Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2026 May 26.
Published in final edited form as: J Health Soc Behav. 2012 Mar;53(1):50–66. doi: 10.1177/0022146512439540

The Social Context of Disablement among Older Adults: Does Marital Quality Matter for Loneliness?

David F Warner 1,2, Jessica Kelley-Moore 1
PMCID: PMC13200694  NIHMSID: NIHMS2175964  PMID: 22382720

Abstract

Prior research has not adequately considered that disablement occurs within a web of relationships that provides socioemotional resources to and/or places demands on older adults. Drawing on the stress process and life course perspectives, we considered the social context of disablement by examining the influence of marital quality on the association between disability and loneliness among married older adults. Using data from the National Social Life, Health, and Aging Project, we found (1) functional limitations were associated with higher levels of loneliness; (2) neither positive nor negative marital quality mediated this association, contrary to the stress-deterioration hypothesis; and (3) positive (but not negative) marital quality moderated this association, consistent with the stress-buffering hypothesis. These associations were similar for women and men. Our findings indicate the importance of the social context of disablement, as interpersonal resources offer protection from the deleterious socioemotional consequences of disability.

Keywords: aging, disability, gender, life course, marital quality, marriage

Older Americans are at great risk of disabling health conditions, which—by challenging routine functioning (Korporaal, van Groenou, and van Tilburg 2008)—may result in a loss of autonomy and potentially hamper social relationships. As social ties are important for socioemotional well-being (Cohen 2004; Thoits 1995), functional limitations that threaten social relationships can be significant stressors in the lives of older adults and have a negative impact on their socioemotional well-being (Bookwala and Franks 2005; Booth and Johnson 1994). Consistent with this stress process perspective (Pearlin et al. 1981) prior studies find that disabled older adults report lower levels of well-being on numerous mental health outcomes, including depressive symptoms (Bierman and Statland 2010; Turner and Noh 1988; Yang 2006), self-esteem (Reitzes and Mutran 2006), and life satisfaction (Ducharme 1994). Yet this work has not adequately situated disablement in its social context—the web of long-term relationships such as marriage that provide resources to and/or place demands on persons over the life course. Despite disabled older adults’ higher levels of loneliness (Korporaal et al. 2008; Savikko et al. 2005), few studies have examined how the quality of the marital relationship affects loneliness for functionally limited older adults.

Therefore, the current study bridges the stress process (Pearlin et al. 1981) and life course perspectives (George 2003) to consider explicitly the social context of disablement and its consequences for socioemotional well-being. We examine the influence of positive and negative marital quality on the association between disability and loneliness using data on married older adults from the nationally representative National Social Life, Health, and Aging Project (NSHAP). We focus on married older adults because marriage organizes social roles and interactions for most persons across the adult life course. The quality of the marital relationship can vary over the years, but it is the product of accumulated events, experiences, and habituated social interaction between partners, making it the prime social context in which to examine the relationship between disablement and loneliness. A self-assessment of the inadequacy of one’s social connections, loneliness may be attenuated in marriages with positive coping strategies and a history of supportive interaction. Loneliness is a notable gauge of well-being, predicting self-rated health, cardiovascular disease, depression, use of health services, and mortality (Cacioppo, Hawkley, and Berntson 2003; Hughes et al. 2004; Savikko et al. 2005). A full understanding of the potential for loneliness among disabled older adults—and the role that marital quality may play—is important for explicating the social context of disability.

BACKGROUND

The Social Context of Disablement, Marriage, and Loneliness

Largely grounded in the stress process perspective (Pearlin et al. 1981), previous research has typically focused on mental health consequences of functional limitations and disablement, such as depression (Turner and Noh 1988; Yang 2006), self-esteem (Reitzes and Mutran 2006), and life satisfaction (Ducharme 1994). However, unlike these indicators of mental well-being, loneliness is a potentially adverse outcome of mounting functional limitations that is inherently embedded in one’s social world. Loneliness is an emotional state encompassing feelings of emptiness, isolation, and abandonment, and it is relatively independent of objective contact with others or social integration (Cornwell and Waite 2009; Savikko et al. 2005). It is an evaluative assessment of the correspondence between one’s need for social resources and the extent to which available resources are meeting those needs (Cornwell and Waite 2009; Hughes et al. 2004).

Loneliness may be an adverse outcome of mid- to late-life disablement because declines in functional ability and autonomy require significant changes in the structure and quality of social relations (Thompson and Heller 1990) that are important for overall well-being (Cohen 2004). Functionally limited older adults often curb participation in activities, experience lower levels of social support, and are less satisfied with their social interactions (Thompson and Heller 1990; Williamson and Schultz 1992; Yang 2006). As a result, the functionally limited older adult may have social and emotional needs that are unmet by the existing social connections. Indeed, prior studies consistently find that older adults in poor health, those with more health problems, and those who have more functional limitations are more likely to be lonely (Essex and Nam 1987; Korporaal et al. 2008; Savikko et al. 2005).

Situating disablement within its social context necessitates reconsidering the ways in which disability may lead to loneliness. As most older adults are married, any examination of the social context of disablement should begin with the disabled person’s spouse—the key figure in a married person’s social network, connecting him or her to others and being the most significant source of companionship and social support (Waite 2005; Waite and Lehrer 2003). Since older adults become disabled overwhelmingly due to chronic illness (Ferrucci et al. 1996), a couple has likely made adjustments over many years in response to the mounting disability of one spouse. By the time such a couple reaches late adulthood, they have established care routines, transferred some roles, and adapted their social outings. The spouse plays an even more central role in the disabled older adult’s social network because he or she is likely the primary caregiver (Spitze and Ward 2000). Consequently, the effect of disablement on socioemotional well-being generally, and loneliness in particular, cannot be considered outside the context of the marital relationship.

Despite the centrality of the spouse for social connectedness and support, few studies have examined the importance of the quality of the marital relationship for understanding how disability affects loneliness (Essex and Nam 1987; Korporaal et al. 2008). Marital quality may be important because simply being married does not necessarily protect against loneliness among the disabled (de Jong Gierveld et al. 2009; Russell 2009). Rather, studies suggest that better marital quality is associated with better mental health overall (Waite 1995). Moreover, the salience of the marital relationship increases with age (Choi and Marks 2008; Umberson et al. 2006), and thus the influence of marital quality on socioemotional well-being may be especially important for older married adults who are simultaneously at greater risk of disability. It is therefore likely that it is not just marriage but having a better quality marriage—one characterized by emotional support, companionship, and a spouse with whom one can discuss problems—that should be associated with lower levels of loneliness among disabled older adults (de Jong Gierveld et al. 2009; Stevens and Westerhof 2006).

Disablement and the Role of Marital Quality: Competing Hypotheses

Marital relationship as susceptible.

Prior studies examining the marital context of disability have largely done so consistent with the stress process perspective’s stress-deterioration hypothesis (Ensel and Lin 1991; Pearlin et al. 1981). This hypothesis argues that disablement is a chronic stressor, disrupting a person’s identity and sense of self-worth, negatively affecting his or her social relationships and consequently his or her health. The effects of disablement on the marital relationship are expected to be pronounced given the primacy of the marital role and because the deleterious effects of stress are likely to be greater when more salient roles are challenged by chronic, diffuse strains such as disability (Thoits 1995).

