Adults with ID/DD live in increasingly small community settings, where the risk of loneliness may be greater. We examined self-reported loneliness among 1,002 individuals with ID/DD from 5 states in relation to community residence size, personal characteristics, social contact, and social climate. One third reported being lonely sometimes and one sixth said they were often lonely, but loneliness was not more common for people living alone or in very small settings. More loneliness was reported by residents of larger community living settings of 7 to 15 people. More social contact and liking where one lives were associated with less loneliness. Social climate variables, such as being afraid at home or in one's local community, were strongly associated with greater loneliness.
In recent decades, American households have decreased in size. By 2003, some 29,431,000 Americans lived alone, constituting 27.6% of all households, whereas in 1970 only 16.2% of households consisted of people living by themselves (Fields, 2004). Likewise, nationally in nonfamily residential settings for people with intellectual disabilities/developmental disabilities (ID/DD) in 2004 (i.e., excluding those who live with parents or other relatives), the average number of residents per setting was 2.8, with about 46% of service users living in households of one to three people (Coucouvanis, Prouty, & Lakin, 2005), whereas in 1992, the average was 5.9 people with ID/DD per residential setting. In 2004, 5 states had 80% or more of service users living in settings with one to three residents (AK, NH, NV, NM, VT), with the average number of residents per setting reaching as low as 1.2 (VT) or 1.3 (AK, NH) in some states.
These trends mean that many service users are living in very small settings or living alone. There are benefits, such as greater choice, associated with very small living settings (Emerson et al., 2001; Lakin et al., in press; Stancliffe, 1997, 2005; Stancliffe, Abery, & Smith, 2000; Stancliffe & Keane, 2000), but there may also be disadvantages. One issue of concern for people with ID/DD is loneliness (Amado, 1993; Chadsey-Rusch, DeStefano, O'Reilly, Gonzalez, & Collet-Klingenberg, 1992; McVilly, Parmenter, Stancliffe, & Burton-Smith, 2006). Loneliness is an important problem in itself (Amado, 1993) and is also associated with significant mental health issues. Among people with intellectual disabilities and in the wider community, loneliness is significantly associated with suicidal ideation (Lunsky, 2004) and depression (Heiman, 2001; Lunsky, 2003).
There is consistent evidence that loneliness is more widespread among people with ID/DD. Margalit (2004) reported that students with intellectual disability reported more loneliness than did their school peers without disability. Likewise, Sheppard-Jones, Prout, and Kleinert (2005) found that adult service users with ID/DD were more likely than the general adult population to be lonely and afraid at home and less likely to have friends.
If loneliness is related to social contact, then adults with intellectual disabilities may be at greater risk of loneliness because they have few significant relationships (Duvdevany & Arar, 2004) and smaller social networks than do adults in the general community (Rosen & Burchard, 1990). Moreover, these networks tend to be dominated by family, disability service staff (especially residential staff), and other service users, such as fellow residents (Robertson et al., 2001). Individuals living in very small settings or alone likely would have less contact with other service users because they have few or no fellow residents and less frequent contact with support staff because many people in such living arrangements have drop-in support from staff for only a few hours daily or even less often (Perry et al., 2006; Stancliffe & Keane, 2000). Under such circumstances, it seems possible that greater loneliness may be associated with living in very small settings or alone.
In this study we examined the relationship between community residence size and loneliness, with a particular focus on very small settings and living alone. Apart from one previous small-scale study (Stancliffe, 2000a), we know of no previous published study looking at the relationship between the size of community living arrangements (and living alone in particular) and loneliness in people with intellectual disabilities. The seminal sociological work, The Lonely Crowd (Reisman, Glazer, & Denney, 1950), showed that, in the general community, being by oneself and encountering loneliness are often very different experiences, so naive assumptions that living alone will necessarily result in loneliness seem ill-advised.
