screen'' potential risk groups and develop remedial interventions.There has been much less emphasis on engaging with the qualitative elements of older people's understanding and experiences of loneliness (see Victor et al. 2009), although the study by Pettigrew (2007) is a notable exception (see Gladstone 1995;Graneheim & Lundman 2010;Hauge & Kirkevold 2010).
The majority of articles on loneliness focus on the precursors to loneliness, the consequences or the correlates of loneliness, with little research examining what it feels like to be lonely (Boyle 2010) and more specifically a widowed person.Qualitative studies of widowhood touch, in passing, on loneliness experiences (Czerenda 2010;Davidson 2001Davidson , 2002)).Davidson (2001Davidson ( , 2002)), for example, reported that for some widowers, motivation for re-partnering was their feelings of loneliness.In the most detailed research, Chambers (2005), in her widows' narratives, identified ''loneliness and despair'' as one theme.These widows mainly reported feelings of isolation and lack of support from family, with occasional references to the absence of the husband.However, although loneliness was identified as a theme, there is only one quote that directly mentions loneliness.Thus, there are few studies which examine in detail what loneliness means to older widowed people and nothing that examines their spontaneous utterances about loneliness and the situations in which they find themselves to be lonely.This article addresses these deficits in our knowledge base by presenting the re-analysis of two existing qualitative and quantitative studies examining the experience of widowhood amongst older people in England.Our aims in undertaking this analysis were to examine: (1) if participants raised the issue of loneliness unprompted Á did it emerge as a spontaneous theme?and (2) how was loneliness following bereavement conceptualised and understood by participants?

Method
The data that formed the basis of this study were derived from two independent studies of older widowed men and women conducted in England.Study 1 was a qualitative study that was designed to explore the experiences of older widowed men and women with an emphasis on understanding the emotional and participatory changes that occurred following spousal bereavement, with data collected in the East Midlands (labelled Mr./Mrs.'letter) (see Bennett & Vidal-Hall 2000).Study 2 focused on gender differences in affect and participation and was a mixed methods study using both in-depth interviews and questionnaire methods, with data collected in Merseyside (labelled Man/Woman'number) (see e.g.Bennett et al. 2005a, b).These two studies combined included 125 widowed men and women aged between 55 and 98 years, who had been widowed between 3 months and 60 years.Table 1 describes our sample.Table 2 shows that there were no statistically significant differences in terms of age, length of time they had been married or duration of widowhood between the lonely and non-lonely groups.There were no differences either between the East Midlands and Merseyside samples.
Participants were all resident in the community, living in their own homes or in sheltered accommodations.They were recruited via a diverse range of formal and informal groups run for or by older people, including organisations run for and by widows, trade unions, organisations supporting older adults and social service departments (see Bennett et al. 2005a for a discussion of selection issues).Interviews were nonprescriptive and semi-structured in nature, as the aims of the original studies were to learn from the widowed people what was important to them using the approach of ''We are the novices and you have the experience.''We asked about life before widowhood, around the time of bereavement, one year after, and at the time of interview.We wanted to know what respondents did and how they felt at these different times, so we used prompts such as ''what did you do?'' and ''what did you feel?''(see Bennett & Vidal-Hall 2000).All respondents gave informed consent and the research teams tape-recorded the interviews, conducted at the respondents' homes, ranging from three-quarters of an hour to an hourand-a-half.Interviews were transcribed verbatim.Ethical approval was obtained from appropriate local ethics committees.

Analysis
Stage 1.This method was the same for both studies.As the interviews were completed, they were transcribed and analysed using a grounded theory approach (Bennett & Vidal-Hall 2000;Charmaz 1995;Smith 1995).More specifically, the interviews were read line-by-line to gain a holistic overview and then re-read and coded.This process was reflexive, and as new codes and themes emerged, the interviews were recoded.Although the two studies were independent, the content of the interviews had many similarities and over 300 codes emerged from the analysis of the 125 interviews.The codes were diverse, including talking to the dead spouse, This was generated when participants spontaneously used the words ''lonely'', ''loneliness'' or ''lonesome'' but not simply the word ''alone''.Fifty-three of the 125 interviews (42%) were coded as lonely.Analysis of the code ''alone'' is presented elsewhere Á in that work, women (but not men) who talked about being ''alone'' were more likely to be coping well (Bennett et al. 2005a).Thus, we believe that being alone has different meanings from being lonely or lonesome.All interviews were coded blind by other members of the original research team so that reliability across interviews could be assessed Á agreement was found to be 80%.Stage 2. Fifty-three interviews, where loneliness was spoken of spontaneously, were analysed in detail with the focus on those passages where participants spoke explicitly of being lonely, lonesome or of loneliness.These passages were coded in a similar manner to stage one, that is, independently by members of the current research team.
Three key themes emerged from this reanalysis: notions of absence, the spatial/temporal aspects of loneliness and unelaborated loneliness.The first two of these themes were further analysed to understand in detail the experience of loneliness.Explanations from an individual were not confined to one theme.

