Disability in U.S. Households, 2000–2010: Findings from the National Health Interview Survey

Abstract

Understanding the demographic structure of households containing members with disabilities is of key importance in policy planning for populations with disabilities at state and national levels. Yet, most, but not all, previous family-level studies of disability have excluded persons living alone or with unrelated persons (e.g., a housemate or an unmarried partner) because they are not considered families. To address this gap, the authors utilize National Health Interview Survey data to produce household-level estimates of disability using a detailed household type variable that includes households omitted from previous reports. Findings indicate that one-person households made up 24.7% of all households with an adult aged 18–64 with a disability, and 42.9% of all households with an adult aged 65 or older with a disability. Including nonfamily households provides a clearer picture of the association between living arrangements and disability in the U.S.

Families have traditionally been the dominant form of support for persons with disabilities, whether the person with a disability lives in the same residence, in the same community, or in nearby communities (Talley & Crews, 2007). Families may provide monetary support, transportation, shopping, assistance dealing with administrative matters, or in some cases daily care. While there is an extensive literature on caregiving in the family setting, less is known about the demography of disability within all American households (Seltzer, 1992; Seltzer & Krauss, 1994; Spillman & Black, 2005; Spillman & Pezzin, 2000; Stone & Kemper, 1989). An understanding of the basic demographic structure of all households containing members with disabilities is of key importance in the ongoing policy planning for populations with disabilities at both the state and national levels (Fujiura, 1998; 2010). Various community programs such as the Medicaid Home and Community-Based service (HCBS) waiver program, which was enacted in 1981, have provided an increasing variety of services that help persons with disabilities live and participate in the community. Additionally, the 1999 Olmsted Decision of the U.S. Supreme Court has facilitated continued movement out of institutions into a variety of living arrangements that began in the 1980s with the deinstitutionalization of persons with intellectual disability (Lakin, Prouty, Polister, & Coucouvanis, 2003; Mann, 2010; O’Connell & Watson, 2001; Olmstead v. L.C., 1999). These changes coincide with wider societal changes in union formation, dissolution, and childbearing that have similarly contributed to more complicated family structures as well as increasing numbers of nonfamily households in the U.S. (Casper & Hofferth, 2007; Lamanna & Riedmann, 2009; Teachman, et al., 2000), underscoring the need for new research to understand the relationship between disability and household structure.

Why is this an issue for persons with disability? Persons with disabilities may have occasions when they need some form of immediate help, yet their living arrangements may not provide that help when needed. In the case of adults with disability, living alone may increase the likelihood that necessary assistance from family or support services may be delayed or unavailable (LaPlante, Harrington and Kang, 2002). This could result in difficulties with personal care (such as dressing, bathing, or eating) or routine needs (such as shopping, running errands, or performing everyday household chores), limitations in social participation, unmet health care needs, or difficulties coping with unpredictable situations such as emergencies or natural disasters (Desai, Lentzner, and Weeks, 2001; Elliot, Painter, and Hudson, 2009; Kailes and Enders, 2007). Additionally, persons living alone have been shown to have lower probabilities of using preventive care than those living with others (Lau and Kirby, 2009). It is also possible that the need for support may determine living arrangements, as when an older adult who has become disabled moves in with his or her adult child. And though more persons with disabilities are able to live in the community, they – and the households in which they live – may be highly dependent on safety nets and social services. By excluding persons with disabilities who are living in nonfamily households, we suggest that family-level analyses of disability are overlooking the proportion of the disabled population that is likely to be most vulnerable to the impact of small changes in the availability of services and more dependent on services provided by publicly funded entities.

To address these issues, we utilized National Health Interview Survey (NHIS) data to produce household-level estimates of disability in the United States during 2000–2010 and to examine the household living arrangements for those identified as disabled. The NHIS obtains basic information on various functional and activity limitations for all family and household members. Using this information, we created several measures that count the number of households containing members with disabilities within particular age groups. We also focused on types of functional limitations or difficulties and living arrangements (i.e., type of household) and the socioeconomic resources available in the household. We asked several questions. First, does the prevalence of disability vary by household type? Second, how does this prevalence change if we take into account the age of the individual with the disability as well as the type of disability? Last, to what extent does disability vary when we control for household socioeconomic status?

Previous family-level studies of disability

Most studies of persons with disabilities have focused on individual-level data but there have been exceptions, including a 1996 National Institute on Disability and Rehabilitation Research (NIDRR) family-level report based on the 1990 NHIS (LaPlante, Carlson, Kaye, & Bradsher, 1996) and a 2005 family-level report based on the 2000 Census (Wang, 2005). However, their analyses omitted persons living alone, with unrelated roommates, or with cohabiting partners because these did not fit the definition of family that has frequently, but not exclusively, guided previous family-level research: that is, a family consists of two or more persons related by blood, marriage, or adoption who reside together (Casper & Hofferth, 2007; Czajka & Denmead, 2008; Lamanna & Riedmann, 2009). Instead, those who lived alone or with unrelated roommates or partners were considered nonfamily households (Goldscheider, 2009).

Fujiura (1998, 2010) and Fujiura, Yamaki, and Czechowicz (1998) also focused on the demography of households containing members of various ages with disabilities using the 1990–1991 Survey of Income and Program Participation (SIPP) and the 2007 American Community Survey (ACS). Using SIPP data, Fujiura (1998) concluded that most children and adults with intellectual or developmental disabilities (ID/DD) lived in family households, that these households were more economically disadvantaged than households without similarly disabled members, and that the degree of this disadvantage varied significantly by family type. Using ACS data, Fujiura (2010) focused on adults with disability and found that 27.1 million adults with disability aged 22 or older lived in family households while 9.7 million adults with disability lived in nonfamily households, and that adults with disabilities were more likely to live in nonfamily households than adults without disabilities. However, though the existence of nonfamily households was certainly acknowledged in Fujiura’s 2010 paper, the primary focus of these analyses was on family households premised on a definition of family that excluded nonfamily households.

Using the Health and Retirement Study (HRS), Waite and Hughes (1999; Hughes & Waite, 2002) created a six-category indicator of family and nonfamily living arrangements: married persons living with spouses only; married persons living with spouses and children; married persons living with their spouses and others (including lineal relatives who may be multi-generational); unmarried persons living alone; unmarried persons living with children only; and unmarried persons living with others (including living in someone else’s household; Waite & Hughes, 1999). Their analyses allowed for both family and nonfamily households, and they distinguished family households in which relationships were nuclear from those with more complex relationships. Additionally, they included types of functional limitations in their analyses and found evidence of different living arrangements by type of functional limitations, including a definite pattern of poorer functioning among persons who were considered to be in the most demanding and least supportive types of family and nonfamily environments. Waite and Hughes (1999, p.10) argued that the household is a critical environment for disability because it not only provides resources to support persons with disability but also defines the roles and tasks the person with a disability is expected to do.