Several investigators have reported data on loneliness and aspects of community living for people with ID/DD. Duvdevany and Arar (2004) and McVilly et al. (2006) found no association between type of living arrangement and loneliness. Perry et al. (2006), Stancliffe (2005), and Stancliffe and Keane (2000) found no difference in loneliness between group home residents (with full-time staff support) and matched individuals living semi-independently (with drop-in staff support), but in the latter two papers, the investigators reported lower social dissatisfaction among semi-independent residents. Social dissatisfaction concerns problems with social relationships and is related to loneliness (Chadsey-Rusch et al., 1992). For example, McVilly et al. (2006) also noted that negative emotional experiences of loneliness can arise from a self-perceived lack of emotional bonding or intimacy with others.
The subjective nature of loneliness means that simply spending time with other people or avoiding being alone is not necessarily a defense against loneliness (Chadsey-Rusch et al., 1992; Stancliffe & Whaite, 1997). For example, Stancliffe and Whaite found that some people experienced quite serious loneliness even though they lived in group homes with 24-hour staffing. Therefore, it is also important to begin to identify factors associated with loneliness that should be considered when planning services or developing interventions to combat loneliness. Consequently, in addition to residence size, we also examined a number of personal characteristics, social contact variables, and social climate factors that may be associated with loneliness.
Findings regarding loneliness and social contact have been mixed. McVilly et al. (2006) reported that there was no relationship between self-reported loneliness and size of one's social network; frequency of contact with friends; or the number, frequency, or duration of activities with friends, thereby reaffirming the subjective nature of loneliness. However, longer duration of time spent with friends was associated with less loneliness (McVilly et al., 2006). On the other hand, Duvdevany and Arar (2004) reported that people with fewer friends and less frequent contact with friends were more lonely.
McVilly et al. (2006) noted that loneliness involves both social and emotional components (i.e., aloneness and sadness) and that loneliness relates to the discrepancy between the person's expectations of relationships and their social experience rather than the absolute number of interactions or size of social networks. Because the experience of loneliness is essentially subjective, it is most appropriate to seek information about loneliness directly from individuals themselves rather than from proxies. Proxies may be able to reliably report relatively objective information about the size of social networks and the frequency of social contact, but the person's experience of loneliness may not be known to a proxy. For example, Chadsey-Rusch et al. (1992) found that staff members' ratings of the person's loneliness were not significantly correlated with the individual's self-reported loneliness. A number of studies have shown that many (but not all) children and adults with intellectual disabilities can reliably report their own experience of loneliness (Chadsey-Rusch et al., 1992; Duvdevany & Arar, 2004; Heiman, 2001; McVilly et al., 2006).
Personal characteristics have usually not been associated with loneliness. McVilly et al. (2006) did report some evidence suggesting that women may be more vulnerable to serious loneliness than men, although gender differences were not evident for their sample as a whole. Chadsey-Rusch et al. (1992) found no difference in self-reported loneliness between adults with mild versus moderate intellectual disabilities. Even so, some personal characteristics may relate to factors affecting loneliness. For example, younger individuals have more family contact (Robertson et al., 2001; Stancliffe, Dew, Gonzalez, & Atkinson, 2001).
In a previous small-scale study, Stancliffe (2000a) found that the loneliness experienced by adults with intellectual disabilities living alone did not differ from that experienced by others living in small-group community settings. However, the sample size in that study was small, so the statistical power to detect differences was limited. In the present study, our intent was to further explore and extend these preliminary findings using a much larger sample of individuals from community living settings in 5 states, excluding those who live with family members or nonfamily live-in caregivers. The specific research questions addressed were: (a) What is the relationship between community residence size and self-reported loneliness? (b) Does the self-reported loneliness experienced by those living in very small groups or alone differ from that experienced by individuals from larger community settings? (c) What personal characteristics, social contact variables, and social climate factors are associated with self-reported loneliness?
Data were collected as part of the National Core Indicators program, a continuing quality assurance effort involving systematic annual evaluation of performance and outcomes of MR/DD service systems in participating states. The National Core Indicators is an ongoing collaboration among participating National Association of State Directors of Developmental Disability Services member state agencies and the Human Services Research Institute.
State selection criteria for the present study were (a) providing regional variation, (b) representing urban and rural states, (c) including states with variations in ethnicity, (d) varying the mix of institutional and community services, and (e) availability of data on residence size for each participant. Selected states were Indiana, Kentucky, Massachusetts, Oklahoma, and Wyoming.
