Interpersonal Relationships and Subjective Well-being among Older Adults in Sheltered Housing

The aim of this article is to examine how experiences with interpersonal relationships contribute to older adults’ well-being in the residential context of sheltered housing. We draw on data collected from sixteen in-depth interviews with older adults living in sheltered housing in a small town in northern Netherlands. Our participants experienced the interaction with their children as of primary importance among their interpersonal relationships, while interactions with other residents were rather superficial. Their children offered emotional support as well as instrumental support and were found to play essential roles in our participants’ wellbeing. Moreover, participants expressed that the social and physical activities organized by the residential care-facility offered them the opportunity to remain physically and mentally active. The help received from housekeepers and caregivers was found to be another important element of interpersonal relationships and so was the reciprocal nature of support exchanged with other sheltered housing residents. We conclude that the benefits of interpersonal relationships in sheltered housing should be considered when designing policy for the well-being of older adults ageing in

nterpersonal relationships are important for older adults' well-being (Bok, 2010;Ritchey, Ritchey, & Dietz, 2001;Baker, Cahalin, Gerst, & Burr, 2005;Diener, Oishi, & Lucas, 2003). Older adults, however, often face a reduction in the number of interpersonal relationships they have (Dupuis-Blanchard, Neufeld, & Strang, 2009;Cornwell, 2011), and the number of social activities they attend (Marcum, 2013), probably as a result of individual health problems and losing friends and family members. As a result of their deteriorating health, older adults are likely to become restricted in terms of out-of-house mobility, and are thus expected to rely more on interpersonal relationships in their direct living environment. Therefore, in order to understand the importance of interpersonal relationships for well-being, it is important to study older adults' interpersonal relationships in relation to their living environment.
In many western societies, the Netherlands among them, the living environment for older people has changed, namely the share of older adults ageing in place has increased. 'Ageing in place' means to remain living in the current living environment, typically the long-time home, throughout the ageing process, with some level of independence (Fausset, Kelly, Rogers, & Fisk, 2011;Cutchin, 2003;Fernández-Carro & Evandrou, 2014;Wiles, Leibing, Guberman, Reeve, & Allen, 2012). Many governments support older adults to continue living in their own home, and discourage ageing in care-facilities. Recently, the Dutch government took measures to further limit the availability of care-facilities to people who need intensive care and assistance (Homan, 2012). As a response to this type of ageing in place policy, intermediate living arrangements for housing and care have emerged. These arrangements combine the opportunity to live independently and have professional care and assistance available (Van Bilsen, Hamers, Groot, & Spreeuwenberg, 2008). For example: Integrated Service Areas (Singelenberg, Stolarz, & McCall, 2014), retirement communities (McHugh & Larson-Keagy, 2005), co-housing communities (Meijering & Lager, 2014) and sheltered housing (Nocon & Pleace, 1999).
Out of all the living arrangements which combine housing and care, we choose to look at sheltered housing because recent policy changes have a large impact on people who live in sheltered housing. The deinstitutionalization measures that are currently implemented will imply that more people with severe health problems will live in sheltered housing, I RASP -Research on Ageing and Social Policy,3(1) 17 whereas more healthy older adults will have to age in place. In the Dutch context, sheltered housing refers to houses built close to, or as part of, a residential care-facility. Sheltered housing offers the possibility to live independently, which is typically desired by older adults (Lawton, Silverstein, & Bengtson, 1994;Silverstein, Chen, & Keller, 1996;Fausset et al., 2011), while the security, services and assistance offered by the carefacility are also available (Van Bilsen et al., 2008;Croucher, Hicks, & Jackson, 2006;Percival, 2001). Sheltered housing also offers the opportunity to participate in activities organized by the care-facility together with other older adults, such as bingo, playing cards or aerobics. These activities are typically positively experienced by sheltered housing residents (Percival, 2001).
Studying the perspectives of participants on the role of interpersonal relationships on well-being could help us to understand the value of intermediate housing arrangements such as sheltered housing for interpersonal relationships at older ages. Moreover, gaining knowledge about the valuable elements of sheltered housing for older adults' well-being could help to make those elements available to older adults ageing in place. Several studies have looked at the advantages of housing arrangements such as sheltered housing for older adults (Van Bilsen et al., 2008;Field, Walker, Hancock, & Orrell, 2005;Percival, 2000;Taylor & Neill, 2009), but none of them investigated how they actually contribute to well-being. The research question addressed in this paper therefore is: How do older adults perceive interpersonal relationships in relation to their well-being in sheltered housing?
Data were collected through sixteen in-depth interviews with older adults living in sheltered housing in a small town in northern Netherlands.

