Social networks and people with intellectual disabilities : A systematic review

BACKGROUND
Despite the importance of social networks for health and well-being, relatively little is known about the ways in which adults with intellectual disabilities in the U.K. experience their social networks.


METHOD
A systematic review was completed to identify research focused on the social networks of adults with intellectual disabilities. Studies published from 1990 to 2019 were identified. Studies were thematically analysed.


RESULTS
Quantitative, qualitative and mixed methods studies were analysed to identify key factors influencing social networks. Experiences of people with intellectual disabilities identified themes of identity, powerlessness, inclusion, family and support. These themes are discussed with reference to theories of stigma and normalisation.


CONCLUSIONS
Stigma and normalisation can be used to better understand the needs, desires and dreams of people with intellectual disabilities for ordinary relationships, from which they are regularly excluded. Implications for policy and practice are discussed in relation to building and repairing often spoiled identities.

find their relationships are enhanced through much-needed supportive networks (Hall & Kramer, 2009) in diverse situations such as living on locked wards (Fish, 2016); being mothers (Stenfert Kroese et al., 2002); or experiences of older age (McCausland et al., 2018). Social networks can also combat loneliness (Callus, 2017), which itself can have a detrimental effect on health and well-being.
For adults with intellectual disabilities and their families, the importance of social networks for positive health, well-being, social inclusion and tackling disadvantage across the life-course cannot be understated (Bele & Kvalsund, 2016). Despite this, non-disabled people rarely consider the importance of social networks for adults with intellectual disabilities (Dunbar, 2015;Emerson & McVilly, 2004;Hall, 2005).
An examination of relevant theories can aid our understanding of the lack of inclusion adults with intellectual disabilities experience. Goffman (1990, p. 15), for example defined stigma as, 'an attribute that is deeply discrediting', and pointed out that when considering the ways in which the self can be presented in society and the social 'norms' that are created in relation to a sense of 'usual' self, identity, institutions and societies, 'it should be seen that a language of relationships, not attributes, is really needed'. In his consideration of human relationships, broadly from the perspective of symbolic interaction, he showed that the power inherent within stigma and labelling can be dynamic and have significant negative impacts on the lives of those who are stigmatised (Goffman, 1991). One critique of Goffman's work is the need to focus not only on what causes this situation, but also on how it can be overcome, in particular when social networks and relationships which, 'emphasise acceptance of differences' are formed between disabled and non-disabled adults (Bogdan & Taylor, 1987, p. 33). One such principle which is used to try to develop positive relationships with self and others is normalisation, developed by Wolfensberger et al. (1972). They suggested that perceived social deviance identified adults with intellectual disabilities in several specific ways: as sub-human organisms; as threatening; as menacing; as objects of pity; as holy innocents; and as eternal children. They proposed that a focus on supporting adults with intellectual disabilities in 'ordinary' day-to-day places, with 'ordinary' adults was more likely to succeed in reducing stigma and tackling the perceived social deviance of adults with intellectual disabilities. This in turn would lead to the opening up of possibilities for adults with intellectual disabilities to build and maintain wider, reciprocal social networks, with a broader range of people. Critiques of normalisation suggest that its absence from sociological and ideological agendas render it inadequate. Challenges include possible genderbias (Williams & Nind, 1999) and that disability-specific spaces can be used to celebrate disabled identities and resist social discourse around what constitutes ordinary vs stigmatised locations. Chappell (1992) suggested normalisation reflects the views of powerful (nondisabled) professionals and their understanding of typical ways of behaving rather than the views of disempowered and impoverished people with intellectual disabilities. A lack of clarity of concept among policymakers and practitioners and of empirical evidence are important concerns, though could be argued to be less important than understanding inclusion as a human rights issue (Culham & Nind, 2003). Despite their flaws, using the concepts of normalisation and stigma highlights that social networks can therefore either protect against or entrench the effects of stigma and segregation for adults with intellectual disabilities. Using these theories adds to our understanding of the possible reasons for a lack of focus on the social networks of adults with intellectual disabilities.
While previous research has considered the effects of poor social networks on the health and well-being of groups of adults with intellectual disabilities, to our knowledge no systematic literature review and synthesis exists that considers the social networks of adults with intellectual disabilities in the U.K. Institutional closures in the 1980s and 1990s, day services closures in the 2000s and austerity policies since 2008 may have impacted the nature of social networks for adults with intellectual disabilities. The aims of this study are therefore to investigate the nature of the social networks of adults with intellectual disabilities and to analyse the reported experiences of adults with intellectual disabilities in relation to those networks.