Individual sample selection
State samples were randomly selected within each state's population of adults (age 18+) with ID/DD receiving institutional, community, or home-based services.
Residence size and living alone
We chose to examine the situation of people living in community settings (i.e., settings with 1 to 15 residents with ID/DD) because these settings are the dominant type of residential service, and we did not want to confound our residence size comparisons with differences by residence type (i.e., community vs. institutional); therefore, we excluded people living in institutional settings with 16+ residents. To ensure that the person who lived “alone” really lived alone, we selected individuals who did not live with family or other people who did not have developmental disabilities and without live-in paid caregivers.
Availability of reliable self-report data on loneliness
The item on loneliness was answered by self-report of the person with ID/DD. Substantial numbers of such individuals were unable to respond to this item because of communication difficulties and were, therefore, excluded from the current study. As part of the protocol for interviewing individuals with ID/DD (Section I of the National Core Indicators Consumer Survey; see below for details), interviewers were asked to provide input about the individual's comprehension of the questions and consistency of responses. As a means of ensuring that self-report data being analyzed were as valid and reliable as possible, we selected from among those interviewed who completed Section I and gave codeable responses to at least half the items, including the loneliness item. In addition, those who were rated by the interviewer as having “very little understanding or no comprehension” or as “did not give consistent and valid responses” were excluded.
Overall, the selection criteria listed above yielded a total of 1,002 individuals with ID/DD living in nonfamily community settings with 1 to 15 residents who provided reliable self-report data on loneliness. Because we selected participants who were judged to have given valid and consistent self-report responses, we expected that the sample would consist predominantly of people with mild or moderate intellectual disabilities. This was the case, with only 13% of sample members having severe or profound intellectual disabilities. Participants' personal characteristics are summarized in Table 1.
Data were collected using the National Core Indicators Consumer Survey. which is composed of three sections: the Background section, Section I, and Section II. In this study we used data from the Background section and Section I.
The Background section requests data on the service user's personal characteristics, functioning, diagnoses, health, problem behavior, living arrangements, and services. These data are obtained from individual records and administrators of the residential setting or case managers. In addition to age and gender, we used the following disability characteristics in our analyses.
Level of intellectual disability
Individuals with developmental disabilities but no labeled intellectual disabilities were grouped together with those labeled as having mild intellectual disabilities.
The person exhibits one of three types of behavior at least once per day: self-injurious, destructive, and uncooperative.
Section I of the National Core Indicators Consumer Survey is more subjective; the service user is asked, for example, about his or her satisfaction/ liking of home, feelings of safety at home and in the community, sense of loneliness, and contacts with friends and family. Section I must be completed by self-report through a direct interview with the individual who has ID/DD; proxy responses are not acceptable. In this paper we focused on several items from Section I, described below.
The item about loneliness served as our dependent variable. This question “Do you ever feel lonely?” was scored as 0 = no—not often, 1 = sometimes, 2 = yes—often feels lonely.
Contact with friends and family
The questions asked were “Can you see friends when you want to see them?” and “Can you see your family when you want to?” and were scored: 0 = no, 1 = sometimes, 2 = yes. Respectively for these questions, individuals who reported not having any friends scored 0, as did those who had no family (or the family did not wish to have contact).
Feeling afraid at home or in the neighborhood
These questions, “Are you ever afraid or scared when you are at home?” and “Are you ever afraid or scared when you are out in your neighborhood?” were scored 0 = no, 1 = sometimes, 2 = yes.
This question was “Do you like your home or where you live?” and was scored 0 = no, 1 = in-between, 2 = yes.
To ensure that all interviewers received consistent training, the National Core Indicators Consumer Survey protocol is supported by a training program for interviewers, which includes training manuals, presentation slides, training videos, scripts for scheduling interviews, lists of frequently asked questions, and picture-response formats. The training includes instruction in basic skills for interviewing persons with ID/DD and question-by-question review of the survey tool.
Multiple tests yielded interrater agreement of 92% to 93%, and a single examination of test–retest reliability resulted in 80% agreement (Smith & Ashbaugh, 2001).
The raw data for loneliness are summarized in Table 2. As shown in this table, half of the respondents reported not feeling lonely, whereas one third said they were sometimes lonely, and one sixth were often lonely. This suggests strongly that loneliness is an issue for many service users.