The stress-deterioration hypothesis posits that marital quality mediates the link between disability and loneliness among older adults (Essex and Nam 1987; Korporaal et al. 2008). Poor health, particularly with respect to mobility limitations, may lead to changes in marital dynamics, including declines in shared leisure activities, social support exchange, and sexual intimacy (Booth and Johnson 1994; Korporaal et al. 2008), and increases in instrumental care provided by one’s spouse (Hafstrom and Schram 1984). Unable to carry on their spousal relationship satisfactorily, disabled older adults may experience diminished marital quality and declines in socioemotional well-being (Cutrona 1996), including increased loneliness (Korporaal et al. 2008).

There is mixed support for the stress-deterioration hypothesis that disablement negatively influences well-being by undermining marital quality. Several studies find that marital quality, and perceived support more generally, mediates the relationship between disability and depressive symptoms (Bookwala and Franks 2005; Taylor and Lynch 2004), but others have not found such mediation between disability and depression (Choi and Marks 2008) or life satisfaction (Ducharme 1994). However, since loneliness indicates an underlying social process, there are limits to what we can learn from mental health outcomes such as depression. Indeed, the findings are inconclusive from the few studies that have explicitly considered whether declines in marital quality link disability and loneliness (Essex and Nam 1987; Korporaal et al. 2008).

Marital relationship as resource.

Conceptualizing marriages as susceptible to deterioration in the face of stressors such as disability neglects the fact that marriages have histories, notably characterized by the accumulation of shared experience and routinized interactions. Viewed from a life course perspective, marital quality is a relatively stable property of long-term relationships (Johnson, Amoloza, and Booth 1992; Kamp Dush, Taylor, and Kroeger 2008), reflecting habituated social interactions, tested coping strategies, mutual investments, and dynamically constructed narratives developed between spouses over the length of their marriage. Later-life disablement is just one of many challenges faced by the marital dyad, likely sparking routinized patterns of coping developed in response to previous stressors. Reports of marital quality among older adults represent the perceived availability of a range of resources, both enacted and in reserve. Therefore, disablement should not deteriorate marital quality; rather, marital quality should buffer the socioemotional challenges that accompany functional limitations in later life.1

Thus, we turn to an alternative hypothesis, also grounded in the stress process perspective: The stress-buffering hypothesis postulates that access to supportive relationships, such as high-quality marriages, can ameliorate the effects of illness and disability on psychological well-being; conversely, the absence of supportive relationships may exacerbate the negative effects of illness (Cohen 2004; Ensel and Lin 1991; Pearlin et al. 1981). Such a buffering effect is especially likely within older adults’ marriages because these marriages tend to be responsive to the social needs of disabled older adults due to the accumulated interpersonal resources and shared history of the marital dyad. Social support and companionship provided by the spouse is one of the key mechanisms by which marriage promotes positive physical and mental health (Waite 2005; Waite and Lehrer 2003), with prior studies demonstrating that perceived social support from intimate relationships is associated with better mental health (Thoits 1995; Turner and Noh 1988; Umberson et al. 1996; Yang 2006). The perceived availability of a confidant (e.g., a spouse) and satisfaction with his or her support (e.g., high marital quality) may buffer the socioemotional consequences of disablement. Having an unsupportive spouse or a marital relationship marked by low levels of marital happiness or above average levels of negative interaction is associated with worse mental health (Choi and Marks 2008; Hawkins and Booth 2005; Kamp Dush et al. 2008; Umberson et al. 2006) and greater loneliness (de Jong Gierveld et al. 2009; Stevens and Westerhof 2006).

Few studies have tested the stress-buffering hypothesis that marital quality mitigates threats to socioemotional well-being among disabled older adults. Although not examining marital relationships specifically, prior research indicates that the availability of a confidant and perceived emotional support from family can ameliorate depressive symptoms and loneliness among older disabled adults (Bierman and Statland 2010; Thompson and Heller 1990; but see Yang 2006). Bookwala and Franks (2005) demonstrated the moderating effect of marital quality in the relationship between disability and depression, finding that negative marital quality exacerbated depressive symptoms among disabled adults older than 60. Positive marital quality, although associated with depression, did not mitigate the effects of disability on depression. Bookwala and Franks’s findings are generally consistent with prior research that negative marital processes have stronger effects than positive processes (Carr and Springer 2010). However, again, we can make only a tenuous generalization of findings about depression to loneliness since these constructs do not share the same underlying process.

We posit that marital quality may be an important buffer from the feelings of social isolation and loneliness that accompany health-related declines— a hypothesis not often considered by prior research, which has not situated both disablement and marriage in the life course, ignoring that the long-term marital relationship is characterized by routinized coping mechanisms and a shared history. This proposition is much more consistent with a moderating relationship for marital quality. Persons in high-quality marriages may continue to receive support (or even additional support) from their spouses in the face of physical disability, lessening the resulting negative consequences; persons in low-quality marriages, however, may not receive adequate levels of support from their spouses, worsening the effects of physical disability.

Older adults in low-quality marriages may be especially at risk for loneliness because prior to disablement these persons may have had alternative (nonspousal) sources of social support. However, older adults in low-quality marriages could be more dependent on their spouses to meet their social needs once disabled. To our knowledge, the current study is the first to examine explicitly whether assessments of marital quality moderate the association between disability and loneliness.

Gender, the Social Context of Disablement, and Loneliness

A life course perceptive on marriage and health highlights the importance of gender for understanding how marriage affects loneliness among the disabled. Men and women have access to different resources, confront different constraints, and frequently have dissimilar experiences within the same institutions (Moen 2001). Much research indicates that men and women have different health experiences within marriage, with men benefiting more from marriage in terms of physical health, mental health, and mortality (Waite 1995; but see Williams 2003). If marital quality moderates the relationship between disablement and loneliness, do wives and husbands reap the same benefits?

Prior studies provide contradictory indications into the potential gendering of the social context of disablement. For instance, men may be less lonely in the face of disablement because they receive more health and social benefits from marriage. Men have smaller networks and are more likely than women to rely exclusively on their spouses for both instrumental and emotional support and connections to others, especially kin (Antonucci and Akiyama 1987; Spitze and Ward 2000). Moreover, wives tend to be more responsive to the emotional and support needs of their husbands, especially when they have chronic illnesses (Hafstrom and Schram 1984), than are husbands to the needs of their wives (Neff and Karney 2005).

However, women are more responsive to the quality of their marriages than are men (Cutrona 1996; Essex and Nam 1987; although see Umberson et al. 1996), particularly negative marital functioning (Almeida and Kessler 1998), so they may be at greater risk of loneliness when disablement occurs. This evidence suggests that any moderating effect of marital quality on the relationship between disability and loneliness may be stronger for women than men. The ambiguity as to how the social context of disablement may be gendered with respect to loneliness is reinforced by the fact that few prior studies have examined these associations with an eye toward gender.

Research Questions

Drawing on the stress process (Pearlin et al. 1981) and life course perspectives (George 2003), the current study advances our understanding of the social context of disablement by examining how marital quality affects loneliness among disabled older adults. To that end, we address three broad research questions:

  1. Consistent with the stress-deterioration hypothesis, does marital quality mediate the association between functional limitations and loneliness?

  2. Consistent with the stress-buffering hypothesis, does marital quality moderate the association between functional limitations and loneliness?

