The Link between Functional Limitations and Depressive Symptoms

Abstract

Having more physical limitations predicts greater depressive symptoms. However, relatively few studies examine self-conceptions as potential explanations for this association. Using ordinary least squares regression on panel data collected in Miami-Dade County, Florida (2001 and 2004, N = 1,362), we examine the effect of functional limitations on five dimensions of the self: self-esteem, mastery, mattering, introspection, and emotional reliance. We find that having more, and increasing, functional limitations diminishes self-esteem and mastery and increases introspection and emotional reliance. These dimensions of the self collectively account for approximately one third of the effect of limitations on depressive symptoms, with self-esteem and mastery having the largest mediating effects. This study builds on prior findings by using panel data to address not only the influence of functional impairment on multiple dimensions of the self-concept but also the role of each in explaining disability’s effect on depressive symptoms. It also responds to the call for the sociology of mental health to give greater attention to self and identity processes.

Keywords

functional limitation self-concept depression

Individuals living with functional limitations experience higher risk of depression, on average, than do their more able-bodied peers. This pattern is found across measures of not only functional limitations, including self-reported levels and objective tests of physical restrictions, but also depression, including self-reported symptoms and clinically based diagnostic measures (Bruce 2001). These findings are not surprising, as limitations in physical functioning are potential stressors—and, for many older adults, ongoing strains—both of which can erode psychological well-being (Pearlin 1999). While physical limitations’ implications for depressive symptoms are well documented, less understood are the explanations for this link.

An underexplored explanation centers on the negative effects of functional limitations on the self-concept—the set of attitudes toward the self that is constructed and maintained through social comparisons with and reflected appraisals from generalized and specific others (Gecas 1982; Mead 1934). Its definition alludes to its relevance to the sociology of mental health. In short, while the self-concept’s manifestation is intrapsychic, its origin is social. Illustrating this social embeddedness, dimensions of the self-concept often either imply comparisons with others who are seen as relevant reference groups (e.g., in the case of mastery or self-esteem) or involve direct assessments of the influence of others on the self (e.g., in the case of mattering, emotional reliance, or introspection). These social processes shaping the self are influenced by the physical body’s capabilities, with greater limitations associated with worse self-conceptions (e.g., Pudrovska 2010; Trzesniewski, Donnellan, and Robins 2003; Yang 2006). While this pattern points to the role of self-conceptions in explaining the higher levels of depression among more disabled individuals, it receives limited attention not only in research on physical limitations but also in the broader sociology of mental health literature. As Thoits (2013:357) observes, “compared to the considerable theoretical emphasis placed on self and identity factors by psychiatrists and clinical and social psychologists, sociology’s dominant etiological approach to mental disorder, stress theory, has given far less theoretical and research attention to self and identity constructs and processes.” She argues that stress researchers’ focus on the external causes of mental health problems would be enriched by addressing issues of self and identity, including their role in mediating the relationship between stress and mental health.

This explanation for the link between functional limitations and depressive symptoms receives support in the literature. Although relatively stable, self-conceptions are influenced by changes, like declines in functional abilities, with most studies finding that greater impairment diminishes self-conceptions (e.g., Pudrovska 2010; Trzesniewski et al. 2003; Yang 2006). In a culture that prizes young, agile bodies, having physical limitations can be “deeply discrediting” and stigmatizing, generating feelings of difference, inferiority, or helplessness that compromise self-conceptions (Goffman 1963:3). Further, diminished self-conceptions are associated with worse mental health (Hirschfeld et al. 1977; Mirowsky and Ross 2003; Overholser 1997; Taylor and Turner 2001; van Gundy and Schieman 2001). The link between self-conceptions and mental health is reflected, as Thoits (2013) notes, in the criteria for some disorders in the Diagnostic and Statistical Manual of Mental Disorders that include, for example, “feelings of worthlessness” and “reduced self-esteem” as depressive symptoms (American Psychiatric Association 2013).