Residential Environment in Old Age in The Netherlands
When looking at the residential environment in which older adults live, four settings can be distinguished. Older adults can live at home (ageing in place), in sheltered housing, in a residential care home, or in a nursing home. In the Netherlands older adults receive a lot of formal care in comparison to other countries. Family plays only a subsidiary role in care provision at older ages, in contrast to Southern European countries (Suanet, Broese van Groenou, & Van Tilburg, 2012). The lower level of care provision by family members is often thought to be related to stronger welfare provisions and the lack of legal obligations with respect to providing care to older family members (Saraceno & Keck, 2010). In the Netherlands, large investments in residential facilities have led to high coverage of formal care. From the beginning of the 1990s onwards policies aiming to increase ageing in place were implemented, which was translated into providing formal care to older adults in their own homes. While other countries stimulated the development of informal care, the Netherlands aimed at offering more home-based care, and made a transition from public care provision to more private provision of formal care (Pavolini & Ranci, 2008;Suanet et al., 2012). More recently, the Dutch government took measures which limit the possibility to move into a care-facility to people who need intensive care and assistance (Homan, 2012). Recently, the demand for informal care became larger in a system where the access to residential facilities is limited and care provision is privatized.
Looking at the figures, the amount of home-based care provided has indeed been increasing substantially, while the amount of older adults living in residential care facilities has been decreasing. The share of older adults (aged 65 and over) that lives in a residential care facility has decreased from eight to four per cent between 1995 and 2013. The share of people living in a residential facility from age 80 onwards has even further decreased, from 25 to 13 per cent in the same time period (Statistics Netherlands, 2015). The most recent figures about residential care situation of older adults in the Netherlands show that, among the people who do not live in institutions, 71 per cent lives in a normal dwelling without receiving care, 19 per cent receives home-based care (including sheltered housing), and the remaining ten per cent lives in some form of sheltered housing, but does not receive formal care (Lijzenga & Van der Waals, 2014).
In the Netherlands, 4.2 per cent of the total housing stock is defined as sheltered housing, and around 18 per cent of the population aged 65 and over lives in such dwellings (Lijzenga & Van der Waals, 2014). In earlier days, older adults who desired to live in sheltered housing could apply for such a dwelling regardless of their physical and mental health. Nowadays, moving into sheltered housing has become restricted to older adults with more severe health problems because the number of places available in institutions is limited. As a result, the share of older adults who need care or assistance has increased and is considerably higher among sheltered housing residents compared to older adults who age in place.