  3. Do the associations between functional limitations, loneliness, and marital quality vary for married women and men?

We expect that contrary to the stress-deterioration hypothesis, marital quality does not mediate the association between disability and loneliness but rather moderates that association in line with the stress-buffering hypothesis. Due to contradictory evidence about the health benefits of marriage and marital quality, we have no clear expectation as to whether (and in what direction) the associations between disability, loneliness, and marital quality differ by gender.

DATA AND METHODS

We used data from the NSHAP (Waite et al. 2007), a nationally representative sample of 3,005 community-dwelling individuals ages 57 to 85 in the contiguous United States. The final weighted response rate was 75.5 percent. Data were collected during two-hour in-home interviews. To minimize respondent burden, NSHAP also used a modularized design so that some questions were included in a leave-behind questionnaire. The return rate for the leave-behind questionnaire was 84 percent (O’Muircheartaigh, Eckman, and Smith 2009). Respondents’ spouses were not interviewed.

Analytic Sample

Our analytic sample was restricted in several ways. First, we limited the analyses to respondents with valid scores on the UCLA Short Loneliness Scale (described below), which was used in the leave-behind questionnaire. We excluded 481 respondents who did not return the questionnaire. We excluded an additional 139 respondents without complete data on the UCLA items. Altogether, we excluded 20.6 percent of the total sample due to missing information on the dependent variable.2 Second, we limited our analyses to married or cohabiting respondents,3 excluding an additional 863 respondents, about one third of the total sample.4 Last, we excluded 22 respondents missing information on at least one of our explanatory variables. Our analytic sample comprised 1,500 married older adults.

Our analytic sample is representative of older men and women (not couples) in long-term first marriages. Respondents were on average 67 years old—respondent age did not significantly differ by gender (not shown). More than two thirds of respondents were in their first marriages. Respondents had been married on average almost 37 years, with those in their first marriages having average marital durations of 44.29 years and those in second and third (or higher order) marriages having average marital durations of 22.81 and 17.01 years, respectively. Women were married about 2 years longer than men on average (37.96 vs. 36.05 years, respectively, t = 2.43, p < .05). The gender difference in marital duration was driven solely by first marriages, as marital durations in remarriages did not significantly differ by gender (not shown). There was no significant difference in the proportion of men and women cohabiting. Overall, the marital composition of the NSHAP data reflects the well-known gender disparities in mortality and remarriage (Waite 1995), with men comprising 56 percent of the respondents in our analytic sample of married older adults.4

Measures

The dependent variable in our analysis is loneliness, measured with the UCLA Short Loneliness Scale, a three-item summated rating scale validated for use in surveys of older persons (Hughes et al. 2004). Respondents were asked to indicate how frequently they felt they lacked companionship, were left out, and were isolated from others. For each question, responses included hardly ever (or never), sometimes, and often (coded 0–2). The summated scale had a range of 0 to 6 and demonstrated good internal consistency (α = .82).5

The primary explanatory variables in our analysis were disability and marital quality. We measured disability with functional limitations, a count of reported difficulty with seven standard activities of daily living expected to last at least three months: walking one block, walking across a room, dressing, bathing or showering, eating, getting in and out of bed, and using the toilet. For each task, respondents indicated whether they had 0 = no difficulty, 1 = some difficulty, or 2 = much difficulty or whether they were 3 = unable to do. As these indicators were highly right skewed, we dichotomized each item so that reports of any difficulty were coded 1, and we summed the seven items for a range of 0 to 7, where higher scores indicated more limitations (KR20 = .81). Approximately 30 percent of the respondents had at least one functional limitation.6 Preliminary analyses indicated that the sum of dichotomous indicators provided a better model fit than alternative coding schemes.

To measure marital quality, we drew on six indicators largely corresponding to the positive and negative aspects identified in prior research (Bookwala and Franks 2005; Johnson et al. 1992). Four items asked about the frequency with which respondents could “talk about. .. worries” with their spouses, could “rely on [their spouses] if [they] have problems,” whether their spouses made “too many demands,” and whether their spouses “criticize[d]” them. Responses were coded 1 = hardly ever (or never), 2 = some of the time, and 3 = often. A fifth item asked whether respondents and their spouses “spend free time doing things together, or doing things separately,” with responses of 1 = together, 2 = some together, some different, and 3 = different/separate things, reverse coded so that higher scores indicated more time spent together. The sixth item was a global assessment of marital happiness, “taking all things together,” where responses ranged from 1 = very unhappy to 7 = very happy. We recoded this item to 1 = Unhappy (1, 2, 3, 4), 2 = Happy (5, 6), and 3 = Very Happy (7) to adjust for the left skew of original responses and to obtain consistent response categories across all six measures.

We subjected these six items to exploratory factor analysis with the principal factor method used to extract the factors. A scree test (Hatcher 1994) suggested two meaningful marital quality factors; accordingly, we retained two factors for oblique rotation. In interpreting the rotated factors, we considered an item with a loading of .35 or greater significant. We designated factor 1 as positive marital quality (α = .62) and factor 2 as negative marital quality (α = .60). The factor pattern loadings and the factor structure correlations suggested factorial complexity with the marital happiness item, and thus we included it in the reliability assessment of both factors. The interfactor correlation was −.54. We included estimated factor scores (Hatcher 1994) as explanatory variables in our models. Alternative solutions did not fit the data as well. Table 1 summarizes the results for the two-factor solution.

Table 1.

Exploratory Factor Analysis of Marital Quality among Married Older Adults: Items and Corresponding Factor Loadings from the Rotated Oblique Factor Pattern Matrix and Factor Structure Matrix for Two-Factor Solution (n = 1,500)

Factor Patterna Factor Structureb Questionnaire Itemc
1 2 1 2
.35* −.05 .38* −.24 1. “Some couples like to spend their free time doing things together, while others like to do different things in their free time. … Do you [and partner] like to spend free time doing things together, or doing things separately?”
.60* .07 .56* −.25 2. “How often can you open up to [partner] if you need to talk about your worries?”
.55* −.01 .56* −.31 3. “How often can you rely on [partner] for help if you have a problem?”
−.04 .55* −.34 .57* 4. “How often does [partner] make too many demands on you?”
.04 .57* −.27 .55* 5. “How often does [partner] criticize you?”
.42* −.24 .55* −.47* 6. “Taking all things together, how would you describe your [marriage/relationship] with [partner] on a scale from … very unhappy [to] very happy?”
Open in a new tab

Source: National Social Life, Health, and Aging Project.

Note: The six items were subjected to exploratory factor analysis with the principal-factor method used to extract the factors. Two factors were retained for oblique rotation, and items with loadings of .35 or greater were considered meaningful. Factor 1 was designated as “positive marital quality” (α =.62), and factor 2 was designated as “negative marital quality” (α = .60). The factor pattern loadings and the factor structure correlations suggested factorial complexity with item 6, and accordingly it is included in the reliability assessment of both factors. The interfactor correlation was r = −.54. Standardized scoring coefficients were used to calculate estimated factor scores (Hatcher 1994).

a

Factor pattern loadings are standardized regression coefficients for the unique contribution of each latent factor to the observed indicator; loadings greater than .35 are designated with asterisks.

b

Factor structure correlations between the observed indicators and the latent factors; correlations greater than .35 are designated with asterisks.

c

See text for details on response options and coding for each item.