| Search strategy and eligibility criteria
This review took place in November 2019. The method used for review aligns with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (Moher et al., 2009). Databases used were Scopus, Web of science (Social Sciences Science citation index), International Bibliography of the Social Sciences (IBSS) and PsycINFO. Inclusion criteria were that articles must be published in peer-reviewed academic journals; written in English; empirical using the responses of adults with intellectual disabilities, or their families or formal care providers; had adults with intellectual disabilities as the focus; were carried out in the U.K.; and were focused on the social networks of adults with intellectual disabilities. No start date was set; the earliest study was published in 1990 and the last in November 2019.
U.K.-only studies were included in order to identify studies which took account of shifts in U.K. policy such as community rather than institutional care, and more recently, policies of austerity unique to the U.K.
Due to the wide timeframe, some early papers did not include specific information regarding their study's ethical procedures. Where findings formed part of a larger study, sometimes ethical review had already been reported in allied publications and so were absent in the reviewed article. Studies were excluded if they did not differentiate between children, young people and adults with intellectual disabilities; included autistic participants or those with mental health conditions without intellectual disabilities as their focus; or were undertaken outside of the U.K. Keywords used in the electronic search were as follows: • (Social networ*) OR (Social support networ*) AND • (Intellectual and Developmental Disabilit*) OR (learning disabilit*) OR (mental retardation) OR (subnorma*) OR (mental handicap) OR (developmental disabilit*) AND • (famil*) OR (family support) OR (informal support) • OR (Identity) • OR (friend*) • OR (relationship*) The asterisk symbol was used to account for possible suffix variation in keywords. Although several of these terms are considered disrespectful currently, they were included so as not to inadvertently exclude relevant studies published when these terms were considered appropriate or those which were conducted in the U.K. but the terminology changed for publishing in international journals based outside the U.K.

| Study selection
In the first stage, 687 results were found of which 78 were duplicates. The remaining 609 records were screened by the first author with 550 excluded by title, abstract or type. In the second stage, the first and second authors independently reviewed the remaining 59 articles against the inclusion and exclusion criteria, which the Centre for Reviews and Dissemination (2009) suggest is good practice. There was an inter-rater agreement of 56/59 (94.9%). Differences were discussed and agreed. Both authors then independently agreed the exclusion of 32 articles. The remaining 27 articles made up the systematic review and synthesis (see Figure 1).

| Quality assessment
The methodological quality of studies was assessed using Sirriyeh et al. (2012) Quality Assessment Tool for Studies with Diverse Designs (QATSDD). The QATSDD had good reliability and validity and included a range of questions designed to judge the quality and rigour of research and therefore make judgements on whether it should be included (Fenton et al., 2015). The thirteen included papers were rated against QATSDD quality criteria on a 4-point scale from 'not at all' (0) to 'complete' (3). Percentage scores were calculated using the actual score and the maximum total score of 42 for qualitative or quantitative studies, and 48 for mixed methods studies. Papers scoring over 75% were rated as being of 'high' quality, those between 50% and 74.9% as of 'good' quality, 25%-49.9% as 'moderate' and below 24.9% as 'poor'. Five papers were scored as being of high quality, 18 papers as good, three as moderate and one as poor.
Seven papers (25.9%) were randomly selected and scored by the second author to determine the inter-rater reliability of the use of QATSDD for the 27 included papers. Overall inter-rater reliability for quality was 97.94%.