Other studies have also suggested the importance of living arrangements for persons with disabilities, and how these arrangements have changed. Gibson and Ludwig (1968, p. 54) utilized data from Social Security disability applications to examine, in their words, breadwinners with disabilities and their families. Using a 20-category typology of household structure, they showed that 66.5% of male breadwinners with disabilities were married or cohabiting (with or without children), whereas 12.2% lived alone. Among female breadwinners with disabilities, the nature of household structures was somewhat different: 44.2% were married or cohabiting (with or without children), whereas 21.6% lived alone. Other complicated arrangements – for example, living with adult relatives, nonrelatives, and/or grandchildren – accounted for the remaining percentages.

Subsequent research has emphasized how family living arrangements have changed as intergenerational support has become a key factor in determining household structure. Although much has been written about the movement of aging and increasingly functionally limited family members into the same residence or in closer proximity to their adult children (Silverstein & Giarrusso, 2010; Smits, Van Gaalen, & Mulder, 2010), there has been a growing literature on the important role that grandparents play in their children’s families, particularly when the grandparent(s) reside(s) with the family and provide(s) care or support for a child (or children) with disability (Silverstein & Giarrusso, 2010; Mitchell, 2007; Hughes, Waite, Lapierro, & Luo, 2007; Schilmoeller & Baranowski, 1998). Another factor in this changing family structural environment has been the notable increase in single parent families, particularly single women with children. Cohen and Petrescu-Prahova (2006, p. 635) showed that children with disabilities were significantly more likely than nondisabled children to live with a single mother, with cohabiting parents, with grandparents or other relatives, or with nonrelatives. Additionally, children with disabilities were also significantly more likely to be foster children than nondisabled children (Cohen & Pestrescu-Prahova, 2006).

Data, measures of interest, and analysis plan

Data from the 2000–2010 NHIS were used for our analysis. NHIS collects information via in-person interviewing about the health and health care of the civilian noninstitutionalized population of the United States from a representative sample of households across the country, and is conducted continuously throughout the year by the Centers for Disease Control and Prevention’s National Center for Health Statistics (NCHS). Persons excluded from the NHIS include patients in long-term care institutions, such as hospitals for the chronically ill or physically or intellectually disabled. Although the vast majority of NHIS households consist of one family, a small number of households (approximately 2%) contain two or more families. When these households are identified, each family within the household is interviewed separately.

The core NHIS questionnaire consists of three main components: the Family Core, which is based on information supplied by a knowledgeable family respondent, and the Sample Adult and Sample Child Cores, which obtain information about a randomly selected adult (the “sample adult”) and child (the “sample child”), respectively. The Family Core yields the Person and Family data files that were used for our analysis. In addition, a Household file is also created to identify responding and non-responding households, living quarters, and geographic information of responding families. Importantly, the NHIS Person and Family files can be merged with the Household file, and the resulting variables may be appropriately analyzed at the household level using the Household file weights.

Definitions of household types

Unlike the definition of family stating that it is comprised of two or more persons related by blood, marriage, or adoption who reside together (Casper & Hofferth, 2007; Czajka & Denmead, 2008; Lamanna & Riedmann, 2009), the NHIS defines a family as “an individual or a group of two or more related persons who are living together in the same occupied housing unit (i.e., household) in the sample. In some instances, unrelated persons sharing the same household may also be considered as one family, such as unmarried couples who are living together” (NCHS, 2011b). Consequently, estimates obtained from the NHIS Family file will not be directly comparable to estimates obtained using social scientists’ definition of family unless some adjustments are made to the data structure and weights (see Czajka & Denmead, 2008, for a discussion of those adjustments).

Accordingly, we used the NHIS Household file and existing family structure variables to create a detailed household type measure that includes separate categories for nonfamily households, adult-only family households, and family households with children (see text box).

Nonfamily households	One person 2 or more unrelated persons (e.g., roommates) Cohabiting couple only
Family households	without children	Married couple only All other related adults
	with children	Married parents Cohabiting parents Single mothers Single fathers Other single adults Extended Other

According to our typology, nonfamily households consist of one person living alone, two or more unrelated persons (e.g., roommates) or families living in the same residence, and cohabiting or unmarried couples. In the case of persons living with nonrelatives, these may be roommates or, alternatively, the householder may be sharing the residence with nonrelatives who are related to one another but not to the householder (Goldscheider, 2009). Importantly, some of these households may also contain children.

On a related note, the NHIS defines children as family members who are 0–17 years of age (emancipated minors living alone were dropped from our analysis) and adults as family members who are 18 years or older. Adult children (those aged 18 or over) are considered related adults regardless of their relationship (biological/adoptive, step, or foster) to their parents.

Family households may or may not have children. Adult-only family households (that is, without children younger than age 18) include married couple only households (i.e., a householder living with his or her spouse) and all other related adult households. The latter consists of two or more related adults sharing a residence; examples might include an older woman living with her adult children (all age 18 or older); a married couple living with a parent of one of the spouses; or siblings (again, age 18 or older) living with one another. Family households with children (biological, adoptive, or stepchildren younger than age 18) include married parents with child(ren) households; cohabiting parents with child(ren) households; single mother with child(ren) households; single father with child(ren) households; other single adult with child(ren) households (the adult is likely to be a grandparent, an aunt, or an uncle); extended households (one or more children living with at least one biological or adoptive parent and a related adult who is not a parent, such as a grandparent or an adult sibling); and other households (one or more child(ren) living with related or unrelated adults who are not biological or adoptive parents, such as children being raised by their grandparents or fostered by unrelated adults).

Disability measures

We used questions from the Family Health Status (FHS) section of the NHIS Person file because the questions in this section were asked about all household members. This section included questions about whether family members have play or work limitations; needed help with personal care or activities of daily living (ADLs); needed help with routine needs or instrumental activities of daily living (IADLs); had difficulty walking without special equipment; and had difficulty remembering or experienced periods of confusion. The Person file also contained a variable indicating family members with a positive response to any one of the disability or limitation questions in the FHS section; this was used to indicate households containing one or more members with any disability or limitation.

Because the FHS questions asked about limitations regarding various age-appropriate activities (such as play for very young children, attending school for older children, and work or employment for adults age 18 years or older), our resulting disability measures take into account the age of household members. We therefore created household-level counts of disability and limitations for three distinct age groups of household members with disabilities: children age 3–17 years, adults age 18–64 years, and adults age 65 years or older. Households containing children under age 3 with disabilities can only be identified by a variable that indicates children with play limitations, but this is a vague indicator of disability that may be more related to developmental process in general rather than to a specific functional limitation. Therefore, households with very young children with disabilities were not specifically identified for this analysis. For all households with one or more members age 3–17 years, we created three separate indicators identifying households with one or more members in this age group who had any disability or limitation; who required help with personal care; or who received special education or Early Intervention Services (EIS). For all households with one or more members age 18–64 years, we created six separate indicators that identified those households with one or more members in this age group who had any disability or limitation; who required help with personal care; who required help with routine needs; who experienced difficulty walking without special equipment; who had difficulty remembering or periods of confusion; or who had any work limitation. Additionally, for all households with one or more members age 65 years or older, we created five separate indicators that identified households with one or more members in this age group who had any disability or limitation; who required help with personal care; who required help with routine needs; who experienced difficulty walking without special equipment; or who had difficulty remembering or periods of confusion.