We included a number of measures of demographic characteristics, socioeconomic status, and social integration as controls in the analysis (Bookwala and Franks 2005; Hughes et al. 2004; Korporaal et al. 2008; Savikko et al. 2005). Demographic characteristics included age, female, race/ethnicity, whether the respondent was cohabiting, and the number of times the respondent was previously married.7 We measured socioeconomic status with indicators for education and household income. Although objective and subjective indicators of social isolation are independent (Cornwell and Waite 2009), we included measures of social integration to control for potential mediators between disability and loneliness (Thompson and Heller 1990; Williamson and Schultz 1992), including whether the respondent was working for pay, frequency of religious service attendance, and network size. Preliminary analyses also included measures for household wealth and various specifications of the number of living children (including separate measures for sons and daughters); however, the findings were unchanged, so we excluded these measures from the final models. In the interest of space, we present the full list of variables, coding, and descriptive statistics in Table 2.

Table 2.

Model Variables, Coding, and Descriptive Statistics: Analytic Sample of Married Older Adults (n = 1,500)

Variable Description and Coding M a SD a r b
Dependent variable
 Loneliness UCLA Short Loneliness Scale. Summated score of three items assessing the frequency that respondent felt “lack [of] companionship,” “left out,” and “isolated from others” with responses to each of 1 = “hardly ever (or never),” 2 = “some of the time,” and 3 = “often” (α = .82). Items were recoded by subtracting 1 (range = 0 to 6). .75 1.29 1.00
Independent variables
 Functional limitations Count of any difficulty with seven activities of daily living, including walking one block, walking across a room, dressing, bathing, eating, getting and out of bed, and using the toilet (KR-20 = .81; range = 0 to 7). .70 1.51 .07**
 Marital quality
  Positive marital quality Estimated factor score; see Table 1 for description (α = .62; range = −3.42 to .69). .00 .77 −.40***
  Negative marital quality Estimated factor score; see Table 1 for description (α = .60; range = −.75 to 2.52). .00 .73 .36***
Control variables
 Age Age of respondent, mean centered (range = −9.94 to 18.06). .00 7.55 −.05†
 Female Female (1 = yes, 0 = otherwise). .44 — .05*
 Race/ethnicity
  Whitec Non-Hispanic white (1 = yes, 0 = otherwise). .85 — −.10***
  Black Black (1 = yes, 0 = otherwise). .07 — .12***
  Hispanic Hispanic (1 = yes, 0 = otherwise). .06 — −.01
  Other Other race/ethnicity (1 = yes, 0 = otherwise). .02 — .06*
Cohabiting Relationship legal status (1 = currently cohabiting, 0 = currently married). .03 — −.02
Times previously married Number of times previously married (range = 0 to ≥2). .41 .68 .08**
Education
 Less than high school Less than high school (1 = yes, 0 = otherwise). .13 — .06*
 High schoolc High school or equivalent (1 = yes, 0 = otherwise). .25 — .01
 Some college Some postsecondary education (1 = yes, 0 = otherwise). .33 — .03
 College and beyond Four-year degree or more (1 = yes, 0 = otherwise). .29 — −.08**
Income
 <$25,000 Household income of <$25,000 last year (1 = yes, 0 = otherwise). .14 — .08**
 $25,000 to <$50,000c Household income of $25,000 to <$50,000 last year (1 = yes, 0 = otherwise). .28 — .04
 $50,000 to <$100,000 Household income of $50,000 to <$100,000 last year (1 = yes, 0 = otherwise). .31 — −.09***
 ≥$100,000 Household income of ≥$100,000 last year (1 = yes, 0 = otherwise). .18 — −.04
 Income missing Income missing (1 = yes, 0 = otherwise). .08 — .04†
Social integration
 Working Working for pay (1 = yes, 0 = otherwise). .38 — .01
 Religious service attendance Frequency attended religious services in the past 12 months; responses coded from 0 = “never” to 6 = “a few times per week.” 3.30 2.25 −.04†
 Network size Number of persons in respondent’s discussion network, excluding spouse (range = 0 to ≥5) 2.77 1.60 −.06*
Open in a new tab

Source: National Social Life, Health, and Aging Project.

Note: α = Cronbach’s α for internal reliability of continuous measures; KR-20 = Kuder-Richardson Formula 20 for internal reliability of dichotomous measures. All estimates were weighted to account for differential probabilities of selection and differential nonresponse.

a

Mean is equivalent to the proportion coded 1 for dummy variables (standard deviations are omitted for dummy variables).

b

Correlation between variable and UCLA Short Loneliness Scale.

c

Serves as the reference category in multivariate analyses.

†

p < .10.

*

p < .05.

**

p < .01.

***

p < .001 (two-tailed tests).

Analytic Strategy

To examine the relationship between disability, marital quality, and loneliness among married older adults, we employed Tobit regression models to account for the censored nature of our dependent variable (Long 1997): 63 percent of respondents scored 0, 23 percent scored 1 or 2, and the remaining 14 percent scored 3 or higher (3–6) on the UCLA Short Loneliness Scale (not shown). In the Tobit model framework, this clustering occurs because the measured categories are not detailed enough to detect variation in the unobserved underlying continuous latent loneliness construct, and consequently some responses are censored at 0—having hardly ever or never experienced loneliness. This response includes persons who truly never experienced loneliness and those who experienced loneliness so infrequently that they did not categorize their experience as sometimes. Failure to recognize the censored nature of the dependent variable by using ordinary least squares techniques yields downwardly biased parameter estimates (Long 1997).

Our Tobit regression models provide maximum likelihood estimates of the theoretically continuous and normally distributed underlying loneliness construct (yi), where the observed UCLA loneliness measure (yi*) is censored at zero (τ):

yi=yi*=xi𝛃+εiifyi*>ττyifyi*≤τ (1)

For cases greater than the censoring value 0, the structural equation in the Tobit model estimates the effect of a vector of covariates (xiβ) on the observed loneliness measure (yi*) using a standard linear model, while for cases at or below the censoring value 0, the probability of being censored is estimated, and this quantity is used in the likelihood equation. While Tobit models provide several outcomes of potential interest, we focused on changes in the latent outcome yi*, where the effect of any covariate may be interpreted in a linear fashion (Long 1997).

We conducted our analysis in three steps corresponding to our research questions. First, we estimated a zero-order model to establish the baseline association between disability and loneliness, and then we added the measures of marital quality to ascertain whether marital quality mediated that association as predicted by the stress-deterioration hypothesis. Second, we examined whether marital quality moderated the association between disability and loneliness as predicted by the stress-buffering hypothesis. To test for the moderating effect, we specified an interaction term between each measure of marital quality and disability (e.g., positive marital quality × functional limitations). Finally, we explored whether the associations between disability, marital quality, and loneliness elucidated in the prior two steps differed for men and women. To do this, we reestimated the previous models stratified by gender and tested for statistically significant differences between the estimated effects for women and men using a t-test for equality of regression coefficients (Clogg, Petkova, and Haritou 1995). All estimates were weighted to account for differential probabilities of selection and differential nonresponse.