Although receiving theoretical and empirical support, questions remain unanswered about the role of self-conceptions in explaining the diminished psychological well-being of more physically impaired individuals. Studies consider few dimensions of the self, with self-esteem and mastery receiving the most attention. We argue that sense of mattering, introspection, and emotional reliance also are important to consider, as physical impairment may reduce mattering while heightening introspection and emotional reliance—all of which are associated with diminished psychological well-being (Hirschfeld et al. 1977; Mirowsky and Ross 2003; Overholser 1997; Taylor and Turner 2001; van Gundy and Schieman 2001). The effects of functional limitations on well-being may be far-reaching, as suggested by studies revealing a “loss of self” accompanying chronic illness (Charmaz 1983). Alternatively, functional limitations may diminish some aspects of the self while leaving others unaffected. Another unanswered question focuses on the timing of the functional limitations’ effects on self-conceptions and its implications for well-being. Limitations may represent a “biographical disruption” (Bury 1982) eroding self-conceptions only in the short term, perhaps even enhancing them in the long term, with successful adaptation to physical challenges. Exploring these possibilities, however, is limited in prior studies, which tend to use cross-sectional data.

We address these limitations of prior studies by using two waves of data spanning three years to examine the impact of functional limitations and change in limitations on five dimensions of the self-concept: self-esteem, mastery, mattering, introspection, and emotional reliance. We examine their relative contributions to explaining disabled individuals’ elevated depressive symptoms. Our study extends the literature by illuminating the scope of physical limitations’ effects on self-conceptions—across dimensions of the self and over time. Our study not only contributes to the literature examining explanations for the lower well-being of physically impaired individuals but also responds to broader calls in the sociology of mental health to expand considerations of self and identity processes that have often been seen as intrapsychic and, therefore, left in the domain of psychologists and psychiatrists (Thoits 2013).

Background

Functional Limitations, Self-conceptions, and Mental Health

While disability’s effect on mental health is well established (Bruce 2001), less is known about the underlying processes, particularly those involving self-conceptions. Few longitudinal studies examine either functional limitations’ effect on self-conceptions or its role in explaining the elevated depressive symptoms of the physically impaired, with two focusing on self-esteem (Orth, Trzesniewski, and Robins 2010; Reitzes and Mutran 2006), two on mastery (Pudrovska 2010; Jang et al. 2009), and one on both dimensions (Yang 2006). Numerous cross-sectional studies do address these associations; however, interpretation of results is complicated by the bidirectional relationship between health and self-conceptions (Reitzes and Mutran 2006).

Self-esteem

Self-esteem, a person’s beliefs of self-worth and competence (Rosenberg 1965), has received the most coverage in studies of disability and the self, with several longitudinal studies concluding that greater impairment predicts lower self-esteem. For example, research using four waves of Americans’ Changing Lives (1986-2002) finds that worsening functional health decreases self-esteem (Orth et al. 2010). Further support is found in a study of over 4,000 adults from the 1986 and 1992 Established Populations for Epidemiologic Studies of the Elderly (EPESE), reporting that accumulating disability predicts decline in self-esteem (Yang 2006). Similarly, data collected in 1992 and 1994 from 700 older North Carolinians reveal better self-esteem among those with fewer limitations (Reitzes and Mutran 2006).

Additional evidence is found in cross-sectional studies (e.g., Reitzes and Mutran 2002; Schieman and Campbell 2001). For example, a study of 1,549 disabled and nondisabled persons finds that having a disability and greater impairment is associated with lower self-esteem (Schieman and Campbell 2001). Underlying processes are illuminated by qualitative studies (e.g., Charmaz 1991; Zitzelsberger 2010). As an illustration, a study of 14 women with congenital physical disabilities reveals challenges shaping their self-conceptions. As one interviewee notes, “it’s a constant struggle with your self-esteem, and your feelings of who you are as a woman.” (Zitzelsberger 2010:395).

Only one study examines whether self-esteem partially explains the higher depression of disabled persons. The study by Yang (2006) using EPESE data finds that self-esteem and mastery play larger explanatory roles than does social support, with comparable effects for these two dimensions. Approximately a quarter of the effect of change in functional limitations on symptoms is accounted for by change in self-esteem. The corresponding figure for mastery is 20 percent.