There was a significant difference in loneliness by residence size, Kruskal-Wallis test, χ2(2) = 7.32, p = .026, with larger settings associated with more loneliness. To identify more specifically where the differences in loneliness were, we conducted a series of pairwise post-hoc Mann-Whitney analyses by residence size. The only significant pairwise difference was between settings with 1 to 2 residents and settings with 7 to 15 residents, Mann-Whitney z = −2.69, p = .007, with residents in larger settings reporting significantly more loneliness. The lack of difference between very small (1 to 2 residents) and small (3 to 6 residents) settings indicates that the overall difference in loneliness by residence size was due to more loneliness in larger settings. Importantly, there was no significant difference between settings with 1 resident (not often = 51.8%, sometimes = 34.4%, often = 13.8%) and those with 2 residents (not often = 53.3%, sometimes = 35.2%, often = 11.6%), Mann-Whitney z = −0.48, p = .63.
It is possible that the reported relation between residence size and loneliness was influenced by other variables associated with loneliness that differ by residence size, such as personal characteristics. Differences in personal characteristics by residence size are listed in the final column of Table 1. There was no significant difference by residence size in age, gender, psychiatric diagnosis, autism diagnosis, cerebral palsy diagnosis, presence of seizures/neurological problems, vision, or mobility. Level of intellectual disability differed significantly, with a higher proportion of participants living in settings of 1 to 2 people having mild intellectual disability (77.1%) versus those in larger settings (46.9% in settings with 7 to 15 people with developmental disabilities had mild disability). Presence of challenging behavior (self-injurious, destructive, and/or uncooperative behavior present at least once per day) was significantly related to residence size, with the lowest percentage (5.5%) with challenging behavior among those living in settings of 1 to 2 people, and the highest percentage (12.5%) among those living in settings with 3 to 6 residents with developmental disabilities.
To examine the relation between residence size and loneliness but also control for the differences by residence size in level of intellectual disability and presence of challenging behavior, we ran ordinal regression with level of intellectual disability, challenging behavior, and residence size (as a continuous variable) as the independent variables, and loneliness as the dependent variable. Some previous researchers have reported a relation between gender and loneliness (McVilly et al., 2006) and between age and family contact (Robertson et al., 2001; Stancliffe et al., 2001), so we also included gender and age as independent variables. The results of this analysis (Regression 1) are shown in Table 3.
The overall regression equation was significant, χ2(5, N = 979) = 15.38, p = .009. Level of intellectual disability, age, and gender were not significant predictors of loneliness, whereas residence size was significant and challenging behavior approached significance. However, the amount of variance explained was very small, Nagelkerke pseudo R2 = .018. Overall, consistent with the univariate results, there was a small effect, with people in larger settings experiencing more loneliness, once level of intellectual disability, challenging behavior, age, and gender were controlled statistically.
Next, we examined the relation between loneliness and social contact and social climate. We expected that loneliness would be negatively related to social contact outcomes (more contact associated with less loneliness) and to liking where you live, but positively related to social climate outcomes, such as being afraid at home and afraid in one's neighborhood (being afraid linked to greater loneliness). Spearman correlations among the variables are shown in Table 4.
As expected, greater loneliness was significantly positively correlated with residence size, being afraid at home, and afraid in one's neighborhood; and it was significantly negatively correlated with social contact (contact with friends, contact with family) and liking where one lives. Residence size was not significantly related to the variables shown in Table 4 except for loneliness (larger residence size associated with more loneliness), contact with friends (larger residence size means less contact), amount of support at home (larger settings were more likely to have 24-hour support), and attending work/day program (residents of larger settings were more likely to attend work/day program).
Because of the interrelationships among these variables, bivariate correlations with loneliness may be somewhat misleading, and it is important to evaluate these variables in a multivariate context. Ordinal regression (Regression 2, Table 3) showed that overall these variables accounted for a medium amount of variance (Nagelkerke pseudo R2 = .129), and the overall regression equation was significant, χ2(12, N = 915) = 106.62, p = .000. Being afraid at home, afraid in one's neighborhood, and residence size were significant positive predictors (being more afraid and larger residence size each associated with greater loneliness); liking where you live, contact with family, contact with friends, and age were significant negative predictors (liking where you live more, seeing family and friends more, and older age each linked to less loneliness). Level of intellectual disabilities, challenging behavior, gender, amount of paid support at home, and attending work/day program all were not significant predictors.