Interpersonal Relationships and Subjective Well-being
Drawing on the work of Buckley and McCarthy (2009), Cooney and colleagues (2014), Cohen and Wills (1985) and Westaway, Seager, Rheeder, and Van Zyl (2005), we conceptualise interpersonal relationships as all social meetings and interactions in which people are involved. This includes, for instance, talking with other people, engageing in shared activities, and visiting places together. The main functions of interpersonal relationships are: emotional support, instrumental support, appraisal support, informational support and social companionship (Cohen & Wills, 1985;Westaway et al., 2005). Emotional support refers to love and empathy. Instrumental support involves practical assistance. Appraisal support constitutes comments from other people that allow a person to reflect on or evaluate what s/he does. Informational support is received through advice or information that can help solve problems. Social companionship involves spending time with others.
Social Production Functions Theory (SPF-theory) is a useful framework for studying the five main functions of interpersonal relationships in relation to well-being, because it allows us to study how activities contribute to subjective well-being (SWB). SWB is a cognitive and emotional evaluation of well-being (Diener, Suh, Lucas, & Smith, 1999) and refers to an individual's assessment of his or her own life situation (Ormel, Lindenberg, Steverink, & Verbrugge, 1999). In SPF-theory, assumptions are formulated about how individuals produce SWB, consisting of physical and social wellbeing, by optimizing achievement of instrumental goals (Table 1; Ormel et al., 1999). Physical well-being is attained by the instrumental goals stimulation and comfort and social well-being is built from the instrumental goals status, behavioural confirmation and affection. Whether or not the instrumental goals are realized depends on the abilities (resources) and absence of abilities (constraints) an individual faces that help or prevent the development of SWB (Ormel et al., 1999;Ormel, Lindenberg, Steverink, & Vonkorff, 1997;Nieboer, Lindenberg, Boomsma, & Bruggen, 2005). In the 20 Herbers & Meijering -Older Adults in Sheltered Housing remainder of the paper we will use the terms well-being and subjective wellbeing interchangeably, referring to SWB as conceptualized in SPF-theory. Table 1 The hierarchy of Social Production Functions Theory (Source: Ormel et al., 1999)  Within our framework, in which we connect interpersonal relationships with SWB, interpersonal relationships are the activities that individuals use to achieve the instrumental goals as defined in SPF-theory, which results in achievement of universal goals and then SWB. Emotional support, for example, is likely to lead to affection as an instrumental goal because people receive love and empathy through emotional support. Similarly, negative experiences with the functions of interpersonal relationships could limit the achievement of goals and consequently relate negatively to SWB. When looking at the current evidence of the relation between interpersonal relationships and SWB, several aspects of interpersonal relationships have been found to contribute positively to SWB. These are: having many interpersonal relationships in general (Baldassare, Rosenfield, & Rook, 1984;Hilleras, Aguero-Torres & Winblad, 2001;Street & Burge, 2012), interpersonal relationships with friends and neighbours (social companionship, emotional support) (Lennartsson, 1999;Helliwell & Putnam, 2004), participation in activities that involve social companionship, or meeting or interacting with other people (Litwin & Shiovitz-Ezra, 2011), engagement in productive activities (e.g. volunteering and providing care), social activities (e.g. attending meetings) and physical activities (e.g. doing sports) (Baker et al., 2005).
Evidence of the link between interpersonal relationships and SWB is not uniform. Children have been found to be important for the SWB of older adults in several studies (Margolis & Myrskylä, 2011;Hansen & Slagsvold, 2012), while others did not find a significant impact of children on SWB (Glenn & McLanahan, 1981;Kohler, Behrman, & Skytthe, 2005). Furthermore, while receiving support can have positive well-being effects, its effectiveness depends on the individual appropriateness of the support (Rowe & Kahn, 1997), and several studies found that receiving instrumental support relates negatively to well-being, probably because people do not like to give up privacy and independence (Connidis, 2010;Reinhardt, Boerner, & Horowitz, 2006). The partly ambiguous evidence from quantitative research on the link between interpersonal relationships and SWB calls for an indepth approach, which further explores and explains this link.

Interpersonal Relationships in Sheltered Housing
In sheltered housing, other residents are potentially available as social contacts, which could facilitate the development of interpersonal relationships. It has been found that older adults in sheltered housing typically have more contact with their neighbours than people who age in place (Field, Walker, & Orrell, 2002;Dupuis-Blanchard et al., 2009). These findings suggest that the function of social companionship is fulfilled relatively easily in sheltered housing. Moreover, sheltered housing residents have the opportunity to participate in many social activities and physical activities facilitated by the care organization (Baker et al., 2005). Through these activities the functions of social companionship, emotional support, and informational support could be fulfilled. Furthermore, residents have instrumental support available, in the form of professional care and assistance, such as with (instrumental) activities of daily living ((I)ADL).
In sheltered housing, family care is usually arranged according to the dual-specialization model: staff is responsible for ADL and IADL, while family members offer emotional support (Litwak, 1985;Gaugler, Anderson, Zarit, & Pearlin, 2004). The focus of family members on emotional support could be an advantage, because emotionally meaningful relationships become more important at older ages (Carstensen, 1995;Carstensen, Fung, & Charles, 2003). A disadvantage of the dual-specialization model is that the role of family in providing care can be ill-defined, which may cause conflicts between staff and family and may consequently have a negative impact on the resident (Gaugler et al., 2004;Schwarz & Vogel, 1990). Sheltered housing ideally provides a flexible amount of assistance with ADL and IADL, based on individual needs. Additionally, family can be involved in providing instrumental support when this is desired by sheltered housing residents (Croucher et al., 2006).
Living in, or moving to, sheltered housing could also have negative consequences for older adults' experiences with interpersonal relationships. In contrast to people who age in place, living in sheltered housing involves a residential move. A move to sheltered housing may disrupt interpersonal relationships and their functions, which the individual had in his or her previous living environment (Heijdam & Hillebrand, 2014). In addition, Percival (2000) found that gossip is a prominent feature of relationships in sheltered housing, and that this gossip is often perceived as a negative aspect of social companionship. Although Percival (2000) pointed out that gossip may help to 'safeguard the reputation' (p. 324) of the individual, he also found that because of gossip, older adults sometimes choose to limit or withdraw from interaction with other people in sheltered housing. Furthermore, a negative aspect of interpersonal relationships in sheltered housing is the fact that residents often meet others who are frailer than themselves. This appears to result in people perceiving themselves as being frailer than they actually are (Golant, 1999;Percival, 2001).

