Participation and Companions for Socially Inclusive Community Activities by U.S. Adults With Intellectual and Developmental Disabilities

Abstract

We investigated socially inclusive participation in mainstream community groups and religious services by U.S. adults with intellectual and developmental disabilities using weighted secondary analyses of 2018–2019 National Core Indicators data. Overall, 34.4% participated in community groups and 42.4% in religious services. Some 45.0% had an unmet desire for community-group participation, whereas most (75.0%) attended a religious service as often as preferred. The type of companion varied by living arrangements and age group. Attending community groups and religious services were each strongly associated with better friendship outcomes but were not related to loneliness. The large unmet demand for community-group participation reveals a major gap. The friendship outcomes underline the benefits of socially inclusive community participation.

Community Participation

Adults with intellectual and developmental disabilities (IDD) participate less in the community than the non-disabled population (Amado et al., 2013; Verdonschot et al., 2009b). Community dwellers experience more community participation than institution residents (Dusseljee et al., 2011; Verdonschot et al., 2009a, 2009b). Even for community dwellers, community participation and inclusion are often limited, especially for those living in provider-owned or operated community homes, such as group homes (Friedman, 2019).

Community participation ranges from mere presence in community settings available to everyone (e.g., shopping centers) to social inclusion involving relationships with community members without disabilities. Many more adults with IDD go to anonymous community locations (e.g., entertainment venues) and encounter strangers, than attend membership-based settings, such as community groups or religious services (Bigby & Wiesel, 2019; Carter et al., 2015).

The current study focuses on participation of adults with IDD in two types of membership-based, mainstream settings: community groups and religious services. Adults typically attend these organizations freely and can go with chosen companions or alone. These settings differ from other socially inclusive adult environments such as community-based employment, because showing up to work is not optional, work activities are mandatory, you cannot invite companions to come to work with you, and work does not have the primary goal of creating community around shared interests.

Community groups are social organizations that people attend voluntarily in their free time. They usually involve some common purpose, shared activities, and typically encompass hobby, craft, or sports groups; faith-based social groups; volunteer organizations; book groups; neighborhood groups; or social clubs (Sandstrom & Alper, 2019). While at the group, people participate voluntarily and can take part in the activities they prefer. Religious services are held at places of worship (e.g., church, synagogue), involve a formal public ceremony with religious meaning, and prescribed rules for participation. Religious service participants typically share similar values and beliefs.

Specific research on participation in community groups by people with IDD shows they are much less likely to take part (Amado et al., 2013; Verdonschot et al., 2009a), whereas Carter et al. (2015) reported a similar but slightly lower level of religious service participation compared to the U.S. general community. Moreover, among people with IDD, participation varies by personal characteristics and demographics. For example, Carter et al. (2015) found attendance at religious services by people with IDD was significantly related to gender, race, level of intellectual disability (ID), Down syndrome diagnosis, autism spectrum disorder (ASD) diagnosis, presence of behavioral challenges, mobility level, communication style, and residence type. Regarding community participation, Amado et al.’s (2013) review of social inclusion and community participation reported that those who are younger, have fewer support needs, live in the community, and live in smaller community settings participate more. However, Amado et al.’s findings were about community participation broadly, rather than participation in community groups specifically.

Person-Centered Community Participation

Just because a person with IDD attends a community group or religious service does not mean that they want to go, but research has mostly overlooked this issue (Amado et al., 2013; Verdonschot et al., 2009a). It is important to investigate how person-centered participation was experienced by comparing each person’s attendance status with their reported wishes about taking part. When desired and current participation match, the situation can be seen as person centered. Where there is a mismatch, this could represent an unmet need (person wants to participate but has no opportunity) or undesired participation (individual attends a group but does not want to).

Social Benefits: Friendship and Loneliness

People with IDD typically have few friends, especially beyond family, staff, and service users, and have high rates of loneliness (Amado et al., 2013; Stancliffe et al., 2007, 2010). Socially-inclusive community participation is a positive in itself, but it is also of interest to know whether presumed social benefits of participation, such as relationships/friendships and less loneliness, are realized.

A relationship involves repeated interaction with a specific individual. Membership-based settings offer better opportunities for developing relationships because the same individuals attend regularly, and participants have shared interests, activities, values, or beliefs (Craig & Bigby, 2015; Sandstrom & Alper, 2019). Research has repeatedly shown that proximity and similarity are two important contextual factors facilitating friendship (Wiener & Schneider, 2002). Proximity is important because friendships are more likely given ongoing contact—you do not become friends with people you have not met or encounter fleetingly. Similarity (of interests, experiences, or attitudes) helps form the basis for an ongoing relationship. These friendship-related factors are more available in membership-based organizations, than in anonymous community settings.

Loneliness is subjective. Being alone is not necessarily experienced as loneliness, just as people can feel lonely when with others (Emerson et al., 2021). Loneliness involves an emotional response (sadness/distress) due to a discrepancy between the person’s expectations of relationships and the perceived quality of their social experience (Hawkley & Cacioppo, 2010). Stancliffe et al. (2007, 2010) found that more contact with friends and with family was associated with less loneliness among adults with IDD. Therefore, we investigated whether having family or friends as companions when attending community groups or religious services was beneficial in minimizing loneliness, compared to attending with other companions or alone, and with nonattendance.

Companions

The identity of companions at community groups and religious services has received little attention apart from Carter et al.’s (2015) study of religious participation, where family and friends were the most common companions, with disability staff the next most common. Verdonschot et al.’s (2009a) review noted that staff and service users were frequent companions for community participation overall but was silent on companions in specific settings, such as community groups.

Participation in community groups and in religious services by people with IDD appears to offer a readily available means of addressing low rates of socially inclusive community participation, having fewer friends, and feeling lonely. Therefore, it is important to understand the details of such community participation to help identify approaches to improve the participation rate and social benefits for those who wish to attend.

Research Questions

We aim to better understand factors related to participation by adults with IDD who use IDD services in socially inclusive community groups and religious services, regarding who has access, whom they go with, and whether participation reflects individual preferences. We also analyze benefits of such participation for friendship and loneliness. We examine five research questions separately for participation as a member of community groups or attendance at religious services.

1. What percentage of adults with IDD who use IDD services participate in socially inclusive community groups or religious services?