| Data analysis
Quantitative studies were analysed for information regarding the reported size and structure of participants' social networks.
Influencing factors were also noted. These are reported in Table 1 On average people were in contact with 2 family members and 2 friends. n = 283 people discharged from institutional settings and a 40% sample followed up after 3 and 6 years. Mixed economy of care was limited so most homes were large private residential homes. Visits from others were infrequent, most people had no friends outside their place of residence. Little choice or opportunity to maximise their potential socially or economically.  Community care: Community-based residential supports compared with residential campuses for people with severe and complex disabilities People living in dispersed housing had larger networks z = 3.30 than those in residential campuses (z not reported).
n = 40 (20 in residential supports and 20 in residential campuses. Family contact was greater for those living in residential campuses, but social network size and composition was greater for those living in dispersed housing. Participants in dispersed housing had people in their networks who were not staff, family or other people with intellectual disabilities. Participants in residential campuses reported none.
Emerson et al. The public sphere was a more likely location for friendship activities to take place than a person's home. Greater levels of adaptive behaviour were linked to greater numbers of friendship activities.
The review of the qualitative aspects of the literature was undertaken using a thematic synthesis approach, as advocated by Thomas and Harden (2008) using Braun and Clarke's (2006) six-step analysis process. This process was also used to consider the process of understanding and coding themes and subthemes. Articles were read repeatedly, preliminary descriptive codes were assigned, patterns and themes were identified across articles, themes were reviewed in line with article content and themes were defined and named.
Coding was based on the research question, 'How do people with intellectual disabilities experience social networks?'. The one article Supported employment Average social network size increased over time from 36 to 42 members.
The most common type of support provided was providing company, invisible support, confiding and support with decision-making. All types of support rose from colleagues significantly after time in an employment setting. Networks were more diverse than is typical for this group.
Forrester-Jones et al.
Community care: Resettlement from long-stay hospitals to small group homes, residential and nursing homes, supported accommodation, and hostels.
Average network size 22. Accommodation type made a difference to the social networks of participants. Those in smaller services were more likely to experience close and companionable relationships than those in residential or nursing homes, but also more likely to experience aspects of relationships that were critical. Relationships also tended to be dense. Grant ( 1993) Community care At baseline: family: 6, friends and neighbours: 1 n = 78 family carers over 2 years. Increased involvement by professionals was linked to policy development and accompanied a decrease in family contacts, usually due to deaths, moves and loss of capacity to provide care. Mothers were most typically the main carer.
Community care Mean size 2 people excluding staff. n = 500. 83% of participants reported a staff member as part of their network, 72% a family member, 54% another person with intellectual disabilities and 30% had members which did not fit into these categories. Staff provided most practical, emotional, informational and close support. Friends with intellectual disabilities provided the most bidirectional reciprocity and were significantly more likely to have been known by the participant for more than 5 years. Networks are affected by the personal characteristics of the person with intellectual disabilities, the types of accommodation, staff ratios, Institutional climate and whether 'active support' is used. Robertson et al. (2007) Person-centred support Not measured Increased social network was associated with having a small network before PCP training was delivered and with living in an area of deprivation. Living closer to family did not link with having more contact with family. Most older people lived in residential care and had more restricted social networks than younger people, and spent less time in enjoyable social interactions. Older people saw more people than the younger cohort, but these were limited to the residential home. Current services do not meet the needs of older people with ID.
Leicestershire Ethical Research Committee.
Quantitative population-based comparison research.
Cross-sectional design so younger and older group not matched Older people have less family and so less family contact. 92% (n = 48) of participants has at least one person in their social network. Mean number was 3.1 (SD = 2.1).
None discussed. Social networks. A range of people filled out the questionnaires. More men than women in the sample. Small community-based homes provided better quality of life than larger new-built campuses, but at greater cost. Family contact was greater for those living in residential campuses, but social network size and composition was greater for those living in dispersed housing.
Not discussed but consent was gained from a range of sources.

Quality and cost correlations
None identified.
People living in smaller homes had greater choice, did more community-based activities and had larger and more diverse social networks. They were also more likely to have their homes vandalised and were considered at greater risk of exploitation.
None given but consent procedures used and research staff from King's College London were involved.