Taken together, these findings show that there is a statistically significant but weak relation between residence size and loneliness. More loneliness was associated with larger residence size. Most personal characteristics had a nonsignificant multivariate association with loneliness (challenging behavior approached significance in Regression 1 and age was significant in Regression 2), whereas social climate variables (being afraid at home or in one's neighborhood, liking where you live) were strongly associated with loneliness. Moreover, social connection variables (contact with family or friends) were also associated with loneliness in expected ways. Overall, Regression 2 shows that in a multivariate context where personal characteristics, social climate, and social connections were taken into account, loneliness continued to be related to residence size, with more loneliness among residents from larger settings.
Comparing Regressions 1 and 2 suggests that social climate and social contact were much more important in accounting for variability in loneliness (pseudo R2 change = .111) than were personal characteristics or residence size.
This study examined self-reported loneliness and residence size in a large random sample of adults with mostly mild to moderate intellectual disabilities from non-family community living settings in five states who participated in the National Core Indicators program. Loneliness was an issue for half the sample, with one third reporting being lonely sometimes and one sixth saying they were often lonely. These findings confirm previous reports that loneliness is an important problem among people with ID/DD (Amado, 1993; Chadsey-Rusch et al., 1992; Margalit, 2004; McVilly et al., 2006; Sheppard-Jones et al., 2005).
Residence size had a significant but weak association with loneliness, with greater loneliness evident in larger settings. This was true for both univariate and multivariate analyses in which differences in personal characteristics by residence size were controlled statistically. Individuals living in settings of 1 to 2 people with ID/DD reported less loneliness than their counterparts from settings of 7 to 15 residents, but did not differ significantly with those from settings with 3 to 6 residents. This indicates that higher levels of loneliness in the larger community residences were the main source of the differences. Importantly, loneliness among people living alone did not differ significantly from loneliness in two-person households, showing that living alone was not associated with more loneliness than other very small community living settings. Therefore, concerns about loneliness should not be a barrier to living alone. Taking our findings together with other studies that report advantages of living in very small community settings (usually three or fewer residents), one may conclude that small community settings are to be preferred because of their superior outcomes in areas such as choice, domestic participation, community participation, and loneliness.
There were a number of other important predictors of loneliness. More contact with family and friends and liking where one lives were significantly associated with less loneliness. Being afraid at home or in one's local community was strongly associated with greater loneliness. Experiencing such fear and not having the support or positive companionship to overcome it may be experienced as loneliness. It appears that positive contacts with friends or family reduce loneliness, whereas loneliness is increased by negative experiences at home and in the community where a person may wish for but not have sufficient support and positive companionship. It seems possible that increases in social contact and/ or decreases in being afraid may result in a reduction in loneliness, but the current cross-sectional study did not reveal cause and effect. Evaluation of the effectiveness of these approaches to reducing loneliness awaits intervention studies or longitudinal research.
The amount of paid support the person received at home was not significantly related to loneliness. Rather than the amount of staff presence, it may be what staff do while present that could assist with this problem. For example, staff support for contact with family and friends may be one way to help alleviate loneliness. Once again, this suggestion needs to be evaluated empirically.
There was clear evidence showing that our loneliness measure yielded valid data. Contact with family and friends should reduce the experience of loneliness and, as expected, these variables were significantly negatively related to loneliness in both univariate and multivariate analyses. Experiencing an unpleasant emotion like loneliness likely is associated with negative feelings about one's living environment, so it was to be expected that loneliness would be significantly negatively correlated with liking where one lives.