Results
Overall, 42% of the participants interviewed in these studies spontaneously talked about loneliness: there was no prompting or mention of the topic in the interview guides.Thus, our remaining analysis focuses on this subsample of 53 interviews.We now present the results of the three key themes.

Absence
Absence was a central theme within our participants' narratives of loneliness.In 26 interviews, participants discussed their loneliness focusing on the absence of the spouse, in particular, or the absence of people in general (50% of those participants who reported being lonely).
Of these participants, more than half (61%, n 016) described their loneliness as the absence of their spouse and this broad theme encapsulates four important components: the absence of the spouse, the ''irreplaceable'' nature of their spouse, the permanence of the loss and the spatial elements of this loss.This notion of loneliness clearly resonates with the ideas of Weiss (1973) about the loss of a ''key'' attachment figure who is ''non-substitutable'' being central to the experience of loneliness.Our participants report that they are surrounded by family and friends, live active social lives and do not describe themselves as ''outsiders'', nor would their communities describe them as lonely.We can see the four dimensions of this theme illustrated as follows: . Loss of their spouse: .The specificity of the loss Á the absence of the specific person who was their spouse and not simply the absence of someone who could be a spouse: Because for me, nobody would ever fill Andrew's shoes.(Mrs.A) . The permanence of the loss: it is not only the absence that leads to loneliness Á it is the permanent absence.Mr. H's wife had been in and out of hospital, sometimes absent for a few months.He had not been lonely then (or at least not in the same way).However, now he knows she is not coming home: Suddenly alone . . .previously when I'd been on my own she was goin' to come back in a couple of months but I knew now she wouldn't.(Mr. H) Although Mr. H uses the term ''alone'' in this quote (and we have more to say about the use of that word later), rather than the term ''lonely'', which he does elsewhere, we have chosen this quote because it demonstrates so well the permanence of his situation.
. The spatial elements of the loss of a spouse.Loneliness was associated not simply with the absence of their spouse, or the permanence of this loss but also with their absence from a particular place in the house/home.Often this was absence from a chair, He wasn't in that chair.(Woman 29) I never sat on my own as my husband was always there.(Woman 9) I was always sat in that chair there and whenever I looked up from here she was there.
But when she'd gone she was not.(Man 8) In 10 interviews (39% of those talking about absences), participants talked more generally about the loneliness associated with the absence of people.This resonates more closely with Weiss's (2007)  There is also the pernicious effect that this lack of social interaction can have as illustrated by Mr. D who feels that he no longer can talk to others, and as a result people stop wanting to talk to him Á a vicious circle: I've lost the art of conversation, people don't have much to do with me.
And the fear that: They wouldn't know if I was dead.(Man 12)

Time and Place
The second key theme, that links with the ideas about absence as an explanation and cause of loneliness, relates to the socio-temporal dimension of loneliness.This theme was present in 18 of the interviews (34% of those reporting to be lonely) and relates to the seasonal and weekly patterns of life rather than time since bereavement.Mrs. M drew attention to the importance of time in the genesis of loneliness when she commented: Sometimes I . . .from when the sun comes in to the next week, I never sort of see anybody apart from when I go out to see people.Nobody goes, ladies won't come out at night, you know.
Night-time was problematic.In many households, couples would have spent the evenings together, or at least have been in the house together, for most of their married lives, whether working or retired.Now our participants were alone in the house and we can see this illustrated in these comments from both men and women: For many women, evenings highlighted their loneliness, making them feel trapped in their own homes.During the day, they could escape the emptiness of their houses but at night there was no escape, unless people visited and as we have seen this was a rare occurrence: This contrasts with most male participants who also spoke of being ''trapped'' within their homes but it was not linked to a specific time of day; in particular, two men (Man 28 and Man 42), used the expression: Only got these four walls.
The temporal patterns of loneliness are not simply diurnal but are overlaid by weekly patterns, with weekends seen as a lonely time as these were the days that others spent with family.Interestingly, this observation was confined to the women in our sample as the comments from Mrs. M and Woman 44, respectively, illustrate: Such a lonely life . ... Saturdays and Sundays are a bit dead for me . ... So long [Sunday] and so lonely.
In addition, again only our female participants observed that the weekend was also seen as a time when other people were seen in couples (n 04) and this emphasised how lonely the women were without their husbands.
As Woman 20 says: I find the weekends very lonely because people are always in couples at the weekends.