Control measures

In addition to household type, we utilized several socioeconomic status (SES) measures to describe household characteristics. Household poverty status was obtained from the 2000–2010 NHIS imputed income files so no observations were dropped due to missing information on income. (Each household’s income was based on the family respondent’s estimate of total family income of all family members from all sources, before taxes, in the last calendar year; if the respondent didn’t know or refused to answer the question, the value of total family income was obtained from the NHIS imputation process. In any given survey year, income will be imputed for roughly 20–25% of families.) The variable on these imputed files is a ratio of the household’s income in the calendar year prior to the interview to the appropriate poverty threshold for that same year as defined by the U.S. Census Bureau. We collapsed this ratio into four mutually exclusive categories – strictly below the federal poverty threshold; 1.00 to less than 2.00 times the federal poverty threshold; 2.00 to less than 4.00 times the federal poverty threshold; and 4.00 times the poverty threshold or more. Note that we did not attempt to determine poverty status for households consisting of two or more unrelated persons (e.g., roommates) or families because the extent to which family income in these households was actually shared is unknown.

Regarding household education, we compared completed education for all adult members of each household and selected the education of the adult in the household with the highest completed level of education; this was further collapsed into four mutually exclusive categories: less than a high school diploma; General Equivalency Diploma (GED) or high school diploma; some college (including Associate’s degree); and college degree (BA, BS, graduate, or professional degree).

Analysis plan

All counts and percentages shown here are weighted, annualized estimates for U.S. households based on an unweighted total of 380,179 households from the 2000–2010 NHIS. Data weighting procedures are described in more detail elsewhere (Botman, Moore, Moriarity, & Parsons, 2000). Pooled point estimates and confidence intervals for this analysis were calculated using the SUDAAN software package (release 10) to account for the complex sample design of the NHIS (Research Triangle Institute, 2008). The Taylor series linearization method was used for variance estimation.

Because we are pooling 11 years of data for our study, it is important to consider whether the prevalence of disability and household composition have changed across the study period. Previous studies have noted changes in both household composition and disability prevalence in the U.S. during the past decade (Cherlin, 2010; Freedman, Martin, & Schoeni, 2002; NCHS; 2011a). In our data, households consisting of married parents with children decreased from 19.9% of all households in 2000 to 16.2% of all households in 2010. Likewise, households containing children age 0–17 with disabilities gradually increased from 10.4% in 2000 to 13.0% in 2010, while households with adults age 18–64 with disabilities years increased from 15.9% in 2000 to 17.4% in 2010. However, households with adults age 65 years or older with disabilities declined from 41.1% in 2000 to 39.7% in 2010. Therefore, percentage estimates shown in our article can be interpreted as either an estimate for the midpoint of the study period or as an average across the study period. Significance of differences between percentages was evaluated using two-tailed t-tests at the 0.05 level. No adjustments were made for multiple comparisons. Terms such as “higher,” “lower,” and “fewer” indicate statistically significant differences. Terms such as “similar,” “comparable,” and “not different” indicate that the statistics being compared were not significantly different.

Results

Distribution of household type

Figure 1 shows the percent distribution of households in the United States during 2000–2010. One-person households made up 28.3% of all households, whereas households consisting of two or more unrelated persons (e.g., roommates) or families made up 1.7% of all households. Adult-only households consisting of married couple only households, cohabiting couple only households, or all other related adult households comprised 22.4%, 2.9%, and 10.9%, respectively, of the household type distribution. Married parents with child(ren) households made up 18% of all households, whereas cohabiting parents with child(ren) households made up 1.7%; single mother with child(ren) households made up 4.6%; single father with child(ren) households made up 0.7%; other single adult with child(ren) households made up 0.3%; extended households made up 7.3%; and other households made up 1.1% of the distribution of households.

Figure 1.

Weighted percent distribution of household type: United States, 2000–2010

Source: Centers for Disease Control and Prevention, the National Center for Health Statistics (2001–2010).

Household type and disability

Table 1 shows percentages of households by household type and whether one or more members of the household had a disability during 2000–2010. During our study period, 25.6% of all households had at least one household member of any age with a disability or limitation, but the percentage of households with one or more members with a disability varied considerably by household type. Among all adult-only households, only 19% of cohabiting couple only households had one or more members with disabilities, compared with 26.4% of married couple only households, 26.6% of one-person households, and 35% of other related adult households. Disability across households also fluctuated when the presence of children was taken into account. Only 16.8% of households consisting of married couples with children contained one or more members with disabilities, whereas 19.9% of households consisting of single fathers with children, 22.4% of households consisting of cohabiting parents with children, 23.4% of households consisting of single mothers with children, 28.9% of extended households, 41.3% of households consisting of single adults with children, and 41.6% of other households had at least one member with a disability.

Table 1.

Weighted Percents (and Confidence Intervals) of Households by Household Members’ Disability Status and Household Type: United States, 2000–2010

	All House- hold Types	One- Person House- holds	House- holds Consisting of 2 or more Unrelated Persons (e.g., room- mates) or Families	Married Couple only House- holds	Cohabiting Couple only House- holds	All Other Related Adult House- holds	Married Parents with Child(ren) House- holds	Cohabiting Parents with Child(ren) House- holds	Single Mother with Child(ren) House- holds	Single Father with Child(ren) House- holds	Other Single Adult with Child(ren) House- holds	Extended House- holds	Other House- holds
Disability Status	Percent (95% confidence interval)
One or more members of any age with disabilities in household	25.6 [25.3, 25.8]	26.6 [26.1, 27.2]	23.2 [21.6, 24.8]	26.4 [26.0, 26.8]	19.0 [18.1, 19.9]	35.0 [34.5, 35.6]	16.8 [16.5, 17.2]	22.4 [21.3, 23.6]	23.4 [22.7, 24.1]	19.9 [18.3, 21.6]	41.3 [38.4, 44.2]	28.9 [28.3, 29.5]	41.6 [39.9, 43.2]
No members with disabilities in household	74.4 [74.2, 74.7]	73.4 [72.9, 73.9]	76.8 [75.2, 78.4]	73.6 [73.2, 74.0]	81.0 [80.1, 81.9]	65.0 [64.4, 65.5]	83.2 [82.8, 83.6]	77.6 [76.4, 78.7]	76.6 [75.9, 77.3]	80.1 [78.4, 81.7]	58.7 [55.8, 61.6]	71.1 [70.5, 71.7]	58.4 [56.8, 60.1]

Source: Centers for Disease Control and Prevention, the National Center for Health Statistics (2000–2010).