Study Setting
The study took place in sheltered housing adjacent to a residential carefacility in a small town in northern Netherlands. Inhabitants of the sheltered housing accommodation all lived in apartments with their own entrance. All residents have an alarm-system in their homes and can use communal facilities such as the garden and the spacious common rooms inside the carefacility. Moreover, inhabitants can use social and care services of the residential care-facility. The care-facility, as well as the sheltered houses, is located beside the largest public garden of the town and many sheltered housing inhabitants have a view on the channel next to the facility. The facility is located at one kilometre distance from the city centre.

Research Method and Participants
We used a qualitative research approach because we were interested in participants' experiences with interpersonal relationships in relation to their well-being. Compared to quantitative research approaches, a qualitative approach allows the researcher to get a more detailed and deeper understanding of people's perceptions and experiences with respect to the studied subject (Hennink, Hutter & Bailey, 2010).
Participants (Table 2) were recruited through gatekeepers from the residential care facility. After conversations with the facility manager, the research project was announced to all inhabitants and employees of the residential facility through announcements on a digital screen in the common area and a notification in the newsletter, which was distributed among all 46 homes of the sheltered housing residence. In consultation with the unit manager of the facility, it was decided to exclude four households from participating in the research because of cognitive impairments. Following the announcements and newsletter, a letter was sent to 42 sheltered housing residents in which the project was described and in which a visit by the first author, with the purpose of recruiting participants, was announced. As a third step, all inhabitants of sheltered housing were visited face-to-face by the first author, who briefly explained the project and asked whether they were willing to participate in the study. Appointments for in-depth interviews were made during the visits with those inhabitants who agreed to participate. Out of the 42 requests that were addressed through the letter, sixteen agreed to participate in the study, twenty-two refused, and we were not able to get in touch with four inhabitants. The main reasons people gave for not participating were that they did not want to talk about their interpersonal relationships or that they were not willing to share aspects of their private life.
In our sample, three participants said they live without any support from the care facility, five participants indicated they received assistance from the facility in most of their daily activities (showering, preparing dinner, housekeeping), while the other eight participants received just some assistance from the care-facility (e.g. housekeeping).

Data Collection & Operationalization
The data used for this paper were collected in 2011 through tape-recorded in-depth interviews, using a semi-structured interview guide which was drawn up based on SPF-theory. The interview-guide was pilot-tested after which mainly probing questions were added. All participants chose to be interviewed in their own homes, which provided a feeling of familiarity and safety for the participants, and insight into the spatial context of houses within sheltered housing for the interviewer. After a brief introduction to the project and interview, questions addressed the importance of interpersonal relationships for older adults and the perceptions and experiences related to these relationships. Interviews lasted between 50 minutes and 2 hours.
During the interview, the interviewer tried to gain insight into the importance of interpersonal relationships for well-being by asking about the value of having contact with other people, perceptions of their relationship with others, whether the participants felt that they had enough social interaction with others, and whether they were satisfied with the nature of the relationships. Through probing, the interviewer attempted to specify the contribution to instrumental goals of different aspects of interpersonal relationships. The interviewer probed whether participants felt loved, respected and supported by others and whether joining social activities was important in order to remain active. Moreover, the interviewer probed about interpersonal relationships that may have been important besides the 26 Herbers & Meijering -Older Adults in Sheltered Housing relationships participants talked about automatically. Some examples are: siblings, housekeepers and, contacts through work, church or sports. Therewith participants were encouraged to talk about all aspects of interpersonal relationships they experienced as important.