Quality and costs correlations
Organisations were not randomised or stratified. Few agencies involved. No random allocation re residential supports. Defined supported living as described by care providers.
Low levels of friendship activity within supported accommodation, most people with ID were more likely to undertake activities with other people who had ID, the public sphere was a more likely location for friendship activities to take place.

Not discussed Friendship activities and supported accommodation
Not possible to determine the response rate or representativeness of the sample.
Cluster housing does not provide the connected community that is often assumed. People living in cluster housing have less social and friendship activities than those in dispersed housing.

Key findings Ethics? Theories Limitations identified by authors
Average network size 22. Accommodation type made a difference to the social networks of participants. Those in smaller services were more likely to experience close and companionable relationships than those in residential or nursing homes, but also more likely to experience aspects of relationships that were critical. Relationships also tended to be more dense rather than dispersed.
Not stated but detailed consent information.
Social network size and type in relation to accommodation type.
Fewer than half of the participant answered all questions. Participants may overstate their network size. Participants had mild to moderate ID.
Mothers tend to be main carers, as they age, other network members or agencies may not step in to provide support. Relationships with people with ID are reciprocal. Earlier intervention and greater information could support mothers to see transferring the care role more positively. Changes in networks happen independently of changes to the needs of people with ID.
Not discussed Transitions in support networks 2 years is a short time period in which to consider change. People with ID not included.
In village communities people are more satisfied with friendships and relationships than people living in residential, but all are happier with accommodation and day services than with friendships. All satisfaction with friendships is 'far from optimal' which may link to isolation, harassment and institutional constraints. More hours of support relates to more satisfaction with friendships. More people with IDD in a network relates to greater satisfaction with friendships. Structural and process factors affect friendship formation. Proximity to people without IDD does not tackle the issues relating to making friends with people without IDD. '

Key findings Ethics? Theories Limitations identified by authors
People with IDD accept fatalistic roles or nonconforming roles in relation to the stigmatising situations in which they find themselves. Many suggestions for change but lack of optimism for change. Self-advocacy as the way forwards.

No information Social Identity, fatalism
It is unsafe to generalise from small studies such as this, or to extrapolate findings from research with people with IDD to other groups.
Social support does not impact psychological life events. Having people being critical of participants was associated with higher levels of psychological problems.
University of Wolverhampton and Bangor University.
Social support and psychological problems.
Participants struggled with quantitative judgements, with judgements that involved time, carers assistance was required with significant events and with financial questions. Questions about sexual abuse were removed. Quality of life remained the same or improved over time. Improvement in particular was found in greater family contact and in a reduction in staff-reported challenging behaviour.
Multi-centre Research Ethics Committee.
Evaluation of resettlement outcomes.
Improvements within hospital before moving may be due to staff training and awareness. More time is needed to review staff training outcomes.
(Continues) rated as poor and the three rated as moderate were used as secondary rather than primary sources when themes were considered.
Themes were checked by the second author.

| Description of studies
Sixteen of the studies reported on quantitative research, six on qualitative and five were mixed methods studies (see Tables 1 and   2). All but two studies (Emerson, 2004;Emerson & McVilly, 2004) measured network size, but often in different ways. Other studies used social networks to evaluate interventions (Donnelly et al. 1997;Hamilton et al., 2017;Robertson et al., 2007); mental health (Hulbert- Williams et al., 2011); the needs and views of caregivers (Heyman et al., 1997;Prosser & Moss, 1996); or compared people who had intellectual disabilities with people who had physical disability (Lippold & Burns, 2009). A consideration of the factors which influenced social networks is now presented. Firstly, quantitative studies will be examined, including the quantitative aspects of the mixed methods studies. These are followed by a thematic analysis of the qualitative studies, including the qualitative aspects of the mixed methods studies.

| Quantitative studies: key findings
Key findings from an examination of the quantitative data within this review are presented below. These findings relate to network size and characteristics, accommodation and location, and activities.