Table 2 shows distributions of household type by disability type for all households with one or more elderly members age 65 or older during 2000–2010. (Three household types – cohabiting parents with child(ren), single mothers with child(ren), and single fathers with child(ren) – were not included in this table due to very small cell counts that yielded percentages of zero.) Note that the first column represents the overall distribution of household type for all households having one or more members age 65 or older, to which we compare the remaining columns showing distributions of household type for households having one or more elderly members with selected disabilities (e.g., any disability or limitation, requiring help with personal care or with routine needs). Although one-person households accounted for 41% of all households having one or more elderly members in 2000–2010, this household type accounted for 42.9% of households containing elderly members with any disability or limitation, 47% of households with elderly members who required help with routine needs, and 43.8% of households with elderly members who had difficulty walking without special equipment. However, the percentages of one-person households containing elderly members requiring help with personal care (31.4%) or having difficulty remembering or periods of confusion (36.1%) were lower than their percentage in the overall distribution.

Table 2.

Weighted Percent Distribution (and Confidence Intervals) of Household Type, for Households with One or More Members Age 65 or Older, by Disability Type: United States, 2000–2010

	All Households Having 1+ Members Age 65+	Households Having 1+ Members Age 65+ with Any Disability or Limitation	Households Having 1+ Members Age 65+ Requiring Help with Personal Care	Households Having 1+ Members Age 65+ Requiring Help with Routine Needs	Households Having 1+ Members Age 65+ with Difficulty Walking without Special Equipment	Households Having 1+ Members Age 65+ with Difficulty Remembering
Household Type	Percent (95% confidence interval)
One-person households	41.0 [40.4, 41.6]	42.9 [42.0, 43.8]	31.4 [29.9, 33.0]	47.0 [45.7, 48.3]	43.8 [42.7, 44.9]	36.1 [34.7, 37.6]
Households consisting of 2 or more unrelated persons (e.g., roommates) or families	0.6 [0.5, 0.6]	0.8 [0.7, 1.0]	1.9 [1.6, 2.3]	1.4 [1.2, 1.6]	0.9 [0.8, 1.1]	1.6 [1.4, 2.0]
Married couple only households	36.1 [35.6, 36.6]	31.8 [31.1, 32.5]	28.9 [27.6, 30.2]	22.4 [21.6, 23.3]	29.1 [28.3, 30.0]	29.0 [27.8, 30.1]
Cohabiting couple only households	1.1 [1.0, 1.2]	0.9 [0.8, 1.0]	0.7 [0.5, 1.0]	0.7 [0.5, 0.9]	0.7 [0.6, 0.9]	0.7 [0.5, 0.9]
All other related adult households	15.4 [15.1, 15.7]	18.1 [17.6, 18.6]	29.2 [27.9, 30.5]	22.7 [21.8, 23.6]	19.6 [18.9, 20.4]	25.4 [24.3, 26.5]
Married parents with child(ren) households	0.3 [0.3, 0.4]	0.2 [0.2, 0.3]	0.2 [0.1, 0.4]	0.1 [0.1, 0.3]	0.2 [0.1, 0.3]	0.2 [0.1, 0.4]
Other single adult with child(ren) households	0.2 [0.2, 0.3]	0.2 [0.2, 0.3]	0.1 [0.1, 0.2 ]	0.2 [0.2, 0.3]	0.2 [0.2, 0.3]	0.1 [0.1, 0.2]
Extended households	4.1 [4.0, 4.3]	3.8 [3.6, 4.1]	6.1 [5.5, 6.7]	4.3 [4.0, 4.7]	4.2 [3.9, 4.5]	5.6 [5.1, 6.2]
Other households	1.1 [1.1, 1.2]	1.1 [1.0, 1.3]	1.4 [1.2, 1.7]	1.0 [0.9, 1.2]	1.2 [1.0, 1.4]	1.2 [1.0, 1.5]
Total	100.0	100.0	100.0	100.0	100.0	100.0

Source: Centers for Disease Control and Prevention, the National Center for Health Statistics (2000–2010).

Notes: Cohabiting parents with child(ren) households, single mother with child(ren) households, and single father with child(ren) households had 0% in each distribution above and were omitted from this table.

Columns may not add to 100.0 due to rounding.

All other related adult households, which accounted for 15.4% of all households with one or more elderly members, were higher in all the distributions of households containing elderly members with various types of disability. On the other hand, married couple only households, which accounted for 36.1% of all households with one or more elderly members, were lower in all the distributions of households containing elderly members with disabilities.

Table 3 shows distributions of household type by disability type for all households with one or more adult members age 18–64 years during 2000–2010. Again, the first column represents the overall distribution of household type for all households having one or more adults age 18–64; the remaining columns show the distributions of household type for households having one or more adults in this same age group with disabilities (e.g., any disability or limitation, requiring help with personal care or with routine needs, any work limitation). Although one-person households accounted for 22.3% of all households with adults age 18–64, their percentages were higher in the distributions of households containing adults with any disability (24.7%), requiring help with routine needs (26.9%), having difficulty walking without special equipment (28.3%), having difficulty remembering or periods of confusion (30.2%), and having any work limitation (24.9%). Married couple only households, which accounted for 19.4% of all households with adults age 18–64, also had higher percentages in the distributions of households containing adults with any disability (22.3%), requiring help with personal care (21.9%), having difficulty walking without special equipment (24.2%), and having any work limitation (22.2%); their percentage was lower in the distribution of households containing adults having difficulty remembering (17.3%). All other related adult households, which comprised 12.8% of all households with adults age 18–64, had higher percentages in all the distributions of households containing adult members with various types of disability.

Table 3.

Weighted Percent Distribution (and Confidence Intervals) of Household Type, for Households with One or More Members Age 18–64, by Disability Type: United States, 2000–2010