Being afraid at home was a strong predictor of loneliness. Unfortunately, we have no data on what participants were afraid of at home. One possibility is that they were afraid of other residents. Hewitt, Larson, and Lakin (2000) found that of those who were afraid of someone in a community living setting, the majority were afraid of roommates. Whaite, Stancliffe, and Keane (1999) reported serious problems of resident incompatibility in community group homes. The Donald Beasley Institute (no date) found 22% of all injuries in residential settings were caused by resident-on-resident assault. In these circumstances, the victim likely would feel alienated from fellow residents and may well experience loneliness. This explanation of the relation between loneliness and feeling afraid at home was not evaluated in our study and so will remain speculative until investigated by future researchers. However, Margalit, Tur-Kaspa, and Most (1999) reported that students with learning disabilities who had an identified “enemy” in class felt more lonely than those without such enemies, but similar research has not been conducted regarding adults with ID/DD in community living settings.
These factors may help to explain why there was less loneliness in very small settings (1 to 2 people) than in larger (7 to 15) community settings. Someone living alone will not encounter compatibility problems with roommates, and there are no fellow residents of whom to be fearful. The chances of such problems likely increase as residential settings become larger, both because there are more people with whom to be incompatible and because in larger settings it is more difficult for providers to give careful attention to incompatibility and resident-on-resident assault. Serious incompatibility may well result in feelings of social dissatisfaction, a component of loneliness (Chadsey-Rusch et al., 1992). As noted previously, Stancliffe (2005) and Stancliffe and Keane (2000) reported lower social dissatisfaction among individuals in smaller settings living semi-independently as compared to group home residents.
In future studies researchers should investigate in various settings the salient differences between people who are lonely and those who are not. How do people who live alone ensure sufficient satisfying social contact? What is the problem for people living in group settings that results in loneliness despite the availability of others with whom to interact?
The strengths of the current study include the use of self-report data and the size and representativeness of the sample, which consisted of random samples of service users from 5 states. This study also has limitations. Our data were derived from interviews with service users, so most participants had mild to moderate intellectual disabilities. This approach eliminates problems of relying on proxy data (Chadsey-Rusch et al., 1992; Stancliffe, 2000b), but it does not tell us about loneliness among individuals with more severe disabilities who cannot self-report reliably. Another limitation is that it was unclear how severe a problem loneliness was for participants, because the National Core Indicators does not assess how strongly participants felt lonely.
Our dependent variable came from a single item on loneliness, but it appears to have good validity in that it was related to contact with friends and family in expected ways. Other studies have shown that a single loneliness item is related strongly to the overall score obtained from a multi-item loneliness scale (Chadsey-Rusch et al., 1992).
Loneliness was certainly an important issue for substantial numbers of sample members, but loneliness was not a more serious issue for people with ID/DD living alone or in very small settings than for individuals living in congregate community residences. Indeed, more loneliness was reported in larger community living settings of 7 to 15 people. Given the documented advantages of living in very small community settings, such as greater choice and self-determination, the present findings show that fears about greater loneliness when living alone are ill-founded and should not be used as a reason to prevent someone from living alone if they wish to. Instead, results of this study suggest that concerns about loneliness may be better addressed by supporting the person to live in a safe neighborhood, alone or with compatible roommates, to be free from fear at home, and to have regular contact with friends and family.
This research was funded by the National Institute on Disability and Rehabilitation Research (H133B031116). The National Association of State Directors of Developmental Disabilities Services (NASDDDS) contributed to the planning, instrumentation, and implementation of this study. We are, of course, deeply indebted to representatives from the 5 states in which interviews used in this study were conducted and to the more than 1,000 individuals with disabilities who participated.
Roger J. Stancliffe, PhD (email@example.com), Associate Professor, Disability Studies, Faculty of Health Sciences, The University of Sydney Cumberland Campus, PO Box 170, Lidcombe NSW 1825, Australia. K. Charlie Lakin, PhD, Director, and Robert Doljanac, PhD, Research Associate, University of Minnesota, Research and Training Center on Community Living, 214 Pattee Hall, 150 Pillsbury Dr. SE, Minneapolis, MN 55455. Soo-Yong Byun, PhD, Research Associate, International Study Center 250b Erickson Hall, Michigan State University, East Lansing, MI 48824-1034. Sarah Taub, MMHS, Senior Policy Specialist and Giuseppina Chiri, MA, Senior Research Analyst, Human Services Research Institute, 2336 Massachusetts Ave., Cambridge, MA 02140