2. What personal characteristics and living arrangements are associated with participation?

3. Controlling for personal characteristics and living arrangements, what is the relationship between current and individually desired participation?

4. Controlling for personal characteristics and living arrangements, with whom do people with IDD attend? Is companion type related to living arrangements and age-group?

5. Controlling for personal characteristics and living arrangements, what is the relationship between friendship, loneliness, and attendance at community groups or religious services?

Method

Data Source

The National Core Indicators (NCI) program is a collaboration of the National Association of State Directors of Developmental Disabilities Services (NASDDDS), the Human Services Research Institute (HSRI), and state developmental disability agencies in participating U.S. states (Bradley & Moseley, 2007). We completed secondary analyses of 2018–2019 National Core Indicators In-Person Survey (NCI-IPS) data (NASDDDS & HSRI, 2018), an annual, cross-sectional survey described next. This survey predated the COVID-19 pandemic.

Participants

NCI-IPS participants are adults aged 18 years or older receiving case management and at least one other service from their state IDD system. Based on a weighted analysis (see Analysis section for details of weighting), there were 21,099 participants with valid data for community group participation, and 21,035 for going to a religious service. Participants came from 37 U.S. states (AL, AR, AZ, CO, CT, DE, FL, GA, HI, IN, KS, KY, ME, MI, MN, MO, NC, NE, NH, NJ, NV, NY, OH, OK, OR, PA, RI, SC, SD, TN, TX, UT, VA, VT, WA, WI, WY). Participating states were each asked to recruit a random sample of at least 400 adult service users. Weighted state samples (community groups’ sample) ranged from 73 (WY) to 1,599 (PA) and averaged 570.3. Average age (community groups’ sample) was 42.54 years (SD = 15.77, range 18–107). To avoid repetition, details of all other participant characteristics and demographic information are presented in Table 1 in the Results section.

Table 1.

Weighted Descriptive Data on Participation by Personal Characteristics and Living Arrangements

Variable/level	Community group			Religious service
	No	Yes	Total	No	Yes	Total
	%	%	n	%	%	n
Personal characteristics
Gender
Male	66.7	33.3	12,154	59.3	40.7	12,095
Female	64.0	36.0	8,806	55.3	44.7	8,804
Other	75.0	25.0	16	75.0	25.0	16
Age-group
18–24	64.7	35.3	2,605	61.1	38.9	2,603
25–34	62.9	37.1	5,519	57.1	42.9	5,509
35–44	64.9	35.1	4,000	53.9	46.1	3,972
45–54	65.7	34.3	3,442	57.0	43.0	3,397
55–64	68.3	31.7	3,233	58.6	41.4	3,250
65+	70.6	29.4	2,112	60.8	39.2	2,122
Race/ethnicity
Non-Hispanic White	66.2	33.8	14,920	60.3	39.7	14,899
Non-Hispanic Black	62.0	38.0	3,268	45.3	54.7	3,251
Hispanic	69.7	30.3	1,330	54.4	45.6	1,310
Non-Hispanic Other	65.3	34.7	883	62.1	37.9	874
Level of ID
Mild	63.6	36.4	7,361	56.4	43.6	7,357
Moderate	62.1	37.9	5,373	52.0	48.0	5,367
Severe	72.1	27.9	2,323	61.0	39.0	2,303
Profound	77.9	22.1	1,459	67.1	32.9	1,445
Behavior challenges
No	63.2	36.8	13,355	54.6	45.4	13,314
Yes	70.1	29.9	6,567	63.4	36.6	6,552
Down syndrome
No	66.1	33.9	17,126	57.9	42.1	17,064
Yes	52.2	47.8	1,782	44.7	55.3	1,766
ASD
No	64.9	35.1	15,996	56.3	43.7	15,967
Yes	67.4	32.6	4,207	63.0	37.0	4,188
Participation self-reported
No	70.7	29.3	9,631	60.8	39.2	9,138
Yes	60.9	39.1	11,227	54.2	45.8	11,460
Communication
Other	75.2	24.8	4,298	66.8	33.2	4,221
Speech	62.9	37.1	16,544	55.1	44.9	16,555
Mobility
Mobile w/out aids	63.1	36.9	16,011	55.4	44.6	15,979
Mobile with aids	70.0	30.0	2.972	62.3	37.7	2,946
Not mobile	77.9	22.1	1,919	67.7	32.3	1,916
Living arrangements
Own home	65.6	34.4	3,746	60.4	39.6	3,735
Family home	60.5	39.5	7,992	52.1	47.9	7,946
Foster/host home	62.0	38.0	1,392	52.6	47.4	1,400
Group 2–3	72.7	27.3	2,132	65.3	34.7	2,120
Group 4–6	69.8	30.2	3,702	62.3	37.7	3,688
Group 7–15	70.8	29.2	1,216	62.6	37.4	1,234
Other	62.8	37.2	172	58.0	42.0	174
Institution	83.5	16.5	448	57.4	42.6	439

Note. Row percentages shown for each dependent variable. ID = intellectual disability; ASD = autism spectrum disorder.

Instrument: National Core Indicators In-Person Survey

The NCI-IPS is an in-person interview with the adult service user and usually a support person/proxy respondent who knows the person well. The survey has three main sections. Background Information is usually answered from case management and service provider records and deals with the service user’s personal characteristics, functional capacities, health issues, diagnoses, residential status, service usage, employment, and daily activities. Section I may only be answered by the person with IDD via an in-person interview, so Section I data are missing for participants unable to respond to these interview questions. Section I covers the person’s views about their home, employment/day activities, feeling safe, friends and family, community participation and leisure, rights and privacy, and satisfaction with services/supports. Section II is answered by the person with IDD where possible, or by family, advocate, or staff (proxy) if the person is unwilling or unable. Section II asks about community inclusion, choices, rights, access to needed services/supports, health and wellness, and self-directed supports.

The Background Section provided data on age, gender, race/ethnicity, level of ID, behavior challenges, Down syndrome diagnosis, ASD status, communication, mobility level, living arrangements, and state. Data on friendship and loneliness came from Section I. Data on each type of community participation, whom the person went with, if they would like to go more, and who answered these items came from Section II.