	All Households Having 1+ Members Age 18–64	Households Having 1+ Members Age 18–64 with Any Disability or Limitation	Households Having 1+ Members Age 18–64 Requiring Help with Personal Care	Households Having 1+ Members Age 18–64 Requiring Help with Routine Needs	Households Having 1+ Members Age 18–64 with Difficulty Walking without Special Equipment	Households Having 1+ Members Age 18–64 with Difficulty Remembering	Households Having 1+ Members Age 18–64 with Any Work Limitation
Household Type	Percent (95% confidence interval)
One-person households	22.3 [22.0, 22.7]	24.7 [24.1, 25.2]	21.2 [19.9, 22.6]	26.9 [25.9, 28.0]	28.3 [27.3, 29.3]	30.2 [29.0, 31.4]	24.9 [24.3, 25.5]
Households consisting of 2 or more unrelated persons (e.g., roommates) or families	2.0 [1.9, 2.1]	2.3 [2.2, 2.5]	2.7 [2.3, 3.2]	2.8 [2.5, 3.2]	2.3 [2.1, 2.6]	3.6 [3.2, 4.1]	2.2 [2.0, 2.4]
Married couple only households	19.4 [19.2, 19.6]	22.3 [21.9, 22.8]	21.9 [20.6, 23.3]	20.0 [19.1, 20.9]	24.2 [23.4, 25.1]	17.3 [16.4, 18.2]	22.2 [21.7, 22.7]
Cohabiting couple only households	3.4 [3.3, 3.5]	3.5 [3.3, 3.7]	3.7 [3.0, 4.5]	3.1 [2.7, 3.5]	3.2 [3.0, 3.6]	3.4 [3.0, 3.8]	3.5 [3.3, 3.7]
All other related adult households	12.8 [12.6, 12.9]	18.9 [18.5, 19.3]	28.5 [27.2, 29.8]	23.4 [22.6, 24.4]	18.4 [17.6, 19.2]	20.0 [19.1, 20.9]	19.4 [19.0, 19.9]
Married parents with child(ren) households	21.5 [21.3, 21.8]	10.5 [10.2, 10.9]	5.9 [5.2, 6.6]	6.9 [6.4, 7.4]	7.8 [7.3, 8.3]	8.0 [7.4, 8.6]	9.9 [9.6, 10.2]
Cohabiting parents with child(ren) households	2.0 [1.9, 2.1]	1.5 [1.4, 1.7]	0.7 [0.5, 1.0]	0.9 [0.7, 1.2]	0.9 [0.8, 1.1]	1.5 [1.2, 1.7]	1.5 [1.4, 1.6]
Single mother with child(ren) households	5.5 [5.4, 5.6]	3.7 [3.6, 3.9]	2.0 [1.7, 2.4]	3.2 [2.9, 3.5]	2.3 [2.1, 2.6]	4.0 [3.6, 4.4]	3.8 [3.6, 4.0]
Single father with child(ren) households	0.9 [0.8, 0.9]	0.6 [0.5, 0.7]	0.4 [0.3, 0.7]	0.5 [0.4, 0.7]	0.5 [0.4, 0.7]	0.4 [0.3, 0.6]	0.6 [0.5, 0.7]
Other single adult with child(ren) households	0.3 [0.3, 0.4]	0.6 [0.5, 0.6]	0.6 [0.4, 0.9]	0.6 [0.5, 0.7]	0.6 [0.5, 0.8]	0.5 [0.4, 0.7]	0.6 [0.5, 0.7]
Extended households	8.6 [8.5, 8.8]	9.5 [9.2, 9.8]	10.4 [9.5, 11.3]	9.7 [9.1, 10.3]	9.1 [8.6, 9.6]	9.4 [8.8, 10.1]	9.4 [9.1, 9.7]
Other households	1.2 [1.2, 1.3]	1.9 [1.8, 2.0]	2.0 [1.6, 2.4]	2.0 [1.7, 2.3]	2.2 [1.9, 2.4]	1.8 [1.5, 2.1]	2.0 [1.9, 2.2]
Total	100.0	100.0	100.0	100.0	100.0	100.0	100.0

Source: Centers for Disease Control and Prevention, the National Center for Health Statistics (2000–2010).

Note: Columns may not add to 100.0 due to rounding.

In addition, extended households, which accounted for 8.6% of all households with adults age 18–64, had higher percentages in the distributions of households containing adults with any disability (9.5%), requiring help with personal care (10.4%), requiring help with routine needs (9.7%), having difficulty remembering (9.4%), and having any work limitation (9.4%). Households consisting of other single adults with children, which accounted for 0.3% of all households with adults age 18–64, had higher percentages in all the distributions of households containing adults with various types of disability. On the other hand, households consisting of married or cohabiting parents with children and single mothers or fathers with children were lower in all the distributions of households containing adults with various types of disabilities compared to their percentages in the overall distribution of households with one or more adults age 18–64.

Table 4 shows distributions of household type by disability type for all households with one or more children age 3–17 years during 2000–2010. Again, the first column is the distribution of household type for all households with one or more children age 3–17; the remaining columns show the distributions of household type for all households with one or more children with disabilities (any disability or limitation, requiring help with personal care, and receiving special education or EIS). The percentages of households consisting of married parents with children, which accounted for 43.7% of all households with at least one child age 3–17, were lower in the distributions of households containing children age 3–17 with any disability (34.2%) and who received special education or EIS (34.6%).

Table 4.

Weighted Percent Distribution (and Confidence Intervals) of Household Type, for Households with One or More Members Age 3–17, by Disability Type: United States, 2000–2010

	All Households Having 1+ Members Age 3–17	Households Having 1+ Member Age 3–17 with Any Disability or Limitation	Households Having 1+ Members Age 3–17 Requiring Help with Personal Care	Households Having 1+ Members Age 3–17 Receiving Special Education or Early Intervention Services
Household Type	Percent (95% confidence interval)
Households consisting of 2 or more unrelated persons (e.g., roommates) or families	0.8 [0.7, 0.8]	1.0 [0.7, 1.5]	0.9 [0.3, 2.9]	1.0 [0.7, 1.5]
Married parents with child(ren) households	43.7 [43.2, 44.2]	34.2 [32.6, 35.8]	39.7 [34.1, 45.6]	34.6 [32.9, 36.4]
Cohabiting parents with child(ren) households	5.0 [4.8, 5.2]	5.5 [4.7, 6.3]	6.0 [3.9, 9.2]	5.6 [4.8, 6.5]
Single mother with child(ren) households	15.0 [14.7, 15.4]	20.6 [19.3, 21.9]	20.7 [16.3, 25.8]	19.9 [18.5, 21.3]
Single father with child(ren) households	3.1 [2.9, 3.3]	3.1 [2.6, 3.8]	1.7 [0.7, 4.1]	3.1 [2.5, 3.8]
Other single adult with child(ren) households	1.4 [1.3, 1.5]	2.3 [1.9, 2.8]	2.5 [1.2, 5.3]	2.1 [1.7, 2.7]
Extended households	27.0 [26.6, 27.5]	27.9 [26.4, 29.4]	24.4 [19.9, 29.5]	27.8 [26.3, 29.4]
Other households	3.9 [3.7, 4.1]	5.3 [4.6, 6.1]	4.1 [2.4, 6.9]	5.7 [4.9, 6.6]
Total	100.0	100.0	100.0	100.0

Source: Centers for Disease Control and Prevention, the National Center for Health Statistics (2000–2010).

Note: Columns may not add to 100.0 due to rounding.

However, single mother with children households, which accounted for 15% of households with at least one child age 3–17, had higher percentages in all the distributions containing at least one child with various disabilities. Other single adult with children households and other households, which accounted for 1.4% and 3.9%, respectively, of all households with at least one child, also had higher percentages in the distributions of households containing children age 3–17 with any disability (2.3% and 5.3%, respectively) and who received special education (2.1% and 5.7%, respectively). In contrast, the percentages of households consisting of single fathers with children having one of the disabilities in Table 4 were not different from the percentage of single fathers with children households in the overall household type distribution. The same was also true for households consisting of two or more unrelated persons (e.g., roommates) or families (containing children), cohabiting parents with children, and extended households.