Ethical Considerations
Informed consent was established before every interview. All participants agreed to a tape-recording of the interview. Confidentiality was guaranteed and data were anonymised in order to protect the identity of participants. In order to make the participants feel comfortable, appointments were made in their familiar home-environment, at times that suited them well. At the time of interviews, the first author was a graduate student, and he introduced himself as such to the participants. He addressed the participants as experts on the social relationships in their lives, whilst the interviewer himself could not draw on this from his own experiences, because of his younger age. This enabled the participants to share their stories with regard to interpersonal relationships as freely as possible. Participants were informed that they were free to stop the interview whenever they wanted to do so, but no one did so. However, several participants found it difficult to talk about certain topics, such as their deceased partner, or arguments with their children.

Data Analysis
The interview data were transcribed verbatim and the transcripts were complemented with notes on the interview process. Data analysis was done in MAXQDA (software program). A codebook with both codes derived from the theoretical framework (deductive codes), as well as inductive codes, was developed before and during data-analysis (Hennink et al., 2010). During and before data collection, literature was used to define categories of interest, while later on, the collected data were used to develop concepts that identify the link between interpersonal relationships and subjective wellbeing. Thus data analysis was an iterative process. The data were first coded with the help of open ended coding techniques, which helped to identify explanatory concepts in the data (Goulding, 2005). The second author coded a selection of the data, to confirm the concepts. In the process of data-analysis, we followed the strategies of description, comparison, categorisation, conceptualization and theory development as defined by Hennink et al. (2010).

Findings
The analysis revealed four aspects of interpersonal relationships that were most important for the participants' well-being: (i) interaction with other people, (ii) social and physical activities, (iii) receiving care or assistance and (iv) providing care or assistance. In the subsequent sections we discuss how the participants' experiences with those relationships relate to their SWB.

Interaction with Other People
Interaction with other sheltered housing residents was typically experienced as superficial and participants kept other residents at a distance. Hardly any participant reported visiting each other at home. Linda, for instance, repeatedly mentioned her need for privacy: Linda: "Yeah, a woman came to live here and she asked me: "Do you want to come over and drink coffee?", but I'll try to prevent that! (…) Some of them always need to be together, well I don't need that at all." (woman, 91 years) Similar to Linda, several other participants found it important to "have their own life" and "let others not interfere too much". Clearly, many participants limited the amount of interaction with other residents, most likely in order to secure their privacy. This confirms an earlier study which suggests that residents of high-density units put high value on privacy in their own apartment (Dupuis-Blanchard et al., 2009).
Gossiping came up as a reason to avoid the amount of interaction with other sheltered housing residents: Other participants illustrate the feeling of being afraid of gossip with phrases such as: "It is not good to talk bad about others!" and, "You have to be careful with gossiping". Many participants did not want to gossip because they were afraid to say things others might disapprove of. This confirms that gossip is a key-feature of interaction in sheltered housing (Percival, 2001). Moreover, Frank told us that he limited the contact with other people because he did not want others give reason to gossip about him. Frank seemed afraid that gossiping would have a negative impact on his status and thereby also on his well-being.
To conclude, the study participants did not seek to develop friendships with fellow-residents. Our analysis revealed that participants consciously limit the amount of interaction with fellow residents, something that was also found in other studies (Dupuis-Blanchard et al., 2009). More specifically, we found that our participants were afraid that social interaction with other residents would contribute negatively to their well-being.
In contrast to interaction with sheltered housing residents, participants expressed that interaction with their children was very important for them. John explained the difference between interaction with children and interaction with other people: John: You see, with my daughters I am able to talk about more intimate topics, compared to other people. With them, contact is more superficial. [...] through work and sport I have contact and those contacts remain good, but with one the relation is better than with the other, that's normal. [...] But most intensive, yeah, that is with family. (man, 71 years) These findings are consistent with previous studies that showed that family members are the most important providers of support later in life (Bengtson 2001;Van Tilburg, 1995). The value of interaction with children was illustrated by expressions such as: "They mean everything to me", "I am happy when my children are around" and "I am able to discuss everything with them". Through interaction with their children participants received emotional support, which contributed to their well-being because the instrumental goal affection was fulfilled.
How contact with children is experienced seems for some participants to be related to the move to sheltered housing. Lisa, for instance, explained that she was not able to "Just go to my children whenever I want" anymore, because the geographical distance between her and her children increased after the move to sheltered housing. Annie told us that she has more contact with her daughter since she had moved closer to her. She said they undertake activities together more often and her daughter provides instrumental support by assisting with shopping and cleaning.
Even though none of the participants complained about the number of times they see their children, it seems that Lisa, for one, would enjoy more interaction with her child; however, said she accepted the situation as it is: Lisa: I see them (her son and his wife) regularly (…). But yes, it is not so easy, they work and have their own children, their children also need to be pleased, because they are working as well and my son needs to babysit every now and then. So they can't be with me all the time, and I don't need that, it's fine like this. (woman, 84 years) It could be that participants do not experience limited contact as negative because at older ages people are better able to adjust their needs (Hansen & Slagsvold, 2012) and tend to report positively about close contacts such as those with children (Fingerman, Hay, & Birditt, 2004;Erber, 2010;Mariske, Franks, & Mast, 2001). Furthermore, some participants experienced limited face-to-face contact with children but indicated that making telephone calls was valuable as well.
Some of our participants did not have much contact with their children, and they in particular appeared to need more close relationships, and sought friendship relations with fellow residents. Since most of the older adults did not desire a lot of interaction with other sheltered housing residents, those participants who did seek contact experienced the 'closed' attitude of fellow residents as negative. Moreover, some participants have lost (some of) the social relationships they had in their previous environment. Two women, Sara (80) and Lisa (84), expressed feeling 'excluded' from the wider society. Compared to their previous, independent living situation, sheltered housing 30 Herbers & Meijering -Older Adults in Sheltered Housing seems to limit the interaction with people from outside the facility, which is likely to be exacerbated by the physical limitations these women have.