Mastery

Studies of functional limitations and mental health also examine mastery, defined as the “extent to which individuals view themselves as personally powerful or influential in affecting their own life outcomes” (Schieman and Turner 1998:170). Three studies use panel data to examine the link between disability and mastery. The study using EPESE data finds that greater disability accumulation leads to greater decline in mastery as well as self-esteem (Yang 2006). Similarly, a panel study using Midlife in the United States (1995-1996 and 2004-2006) finds that greater functional limitations, chronic illness, and pain diminish mastery (Pudrovska 2010). Another study, using data from 141 older Korean Americans interviewed in 2003 and 2005, finds that greater disability at wave 1 (but not increasing disability between waves) predicts lower mastery at wave 2 (Jang et al. 2009).

Cross-sectional studies report similar findings (e.g., Charmaz 1991; Schieman and Turner 1998). For example, a study of disabled (n = 730) and nondisabled (n = 850) Canadians finds that having a disability and greater impairment is associated with lower mastery (Schieman and Campbell 2001). Providing a qualitative example, a study of 22 people with multiple sclerosis in the Netherlands and Belgium finds that the gradual loss of functioning diminishes perceived control over normal activities, such as maintaining finances and playing with grandchildren (Boeije et al. 2002).

Two studies provide evidence that mastery partially explains disabled individuals’ worse mental health. The panel study using EPESE data found that mastery (and self-esteem) play mediating roles, with comparable effects observed for these two dimensions (Yang 2006). A cross-sectional study of physically disabled Canadians finds that higher anxiety among those with greater impairment is partially explained by their lower mastery (Schieman 1999). In fact, mastery was the strongest mediator among those examined (e.g., health care efficacy, education).

Mattering

Studies of disability’s effect on psychological well-being give limited attention to mattering, a sense of belonging, meaningfulness, and relevance to close others (Rosenberg and McCullough 1981). However, disability may be particularly consequential for mattering, as it is tied to positive affirmation from others. Several studies suggest that greater impairment may diminish mattering. For example, a study of 208 college students reports in a correlation matrix that poorer physical health is associated with lower mattering (Raque-Bogdan et al. 2011). Also lending support, a study of retirement community residents (N = 167) found that higher overall wellness is associated with greater mattering (Dixon 2007). Qualitative studies describe how disability can erode mattering. For example, a study of wheelchair users finds they often experience “civil inattention” (Cahill and Eggleston 1995). Providing further support, a study of individuals with multiple sclerosis suggests that declining ability to perform tasks, like home repairs or cooking, diminished mattering. One man noted that his inability to do home repairs, tasks eventually taken on by his wife, left him feeling diminished and irrelevant (Boeije et al. 2002).

Introspection

Introspection, the degree to which one devotes attention either inward or toward the outside world (Mechanic and Hansell 1987), has received moderate attention in disability research. Cross-sectional studies suggest that greater disability is associated with higher introspection. For example, a study of college students (n = 1,014) and older adults (n = 48) reports that greater physical symptoms are associated with higher introspection (Mechanic and Hansell 1987). Building on this work, two studies of disabled and nondisabled adults in Canada (N = 1,567) report mixed findings. One study finds that having a disability is associated with greater introspection (van Gundy and Schieman 2001), while another reports that greater impairment, pain, and worse health are associated with lower introspection (Schieman and Campbell 2001). Qualitative studies highlight the processes through which impairments lead to greater introspection. For example, a study by Charmaz (1991) finds that chronic illness can lead to social isolation and immersion in thought about self and body. Providing further support, a qualitative study of men with spinal cord injuries finds that pain can be isolating and produce an inward focus. Recalling experiences in traction, one respondent explained that pain led him to “go further into [his] shell,” while another described being unable to “push [his] body out of [his] mind.” (Sparkes and Smith 2008:686).

We identified only one study examining whether introspection partially explains the elevated depressive symptoms of those with physical disabilities. While not their study’s focus, van Gundy and Schieman (2001) report in cross-sectional regression models of depression that adjusting for introspection reduced the coefficient for disability by a very modest amount (i.e., 4 percent). Although results suggest a limited role for introspection, further study is needed using panel data and exploring multiple dimensions of the self-concept.