| Network size and characteristics
Network size in the 11 quantitative studies which measured this specifically, ranged from 0 (Dagnan & Ruddick, 1997)

| Accommodation and location
A key positive influence on social network size and/or satisfaction with social network members was accommodation type and location, which were usually inter-linked, most often as a result of U.K.
Community Care policies which resulted in the closure of long-stay hospitals or other institutional settings. Studies found that larger settings, such as large residential homes, related to larger numbers of contacts with other adults with intellectual disabilities (Cooper, 1998), with few or no friends reported outside of the home (Donnelly et al., 1997). Cluster housing was reported to be related to smaller network size than accommodation which was more dispersed (Emerson, 2004).
Smaller homes based in the community related to networks which were less dense (Dagnan & Ruddick, 1997; and more likely to be considered close and companionable (Forrester-Jones et al., 2006). Grant (1973Grant ( /1993 found that when adults with intellectual disabilities lived with parents, networks were typically small (six family and one friend or neighbour) with mothers providing most care. The structure of networks in this study changed over 2 years as policies were introduced which led to professional workers becoming more involved in the lives of adults with intellectual disabilities and their parent carers. This tended to replace the one friend or neighbour with one professional worker, leaving network size unchanged. Perry et al. (2011) offered an alternative view of the influence of accommodation and location. In their study of adults moving from institutions into community settings, they found that staff training may have been an influencing factor in the higher number of activities undertaken by adults before they left the institutional setting. Hospital staff knew that adults were going to be moving into community settings, and the institution day service closed; suggesting that while important, smaller community-based accommodation could be even further enhanced. Robertson et al. (2001) similarly found that while accommodation was a clear factor in network size and structure, personal characteristics, staff ratio, institutional climate and the use of 'active support' also affected these aspects of the social networks of adults with intellectual disabilities.

| Activities
Over the timeframe of these studies, it was suggested that public spaces where friendship activities were found to be more likely to occur (Emerson & McVilly, 2004). Forrester-Jones et al., (2004) findings concurred; they concluded that being employed was a significant factor in increasing social network size and diversity. Day services, where these were available and accessed, were spaces where adults with intellectual disabilities could engage with others outside their home environment, but networks remained small overall (Bhardwaj et al., 2018;Cooper, 1998). Many studies highlighted that even though policies with their basis in normalisation principles, such as community care and person-centred planning, had had a beneficial impact on the lives and networks of adults with intellectual disabilities, the aim of full inclusion in terms of community presence and relationships that early authors had argued for had yet to be achieved (Cooper, 1998;Donnelly et al., 1997;Robertson et al., 2001Robertson et al., , 2007. They found that although activities had shifted location from institutional to community settings, these were still segregated from 'ordinary' activities enjoyed by non-disabled people. Activities often took place as a group of adults with one staff member all undertaking the same necessary household activity such as shopping, or took place in day service settings. Both these types of activity, while important, provided adults with limited opportunity to create relationships or build networks as they were unlikely to either meet new people, or to meet a diverse range of people (Lippold & Burns, 2009

Summary
An examination of the quantitative data has suggested that social network size and structure can be affected by policy. Policies based on normalisation from the 1990s led to adults with intellectual disabilities being moved from institutional to a range of community settings. This provided opportunities for social networks to be developed and maintained. Differences in social networks were found between accommodation type and location, and activity type.
Networks remained smaller than that of the wider non-disabled population and were typically comprised of other people with intellectual disabilities, staff and family. Using thematic analysis to consider the experiences of adults with intellectual disabilities from their own perspective, an aspiration for ordinariness, of 'normality' was found throughout the qualitative literature, yet opportunities for the development of broader social networks were often thwarted (Hamilton et al., 2017;Jahoda et al., 1990). This was the main theme which emerged in the analysis of the qualitative data. The following subthemes were also identified: networks and identity; networks and power; networks and inclusion; networks and family; and networks and support services (see Figure 2). These are presented in turn below.