Inclusive Community Participation Items

Community group participation and religious-service attendance items are from the Community Inclusion section of NCI-IPS Section II. Instructions to interviewers specify that the survey is intended to assess whether “the person participates in integrated community activities (activities that include people with and without disabilities)” (NASDDDS & HSRI, 2018, p. 49).

Community Groups.

The question states “Are you a part of any community groups? (This includes church groups, book clubs, knitting groups or any other formal or informal community group in an inclusive setting.)” (NASDDDS & HSRI, 2018, p. 50). Interviewers are trained to use this wording but can also provide other individually relevant examples of community groups. The next question, “Who did you participate in community groups with?” (p. 51), has six response options: (a) Alone, (b) Friends, (c) Family, (d) Housemates or coworkers, (e) Staff, and (f) Others not listed. Respondents may choose all that apply. The items about companions are only asked of individuals who answered yes to being part of a community group; all others are coded as not applicable regarding companions and the data treated as missing. The final question, asked of all participants, is “Do you want to be a part of more groups in your community?” with response options of yes or no (p. 49). There is no item about the frequency of community group participation.

Religious Services.

The NCI-IPS asks “How many times did you go out to a religious service or spiritual practice in the past month? (Examples: church, synagogue, study, or other place of worship)” (NASDDDS & HSRI, 2018, p. 53), with four mutually exclusive response options of (a) Did not go, (b) 1–2 times, (c) 3–4 times, and (d) 5 or more times. For most analyses, we recoded religious-service attendance variable into a yes/no variable. The question about companions has a similar format to this question for community groups. The final question asks, “Would you like to go to religious services or spiritual practices more, less, or the same amount as now?” (p. 54).

Self-Reporting.

For the items on participation in community groups and in religious services, the NCI-IPS interviewer records who the respondent was (person with IDD, family/friend, staff, other) in each case. We recoded these data into a binary yes/no self-reporting variable, with self-reports coded as yes (= 1) and reports from others as no (= 0).

Personal Characteristics and Environmental Variables

Personal Characteristics.

The NCI-IPS includes items on date of birth and gender. We recoded the race/ethnicity item into a four-category variable: non-Hispanic White, non-Hispanic Black, Hispanic, and non-Hispanic Other. Level of ID is recorded as mild, moderate, severe, or profound. We coded as missing data individuals who did not have ID, whose level of ID was unspecified or whose ID diagnosis was unknown. The behavior challenges, Down syndrome, and ASD items each ask whether this condition is noted in the person’s record. The communication item asks about the person’s preferred means of communication. We recoded these responses into two categories—speech and other. The mobility item has three response options, (a) mobile without aids, (b) mobile with aids (e.g., uses wheelchair independently), or (c) not mobile (always needs assistance).

Living Arrangements.

The NCI-IPS asks about 14 different mutually exclusive types of living arrangement. We recoded these into the eight categories listed in Table 1.

Friendship and Loneliness

We analyzed three items about friendship—(a) Do you have friends you like to talk to or do things with?, (b) Do you have a best friend, or someone you are really close to?, and (c) Can you see your friends when you want to? There was a single item about loneliness (Do you ever feel lonely?).

Analyses

Because raw state NCI-IPS sample size is not necessarily proportional to the state’s total population of adult recipients of IDD services, raw data analyses can result in less accurate prevalence estimates. We therefore used weighted analysis. Based on each state’s number of valid surveys and total survey-eligible population, NCI-IPS 2018–2019 weights ensure that each state’s weighted contribution is proportional to its adult service population (National Core Indicators, n.d.). Weights for each state varied from 0.235 (Nevada) to 3.466 (New York). Data for each participant from a particular state is weighted by the weight for that state. SPSS uses simulated replication to give individual cases different weights for statistical analysis. If the weighted number of cases exceeds the raw sample size, significance tests are inflated; if smaller, they are deflated. In the present study, the size of the total raw sample and the total weighted sample was identical, so significance tests were unaffected by weighting.

We addressed each research question in turn. The dependent variables were categorical, so we used chi square, logistic regression, and multinomial logistic regression as appropriate. We report the results of 24 separate analyses (note that Tables 5 and 6 report different aspects of the results from the same 12 analyses), so we used Bonferroni adjustment of alpha, resulting in a revised alpha of .05/24 = .0021. In all analyses, we analyzed complete cases only.

Table 5.

Weighted Raw Percentage of Participants With Each Type of Companion by Living Arrangement and Results of Logistic Regressions

Community activity Companions	Living arrangement								Total	Logistic regressions
	Own home	Family homea	Foster/Host home	Group 2–3	Group 4–6	Group 7–15	Other	Institution		Overall model χ2(26) = Nagelkerke R2	p Living arrangements
Community-group participation
Alone	21.5***	8.2a	8.0	7.3	7.0	3.3	19.5	3.2	10.2	391.56*** R2=.168	p <.001***
Friends	47.0	42.3a	42.4	28.2***	25.7***	37.2	47.5	14.5	38.7	269.45*** R2=.076	p<.001***
Family	26.2***	67.4a	31.1***	18.5***	15.5***	9.1***	10.0***	9.7***	41.5	1,246.31*** R2=.316	p<.001***
Housemates/co-workers	11.1***	3.1a	16.3***	28.6***	41.6***	52.9***	14.6**	51.6***	16.2	883.75*** R2=.294	p<.001***
Staff	47.9***	27.0a	53.9***	69.5***	71.5***	69.4***	35.0	85.2***	45.9	800.09*** R2=.211	p<.001***
Others	5.2	4.7a	7.3	5.8	4.3	4.1	5.0	6.5	5.1	26.96 R2=.018	p=.428
Religious-service attendance
Alone	19.3***	3.3a	5.3	5.6	5.8**	4.5	22.7***	3.2	6.9	526.89*** R2=.219	p<.001***
Friends	24.2***	10.3a	15.4	15.1	11.3	19.8***	25.0	8.9	14.0	256.05*** R2=.076	p<.001***
Family	44.5***	91.3a	47.4***	25.4***	29.1***	23.2***	25.6***	8.9***	59.3	2,633.05*** R2=.483	p<.001***
Housemates/co-workers	8.5***	0.5a	17.2***	23.9***	36.5***	41.2***	9.1***	59.0***	14.0	1,379.85*** R2=.372	p<.001***
Staff	32.5***	8.9a	49.5***	66.6***	66.1***	64.4***	37.2***	94.2 ***	34.7	2,200.18*** R2=.427	p<.001***
Others	2.4	1.3a	3.3	1.4	2.2	2.3	7.0	3.2	1.9	49.04 R2=.047	p=.022

Note. Row percentages shown, indicating participants who answered yes to that type of companion. Only individuals who were included in the relevant logistic regression analysis were included when calculating the percentages shown in each row of Table 5. Respondents may choose multiple types of companions, so percentages for each living arrangement add to exceed 100.