Emotional Reliance

Few studies have examined the relationship between limitations and emotional reliance, defined as a need for emotional attention, attachment, and approval from close others (Hirschfeld et al. 1977). Findings suggest that physical dependence on others generated by limitations can set the stage for greater emotional dependence. For example, a study using a sample of 1,768 ethnically diverse disabled and nondisabled adults finds differences in average levels of emotional reliance by disability status, though patterns varied by suicidal ideation (Russell, Turner, and Joiner 2009). Among those with suicidal thoughts, disabled persons reported greater emotional reliance than nondisabled persons, but no difference by disability status was found among individuals without suicidal thoughts. The study also suggests that this process partially explains the worse mental health of disabled individuals. The positive relationship between disability status and lifetime suicidal ideation is reduced by 10 percent when dimensions of the self-concept—that is, mastery, self-esteem, and emotional reliance—are included in the model.

Several studies have examined interpersonal dependency, of which emotional reliance is a component. Interpersonal dependency is “the need to associate closely with, interact with, and rely upon valued other people” (Hirschfeld et al. 1977:610). Studies report that poorer physical health is associated with greater dependency. For example, a study of older adults (N = 105) finds that lower mobility is associated with higher interpersonal dependency (Gardner and Helmes 2006). Similar conclusions are drawn from a study of primarily African American women (N = 110) finding that chronic illness and impairment are associated with overdependence and detachment (Porcerelli et al. 2009). Qualitative studies also highlight how having disabilities can increase emotional reliance. For example, in a study of 25 disabled adults, one man mentions, “Because I am disabled, it gives you the worry about getting a girlfriend, you hold onto it for dear life” (Liddiard 2014:121).

Our review of studies examining the relationship between functional limitations and self-conceptions draws into focus several gaps in the literature. The limited use of panel data constrains our understanding of how changes in impairment affect self-conceptions. Studies also center on self-esteem and mastery, often neglecting other dimensions. Further, few studies address whether self-conceptions partially explain the impact of disability on psychological well-being. Addressing these limitations, this study uses panel data to assess the impact of functional limitations and change in limitations on these five dimensions of self-conceptions. We address whether self-conceptions partially explain the effects of functional impairment on depressive symptoms. We test two hypotheses. The first predicts that higher levels of functional limitations and greater increases in limitations over time predict greater declines in self-esteem, mastery, and mattering and greater increases in introspection and emotional reliance. Second, we hypothesize that differences in the initial level of and change in self-conceptions provide an explanation for the higher depressive symptoms of individuals with greater limitations, compared with their more able-bodied peers.

Data and Methods

Data

Data are drawn from two waves of the Physical Disability, Mental Health, and Drug Use Study conducted in 2001 and 2004 in Miami-Dade County, Florida (Turner, Lloyd, and Taylor 2006). The study was designed to assess the impact of social factors on the mental health of disabled people. Using a sampling frame of 10,000 households, the study had two goals: a matched sample of 1,000 respondents with and 1,000 without a disability and a race-ethnic distribution consistent with area demographics (i.e., 25 percent of each of the following groups: non-Hispanic white, African American, Cuban, and other Hispanic). The response rate for Wave 1 was 82 percent, yielding a sample size of 1,986. After exclusion of 100 respondents who died and 59 too ill for reinterview, 1,827 persons remained for interview in the second study. With a follow-up response rate of 82 percent, the sample size at Wave 2 is 1,495. Compared with those interviewed at both waves, the 332 respondents interviewed only at Wave 1 had lower rates of homeownership; lower socioeconomic status, social support, self-esteem, mattering, mastery, and introspection; and greater functional limitations and depressive symptoms.

Of the Wave 2 sample of 1,495, some respondents were excluded from the analytic sample, including 92 reporting “other” race-ethnicity and 41 answering fewer than half of the items measuring functional limitations, self-conceptions, social support, or depressive symptoms. Compared with those in the analytic sample (n = 1,362), those excluded (n = 133) had significantly lower introspection at Wave 1 and lower social support at both waves.

Measures

Development of measures was informed by factor analyses. The models, using oblique rotation, included all items from the scales tapping depressive symptoms, self-conceptions, and social support. Results indicated some redundancy and low factor loadings, leading to omission of some items from the scales. Items used in scales are described in the appendix.