| Networks and identity
The ways in which people with intellectual disabilities made meaning of their experiences were directly affected by how they understood themselves in terms of their relationships with others (Heyman et al., 1997;Head et al., 2018). This could be confusing (Dagnan & Ruddick, 1997;Head et al., 2018) and work both positively and negatively for individuals. The dichotomy that adults with intellectual disabilities experienced was knowing that even while striving to be ordinary, they were being treated as other than ordinary, which could also negatively impact one's sense of self and one's relationships (The Positive relationships and networks were found to be essential for the formation of a positive sense of identity Heyman et al., 1997;Jahoda et al., 1990).

| Networks and power
Thwarted opportunities also related to power, especially the powerlessness experienced by adults with intellectual disabilities.
When relationships with staff worked well, this could outweigh People also experienced little power over ending of relationships, especially when decisions about where they live, with whom, and whether and when they move were made at short notice, often without information or their input. Often the importance of maintaining relationships was thwarted by staff who were nevertheless ideally placed to support that maintenance (Jahoda et al., 1990;McConkey et al., 2003). When relationships, which often helped people to feel safe, were so frequently ended (especially relationships with staff), people were left feeling bereaved as they knew and were known by so few people (Hamilton et al., 2017;Head et al., 2018).
People reported finding a range of systems difficult or impossible to negotiate (Power & Bartlett, 2019). Many people did not have access to digital systems and many could not read or write, so needed support from people in their networks without intellectual disabilities in order to access and negotiate systems which are vital for well-being (such as welfare benefits, housing and finances). Policies introduced to promote adults with intellectual disabilities as employers of their own staff, in principle appeared to afford the opportunity to move into more powerful roles. Yet people still experienced stigma and powerlessness even in these relationships (Hamilton et al., 2017), thereby thwarting possible opportunities for social network development.

| Networks and inclusion
The composition of people's social networks was far from inclusive of non-disabled people, with most people with intellectual disabilities having very few close relationships of any kind and of those, most being people with intellectual disabilities.
People who had moved from institutional life to living in communities were often, 'deeply disappointed' (Jahoda et al., 1990, p. 138) with their lack of integration into 'ordinary' community life and their lack of expected friendships with non-disabled people. Sadly, lower expectations were indicative of people's higher satisfaction with their move Head et al., 2018). Lack of community acceptance appeared to occur irrespective of a person's previous or current accommodation (Murphy et al., 2017).
Lack of meaningful daytime activity was also an issue (Murphy et al., 2017) and when staff attempted to involve people with intellectual disabilities in their communities through activities, these F I G U R E 2 Diagram of the theme and subthemes indicated by qualitative data  (Jahoda et al., 1990, p. 139).
This stigma was also seen in 'ordinary' activities and spaces.
Adults with intellectual disabilities and the people who supported them were aware of often erroneous reasons for resistance from employers for example (Hamilton et al., 2017;Murphy et al. 2017), even when people performed above the standard of non-disabled colleagues (Hamilton et al., 2017).  2004;Head et al., 2018;Heyman et al., 1997;Jahoda et al., 1990;McConkey et al., 2003).