RESULTS

Descriptive Statistics

The mean (M) and standard deviation (SD) for each study variable as well as the bivariate correlation (r) with the loneliness measure are presented in Table 2. As expected in a sample of married older adults, respondents reported low levels of loneliness (M = .75) and few functional limitations (M = .70), and the two were weakly correlated (r = .07, p < .01). Most respondents indicated that their marriages were characterized by high levels of positive marital quality and low levels of negative marital quality. The measures of marital quality were inversely correlated with loneliness. Neither measure of marital quality was associated with functional limitations (not shown). The correlation between remaining study variables and loneliness was generally consistent with prior studies—age, male, white, education, income, religious service attendance, and network size were all negatively correlated with loneliness, while black and number of times previously married were positively correlated with loneliness.

The Role of Marital Quality in the Disability-Loneliness Association

The results from the Tobit regression models are presented in Table 3. The model (model 1) regressing loneliness on functional limitations shows an association twice as strong as that implied by the correlation coefficient where the censored nature of the dependent variable was ignored (b = .16, p < .01). Prior studies employing ordinary least squares regression and ignoring the censored nature of loneliness have underestimated the strength of the association with disability.

Table 3.

Association between Functional Limitations, Marital Quality, and Loneliness among Married Older Adults: Tobit Regression Estimates (n = 1,500)

Variable Model 1 Model 2 Model 3 Model 4 Model 5
Functional limitations .16** .16*** .15** .12** .12**
Positive marital quality −.97*** −.95*** −.93*** −.92***
Negative marital quality .67*** .68*** .69*** .68***
Ageb −.03** −.04*** −.03**
Female .19 .13 .17
Race-ethnicity
 Black .89*** .73** .74**
 Whiteb — — —
 Hispanic .06 −.18 −.20
 Other 1.49*** 1.51*** 1.49***
Cohabiting −.42
Times previously married .33** .32** .30**
Education
 Less than high school .38† .38†
 High schoolb — —
 Some college .03 .05
 College and beyond −.17 −.14
Income
 <$25,000 .01 .00
 $25,000 to <$50,000c — —
 $50,000 to <$100,000 −.32* −.34*
 ≥$100,000 −.45* −.48*
 Income missing .11 .09
Social integration
 Working .12
 Religious service attendance −.01
 Network size −.05
Intercept —1.02*** −.94*** —1.27** −1.05*** −.92***
Sigma 2.76 2.43 2.38 2.36 2.36
Model fitc
Log likelihood −2079.09 −1952.22 −1932.11 −1923.53 −1922.60
Δ model χ2 8.84*** 253.74*** 40.22*** 17.16* 1.86
Δ df 1 3 7 7 3
Open in a new tab

Source: National Social Life, Health, and Aging Project.

Note: All estimates were weighted to account for differential probabilities of selection and differential nonresponse.

a

Variable is mean centered.

b

Serves as the reference category.

c

Δ model χ2 is improvement in model fit, and Δ df is the change in degrees of freedom relative to the preceding model; for model 1, the comparison is to the null model without any predictors (not shown).

†

p < .10.

*

p < .05.

**

p < .01.

***

p < .001 (two-tailed tests).

As expected given the lack of bivariate association between marital quality and functional limitations, and contrary to the stress-deterioration hypothesis, the introduction of positive and negative marital quality indicators did not alter the association between functional limitations and loneliness (see model 2). The absence of a mediating effect of marital quality is consistent with some prior studies of loneliness (Korporaal et al. 2008) and depression (Bookwala and Franks 2005; Taylor and Lynch 2004). Marital quality was inversely associated with loneliness such that a one-unit increase in positive marital quality reduced loneliness by almost one point (b = −.97, p < .001) and a one-unit increase in negative marital quality increased loneliness by two thirds of a point (b = .67, p < .001).

Overall, the association between functional limitations and loneliness was robust to controls for demographic characteristics, socioeconomic status, and social integration (see models 3–5, respectively, and Table 3). The inclusion of indicators for income in model 4 partially attenuated the effect of functional limitations, as persons with higher incomes have fewer functional limitations and feel less lonely. With the other indicators in the model, measures of social integration—which reflect opportunities for social interaction and support apart from marriage—were not associated with loneliness, consistent with prior research (Cornwell and Waite 2009).

Employing a life course perspective on marriage, we expected that marital quality would be a stress-buffering resource and moderate the effect of disability on loneliness. We specified interaction terms between each indicator of marital quality and functional limitations to test this proposition. As Table 4 shows, we found partial support for our hypothesis. Positive marital quality significantly moderated the effect of functional limitation on loneliness (b = −.19, p < .05), but negative marital quality did not. The absence of a moderating effect of negative marital quality differs from prior research examining depression (Bookwala and Franks 2005), a point we consider fully in the Discussion.

Table 4.

Moderating Effect of Marital Quality on the Association between Functional Limitations and Loneliness among Married Older Adults: Tobit Regression Estimates (n = 1,500)

Variable Modela
Functional limitations .12**
Positive marital quality −.79***
Negative marital quality .77***
Marital quality interactions
 Positive marital quality × functional limitations −.19*
 Negative marital quality × functional limitations −.12
Intercept −.88***
Sigma 2.51
Model fitb
Log likelihood −1810.10
Δ model χ2 225.00***
Δ df 2
Open in a new tab

Source: National Social Life, Health, and Aging Project.

Note: All estimates were weighted to account for differential probabilities of selection and differential nonresponse.

a

Model also includes controls for age, female, race/ethnicity, whether current partnership is a cohabitation, the number of times previously married, education, income, working for pay, religious service attendance, and size of network (excluding spouse).

b

Δ model χ2 is improvement in model fit relative to the model excluding the interaction terms (see model 5 in Table 3).

*

p < .05.

**

p < .01.

***

p < .001 (two-tailed tests).

To aid our interpretation of the interaction term, we followed the procedure detailed by Roncek (1992) and calculated predicted values for the effect of functional limitations on loneliness for below average (–1), average, and above average (+1) levels of positive marital quality, with all other covariates set to zero (not shown). These calculations demonstrate that respondents with above average positive martial quality are effectively buffered from the deleterious effects of disability on loneliness, as the slope of functional limitations is essentially zero and a line of predicted values flat. However, the effect of functional limitations on loneliness is exacerbated for respondents whose marriages are below average in positive marital quality. Although the cross-sectional nature of the data precludes definitive conclusions, these interaction effects suggest that functionally impaired older adults in supportive marriages are somewhat insulated from feeling socially isolated, while those whose marriages lack positive emotional and social support are especially at risk of loneliness when confronted with functional limitations.

Gender Differences

We explored whether the associations between functional limitations, marital quality, and loneliness differed for men and women by estimating gender-stratified models and testing for significant differences in the coefficients between the models. We generally did not detect any significant gender differences in the effect of functional limitations, marital quality, or their interaction on loneliness (not shown). There was some indication that negative marital quality may be marginally more consequential for loneliness among married men than women (t = 1.62, with df > 40, p < .06), supportive of the idea that poor marital relations are more detrimental for men’s mental health given that they are more likely than women to rely exclusively on their spouses for support (Antonucci and Akiyama 1987; Umberson et al. 1996). Nevertheless, our findings suggest that marital quality moderates the association between functional limitations and loneliness similarly for men and women.

DISCUSSION

The current study integrated the stress process (Pearlin et al. 1981) and life course perspectives (George 2003) to situate disablement in its social context, examining whether marital quality matters for the association between functional limitations and loneliness. Loneliness is a potential adverse outcome of functional limitations, inherently embedded in one’s social world and reflecting the subjective evaluation of that social context. As marriage is one of the primary and enduring adult social relationships, we focused on married older adults. We theorized that the quality of the marriage would be of particular importance because it reflects, in part, perceptions of the adequacy of social support from an intimate partner, grounded in the routinized social interactions and coping strategies, mutual investments, and dynamically constructed narratives developed over the course of the marriage.