Independent variables in logistic regressions: Gender, Age group, Race/ethnicity. Level of intellectual disability, Behavior challenges, Down syndrome, Autism spectrum disorder, Participation self-reported, Communicates by speech, Mobility, Living arrangements.

^aReference category.

^***p < .001.

^**p < .0021. In the living arrangements columns, significance asterisks denote the results of simple contrasts with the reference category within each row where the overall logistic regression model for that row was significant and the overall living arrangements variable was significant at p < .0021.

Table 6.

Weighted Raw Percentage of Participants With Each Type of Companion by Age-Group and Results of Logistic Regressions

Community-activity Companions	Age-group						Total	Logistic regressions
	18–24	25–34a	35–44	45–54	55–64	65+		Overall model χ2(26) = Nagelkerke R2	p Age group
Community-group participation
Alone	8.0	10.5a	10.0	7.4	12.3	13.5	10.1	391.56*** R2=.168	p=.007
Friends	41.1	43.0a	41.9	35.7	32.2	32.8	38.7	269.45*** R2=.076	p=.003
Family	59.5	50.0a	49.6	35.9	22.5***	19.4	41.5	1,246.31*** R2=.316	p<.001***
Housemates/co-workers	8.4	10.2a	13.2	20.3	26.2	25.1	16.2	883.75*** R2=.294	p=.378
Staff	33.5	40.8a	40.0	49.9	55.2	65.4	45.9	800.09*** R2=.211	p=.002**
Others	5.9	5.0a	5.2	3.7	5.3	5.6	5.0	26.96 R2=.018	p=.415
Religious-service attendance
Alone	4.2	5.5a	7.3	6.3	9.3	8.2	6.7	526.89*** R2=.219	p=.179
Friends	7.5	13.8a	13.5	15.1	14.5	18.5	13.9	256.05*** R2=.076	p=.017
Family	81.4	74.6a	65.7	53.9	36.2***	24.7***	59.2	2,633.05*** R2=.483	p<.001***
Housemates/co-workers	5.8	7.3a	11.3	17.3	23.0	28.2	14.1	1,379.85*** R2=.372	p=.008
Staff	18.9	23.1a	28.9	38.9	52.8***	62.4***	34.9	2,200.18*** R2=.427	p<.001***
Others	1.9	1.8a	1.5	1.8	2.1	3.3	1.9	49.04 R2=.047	p=.599

Note. Row percentages shown, indicating participants who answered yes to that type of companion. Only individuals who were included in the relevant logistic regression analysis were included when calculating the percentages shown in each row of Table 6. Respondents may choose multiple types of companions, so column percentages for each age group add to exceed 100.

Independent variables in logistic regressions: Gender, Age group, Race/ethnicity. Level of intellectual disability, Behavior challenges, Down syndrome, Autism spectrum disorder, Participation self-reported, Communicates by speech, Mobility, Living arrangements.

^aReference category.

^***p < .001.

^**p < .0021. In the age group columns, significance asterisks denote the results of simple contrasts with the reference category within each row where the overall logistic regression model for that row was significant and the overall age group variable was significant at p < .0021.

Results

Research Question 1: What Percentage of Adults With IDD Who Use IDD Services Participate?

Overall, 34.4%, 99% CI [33.6, 35.3] of the sample participated as a member of a community group and 42.4%, 99% CI [41.5, 43.3] went to a religious service monthly or more often. Nearly a quarter (23.5%) of people did both, 46.5% did neither, and 30.0% did one but not the other, χ²(1) = 2925.13, p < .001, N = 20,647. Between-state variation for community group participation ranged from 7.5% (KY) to 49.4% (GA), and for religious services from 22.0% (KY) to 67.1% (SC). Further analysis of state variation was beyond the current study’s scope.

Research Question 2: What Personal Characteristics and Living Arrangements Are Associated With Participation?

Table 1 presents the number of participants with various characteristics and the percentage with each characteristic who did/did not participate in a community group and in a religious service. The independent variables in Table 1 are interrelated. For example, individuals living in their own home tend to have milder ID. Therefore, respectively for community-group participation and religious-service attendance, we used separate logistic regressions with the same independent variables (Table 2). This analytic approach allowed us to evaluate the association of each characteristic independent of other demographic factors. The self-reporting variable was included primarily to control for data source, because the items on participation in community groups and religious services could each be answered by the person with IDD self-reporting or by a proxy.

Table 2.