Selected socioeconomic status characteristics of households

Households can, of course, vary by other characteristics in addition to household type. Accordingly, the left-most sub-column in Table 5 shows the overall distributions of poverty status for households with and without members with disabilities in 2000–2010, whereas the remaining sub-columns show distributions of poverty status for each household type. Patterns of association between disability status and poverty status were clearly apparent in the table. Overall, 20.7% of households having one or more members with disabilities were below the federal poverty threshold, compared with 10.4% of households having no members with disabilities. Conversely, 22.2% of households with one or more members with disabilities had household incomes that were four or more times the poverty threshold, compared with 42.6% of households not having any member with a disability. Poverty status varied even more so when disaggregated by household type. Among adult-only households with one or more members with disabilities, one-person households (31.2%) were much more likely to be below the poverty threshold than the remaining adult-only households. In particular, note that only 7.2% of married couple only households with one or more members with disabilities were below the poverty threshold. However, among adult-only households without any members with disabilities, 13.5% of one-person households and just 2.9% of married couple only households were below the poverty threshold.

Table 5.

Weighted Percent Distributions (and Confidence Intervals) of Poverty Status, by Household Members’ Disability Status and Household Type: United States, 2000–2010

	All House- hold Types	One- Person House- holds	House- holds consisting of 2 or More unrelated Persons (e.g., room- mates) or Families	Married Couple Only House- holds	Cohabiting Couple 0nly House- holds	All Other Related Adult House- holds	Married Parents with Child(ren) Households	Cohabiting Parents with Child(ren) House- holds	Single Mother with Child(ren) House- holds	Single Father with Child(ren) House- holds	Other Single Adult with Child(ren) House- holds	Extended House- holds	Other House- holds
Household Characteristic	Percent (95% confidence interval)
	Households having one or more members with disabilities
Poverty status relative to federal poverty thresholds
pov threshold < 1.00	20.7 [20.2, 21.2]	31.2 [30.3, 32.1]	n/a	7.2 [6.8, 7.7]	16.6 [14.7, 18.7]	12.9 [12.3, 13.5]	13.0 [12.2, 13.7]	30.7 [27.9, 33.7]	51.8 [50.0, 53.6]	26.9 [22.4, 31.8]	59.4 [54.7, 63.9]	23.3 [22.3, 24.4]	27.3 [24.9, 29.8]
1.00 to < 2.00 pov threshold	26.4 [26.0, 26.9]	32.2 [31.5, 32.9]		21.5 [20.8, 22.3]	26.3 [24.0, 28.7]	24.7 [23.9, 25.6]	21.0 [20.1, 21.9]	31.8 [29.0, 34.6]	27.7 [26.0, 29.5]	27.0 [23.0, 31.4]	27.1 [23.0, 31.6]	27.6 [26.5, 28.8]	33.0 [30.5, 35.7]
2.00 to < 4.00 pov threshold	30.6 [30.2, 31.1]	25.5 [24.7, 26.2]		37.9 [37.1, 38.7]	30.1 [27.7, 32.7]	33.0 [32.1, 33.9]	34.5 [33.5, 35.6]	26.4 [23.8, 29.2]	15.8 [14.5, 17.1]	31.2 [26.8, 36.0]	11.6 [8.9, 15.1]	30.6 [29.3, 31.9]	27.1 [24.7, 29.6]
4.00 and above pov threshold	22.2 [21.8, 22.7]	11.2 [10.6, 11.8]		33.4 [32.5, 34.3]	27.0 [24.7, 29.4]	29.5 [28.5, 30.4]	31.5 [30.4, 32.6]	11.1 [9.3, 13.2]	4.8 [4.1, 5.6]	14.9 [11.7, 18.9]	1.9 [1.0, 3.8]	18.4 [17.4, 19.5]	12.6 [10.6, 15.0]
Total	100.0	100.0		100.0	100.0	100.0	100.0	100.0	100.0	100.0	100.0	100.0	100.0
	Households having no members with disabilities
Poverty status relative to federal poverty thresholds
pov threshold < 1.00	10.4 [10.2, 10.7]	13.5 [12.9, 14.0]	n/a	2.9 [2.8, 3.1]	6.1 [5.4, 6.7]	6.3 [5.9, 6.6]	6.9 [6.7, 7.2]	19.0 [17.7, 20.3]	34.8 [33.7, 35.9]	10.9 [9.6, 12.4]	35.6 [31.8, 39.6]	13.6 [13.0, 14.1]	17.7 [16.0, 19.4]
1.00 to < 2.00 pov threshold	16.2 [16.0, 16.4]	19.2 [18.8, 19.6]		9.2 [8.8, 9.6]	12.0 [11.2, 12.9]	12.9 [12.5, 13.4]	14.8 [14.4, 15.2]	26.6 [25.2, 28.1]	28.8 [27.8, 29.7]	21.7 [19.8, 23.7]	33.7 [29.9, 37.7]	22.6 [21.8, 23.3]	26.0 [24.1, 28.0]
2.00 to < 4.00 pov threshold	30.7 [30.4, 31.0]	33.4 [33.0, 33.9]		26.1 [25.6, 26.6]	29.0 [27.8, 30.2]	29.4 [28.6, 30.1]	32.5 [31.9, 33.1]	34.0 [32.5, 35.5]	26.1 [25.1, 27.1]	38.4 [36.0, 40.9]	22.8 [19.6, 26.3]	34.6 [33.6, 35.5]	33.0 [30.9, 35.2]
4.00 and above pov threshold	42.6 [42.2, 43.1]	33.9 [33.3, 34.5]		61.8 [61.2, 62.4]	53.0 [51.6, 54.3]	51.5 [50.6, 52.3]	45.9 [45.1, 46.6]	20.4 [19.1, 21.8]	10.4 [9.6, 11.1]	29.0 [26.7, 31.4]	8.0 [6.1, 10.4]	29.3 [28.5, 30.2]	23.3 [21.3, 25.5]
Total	100.0	100.0		100.0	100.0	100.0	100.0	100.0	100.0	100.0	100.0	100.0	100.0

Source: Centers for Disease Control and Prevention, the National Center for Health Statistics (2000–2010).

Notes: Columns may not add to 100.0 due to rounding.

pov = poverty.

Among households with children, married parents with children and at least one member with a disability (13%) were much less likely to be below the poverty threshold than the remaining households containing children and at least one member with a disability. Again, though, married parents with children households without any members with disabilities were even better off economically: only 6.9% of these households were below the poverty threshold.