Using data on 1,500 married older adults from the NSHAP, we found that functional limitations were associated with greater loneliness and that—while positive and negative marital quality were inversely associated with loneliness—neither positive nor negative marital quality mediated the association between disability and loneliness. Thus, we found no support for the stress-deterioration hypothesis that disablement erodes marital quality and leads to greater loneliness. Rather, consistent with the stress-buffering hypothesis and a life course perspective on marriage, we found that positive marital quality moderated the association between functional limitations and loneliness, offering protection from the deleterious socioemotional consequences of disablement. For persons with high positive marital quality, loneliness is very low and independent of functional limitations.

Surprisingly, low levels of positive marital quality, rather than high negative marital quality, increased loneliness in the face of functional limitations. Older adults who reported less openness in their relationships, were less able to rely on a spouse to deal with problems, and/or spent little time in shared activities had incrementally higher levels of loneliness with each additional functional limitation. For those who were not functionally limited, having lower levels of positive marital quality was associated with greater loneliness in general—but older adults with greater functional limitations and marriages characterized by less positive marital quality had even higher levels of loneliness. One possible explanation for this increase in loneliness with functional limitations is that persons in long-term marriages that are not of the highest quality may have developed social resources outside of the marital dyad that partially compensated for the lack of perceived social support from the spouse. Yet upon disablement, these extramarital social resources may be less accessible, and the functionally limited older adult may look more toward his or her spouse to fill these social needs. Without the accumulated patterns of positive interaction within the marriage, the spouse may be less able (or less willing) to meet these social needs adequately, and the disabled older adult may perceive an even greater disjuncture between needed and received social interaction.

Interestingly, we did not find negative marital quality to moderate the effect of functional limitations on loneliness. Negative marital quality tends to be a stronger predictor of physical health and socioemotional well-being than positive marital quality (Carr and Springer 2010). Indeed, Bookwala and Franks (2005) found negative marital quality exacerbated feelings of depression among functionally limited older adults, while positive marital quality had no moderating effect. Rather than our findings contradicting Bookwala and Franks’s, however, the absence of any moderating effect of negative marital quality on loneliness may reflect that loneliness and depressive symptoms arise from different underlying processes (Cacioppo et al. 2006), or it may be that negative marital quality is less consequential for loneliness among disabled older adults.

We acknowledge that the absence of a moderating effect of negative marital quality may also be due to measurement differences between this study and others. For example, our measure of negative marital quality differs from that used by Bookwala and Franks (2005) because it does not contain an item about frequency of disagreements. This omission may be consequential for our finding that marital quality affects the association between functional limitations and loneliness similarly for women and men, as women find arguments with the spouse to be more upsetting than do men (Almeida and Kessler 1998). Unfortunately, we cannot assess either of these potential explanations within the limits of the NSHAP data.

This study contributes to our understanding of the social context of disablement and the consequences for socioemotional well-being in several ways. First, we integrated the stress process and life course perspectives to examine how the social embeddedness of disability has consequences for loneliness. Second, and related, we tested whether marital quality moderated the effects of disablement on loneliness, casting marital quality as a relationship-specific resource. Few prior studies have explicitly examined the importance of marital quality for understanding how disability affects loneliness, and those that have done so have tended to apply, implicitly or explicitly, the stress-deterioration hypothesis. We tested both the stress-deterioration and the stress-buffering hypotheses in our examination of marital quality. Finally, the current study applied a gendered life course perspective to explicate further the social context of disablement. We did not find any systematic gender differences, indicating that marital quality is an important resource for socioemotional well-being among disabled older men and women—despite the fact that men and women have, on average, different social resource profiles. That the moderating effect of positive marital quality appears to operate similarly for women and men is consistent with findings of gender parity in the effects of marital quality on depression and life satisfaction (Williams 2003).

Future research is needed to address a number of remaining questions about the social context of disablement. We focused on married older adults because spouses are central to meeting the social needs of disabled older adults—given the shared history of the marital dyad. However, older adults are embedded in a web of interpersonal relations, and support from nonspouse family and friends may also be important resources. Additional studies are needed to examine how these other types of social relations affect socioemotional well-being among disabled older adults. One question to consider is whether support from other family and/or friends can offset the negative effects of low positive marital quality we have identified here. How these other sources of social support affect loneliness among the nonmarried is another avenue for exploration in future studies (Carr and Springer 2010).

Future studies should also consider the spouse’s health status and its consequences for one’s socioemotional well-being. We focused solely on how marital quality may moderate the effect of one’s own functional limitations on loneliness. Yet in accordance with the life course perspective’s attention to linked lives, it is important to examine how nondisabled spouses’ well-being is affected by the physical limitations of the other and whether persons in high-quality marriages are better able to adjust. A study of Dutch older adults (Korporaal et al. 2008) suggested that both one’s spouse’s and one’s own disability are associated with increased loneliness (and in different ways for men and women) but did not test whether marital quality moderated these relationships. Unfortunately, we cannot test this proposition as spouses were not interviewed in NSHAP.

Finally, we note a number of prior longitudinal studies have found high levels of negative marital quality and low levels of positive marital quality are associated with increased functional limitations and worse self-rated health over time (Choi and Marks 2008; Hawkins and Booth 2005; Umberson et al. 2006). Our findings are limited by the cross-sectional nature of this study. Our interpretation that positive marital quality is an important resource for buffering the effect of marital quality on loneliness is bolstered by the absence of any direct correlation between functional limitations and either dimension of marital quality. Nevertheless, we cannot rule out alternative causal interpretations within the constraints of our data.

Overall, this study illustrates the importance of considering the social context in which disablement occurs. While most older adults are married (or partnered), these relationships vary in quality, and this has implications for well-being (Hawkins and Booth 2005; Kamp Dush et al. 2008). The marital relationship is a potential resource for coping with functional limitations (Cutrona 1996); however, given normative expectations that spouses provide physical and emotional support, disabled older adults in lower-quality marriages may be at risk of negative socioemotional outcomes. This risk is likely to grow among future cohorts of married older adults as institutional supports for marriage and social connections to others have waned at the same time that marriage and relations between husbands and wives have become more individualistic (Waite and Lehrer 2003). Understanding how relationship characteristics, beyond simple marital status, are associated with physical, mental, and social well-being is becoming ever more important as older adults’ relationship histories and experiences become increasingly diverse.

ACKNOWLEDGMENTS

We thank Scott Adams for research assistance and Jamila Bookwala, Melissa Hardy, and Tara D. Warner for comments on previous drafts. We extend special thanks to Linda Waite for helpful guidance.

FUNDING

This research was partially supported by the National Center for Family and Marriage Research (NCFMR) at Bowling Green State University, Bowling Green, Ohio. The NCFMR is supported with a grant from the U.S. Department of Health and Human Services, Office of the Assistant Secretary for Planning and Evaluation, grant 1 U01 AE000001-01. The opinions and conclusions expressed herein are solely those of the authors and should not be construed as representing the opinions or policy of any agency of the federal government.