Weighted Logistic Regressions

Independent variable level	Community group N = 14,569					Religious service N = 14,346
	B	Sig.	OR	99% CI		B	Sig.	OR	99% CI
				LL	UL				LL	UL
Personal characteristics
Gender femaleab	0.116	.002**	1.12	1.02	1.24	.153	<.001***	1.17	1.06	1.28
Age group		.662					.002**
18–24	−0.072	.294	0.93	0.78	1.11	−.161	.017	0.85	0.72	1.01
25–34c
35–44	−0.075	.175	0.93	0.81	1.07	.119	.027	1.13	0.98	1.30
45–54	0.001	.990	1.00	0.86	1.17	.048	.407	1.05	0.90	1.22
55–64	−0.035	.583	0.97	0.82	1.14	−.043	.485	0.96	0.82	1.12
65+	−0.068	.371	0.93	0.77	1.14	.030	.682	1.03	0.85	1.24
Race/ethnicity		<.001***					<.001***
Whitec
Black	0.169	<.001***	1.19	1.04	1.35	.625	<.001***	1.87	1.65	2.12
Hispanic	−0.197	.013	0.82	0.67	1.01	.089	.232	1.09	0.90	1.33
Other	0.158	.089	1.17	0.92	1.49	.058	.528	1.06	0.84	1.34
Level of ID		.018					.002**
Mildc
Moderate	0.082	.064	1.09	0.97	1.22	.143	.001**	1.15	1.03	1.29
Severe	−0.083	.220	0.92	0.77	1.10	.073	.262	1.08	0.91	1.27
Profound	−0.132	.161	0.88	0.69	1.12	−.069	.432	0.93	0.74	1.17
Behavior challengesd	0.146	<.001***	1.16	1.04	1.29	.214	<.001***	1.24	1.11	1.38
Down syndromeb	0.457	<.001***	1.58	1.34	1.86	.377	<.001***	1.46	1.24	1.72
ASDb	0.088	.102	1.09	0.95	1.26	−.121	.022	0.89	0.77	1.02
Participation self-reportedb	0.343	<.001***	1.41	1.25	1.59	.182	<.001***	1.20	1.07	1.35
Communicates by speechb	0.142	.021	1.15	0.98	1.35	.280	<.001***	1.32	1.14	1.54
Mobility		<.001***					<.001***
Mobile without aids	0.411	<.001***	1.51	1.21	1.87	.372	<.001***	1.45	1.19	1.77
Mobile with aids	0.291	.002**	1.34	1.05	1.70	.142	.097	1.15	0.93	1.44
Not mobilec
Living arrangements		<.001***					<.001***
Own home	−0.300	<.001***	0.74	0.64	0.86	−0.448	<.001***	0.64	0.55	0.74
Family homec
Foster/host home	−0.019	.786	0.98	0.82	1.18	−.037	.607	0.96	0.80	1.16
Group 2–3	−0.514	<.001***	0.60	0.50	0.71	−.531	<.001***	0.59	0.50	0.70
Group 4–6	−0.292	<.001***	0.75	0.65	0.87	−.385	<.001***	0.68	0.59	0.79
Group 7–15	−0.155	.195	0.86	0.63	1.17	.061	.596	1.06	0.79	1.43
Other	−0.042	.841	0.96	0.56	1.64	−.431	.035	0.65	0.38	1.10
Institution	−0.681	<.001***	0.51	0.35	0.74	.303	.013	1.35	.99	1.85
Constant	−0.712	<.001***	0.49			−0.268	<.001***	0.77

Note. Nagelkerke R² = .052 (Community group) and = .070 (Religious service)

Comparisons of levels of each independent variable involved simple contrasts with the reference category. CI = confidence interval; LL = lower limit; UL = upper limit.

^aParticipants with other gender omitted.

^bReference category = no.

^cReference category.

^dReference category = yes.

^***p < .001.

^**p < .0021.

Independent of other variables, females, who were Black, without behavior challenges, with Down syndrome, who self-reported participation, were mobile without aids, and lived in their family home were significantly more likely to participate in community groups (Table 2). The overall model was significant, χ²(26) = 511.85, p < .001, N = 14,569, Nagelkerke R² = .052.

For religious service attendance, controlling for other variables, individuals who were female, Black, with moderate ID, without behavior challenges, with Down syndrome, who self-reported participation, communicated using speech, were mobile without aids, and lived in the family home were significantly more likely to attend religious services. The logistic regression model was significant, χ²(26) = 702.19, p < .001, N = 14,346, Nagelkerke R² = .070.

There were multiple similarities in the findings for community groups and religious services. Females who were Black, without behavior challenges, with Down syndrome, who self-reported participation, were mobile without aids, and lived with family experienced higher participation rates in community groups and in religious services. The fact that self-reporting was significantly associated with higher participation rates for both outcomes underlines the importance of controlling for data source when analyzing data that contains a mix of self-reports and proxy reports.

Research Question 3: What Is the Relationship Between Current and Individually Desired Participation?

Community Groups

Most (63.0% of the total sample) wanted to be part of more community groups (Table 3). Some 5,281 (44.7%) had an unmet desire to be in a group, more than currently participated in one. A further 2,154 (18.2%) already attended a group but wanted involvement with more groups. We used weighted logistic regression to analyze wanting to take part in more groups and to control for all the independent variables listed in Table 1, overall model, χ²(27) = 864.29, p < .001, N = 11,809, Nagelkerke R² = .096. We found nonparticipants in community groups were significantly more likely to want to take part in more groups than those already participating, OR = 2.30, p < .001, 99% CI [2.07, 2.56].

Table 3.

Weighted Frequency of Wanting to Be Part of More Community Groups by Community-Group-Participation Status (N = 11,808)

Want to be part of more community groups?	Community group participation?
	No n = 7,482		Yes n = 4,326		Total
	n	%	n	%	n	%
Yes	5,281	70.6	2,154	49.8	7,435	63.0
No	2,201	29.4	2,172	50.2	4,373	37.0

Note. Column percentages shown. Only individuals who were included in the logistic regression analysis of wanting to take part in more groups were included in Table 3.

Religious Services

We examined current religious attendance and the person’s preferred frequency (Table 4). Most people (74.6% of the total sample) attended a religious service as often as they preferred. Only 25.4% had discrepant current and preferred attendance rates, with most (22.2%) wanting to go more often, and only 3.3% preferring less often.

Table 4.

Weighted Crosstabulation of Current and Preferred Frequency of Religious-Service Attendance (N = 11,685)

Like to go more, less or same amount?	Current frequency of religious-service attendance: How many times in past month?								Total
	0 n = 5,980		1–2 n = 1,746		3–4 n = 2,720		5+ n = 1,239
	n	%	n	%	n	%	n	%	n	%
More	1,252	20.9	492	28.2	555	20.4	292	23.6	2,591	22.2
Less	140	2.3	99	5.7	99	3.6	44	3.6	382	3.3
Same as now	4,588	76.7	1,155	66.2	2,066	76.0	903	72.9	8,712	74.6

Note. Column percentages shown. Only individuals who were included in the multinomial regression analysis of preferred frequency of religious-service attendance were included in Table 4.

Overall, the weighted multinomial logistic regression model of preferred frequency of religious service attendance was significant, χ²(58) = 463.76, p < .001, Nagelkerke R² = .053. The analysis included all the variables listed in Table 1 to control for differences in personal characteristics and living arrangements. Overall, the item on current frequency of religious-service attendance was significant, χ²(6) = 97.08, p < .001. Wanting to attend more often was related to current religious attendance. Those who attended 1–2 times in the last month were more likely to want to go more often (OR = 1.29, p = .004), whereas those who attended 3–4 times were less likely to want to go more (OR = 0.79, p = .004). Among the few who wanted to go to religious services less, individuals who went 1–2 times in the last month were more likely to want to go less (OR = 1.90, p < .001).