Unelaborated Loneliness
Fifteen per cent of these participants (n 08) described themselves as being lonely but offered no further elaboration.These participants seem to assume that the interviewer understood what they meant by being lonely and that it needed no additional explanation.Perhaps, they felt that the experience and the underpinning meanings were self-evident.A further 4% simply described the emotional experience of loneliness (n 08): The worse thing to endure in life is to be lonely.(Man 43) I've never been so lonely in my life . .., it broke my heart . . . .So you do get a depth of loneliness.(Mrs.I)

Discussion
Significant loneliness is consistently reported by 8Á10% of the population aged 65 and above across Britain, Western/Northern Europe, North America and Australasia, with the widowed consistently identified as being at increased risk of experiencing loneliness across a range of studies and countries (Dykstra 2009;Yang & Victor 2011).Rates of reported loneliness are approximately double for the widowed compared with the general population of elders and several orders of magnitude higher than those who are married.Although there are a number of studies that have examined the statistical association between widowhood and loneliness, there are comparatively few studies that have focused on this relationship from a qualitative perspective.Do older widowed people He wasn't in that chair talk about loneliness?How do they conceptualise and understand loneliness and can these understandings help us shape potential interventions to alleviate loneliness both for the widowed and others?Almost half (42%) of our 125 participants spoke freely and openly about the experience of loneliness without any prompting.Whilst this is a specific population, it suggests that, as Victor et al. (2005b) report, older people are willing to talk about loneliness and that there is scope for more qualitative studies examining this aspect of old age in-depth.We might speculate that, for older people and, more specifically, older widows, the high prevalence of bereavement reduces the ''stigma'' of loneliness that may be more evident at earlier phases of the life-cycle.How do older widows understand loneliness?We can identify three elements of our participants' understandings of loneliness: (1) an intuitive understanding that needed no explanation or which was so profound that participants could not articulate it with precision as illustrated thus ''You felt as though there were something missing but you didn't know what was missing'' (Woman 11); (2), notions of absence; and (3) the importance of time (and place).
It is not surprising that respondents frequently discussed their loneliness in terms of absence, since the popular view of loneliness focuses upon the lack of social contacts (rarely considering that the quality of social contacts might also be important).Indeed, dictionary definitions of loneliness focus on being alone and in solitude rather than either the quality of social relationships or the relationship with a specific ''key'' person.Our respondents were very precise in defining the absence that underpinned their loneliness and it did not, for example, often concern social isolation.Instead, it was identified as the absence of a spouse or indeed the absence of a spouse from a particular location within the home or domestic environment.This focus upon the loss of a key ''attachment'' relationship supports the ideas of Weiss (1973), but our respondents linking the absence of an individual with their absence from a specific location, develops these ideas further.Loneliness is clearly located within both a spatial and temporal context that encompasses the daily and weekly rhythm of life.For our respondents, night-time is clearly a time of vulnerability at which they are most likely to experience loneliness.This resonates with survey results of Victor et al. (2005a), where 64% of those who reported loneliness experienced it at specific times, with almost half (46%) reporting they felt most lonely during the evenings.Loneliness at weekends was only described by women in this study but this again validates the survey findings of Victor et al. (2009) who reported that 75% of women and 52% of men reported that loneliness was linked to the weekends.We may explain the gender differential in terms of the link between loneliness and weekends by reference to the work of Bennett et al. (2003) who observed that widowers are more likely to be taken care of by their families at weekends.For example, they are more likely to be invited for a meal on a Sunday by offspring or other relatives than are widows who are perceived as more ''domestically competent''.This may reflect attitudes about the perceived relative abilities (or otherwise) of widowed men and women to care for themselves.Our results show, for the first time, that these social relationships are played out within the context of time, space and place.Our participants' comments demonstrated how these are explicitly and inextricably linked.We did examine our data to see if there was evidence of patterns of loneliness that were related to time-since-bereavement.However, there was no clear evidence suggesting time-since-bereavement was an important factor in the discussions of loneliness.
Our finding about the different types of loss/absence that underpin loneliness Á the death of a spouse and/or the denuding of the social network and the temporal nature of the loneliness (and the variability between men and women) Á have implications for both the development and delivery of interventions to alleviate loneliness in general and after bereavement.Some interventions can compensate for some losses that arise from bereavement by providing opportunities for social participation.Interventions could be timed to respond to the time/days of week when feelings of loneliness are most extreme but this suggests the need for flexibility in service delivery timetables that are challenging for statutory service providers.The data also suggest that interventions may need to be scheduled differently for men and women.However, voluntary and ''selfhelp'' organisations may be more appropriate and perceptive in developing services/interventions that may mesh more sensitively with expressed needs of widows.For example, services like the Widow-to-Widow's programme in the United States are more likely to understand the importance of the timing of interventions (Silverman 1986).Many groups for widows meet on a Sunday, as indeed do the widows from the East Midlands in this study.Thus, ''night-time'' and/or weekend telephone befriending services may be more appreciated than day-time services.However, the majority of our participants report that it is the loss or absence of the spouse that it is the root cause of loneliness and this is more difficult to alleviate and to design interventions for.Some widowed people do put in place strategies to reduce those feelings of absence.For example, widowed women report rearranging the bedroom or the living room and changing the position of ''that chair''.Others avoid coming home at particular times or go out to avoid reminders of their lost attachment.As with the Widow-to-Widow's programme or the East Midlands's widowed groups, this type of advice might be more readily received from other men and women in the ''same boat'', rather than from more formal agencies.