We obtained similar findings with regard to household education. Table 6 shows the education of the adult in the household with the most education by household disability status and household type. As with Table 5, the left-most sub-column shows the overall distributions of household education for households with and without members with disabilities in 2000–2010, whereas the remaining sub-columns show distributions of household education for each household type. Among households having one or more members with disability, 16.5% had no adults with a high school diploma or more, compared with only 8.5% of households having no members with disabilities. Similarly, 22.6% of households having one or more members with disability contained at least one adult with a college degree, compared with 37.8% of households without any members with disabilities. When we took into account household type, greater variation in household education became apparent. Among households having one or more members with disabilities, 36.3% of married parents with children households included at least one adult with a college degree compared with 11.5% of cohabiting parents with children households, 12.3% of single mother with children households, 13% of single father with children households, 6.2% of other single adult with children households, 23.8% of extended households, and 16% of other households. Among households not having members with disabilities, nearly 48% of married parents with children households included an adult with a college degree, compared with 14.5% of cohabiting parents with children households, 18.3% of single mother with children households, 24.7% of single father with children households, 12.2% of other single adult with children households, 30.5% of extended households, and 20.3% of other households.

Table 6.

Weighted Percent Distributions (and Confidence Intervals) of Household Education, by Household Members’ Disability Status and Household Type: United States, 2000–2010

	All House- hold Types	One- Person House- holds	House- holds Consisting of 2 or More Unrelated Persons (e.g., room- mates) or Families	Married Couple Only House- holds	Cohabiting Couple Only House- holds	All Other Related Adult House- holds	Married Parents with Child(ren) Households	Cohabiting Parents with Child(ren) House- holds	Single Mother with Child(ren) House- holds	Single Father with Child(ren) House- holds	Other Single Adult with Child(ren) House- holds	Extended House- holds	Other House- holds
Household Characteristic	Percent (95% confidence interval)
	Households having one or more members with disabilities
Education of adult with most education in household
Less than HS diploma	16.5 [16.2, 16.9]	29.4 [28.7, 30.2]	9.3 [7.6, 11.3]	11.8 [11.3, 12.4]	11.1 [9.7, 12.6]	9.9 [9.3, 10.4]	7.1 [6.6, 7.7]	14.2 [12.3, 16.2]	22.8 [21.5, 24.1]	19.6 [16.0, 23.8]	40.4 [36.3, 44.7]	8.7 [8.1, 9.3]	14.4 [12.7, 16.3]
GED/HS graduate	29.4 [29.0, 29.8]	30.9 [30.3, 31.6]	21.3 [19.0, 23.9]	30.5 [29.8, 31.3]	30.7 [28.4, 33.0]	30.0 [29.2, 30.9]	21.7 [20.9, 22.6]	38.0 [35.3, 40.8]	28.8 [27.4, 30.2]	33.6 [29.4, 38.1]	28.2 [24.4, 32.4]	29.1 [28.0, 30.2]	34.4 [31.9, 37.1]
Some college	31.6 [31.2, 32.0]	25.5 [24.8, 26.1]	42.9 [40.0, 45.9]	30.8 [30.1, 31.6]	35.4 [33.1, 37.9]	34.4 [33.5, 35.2]	34.8 [33.8, 35.9]	36.3 [33.7, 39.1]	36.2 [34.7, 37.8]	33.8 [29.6, 38.4]	25.1 [21.5, 29.1]	38.5 [37.4, 39.6]	35.2 [32.9, 37.7]
College degree (BA, BS)	22.6 [22.1, 23.0]	14.2 [13.6, 14.8]	26.5 [23.9, 29.2]	26.9 [26.0, 27.8]	22.9 [20.7, 25.2]	25.8 [24.9, 26.7]	36.3 [35.1, 37.5]	11.5 [9.7, 13.5]	12.3 [11.1, 13.5]	13.0 [10.2, 16.4]	6.2 [4.4, 8.8]	23.8 [22.7, 24.9]	16.0 [14.1, 18.1]
Total	100.0	100.0	100.0	100.0	100.0	100.0	100.0	100.0	100.0	100.0	100.0	100.0	100.0
	Households having no members with disabilities
Education of adult with most education in household
Less than HS diploma	8.5 [8.3, 8.7]	12.1 [11.7, 12.5]	5.7 [4.9, 6.5]	5.0 [4.7, 5.2]	4.6 [4.2, 5.1]	5.5 [5.2, 5.8]	6.5 [6.2, 6.8]	12.7 [11.6, 13.9]	16.0 [15.3, 16.7]	11.4 [10.0, 13.0]	29.0 [25.9, 32.3]	9.7 [9.2, 10.1]	13.9 [12.6, 15.4]
GED/HS graduate	22.9 [22.6, 23.3]	26.7 [26.2, 27.3]	13.0 [11.6, 14.5]	22.5 [22.0, 23.1]	20.2 [19.2, 21.2]	21.2 [20.6, 21.8]	17.1 [16.7, 17.5]	34.1 [32.6, 35.6]	29.0 [28.1, 29.9]	33.2 [31.0, 35.4]	31.9 [28.6, 35.5]	22.9 [22.2, 23.6]	30.8 [28.8, 32.8]
Some college	30.8 [30.5, 31.1]	29.8 [29.2, 30.4]	45.3 [42.4, 48.2]	27.0 [26.5, 27.5]	32.1 [30.9, 33.3]	34.9 [34.2, 35.5]	28.5 [28.0, 29.0]	38.7 [37.2, 40.3]	36.7 [35.8, 37.6]	30.7 [28.7, 32.9]	26.9 [23.9, 30.2]	37.0 [36.2, 37.7]	35.0 [33.0, 37.2]
College degree (BA, BS)	37.8 [37.4, 38.3]	31.4 [30.8, 32.1]	36.1 [33.7, 38.5]	45.5 [44.8, 46.2]	43.2 [41.8, 44.6]	38.4 [37.6, 39.2]	47.9 [47.2, 48.7]	14.5 [13.4, 15.8]	18.3 [17.5, 19.1]	24.7 [22.7, 26.8]	12.2 [9.7, 15.3]	30.5 [29.7, 31.3]	20.3 [18.4, 22.2]
Total	100.0	100.0	100.0	100.0	100.0	100.0	100.0	100.0	100.0	100.0	100.0	100.0	100.0

Source: Centers for Disease Control and Prevention, the National Center for Health Statistics (2000–2010).

Notes: Columns may not add to 100.0 due to rounding.

HS = high school.

Limitations of the data

The NHIS obtains information from respondents via an interviewing process that averages about an hour. Information collected during the FHS portion of the interview is obtained from a family respondent who stated that he or she was knowledgeable about other family members’ health status, disabilities, and limitations. However, the NHIS interviewer has no way of verifying whether this respondent is knowledgeable. In addition, respondents may experience recall problems that could result in inaccurate responses. As with all surveys, respondents may simply underreport characteristics or conditions that they consider undesirable.

In addition, the NHIS is a cross-sectional survey that does not obtain retrospective information on marital history, union formation, or transition into or out of particular types of living arrangements. Thus, household type, as measured in this analysis, reflects the living arrangements of household members at the time of interview. Also, all relationships within NHIS families or households are recorded relative to a reference person, so while we know the relationship of other household members to this person, it is not always possible to know, particularly in the case of more complicated families, how all family members are related to one another. Last, though we can identify households containing members with disabilities, we have no way of knowing whether or to what extent these disabled members actually received help from others within the household; we can only assume that the potential for such help existed. Nor do we have any information whatsoever regarding assistance (either formal or informal) that disabled family members may have received from neighbors, friends, or family who lived nearby but in a different residence than the family interviewed for the NHIS.