Biography

David F. Warner is an assistant professor of sociology at Case Western Reserve University and a visiting scholar at the National Center for Family and Marriage Research at Bowling Green State University. His research examines gender and racial-ethnic disparities in health, work, and retirement, with special attention to marriage and family relationships. This research has appeared in Journal of Health and Social Behavior, Social Science & Medicine, and Journal of Gerontology: Social Sciences.

Jessica Kelley-Moore is an associate professor of sociology at Case Western Reserve University. Her primary area of specialization is the social aspects of health and disablement in the second half of the life course. Her work has appeared in American Sociological Review, Journal of Health and Social Behavior, and Journal of Gerontology: Social Sciences.

Footnotes

An earlier version of this article was presented at the NSHAP Early Results Conference, January 28 to 29, 2009, at the University of Chicago.

1.

Health-related stressors may have deleterious effects on marital quality among younger adults because such experiences are nonnormative (Hatch and Bulcroft 2004) and these marriages of comparatively shorter durations would not have the benefit of sustained patterns of interaction, tested coping strategies, and mutual investments that we theorize make marital quality a potential resource among older adults.

2.

Preliminary analyses revealed that male, black, Hispanic, unmarried, less-educated, low-income, and working respondents were less likely to return completed leave-behind questionnaires; these factors are controlled in our multivariate models. None of our primary explanatory variables—functional limitations, positive or negative marital quality—were significantly associated with nonresponse.

3.

About 3 percent of respondents were in cohabiting relationships. Preliminary analyses indicated that the inclusion of nonmarried cohabiting persons had no substantive influence on the results presented. For simplicity, we refer to married persons and marital quality throughout the text.

4.

Consistent with women’s well-known lower mortality and remarriage rates (Waite 1995), in the full NSHAP sample, 78 percent of men were married compared to 55 percent of women (t = 13.48, p < .001).

5.

While some prior studies of loneliness distinguish between the perceived adequacy of one’s social network and intimate companionship (e.g., de Jong Gierveld et al. 2009), the UCLA Short Loneliness Scale used here is a more general indicator corresponding to overall feelings of emptiness, isolation, and abandonment (Cornwell and Waite 2009; Hughes et al. 2004).

6.

The most prevalent limitations were difficulty walking one block (≈ 20 percent) and difficulty dressing (≈ 14 percent). A total of 8 to 9 percent of respondents had difficulty with each of the following: crossing a room, bathing or showering, bed transfers, and toileting. Less than 3 percent of the sample indicated difficulty with eating. Men and women did not statistically differ in their total count of limitations, and there were few differences in the prevalence of individual activities of daily living items. Men were more likely to have difficulty dressing (M = .16 compared to M = .12, t = 2.08, p < .05), while women were more likely to have difficulty toileting (M = .11 compared to M = .09, t = 2.91, p < .01).

7.

Marital history and duration are highly correlated (r = −.70, p < .001); preliminary analyses indicated that marital history alone provided a better model fit (not shown).