Loneliness Versus Alone
Statistically, loneliness is associated with a range of negative health and quality of life outcomes (see Luanaigh & Lawlor 2008).We can see these negative consequences articulated by our participants who observed that ''loneliness is a terrible thing'' (Man 37) and that the experience of loneliness could explain '' . . .why people do silly things to themselves when they're lonely'' (Woman 8).Conceptually, it is important to distinguish between being lonely and being alone.Woman 11 points out the distinction between choosing to be alone and the loneliness which is imposed by the loss of a spouse thus ''Loneliness forced upon you Á I am lonely.I am also alone but that's totally different.''Indeed being alone and adjusted to this way of life can have positive psychological benefits as Bennett et al. (2005a) found: older widowed women who reported being comfortable alone were more likely to be coping than women who did not, or men in general.

Conclusion
Our data demonstrate that older people are willing and able to talk about loneliness.In our case, participants spontaneously raised this topic.Hence, we feel confident that we can address this issue in-depth within the International Journal of Ageing and Later Life context of qualitative interviews focusing on social engagement and participation in later life as the work of Victor et al. (2009) also suggests.The study also demonstrates that widowed people are not ashamed, embarrassed or upset about talking of loneliness.To them, and this is illustrated by those who do not elaborate on what it means, loneliness is a common and indeed potentially normative aspect of widowhood.Thus, it seems reasonable for studies to be more direct in discussing loneliness with older people or, indeed, other age groups.It is after all no more sensitive than asking widowed people about their bereavement.
This study provides evidence to support both the cognitive deficit and loss conceptualisations of loneliness illustrating the complex and dynamic nature of loneliness.Widowed people who talk about being lonely are making (implicit) comparisons between what social contact they would like, often with their spouse, and what they actually have, and this is often in the context of the presence of more general social contacts.This is true also of the temporal context of their loneliness, for example, wishing for company at night or on Sunday.At the same time this evidence also supports the deficit approaches Á but is more specific than the theory might imply.Our widowed people report, in general but not always, a specific deficit Á the loss of a spouse Á the location of that within a specific space and/or a specific temporal context, such as night-time or weekends.Rather than seeing these models as in competition our data suggest that they complement each other but that we also need to build on ideas about space and time when developing our theoretical understanding of loneliness and generating interventions.A model of widowhood and loneliness should recognise that widowed people (a) may experience reductions in actual social contacts; (b) may experience incongruence between actual and desired quality and quantity of social contact; (c) may experience losses in the relationships of community; and (d) almost always experience the loss of a single and significant attachment figure.And whilst they may experience all or few of them, they are not static experiences, they are dynamic and any model needs to recognise that these experiences are influenced by time and place.Thus, we would argue that there is a rich research agenda to be pursued in developing our understanding of loneliness and the focus on widowhood enables us to identify key issues that are, perhaps, more obscure in studies of the general population.Our study, based upon the re-analysis of existing interview data, provides new insights into the issue of loneliness in general and loneliness and widowhood more specifically.Older widowed people do spontaneously talk about what loneliness means to them and how they experience it.

I
'm lonely of a night.(Man 16) Night-time the worst time.(Woman 31) Of a night you're lonely.(Woman 12)

I
shut myself in at night and that's it . . . it begins to worry you.(Mrs.M) It's when darkness . . .that you're lonely.(Mrs.C)

Table 1 .
Demographic information distinguishing between sub-groups of lonely and those not spontaneously reporting loneliness

Table 2 .
T-test to show non-significant differences between sub-groups of lonely and those not spontaneously reporting loneliness