Physical and Social Activities
Instead of interacting socially with other people in the home environment, many older adults met in the common room of the facility or during activities organized by the residential facility such as card games, bingo, aerobics, or playing shuffleboard: Other participants also experience benefits from the activities: "I remain fresh because of the activities", "I will stay active by going there", and "It is a nice change in the daily routine". Annie expressed that the move to sheltered housing gave her new opportunities to interact with other people, which she did not have when she lived independently in the community. Social companionship seems to be the major function of the activities, and the activities helped participants to maintain well-being through stimulation.
Whether participants experienced the joint activities as positive seemed to depend on the number of years they lived in sheltered housing. Older adults who had recently moved into sheltered housing reported difficulties in developing social relationships with fellow sheltered housing residents. Other studies also found difficulties with integrating among older adults moving into sheltered housing residences (Croucher et al., 2006;Stacey-Konnert & Pynoos, 1992). An older couple who had been living in the sheltered housing for two months mentioned how they saw the activities and meetings in the common room as a good occasion to develop relations, but they experienced difficulties to get in touch with new people when they went to the common room to drink coffee. Other sheltered housing residents seemed to limit the amount time spent in the common room, and the people who are most often present in the common room are mainly those living permanently in the residential care facility, who suffer from more severe health problems.
Sheltered housing residents do join activities, but the opportunities to develop relationships are fewer during activities than during meetings in the common room. As a result, sheltered housing residents who did feel the need to develop social relationships with fellow-residents experienced that these were difficult to establish, which they said influenced their well-being negatively.

Receiving Care or Assistance
Several participants experienced it as positive that personnel from the facility is available when necessary, for example Lisa: The availability of care and assistance from the nearby care-facility seems to offer comfort to the sheltered housing residents. Knowing that someone is available in case of an emergency seemed to offer feelings of security, which is consistent with previous research that found perceived support to be more important for well-being than received support (Reinhardt et al., 2006). When participants talked about the help they received, they experienced the presence of housekeepers and caregivers as pleasant. Positive experiences with housekeepers were often reported: "She is a really nice person", "We always have a good time when she is here", "She is important to me". In contrast, Susan explained that the high pressure on caregivers prevented them from having talks and spending more time with their clients. Our participants generally had the same housekeeper for a long period of time, which made it worthwhile for them to invest in a social relationship.