REFERENCES

  1. Almeida David M. and Kessler Ronald C.. 1998. “Everyday Stressors and Gender Differences in Daily Distress.” Journal of Personality and Social Psychology 75:670–80. [DOI] [PubMed] [Google Scholar]
  2. Antonucci Toni C. and Hiroko Akiyama. 1987. “AnExamination of Sex Differences in Social Support among Older Men and Women.” Sex Roles 17:737–49. [Google Scholar]
  3. Bierman Alex and Statland Denise. 2010. “Timing,Social Support, and the Effects of Physical Limitations on Psychological Distress in Late Life.” Journal of Gerontology: Social Sciences 65B:631–639. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Bookwala Jamila and Franks Melissa M.. 2005. “Moderating Role of Marital Quality in Older Adults’ Depressed Affect: Beyond the Main-Effects Model.” Journal of Gerontology: Psychological Sciences 60B:P338–P341. [DOI] [PubMed] [Google Scholar]
  5. Booth Alan and Johnson David R.. 1994. “Declining Health and Marital Quality.” Journal of Marriage and Family 56:218–23. [Google Scholar]
  6. Cacioppo John T., Hawkley Louise C., and Berntson Gary G.. 2003. “The Anatomy of Loneliness.” Current Directions in Psychological Science 12:71–74. [Google Scholar]
  7. Cacioppo John T., Hughes Mary Elizabeth, Waite Linda J., Hawkley Louise C., and Thisted Ronald A.. 2006. “Loneliness as a Specific Risk Factor for Depressive Symptoms: Cross-Sectional and Longitudinal Analyses.” Psychology and Aging 21:140–51. [DOI] [PubMed] [Google Scholar]
  8. Carr Deborah and Springer Kristen W.. 2010. “Advances in Families and Health Research in the 21st Century.” Journal of Marriage and Family 72:743–61. [Google Scholar]
  9. Choi Heejeong and Marks Nadine F.. 2008. “Marital Conflict, Depressive Symptoms, and Functional Impairment.” Journal of Marriage and Family 70:377–90. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Clogg Clifford C., Petkova Eva, and Haritou Adamantios. 1995. “Statistical Models for Comparing Regression Coefficients between Models.” American Journal of Sociology 100:1261–93. [Google Scholar]
  11. Cohen Sheldon. 2004. “Social Relationships and Health.” American Psychologist 59:676–84. [DOI] [PubMed] [Google Scholar]
  12. Cornwell Erin York and Waite Linda J.. 2009. “SocialDisconnectedness, Perceived Isolation, and Health among Older Adults.” Journal of Health and Social Behavior 50:31–48. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Cutrona Carolyn E. 1996. Social Support in Couples:Marriage as a Resource in Times of Stress. Thousand Oaks, CA: Sage. [Google Scholar]
  14. de Jong Gierveld Jenny, van Groenou Marjolein Broese,Hoogendoorn Adriaan W., and Smit Johannes H.. 2009. “Quality of Marriages in Later Life and Emotional and Social Loneliness.” Journal of Gerontology: Social Sciences 64B:497–506. [DOI] [PubMed] [Google Scholar]
  15. Ducharme Francine. 1994. “Conjugal Support, Coping Behaviors, and Psychological Well-Being of the Elderly Spouse: An Empirical Model.” Research on Aging 16:167–90. [Google Scholar]
  16. Ensel Walter M. and Lin Nan. 1991. “The Life Stress Paradigm and Psychological Distress.” Journal of Health and Social Behavior 32:321–41. [PubMed] [Google Scholar]
  17. Essex Marilyn J. and Nam Sunghee. 1987. “Marital Status and Loneliness among Older Women: The Differential Importance of Close Family and Friends.” Journal of Marriage and Family 49:93–106. [Google Scholar]
  18. Ferrucci Luigi, Guralnik Jack M., Simonsick Eleanpr,Salive Marcel E., Corti Chiara, and Langlois Jean. 1996. “Progressive versus Catastrophic Disability: A Longitudinal View of the Disablement Process.” Journal of Gerontology: Medical Sciences 51A:M123–M130. [DOI] [PubMed] [Google Scholar]
  19. George Linda K. 2003. “What Life-Course Perspectives Offer the Study of Aging and Health.” Pp. 161–88 in Invitation to the Life Course: Toward New Understandings of Later Life, edited by Settersten RA Jr. Amityville, NY: Baywood. [Google Scholar]
  20. Hafstrom Jeanne L. and Schram Vicki R.. 1984. “ChronicIllness in Couples: Selected Characteristics, Including Wife’s Satisfaction with and Perception of Marital Relationships.” Family Relations 33:195–203. [Google Scholar]
  21. Hatch Laurie Russell and Bulcroft Kris. 2004. “DoesLong-Term Marriage Bring Less Frequent Disagreements? Five Explanatory Frameworks.” Journal of Family Issues 25:465–95. [Google Scholar]
  22. Hatcher Larry. 1994. A Step-by-Step Approach to Using SAS® for Factor Analysis and Structural Equation Modeling. Cary, NC: SAS Institute. [Google Scholar]
  23. Hawkins Daniel N. and Booth Alan. 2005. “UnhappilyEver After: Effects of Long-Term, Low-Quality Marriages on Well-Being.” Social Forces 84:451–71. [Google Scholar]
  24. Hughes Mary Elizabeth, Waite Linda J., Hawkley Louise C., and Cacioppo John T.. 2004. “A Short Scale for Measuring Loneliness in Large Surveys.” Research on Aging 26:655–72. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Johnson David R., Amoloza Teodora O., and Booth Alan. 1992. “Stability and Developmental Change in Marital Quality: A Three-Wave Panel Analysis.” Journal of Marriage and Family 54:582–94. [Google Scholar]
  26. Dush Kamp, Claire M, Taylor Miles G., and Kroeger Rhiannon A.. 2008. “Marital Happiness and Psychological Well-Being across the Life Course.” Family Relations 57:211–26. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Korporaal Marga, Broesse van Groenou Marjolein I., and van Tilburg Theo G.. 2008. “Effects of Own and Spousal Disability on Loneliness among Older Adults.” Journal of Aging and Health 20:306–25. [DOI] [PubMed] [Google Scholar]
  28. Long J. Scott. 1997. Regression Models for Categorical and Limited Dependent Variables. Thousand Oaks, CA: Sage. [Google Scholar]
  29. Moen Phyllis. 2001. “The Gendered Life Course.” Pp. 179–96 in Handbook of Aging and the Social Sciences, 5th ed., edited by Binstock RH and George LK. San Diego, CA: Academic Press. [Google Scholar]
  30. Neff Lisa A. and Karney Benjamin R.. 2005. “Gender Differences in Social Support: A Question of Skill or Responsiveness?” Journal of Personality and Social Psychology 88:79–90. [DOI] [PubMed] [Google Scholar]
  31. O’Muircheartaigh Colm, Eckman Stephanie, and Smith Stephen. 2009. “Statistical Design and Estimation for the National Social Life, Health, and Aging Project.” Journal of Gerontology: Social Sciences 64B:i12–i19. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Pearlin Leonard I., Menaghan Elizabeth G., Lieberman Morton A., and Mullan Joseph T.. 1981. “The Stress Process.” Journal of Health and Social Behavior 22:337–56. [PubMed] [Google Scholar]
  33. Reitzes Donald C. and Mutran Elizabeth J.. 2006. “Self and Health: Factors that Encourage Self-Esteem and Functional Health.” Journal of Gerontology: Social Sciences 61B:S44–S51. [DOI] [PubMed] [Google Scholar]
  34. Roncek Dennis W. 1992. “Learning More from Tobit Coefficients: Extending a Comparative Analysis of Political Protest.” American Sociological Review 57:503–507. [Google Scholar]
  35. Russell David. 2009. “Living Arrangements, SocialIntegration, and Loneliness in Later Life: The Case of Physical Disability.” Journal of Health and Social Behavior 50:460–75. [DOI] [PubMed] [Google Scholar]
  36. Savikko N, Routasalo P, Tilvis RS, Standberg TE, and Pitkälä KH. 2005. “Predictors and Subjective Causes of Loneliness in an Aged Population.” Archives of Gerontology and Geriatrics 41:223–33. [DOI] [PubMed] [Google Scholar]
  37. Spitze Glenna D. and Ward Russell. 2000. “Gender, Marriage, and Expectations for Personal Care.” Research on Aging 22:451–69. [Google Scholar]
  38. Stevens Nan and Westerhof Gerben J.. 2006. “Marriage, Social Integration, and Loneliness in the Second Half of Life: A Comparison of Dutch and German Men and Women.” Research on Aging 28:713–29. [Google Scholar]
  39. Taylor Miles G. and Lynch Scott M.. 2004. “Trajectories of Impairment, Social Support, and DepressiveSymptoms in Later Life.” Journal of Gerontology: Social Sciences 59B:S238–S246. [DOI] [PubMed] [Google Scholar]
  40. Thoits Peggy A. 1995. “Stress, Coping, and Social Support Processes: Where Are We? What Next?” Journal of Health and Social Behavior 35:53–79. [PubMed] [Google Scholar]
  41. Thompson Mark G. and Heller Kenneth. 1990. “Facets of Support Related to Well-Being: Quantitative Social Isolation and Perceived Family Support in a Sample of Elderly Women.” Psychology and Aging 5:535–44. [DOI] [PubMed] [Google Scholar]
  42. Turner R. Jay and Noh Samuel. 1988. “Physical Disability and Depression: A Longitudinal Analysis.” Journal of Health and Social Behavior 29:23–37. [PubMed] [Google Scholar]
  43. Umberson Debra, Chen Meichu D., House James S., Hopkins Kristine, and Slaten Ellen. 1996. “The Effect of Social Relationships on Psychological Well-Being: Are Men and Women Really So Different?” American Sociological Review 61:837–57. [Google Scholar]
  44. Umberson Debra, Williams Kristi, Powers Daniel A., Liu Hui, and Needham Belinda. 2006. “You Make Me Sick: Marital Quality and Health over the Life Course.” Journal of Health and Social Behavior 47:1–16. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Waite Linda J. 1995. “Does Marriage Matter?” Demography 32:483–507. [PubMed] [Google Scholar]
  46. Waite Linda J. 2005. “Marriage and Family.” Pp. 87–108 in Handbook of Population, edited by Poston DL and Micklin M. New York: Springer. [Google Scholar]
  47. Waite Linda J., Laumann Edward O., Levinson Wendy, Lindau Stacy Tessler, McClintock Martha K., O’Muircheartaigh Colm A., and Schumm L. Philip. 2007. “National Social Life, Health, and Aging Project (NSHAP)” [Computer File: Icpsr20541-V1]. Ann Arbor, MI: Inter-university Consortium for Political and Social Research. [Google Scholar]
  48. Waite Linda J. and Lehrer Evelyn L.. 2003. “The Benefits from Marriage and Religion in the United States:A Comparative Analysis.” Population and Development Review 29:255–75. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Williams Kristi. 2003. “Has the Future of MarriageArrived? A Contemporary Examination of Gender, Marriage, and Psychological Well-Being.” Journal of Health and Social Behavior 44:470–87. [PMC free article] [PubMed] [Google Scholar]
  50. Williamson Gail M. and Schultz Richard. 1992. “Pain,Activity Restriction, and Symptoms of Depression among Community-Residing Elderly Adults.” Journal of Gerontology: Psychological Sciences 47:P367–P372. [DOI] [PubMed] [Google Scholar]
  51. Yang Yang. 2006. “How Does Functional Disability Affect Depressive Symptoms in Late Life? The Role of Perceived Social Support and Psychological Resources.” Journal of Health and Social Behavior 47:355–72. [DOI] [PubMed] [Google Scholar]

RESOURCES