| Networks and family
Most studies highlighted that a key relationship in many people's lives was the relationship they had with members of their family. As with other relationships in the literature, family relationships (and their lack) could have an empowering or disempowering effect in people's understandings of themselves and their network development.
When people with intellectual disabilities received positive regard from people in their social networks, particularly family, (sometimes described as, 'best friend' (Bhardwaj et al., 2018, p262)), it boosted their self-esteem and self-image. Being close to family both figuratively and literally could improve the way people with intellectual disabilities felt about themselves (Head et al., 2018;Heyman et al., 1997;Jahoda et al., 1990). Conversely, physical and emotional distance or abuse could have devastating effects (Power & Bartlett, 2019;Grant, 1973Grant, /1993. While families could be a source of valuable and valuing support, Murphy et al. (2017, p. 964) warned professionals not to make assumptions about family settings equating to positive solutions, as when family were perceived as overprotective or abusive (Power & Bartlett, 2019) people experienced limitations in the rest of their relationships and network (Grant, 1973(Grant, /1993Bhardwaj et al., 2018), thwarting opportunities to develop supportive networks.
Just as authors suggested that policies based in normalisation, such as personalisation and community care, could be perceived as mere rhetoric when not backed by resources, families too were fearful of their family members falling through the gaps in any policy delivery (Prosser & Moss, 1996;Grant, 1973Grant, /1993, so strived to protect them from the 'irresponsible' 'drive towards increased independence and choice' at any cost (Hamilton et al., 2017, p. 295).
In situations where family were absent, the importance of being connected socially, including being connected to charitable services, became essential for well-being, relationships, social justice and rights (Hamilton et al. 2017;The Money, Friends, & Making Ends Meet Research Group c/o Liz Tilly, 2012). These adults with intellectual disabilities were often reliant on a Social Worker or Care Manager to make decisions on their behalf, making their struggle harder than for those who had family support (Hamilton et al., 2017;Power & Bartlett, 2019).
Both people with intellectual disabilities and their (often olderaged) parents could feel trapped in situations where their own expectations around coping were unrealistically high yet support for improvement was desperately needed but either not trusted or not forthcoming, leading people with intellectual disabilities and their wider family members to crisis point (Prosser & Moss, 1996). Policies which promoted the far earlier intervention by professionals, respected the reciprocal nature of care, and had a psychological understanding of the functioning of families and social networks, could better enable the development of wider social network structures and community participation (Grant, 1973(Grant, /1993).

| Networks and support services
Where budgets. As Michael explained, 'they're trying to stretch people as far as they can without spending any money to give people support' (Hamilton et al., 2017, p. 294).
Proactive support and reciprocal social network relationships were required to prevent problems occurring in the first instance, with for example reading letters, paying essential bills, shopping, ac-