Research Question 4: With Whom Do People With IDD Attend? Is Companion Type Related to Living Arrangements and Age Group?

Companions

For community groups overall, companions were: staff (45.9%), family (41.5%), friends (38.7%), housemates/co-workers (16.2%), alone (10.2%), and others not listed (5.1%; Table 5, Total column). For religious service attendance the pattern of companions differed somewhat; family (59.3%) were the most common companions, then staff (34.7%), housemates/co-workers (14.0%), friends (14.0%), alone (6.9%), and others (1.9%; Table 5, Total column). Some individuals went with companions from more than one category, so column percentages add to well over 100%. Because respondents could choose all types of companions that applied, we needed to conduct separate logistic regression analyses for each type of companion (Tables 5 and 6, Logistic Regressions column).

Companions and Demographic Variables

Companions and Living Arrangements.

Controlling for personal characteristics using logistic regression, the pattern of companions was significantly related to the type of living arrangement (Table 5). For both activities, the highest percentage going alone was own-home residents, with very low percentages in most other living arrangements. Community-group participation with friends was more common among own home, family home or foster/host home residents, but lower for most staffed group-living settings. Religious-service attendance with friends was less common than for community groups, with own-home residents more likely to do so. For both activities, going with family was most common for family-home dwellers, and much less prevalent for all other settings. Family companions were very common for religious-service attendance. For both activities, going with housemates/co-workers, was much more common for staffed group-living settings. Staff companions were the most common for every setting except own home (religious services only) and family home. People could attend with more than one type of companion, so going with staff and friends may have been common for residents of their own home, and foster homes. Going with staff and housemates/co-workers was more common for larger group residential settings.

Companions and Age Group.

Controlling for other personal characteristics and living arrangements using logistic regression, there were significant age-related differences for family and staff companions (Table 6, final column). Attending community groups or religious services with family each diminished markedly with age, with the opposite trend evident for staff.

Research Question 5: What Is the Relationship Between Friendship, Loneliness, and Attendance at Community Groups or Religious Services?

Friendship

Weighted multinomial regression of having friends was significant overall, χ²(54) = 475.47, p < .001; Nagelkerke R² = .069 and community group attendance was significantly related to having friends, χ²(2) = 133.10, p < .001. Compared to nonparticipants, community-group members were significantly more likely to have friends who are not staff or family, OR = 2.49, p < .001, 99% CI [1.98, 3.12], and to have friends who are all staff or family, OR = 1.78, p < .001, 99% CI [1.35, 2.34], (Table 7). The results for having friends and religious service attendance were similar, with multinomial regression significant overall, χ²(54) = 438.86, p < .001; Nagelkerke R² = .064, and attending religious services was related to having friends, χ²(2) = 112.02, p < .001. Religious service attendees were more likely to have friends who are not staff or family, OR = 2.27, p < .001, 99% CI [1.84, 2.80], and to have friends who are all staff or family, OR = 1.85, p < .001, 99% CI [1.43, 2.40].

Table 7.

Weighted Multivariate Analysis of Friendship by Community-Group or Religious-Service Attendance

Variable Level	Community group				Religious service
	No	Yes	Total	Regressiona	No	Yes	Total	Regressiona
	%	%	n	OR Comm. Gp Yes	%	%	n	OR Relig. Srv. Yes
Have friends?
No friendsb	11.2	4.9	809		11.7	5.7	806
Yes (not staff/family)	75.7	84.9	7,381	2.49***	76.2	83.5	7,369	2.27***
Yes (all staff/family)	13.1	10.2	1,108	1.78***	12.1	10.9	1,063	1.85***
Have a best friend?
Nob	30.5	17.3	2,234		29.3	20.0	2,173
Yes	69.5	82.7	6,623	2.03***	70.7	80.0	6,611	1.64***
See friends when you want?
Nob	6.4	4.4	449		6.8	4.5	451
Sometimes can’t	15.2	13.5	1,169	1.29	14.7	14.4	1,169	1.51***
Yes	78.4	82.2	6,474	1.55***	78.4	81.1	6,401	1.58***

Note. Column percentages shown. Only individuals who were included in the relevant regression analysis were included when calculating the percentages shown in Table 7.

Independent variables in regressions: Community group/Religious service attendance, Gender, Age group, Race/ethnicity. Level of ID, Behavior challenges, Down syndrome, ASD, Participation self-reported, Communicates by speech, Mobility, Living arrangements.

^aMultinomial regression used for Have friends? and for See friends when you want? Logistic regression used for Have a best friend?

^bReference category.

^***p < .001.

Weighted logistic regression analysis of having a best friend or not, was significant for community groups overall, χ²(27) = 322.24, p < .001; Nagelkerke R² = .053 and for religious services, χ²(27) = 233.37, p < .001; Nagelkerke R² = .039. The odds ratios shown in Table 7 indicate that those who participated in community groups, OR = 2.03, p < .001, 99% CI [1.77, 2.34] or in religious services, OR = 1.64, p < .001, 99% CI [1.44, 1.88] were significantly more likely to have a best friend.

Multinomial regression of being able to see friends when you want was significant overall, χ²(54) = 221.41, p < .001; Nagelkerke R² = .038 and community group attendance was significantly related to seeing friends, χ²(2) = 23.37, p < .001; Table 7. Community group participants were more likely to be able to see friends when they want (i.e., answer yes), OR = 1.55, p < .001, 99% CI, [1.18, 2.03]. Findings for religious-service attendees were similar—overall, χ²(54) = 216.89, p < .001; Nagelkerke R² = .037 and for religious service attendance, χ²(2) = 20.38, p < .001, with OR = 1.58, p < .001, 99% CI [1.21, 2.05], for answering yes, and OR = 1.51, p < .001, 99% CI [1.12, 2.03], for answering sometimes. The odds ratios (Table 7) all show that community group and religious service participation each are consistently associated with better outcomes for all three friendship items.

Loneliness

For the reasons outlined previously in the section headed “Social Benefits: Friendship and Loneliness,” community-group and religious-service participants were each separated into two groups—those who went with friends/family versus with others/alone. Under multinomial regression, for community groups, χ²(4) = 3.03, p = .55 and religious services, χ²(4) = 7.50, p = .11, there was no significant relationship between loneliness and attendance with friends/family, with others/alone, or not attending.