Herbers & Meijering -Older Adults in Sheltered Housing
Some participants have to give up their independence and privacy mainly because of their physical impairments. They found it difficult to cope with tight schedules of care professionals, and that they were dependent on others for starting the day. For example, Ed (86) explained that he feels he lost his privacy now that he and his wife are dependent on caregivers for getting out of bed and having a shower. Moreover, one woman reported negative experiences with caregivers. It happened that caregivers walked in and out of her house without ringing the bell or saying 'hello', which she experienced as a violation of her privacy. In these cases, receiving care thus seems to relate negatively to well-being through the instrumental goal of comfort.
In the context of sheltered housing, children can offer instrumental support if it is desired by the parent and when children are able to fulfil these tasks. Assistance offered by children is experienced as positive by several participants. While children provide assistance, the time spent together also provides the opportunity to talk. Thus, along with emotional support, children provide instrumental support and social companionship in the context of sheltered housing. Anne is a nice example: her son comes weekly to help with gardening and shopping, which allows them to spend time together while Anne also receives instrumental support. This construction allows older adults to live independently, to decide for themselves who provides care and assistance and to what extent.

Providing Care or Assistance
Several participants stressed the importance of the reciprocal nature of interpersonal relationships. Two of our male participants offered practical assistance to fellow-residents by working in the garden. Among the participants, we found people with limitations who offered support as well, which is consistent with a previous study which has shown that people with limitations are often still able to provide support, but at a different level than people without limitations (Boerner & Reinhardt, 2003). Martin was physically limited but still found it important to help his neighbour with practical matters such as replacing light bulbs. Ellen gave an example of the emotional support she provided during activities: Ellen: "During the activity they put her (another resident from the facility) next to me and during the entire hour the woman was holding my hand.
[…] You don't need to say anything. She also enjoys it, and that gives me a good feeling. To do something for them, just sit next to them. Holding hands might be important to people who are not able to communicate." (woman, 75 years) Ellen's example shows how support can also be given and found in small things in daily life. Other participants felt they were not able to provide support because of their own limitations. If we take a close look at the help that is offered we found that it contributes to caregiver well-being as well (see also: Van Willigen, 2000;Thomas, 2009;Morrow-Howell, Hinterlong, Rozario, & Tang, 2003). We found that providing help to others led to behavioural confirmation for several participants: they told us how it gave them a good feeling to do something for someone else. One of the participants told us that he wanted to be respected and appreciated for what he did for other people: Frank: "If they behave normally towards me, as they're supposed to? Yeah, […] then they are allowed to wake me in the middle of the night.
[…] If the treat me right, I will treat them right. That's how simple it is" […] "I don't want all the old people to break their legs. They are allowed to wake me up in the middle of the night, but I want to be treated with respect. If the neighbour is complaining and commenting on everything, then I won't bother anymore!" (man, 80 years, 7 children) Frank seemed to find it more important to show excellence and unique skills, instead of just 'doing the good thing'. For him it was crucial to get appreciation and respect for the support he provided, and thus, his support contributed to behavioural confirmation.
The context of sheltered housing offers opportunities to volunteer and provide support to others. Our participants lived close to other sheltered housing residents, which seemed to facilitate the process of both receiving and providing support, and providing support in particular contributed to the participants' well-being.