Discussion

Most, but not all, previous family-level studies of disability have omitted individuals living alone or with unrelated other persons because they were not considered families (that is, two or more persons related by blood, marriage or adoption who reside together). According to our estimates, roughly one-third of all households may be excluded by such an approach. Moreover, by excluding persons who are living alone or with nonrelatives such as roommates or unmarried partners, family-level analyses of living arrangements, community participation, or caregiving of persons with disabilities overlook the proportion of the population with disabilities that may be more susceptible to the impact of changes in the availability of services and more dependent on services provided by publicly funded entities. Persons with disabilities live in a wide variety of household settings, especially after the Olmstead Decision of the U.S. SupremeCourt (Olmstead v. L.C., 1999), which has enabled persons with disabilities to move out of institutions and into living arrangements with related or unrelated persons. Family-level studies of disability that exclude nonfamily households will not be able to describe the extent of the diversity in the living arrangements of this population, particularly as persons with disabilities leave or seek to avoid institutional settings.

Our findings show that while approximately 25.6% of all households in the United States had at least one member with a disability, prevalence of disability varied by household type. Moreover, we found that living arrangements varied according to the type of disability and age of household members (see also Waite & Hughes, 1999). Relative to their percentages in the overall distribution of household type for all households with one or more elderly members, one-person households consisting of elderly persons with any disability who required help with routine needs or had difficulty walking without special equipment were higher, whereas one-person households consisting of elderly persons who required help with personal care or had difficulty remembering were lower. These results are not surprising because there are a variety of special devices available to assist with mobility issues, and there are also easier ways to fulfill routine needs than were previously possible (shopping, banking, and paying bills can now be done on-line, for example). However, no such special equipment is available to assist with memory or cognitive difficulties, and not being able to eat, dress, bathe, or use a bathroom without assistance is generally not possible without the partial presence of others for portions of the day.

All other related adult households containing elderly adults with disabilities were also higher than their percentages in the overall distribution of household type in all the disability distributions that we examined, whereas married couple households contained fewer elderly members with disabilities (this was true for every type of disability or limitation that we examined). This may occur because older married couple households enlarged to include more family members – that is, became households consisting of either other related adults or extended family (including three or more generations) – as one or both spouses became more disabled and needed help from co-resident family members. In terms of Waite and Hughes (1999), it is an example of a less complex household becoming more complex. Although greater household complexity may be associated with more familial stress and fewer resources and assets, as Waite and Hughes argued, we suggest that larger and more complex households may also be advantageous simply because they have more members on hand to provide care to members with disabilities, thus reducing the possibility of institutionalization. Additionally, in terms of resources, we found that only 12.9% of all other related adult households with at least one member with a disability were below the federal poverty threshold in the previous calendar year (only married couple households with one or more members with disabilities had a lower percentage); all other related adult households were also comparatively “well off” with respect to household education.

As with elderly adults with disability in one-person households, the percentages of one-person households consisting of persons age 18–64 with various disabilities (requiring help with routine needs, difficulty walking, difficulty remembering, or work limitations) were all higher than their percentage in the overall distribution of household type for all households with at least one adult in this age group. However, unlike married couple households containing elderly adults with disabilities, households with married couples containing one or more members age 18–64 with any disability or limitation, who required help with personal care, had difficulty walking without special equipment, or had any work limitation were higher than their percentage in the overall distribution of household type for all households with at least one adult member. In addition, compared to their percentages in the overall distribution of household type for all households, all other related adult households, other single adult with children households, and extended households were higher in all – or, in the case of extended households, nearly all – the distributions containing one or more adults age 18–64 with any type of disability we examined. On the other hand, married or cohabiting parents with children households as well as single mothers or fathers with children households were less likely to have adults age 18–64 with disability, given their percentages in the overall distribution of household type for all households with one or more adults age 18–64. The latter effect is almost certainly related to age: extended households and households consisting of other single adults with children are more likely to contain older adults (i.e., grandparents) whereas married and unmarried couples and single mothers and fathers raising children are more likely to be younger adults of childbearing age.

Regarding households containing children age 3–17 with disabilities, those consisting of married parents with children were less likely to contain children with any disability or children who received special education than their percentages in the overall distribution of household type for all households with children in this age group. On the other hand, households consisting of single mothers and children with any disability, who required help with personal care, or received special education were higher than their percentage in the overall distribution of household type for all households with children. Other households (which included children being raised by relatives other than their parents or by nonrelatives in the case of foster children) and households consisting of other single adults with children were also more likely to have children with any disability or who received special education than their percentages in the overall distribution of household type for all households with children would suggest. Our results were therefore consistent with those of Cohen and Petrescu-Prahova (2006), who found that children with disabilities were significantly more likely than nondisabled children to live in non-nuclear families. Moreover, as other researchers have demonstrated, children with disabilities who live in non-nuclear families are more economically disadvantaged than children with disabilities in nuclear families (Fujiura, 1998; Hogan, 2012). We obtained consistent results: non-nuclear households with children and at least one member with a disability of any age were more likely to be below the federal poverty threshold in the previous year than nuclear households (i.e., married parents with children) and at least one member with a disability. Households consisting of single mothers or single adults with children that had at least one member with a disability were also likely to be particularly disadvantaged with respect to household education, especially when compared with nuclear households.

From a more general perspective, the Fujiura (2010) analysis of the ACS data makes an excellent case for family-level policy considerations. However, the omission of nonfamily households conceals the true size of the problem that challenges the system capacity and ignores the potentially greater risk faced by one-person households should disabilities worsen, programs change, or when emergency situations such as storms like Sandy or Katrina occur. Our analysis demonstrated that among one-person households containing an adult of any age with a disability, 31% had incomes under the poverty threshold and only 50% owned their homes (results not shown but available upon request). In contrast, higher percentages of married couple households, married couple with children households, and all other related adult households having members with a disability were above the poverty level, and higher percentages of these households owned their homes. By comparison, clearly, the most seriously disadvantaged households, as identified by Fujiura (1998) and Hogan (2012), were those in single mother with children or other single adult with children households, which were most likely to have incomes below the poverty line (51.8% and 59.4%, respectively). Moreover, only 29.1% of single mother with children households owned their homes (results not shown but available upon request).

In conclusion, researchers wishing to understand the living arrangements of persons with disabilities need to include both family and nonfamily households. Many, but not all, previous researchers have relied on a definition of family that omitted persons living alone or with nonrelatives, thus providing policy makers with an incomplete picture of the living arrangements of persons with disabilities. It is more important than ever that policymakers understand the complex relationship between living arrangements and disability, as both the federal government and many state governments have been enabling persons with disabilities to leave or avoid institutions and remain in the community.