Discussion

We examined participation by adults with IDD who use IDD services in mainstream (i.e., open to the general community) community groups and religious services, using weighted multivariate analysis to control for multiple personal characteristics and living arrangements. Both types of community participation were positively associated with better friendship outcomes, suggesting that inclusive community participation may support existing friendships and facilitate new ones. This additional benefit of inclusive community participation is especially important because people with IDD typically have fewer friends (Amado et al., 2013).

Unmet desire for participation differed markedly by community participation type. Those wishing to take part in religious services mostly did so, indicating little unmet demand. Likewise, few people went to a religious service unwillingly. By contrast, there was a vast unmet demand for participation in a mainstream community group revealing an enormous disparity to be addressed.

Our study is one of the first to examine in detail companions for inclusive community participation. We found differing patterns of companions according to the type of community activity, the person’s living arrangements, and age group. For example, many who lived in disability provider settings attended these community groups with peers and staff, likely reflecting a group approach that was not individually person centered. By contrast, there was a very strong family-based character to religious service attendance. We contend that our findings are more robust and nuanced because they involved two different types of socially inclusive community settings.

Comparison With the General Community

One important way to interpret participation rates of people with IDD is through comparison with the general community. The NCI-ACS only surveys adults with IDD, so we had no general community data to analyze. Instead, we relied on comparisons with recent research findings on general community participation in community groups and religious services.

Our results suggest fewer U.S. adult service users with IDD (34.4%) take part in socially inclusive community groups than the U.S. general community (57%, Sandstrom & Alper, 2019). This is consistent with lower community participation by adults with IDD, especially participation in socially inclusive settings (Amado et al., 2013; Dusseljee et al., 2011; Verdonschot et al., 2009a, 2009b).

In the U.S. adult general community, more people aged 65+ (66%) participate in community groups than younger adults (55%; Sandstrom & Alper, 2019), whereas we found the 65+ age group had lowest participation rate (29.4%, Table 1). Stancliffe et al. (2018) reported a strong age-related rise in participation rate in segregated IDD day activities, reaching 50% of those aged 65+. Many mainstream seniors’ community groups meet during the day on weekdays (Stancliffe et al., 2015), so seniors with IDD attending weekday disability activities are unavailable for these community groups. More participation in disability groups, and less in inclusive community groups both reveal a notable bias toward segregated activities in older age.

The weighted religious-service attendance rate by adults with IDD was 42.4%, similar to Carter et al.’s (2015) unweighted rate of 48.3%. Attendance at religious services monthly or more often by general-community U.S. adults in 2018–2019 was 45% (Pew Research Center, 2019), with higher rates among Black Americans, women, and older age groups. This is one of the few areas of socially inclusive community life where adults with IDD achieve similar participation rates to the general community, although we had no data on the quality of the social inclusion achieved. One important factor appears to be family support. The very high rate of going with family (91.3%, Table 5) among those who live with family, suggests that religious attendance is frequently a family activity. Even for those living in their own home or foster home, attendance with family exceeded 40%.

The age profile of religious-service attendance differed from the general community. People with IDD aged 65+ had a lower attendance rate (Table 1), whereas the general community had a substantially higher rate in the oldest age groups (Pew Research Center, 2019). The central role of family may help explain this disparity. For older adults with IDD, parental health problems or death likely meant that family companionship or transportation for religious-service attendance was no longer available. The marked drop in family companions with increasing age (Table 6) is consistent with this notion. An important implication concerns transition from family to disability staff support. Lower religious attendance in older age may be partly due to unsuccessful transition to staff support.

Met and Unmet Desire to Participate

Socially inclusive participation is considered a human and civil right, and conforms to current U.S. Home and Community Based Services (HCBS) policies (Riesen & Snyder, 2019). However, such participation is considered to be person-centered only if the person wants to take part.

Community Groups

We found a very large unmet wish to attend socially inclusive community groups. Issues likely include practical barriers (transportation and/or support), informational barriers (knowledge about local community groups), and attitudinal barriers/low expectations, such as beliefs that socially inclusive participation is not possible or appropriate despite our clear evidence to the contrary (Bigby et al., 2011). However, Australian research involving sheltered workshop employees with IDD who individually participated in a mainstream community group as part of transitioning to retirement showed that, where these barriers existed, they were overcome (Stancliffe et al., 2015). Moreover, Stancliffe et al.’s participants experienced multiple benefits, such as long-term socially inclusive participation, new relationships, and greater social satisfaction (receiving social support from others, and having friends and people to talk to).

The numbers we reported who were already involved in U.S. community groups indicate the scope for participation at scale and show that many groups exist that have members with IDD. Data limitations meant that the degree to which current participants want to attend their community group could not be determined unambiguously, but the fact that 49.8% (Table 3) want to attend more such groups speaks to many enjoying their current participation and community connection.

Religious Services

We extended Carter et al.’s (2015) findings by providing weighted attendance rates and by analyzing people’s wishes regarding frequency of religious attendance. Three fourths went as often as they preferred. Where current and preferred attendance differed, most wanted to go more often, with only 3.3% preferring to go less, indicating that few people went to a religious service unwillingly. The more widespread issue was attending less often than desired. Overall, unmet desire for religious-service attendance was far less common than an unmet wish for community group participation. Interestingly, most religious services occur on weekends, when formal out-of-home disability day activities are usually not available and do not “compete.”

Person Centeredness

Participation in mainstream community groups and religious services is necessary for inclusion in these settings. Lack of such opportunities was evident in the very large unmet wish for participation in community groups. Although necessary, mere presence is not likely to be sufficient for meaningful inclusion. Issues of choice of participation and companions and the availability of appropriate support are all likely important. Unfortunately, data limitations meant that we had no information on the quality or extent of meaningful social inclusion, so we were unable to explore these issues empirically. A constraint on genuinely inclusive and person-centered community participation is that service users are often required to attend in groups, regardless of individual interest in the activity or preferences for companions (Bigby & Knox, 2009; Stancliffe et al., 2022). We found substantial levels of fellow service users and staff as companions, particularly among those living in congregate residential settings, indicating that group treatment likely was at play.