Discussion
This article examined how older adults perceive their interpersonal relationships in relation to their subjective well-being in the context of sheltered housing. Our findings show that sheltered housing residents' interpersonal relationships are important for their SWB. Interaction with other sheltered housing residents is often superficial, while interaction with children was experienced as meaningful: children as a source of both emotional and practical support. Thus, social interaction with children plays a key role in the well-being of older sheltered housing residents. Participation in activities contributes to well-being because it enables older adults to remain physically and mentally active. Moreover, having care and assistance available when necessary is an aspect of interpersonal relationships that positively relates to well-being. Furthermore, reciprocity in receiving and offering support and assistance is experienced as a positive aspect of interpersonal relationships.
From the data we induced how different functions of interpersonal relationships relate to instrumental goals of SWB. We conclude that interpersonal relationships in sheltered housing contribute to well-being through all instrumental goals, and based on our findings we propose a new model for conceptualizing the relation between interpersonal relationships and subjective well-being (Table 3). This model can help to deepen our understanding of the complex relations between interpersonal relationships and SWB in sheltered housing. We derived that the main function of interaction with children is emotional support and contributes to SWB through affection. The function of participation in social or physical activities is mainly social companionship. Participation in activities seems to stimulate and activate the sheltered housing residents, which makes them feel better and thus enhances their well-being. Instrumental support is the main function of receiving care or assistance. Receiving care or assistance contributes to sheltered housing residents' SWB through comfort. Offering support or assistance seems to lead to appraisal support. Doing something good for others appears to be the main reason for offering support and thereby contributes to SWB through behavioural confirmation. Despite the positive experiences with interpersonal relationships, we should be aware of some negative aspects as well. These negative aspects relate first and foremost to contrasting expectations and needs that participants have. On the one hand, many participants seemed to limit the interaction with other sheltered housing residents in order to prevent these relationships having a negative influence on their well-being. Relations with other sheltered housing residents typically remain superficial and several older adults are afraid that too much interference will limit their privacy and influence way others talk about them. On the other hand, several participants indicated that they would like to have more and closer social relationships with other residents, especially those who have limited interaction with family members, and who recently moved to sheltered housing. Moreover, not all older adults report positively about the care they receivesome experienced problems with privacy, while others did not like to depend on professional help. Also, some participants have lost (some of) the social relationships they had in their previous environment, which can be seen as a negative aspect of interpersonal relationships in sheltered housing.
With respect to the discussion on the relative advantages of ageing in place versus housing schemes such as sheltered housing, we believe that several positive experiences with interpersonal relationships relate to the context of sheltered housing. First, the amount of care received is flexible in sheltered housing, so children can also contribute to well-being of older adults by offering primarily emotional support (see also the dual-36 Herbers & Meijering -Older Adults in Sheltered Housing specialization model: Gaugler et al., 2004;Litwak, 1985), but also practical support and social companionship. Second, it seems that having the opportunity to participate in social or physical activities organized by the care-facility is a positive aspect of interpersonal relationships in sheltered housing. Older adults who wish to attend activities have them available close by. Third, the feeling of safety and security that comes with the availability of professional assistance is a positive aspect.
The most important limitation of this study relates to the study participants. There were only few men and couples among the participants, and therefore we are unable to comment on the differences with respect to experiences with interpersonal relationships between those subgroups. Gender differences might exist because men and women have different types of social needs. Men, for instance, more often resist participation in social groups (Milligan, Payn, Bingley, & Cockshott, 2015), and participation in social activities might therefore be less important to their well-being. Furthermore, the study lacks a comparison between sheltered housing residents and people who age in place. Such a comparison could help to better understand the advantages and disadvantages associated with both residential contexts. Another limitation relates to the fact that older adults who have negative experiences with interpersonal relationships might be underrepresented in the sample since they might be not willing to talk about their interpersonal relationships, and therefore refused to participate in the study. Future studies could focus on contrasting the importance of interpersonal relationships for SWB between different subgroups: young-old and old-old, men and women, healthy and disabled older adults, and older adults who receive care at home, or live in some form of assisted living.
Housing and care policy should consider the advantages of interpersonal relationships in housing schemes such as sheltered housing. Current policy in the Netherlands, as in other European countries, is directed towards further de-institutionalization, mainly as a way to reduce costs. From 2013 onwards, the Dutch government introduced additional measures to provide care to people without allowing them to live in a long term care facility -socalled extramural care. This type of care is offered at home, in the community (Homan, 2012). As a result, people suffering from more severe functional and cognitive impairments will inhabit dwellings in housing schemes such as sheltered housing. At the same time, older adults who would ideally like to live in sheltered housing, in order to have the reassurance of care in case of emergency, will likely remain living in the community for a longer period. Even though it is often thought that older adults wish to receive care and assistance in their own house for as long as possible (Fausset et al., 2011;Gitlin, 2003;Tang & Lee, 2011), some people might become isolated or feel lonely when ageing in place (Sixsmith & Sixsmith, 2008). We conclude that intermediate housing schemes such as sheltered housing, or at least a residential context with comparable opportunities for interpersonal relationships, are essential and valuable for specific groups of older adults.
We recommend that both service providers and policy makers consider the importance of interpersonal relations for older adults' well-being. They could do so by looking at the quality and availability of interpersonal relationships with family members, friends, or neighbours. These could become a criterion for entering housing schemes such as sheltered housing, given the opportunities for interpersonal relationships in such environments. Moreover, several of our participants experience the time restrictions of care providers as negative in relation to their well-being. Policy makers could give care providers and housekeepers the opportunity to spend more time with older adults who wish some more interpersonal contact. In general, we believe that a transition from public provision of care to more informal care provision, always requires a careful examination of the individual situation.