| DISCUSS ION
When considering the experiences of adults with intellectual disabilities highlighted by this literature review it is possible to illuminate the results further by synthesising Goffman's theory of stigmatisation and Wolfensberger's concept of normalisation.
In this way, it is possible to begin to consider not only the ways in which people with intellectual disabilities are perceived by societies more widely, but also the effects this could have on people's sense of identity, their social networks and their experience of relationships.
The construction of an ordinary identity through social networks can be viewed and understood from the perspective of symbolic interactionism. Using this approach, various aspects of self, such as self-worth, are linked to the ability to understand and play valued social roles, in socially accepted ways (Goffman, 1990a(Goffman, , 1990b. Crucially, the understanding of self and identity is created in one's social interactions with others and is therefore significantly impacted by one's social networks, or their lack. Yet this study has shown that the opportunity structures Forrester-Jones and Grant suggested in their study in 1997 do not appear to have developed into wider interpersonal relationships which authors argued enhance the identities of people with intellectual disabilities (van Asselt-Goverts et al., 2013;Sullivan et al., 2016). The desire for an ordinary identity can be at least partially explained by an awareness that one both is and is not perceived by others as ordinary (Wolfensberger et al., 1972); this review has highlighted that little has changed in terms of opportunities for this ordinariness to be reflected in social networks . There were clear links to Goffman's consideration of stigma as a tool to identify socially constructed deviance and to ensure identity creation is embedded in the understandings of both stigmatisers and those being stigmatised (), which has been highlighted throughout the timespan of this literature review (Hamilton et al., 2017;Jahoda et al. 1990). Wolfensberger et al. (1972) argued that this understanding required people with intellectual disabilities to be immersed in, rather than isolated from communities. The creation of identity then can be understood as a joint, symbiotic activity; identity becomes both a verb and a noun in the context of social networks. Yet adults with intellectual disabilities reported deep disappointment with a lack of community acceptance and engagement which did not appear to change over time (Jahoda et al., 1990;Murphy et al., 2017).
Without the support of social networks, extreme marginalisation could further affect people's poor sense of self, which could be devastating (Goffman, 1990a(Goffman, , 1990bHamilton et al. 2017;Hulbert-Williams et al., 2011;Wolfensberger et al., 1972). Yet typical service approaches to integration, which are assumed to lead to wider networks including friendships and relationships with nondisabled people and therefore address the power imbalances of societies, often did not deliver positive change, as Wolfensberger et al. (1972) warned and both Robertson et al. (2001) and The Money Friends and Making Ends Meet Research Group c/o Liz Tilly (2012) re-iterated. In addition to experiencing powerlessness in their relationships and networks, stigma is also experienced by people with intellectual disabilities in their desire for ordinary patterns of behaviour such as being employed (Goffman, 1990a(Goffman, , 1990b. A range of policies and programmes highlight the opportunities provided by employment (Parkin et al., 2020) which could build a less-stigmatised self-identity; but this too remained a dream rather than a reality for many people with intellectual disabilities (Hamilton et al., 2017).
In terms of inclusion, Goffman (1990) explained that often people are aware of the stigmatised characteristic they embody and can feel shame, as Jahoda et al.'s (1990) study showed. Stigma could extend to the places and spaces individuals with intellectual disabilities inhabit and despite critiques of normalisation (Donnelly et al., 1997) and stigma, this did not appear to change over time Heyman et al., 1997). Just as actors in a drama must maintain an image of self which they feel will be accepted by their peers and society more widely, so people with intellectual disabilities could feel they must hide their true 'selves' by managing their spoiled identities and adapting their behaviours (Emerson & McVilly, 2004).
Stigma therefore created a negative dynamic relationship which also thwarted social network creation; this too did not appear to change over time (Head et al., 2018;Jahoda et al., 1990).
Expectation was an important consideration for people with intellectual disabilities and their families but was only part of the story in a stigmatising society where identity, networks and relationships can be spoiled (Goffman, 1990a(Goffman, , 1990bHeyman et al., 1997). This can apply in all relationships and can be particularly complex and nuanced in families, who can often experience the effects of courtesy stigma from others. Family members of people with intellectual disabilities can also internalise the stigma they experience, blaming themselves for their family member's condition and situation (often known as affiliate stigma). As this review has shown, unrealistic expectations of and by families could lead to parents taking on the caring 'burden' and being unwilling to share that for many reasons (Prosser & Moss, 1996;Grant 1973Grant /1993McConkey et al., 2003).
Often the opportunity for family members to create and maintain what Goffman (1991) referred to as 'ordinary' social identities as individuals rather than the 'virtual' stigmatised identities they inhabit was not available, especially in settings where segregation and isolation were the norm, creating additional pressure for everyone involved. This too did not appear to change over time (Bhardwaj et al., 2018;Emerson, 2004;Jahoda et al.,1990;Power & Bartlett, 2019).
Until the stigmatising perceptions of people with intellectual disabilities were challenged and addressed, with a strong emphasis on advocacy, Wolfensberger et al. (1972) argued that they would persist unchecked leaving people with intellectual disabilities to exist as one-dimensional, devalued citizens with devalued social roles. Wolfensberger et al. (1972) went on to argue that support could lead to the benefits inherent in being part of an inclusive society which encourages social support networks, relationships and support systems between people with and without intellectual disabilities. In a time of austerity policies leading to significant cuts to services, less statutory support is available for people with intellectual disabilities, and the voluntary sector is being relied upon to provide the most basic of support (NIHR SSCR, 2020 (Hamilton et al., 2017, p. 302).
This literature review and synthesis has shown that the voices and experiences of people with intellectual disabilities, parent carers and staff can illuminate the ways in which the day-to-day effects of policy were experienced in terms of social networks. It is this holistic approach which is needed in order that policy can be informed, created, funded and implemented more effectively, benefitting not only people with intellectual disabilities, but communities and societies also, so that an 'ordinary' life with supportive social networks can be grasped and embraced (Cooper, 1998;Emerson & McVilly, 2004;Sango & Forrester-Jones, 2018).

| Limitations
There

CO N FLI C T O F I NTE R E S T
No conflicts of interest are declared.

DATA AVA I L A B I L I T Y S TAT E M E N T
Data sharing not applicable -no new data generated. Data sharing is not applicable to this article as no new data were created or analyzed in this study.