As noted, Australian research has reported significant social benefits of individual participation (i.e., alone, without companions) in mainstream community groups with support from group members (Stancliffe et al., 2015). By contrast, most of our participants went with companions to mainstream community groups (89.8% with companions) or religious services (92.1%). As is discussed later in more detail, we found better friendship outcomes for community group and religious service participants. Taken together, these findings indicate that attending mainstream community activities alone or with chosen companions both likely yield social benefits, whereas being required to attend as a group may not and is clearly not person centered (Bigby & Knox, 2009).

Personal Characteristics of People With IDD

Certain characteristics were strongly associated with socially inclusive participation. Higher participation rates among women and African Americans reflect broader societal trends for greater involvement by these groups (Pew Research Center, 2019; Sandstrom & Alper, 2019). Disability-related characteristics such as behavior challenges, Down syndrome, mobility, and communication difficulties were significantly related to participation, with individuals with milder disability more likely to participate (Carter et al., 2015). These characteristics likely relate to support needed to travel to and participate in the activity, and possibly to the members’ willingness to accept and support the person. However, we had no specific data about support needs or acceptance. Some with severe disability did participate, albeit at a lower rate (see Carter et al., 2015), with 22% (community group) and 33% (religious service) of individuals with profound ID taking part. Individuals with severe or multiple disability can participate, but practical issues like support needs, physical accessibility, and health/fatigue may all influence feasibility.

Living Arrangements

For community groups, institution residents had by far the lowest participation rate, and those from family homes or host/foster homes had the highest (Table 1). Family home or host/foster homes residents had significantly higher religious-attendance rates, likely reflecting the strongly family-based nature of going to a religious service.

Predominantly, nonparticipants were people with various disabilities (behavior, communication, mobility) who live in group disability accommodation (Table 2). Finding ways to effectively support such individuals to engage in inclusive community activities if they wish represents an important ongoing challenge.

Table 2 shows that multiple personal characteristics and living arrangements were significantly related to participation in community groups and religious services. These findings underline the importance of using multivariate analyses to control for these variables when examining companions, friendship, and loneliness in the sections which follow.

Companions

Often, companions for inclusive participation were those available in the living setting and/or in the person’s close social network, reflecting the principal of proximity (Wiener & Schneider 2002), meaning people who are frequently nearby are the most likely companions. Thus, people living with family were much more likely to have family companions, whereas individuals living in staffed congregate settings mostly had staff as companions (Table 5). Service systems that congregate people with IDD together with continuous support from staff create the conditions for social networks to be dominated by staff and fellow service users, partly due to their proximity.

Congruent with an age-related pattern of people moving out of the family home and into their own home or a staffed living setting, family as companions fell consistently with age for both types of community participation, whereas staff involvement grew significantly with increasing participant age. Age-related parental health problems and death likely contributed to changes in availability of family companions. These findings are consistent with a family life-cycle approach where changes over time in family members’ circumstances and the person with IDD’s living and support arrangements strongly influence whom they spend time with (Bigby & Knox, 2009), including when going to community groups or religious services.

Friendship and Loneliness

Attending community groups and religious services were both strongly and consistently associated with better friendship outcomes. Friendships and social participation can be synergistic. Accepting a friend’s invitation and going together to a community group or religious service likely strengthens an existing friendship. Relationships may also arise from meeting a new friend at a community group or religious service. Data limitations meant that we were unable to determine the relative proportions of existing and new friendships related to such participation. Our findings on friendship are similar to results from a controlled intervention by Stancliffe et al. (2015) examining benefits of individual participation in mainstream community groups, which included people with IDD each developing several new relationships with group members without disability.

Loneliness was not related to community group participation and attending religious services. As noted in the section on “Social Benefits: Friendship and Loneliness”, loneliness is subjective, and arises when the person’s expectations of relationships are not reflected in their experience (Hawkley & Cacioppo, 2010). Here too, our findings parallel Stancliffe et al.’s (2015) results which showed no change in loneliness after 6 months of weekly attendance at a mainstream community group. Overall, these findings suggest loneliness is difficult to change and is not strongly related to social engagement/isolation (Emerson et al., 2021). This may be because of its subjective nature, or because attendance for a few hours per week (or less) at a community group/religious service is too little to affect loneliness. Alternatively, researchers may well need to evaluate the subjective experience of people with IDD who attend mainstream groups to understand their sense of belonging and its relationship with loneliness.

Future Research

The large between-state variability in participation merits closer scrutiny, because factors not examined in the present study may affect the availability and uptake of participation. Community groups with certain characteristics may be more supportive of participation and/or a sense of belonging by people with IDD, so researchers should evaluate the relationship between group characteristics and outcomes experienced by people with IDD, including loneliness.

Participation in community groups and religious services was significantly related, but only 23.5% of participants did both, whereas 30.0% did one but not the other, suggesting selectivity in the activity in which individuals chose to participate. Together with the differing patterns of companions and of unmet need, it seems that these two types of socially inclusive participation differ in important ways, suggesting that they should continue to be studied separately in future research.

Data limitations meant we could not examine the nature or quality of socially inclusive participation. Future research should study the quality and extent of participation and inclusion, and the enjoyment and sense of belonging experienced by participants with IDD. Transition from family support to staff support for community participation deserves careful longitudinal attention.

Limitations

Several multivariate models for community groups and religious services (e.g., Table 2) predicted a small proportion of variance indicating that much variance was not explained by the independent variables examined. Analysis of the substantial between-state variability was beyond our study’s scope but added statistical noise that likely attenuated the effect size of the variables analyzed. Our cross-sectional study could not evaluate cause and effect. We had no data on frequency of attendance at community groups. Nor was data available on the characteristics of the group, so we could not evaluate whether some types of groups were more supportive. Loneliness and friendship data came from NCI-IPS Section I which may only be answered by the person with IDD. Thus, there were smaller samples for these analyses because many participants with severe/profound ID and/or communication difficulties were unable to respond and had missing data for these items.

Conclusions

Sharply lower participation in socially inclusive community groups than the general community, and a very large unmet demand by adults with IDD for such participation reveal a major gap for disability services and the community to address. By contrast, religious service attendance was mostly aligned to general community participation rates and to individual preferences. The friendship outcomes further underline the benefits of socially inclusive community participation.