Functional limitations (Wave 1, α = .88; Wave 2, α = .87) is a scale averaging seven responses on the Stanford Health Assessment Questionnaire, tapping ability to perform daily activities, such as walking a quarter of a mile (Fries et al. 1980). Responses range from 1 (easy to do) to 4 (unable to do). We examine baseline level of limitations and change between waves (i.e., Wave 2 – Wave 1 mean scale). Higher values on the Wave 1 measure reflect greater impairment, and higher values on the change score indicate greater increases in impairment. Our selection of this measure is informed by research finding that scales combining a greater variety of indicators perform better as predictors of inability to work and activities-of-daily-living disability than do dichotomous scales, scales measuring each component separately, those differentially weighting indicators, and those distinguishing between upper- or lower-body functionality (Long and Pavalko 2004).

Depressive symptoms (Wave 1, α = .93; Wave 2, α = .91) are measured using a mean scale derived from a modified version of the Center for Epidemiologic Studies Depression Scale (CES-D; Radloff 1977), consisting of 20 statements asking how respondents felt in the past month (e.g., “You felt you could not shake off the blues”). The CES-D version contained in the dataset differs from the original in several ways. Rather than asking the number of days in the past week each symptom was experienced, the modified scale refers to the past month and has response categories of (1) not at all, (2) occasionally, (3) frequently, and (4) almost all the time. Our scale also differs because it contains 16 rather than 20 items. We omit four—all referring to positive affect—that we found using factor analysis to be redundant with self-esteem. Factor analysis also allowed us to identify redundancy among items measuring functional limitations and those measuring depressive symptoms, particularly somatic ones. Results revealed that items measuring depressive symptoms and functional limitations tap two distinct constructs. CES-D is positively skewed, so we analyze the logged score. Analyzing the logged score reduces skewness at Wave 1 from 1.84 to 1.07 and at Wave 2 from 1.63 to 0.94. Logging also reduces kurtosis at Wave 1 from 6.64 to 3.51 and at Wave 2 from 5.80 to 3.21.

We examine five dimensions of the self-concept. Each is tapped using two measures: mean scale at Wave 1 and change between waves (i.e., Wave 2 – Wave 1 mean scale). Higher values on the Wave 1 measure reflect higher levels of the construct, and higher values on the change score indicate greater increases in levels.

Self-esteem (Wave 1, α = .79; Wave 2, α = .78) is measured using responses to 6 of 10 items on the Rosenberg (1965) scale (e.g., “You feel that you are a person of worth”). Only 6 of 10 items from the original scale were asked in the survey. Responses range from 1 (strongly disagree) to 5 (strongly agree).

Mastery (Wave 1, α = .80; Wave 2, α = .81) is measured using responses to five of seven items on the Pearlin and Schooler (1978) scale (e.g., “You have little control over the things that happen to you”). We omitted two of the original items due to low factor loadings—likely stemming from the phrasing in an opposite (i.e., positive) direction. Responses range from 1 (strongly agree) to 5 (strongly disagree).

Mattering (Wave 1, α = .88; Wave 2, α = .89) is measured using responses to 9 of 10 items from Rosenberg and McCullough’s (1981) General Mattering Scale (e.g., “My well-being matters to others”). We exclude one item because of its low factor loading, which might stem from its phrasing in an opposite (i.e., negative) direction. Responses range from 1 (not at all) to 4 (a lot).

Introspection (Wave 1, α = .86; Wave 2, α = .90) is based on Mechanic and Hansell’s (1987) introspection scale. The original scale includes 12 items; however, only 4 items were included in the survey, all of which were used in our scale (e.g., “How often do you try to figure yourself out?”). Responses range from 1 (very little) to 5 (very much).

Emotional reliance (Wave 1, α = .68; Wave 2, α = .68) is measured using a shortened version of Hirschfeld and colleagues’ (1977) scale. The original scale includes 18 items; however, only 4 items were included in the survey, all of which were used in our scale (e.g., “I would be completely lost if I didn’t have someone special”). Responses range from 1 (strongly disagree) to 5 (strongly agree).

We control for factors associated with mental health, including age, gender, race-ethnicity, socioeconomic status, married/cohabiting status, and social support (Bierman and Pearlin 2011; Mirowsky and Ross 1992, 2003; Riolo et al. 2005; Thoits 2011). All are measured at Wave 1. Age is centered at the mean. We include an age-squared term to capture the nonlinear effect of age on depression (Mirowsky and Ross 2003). Gender is coded 1 for women and 0 for men. Self-reported race-ethnicity is measured using four dichotomous variables: non-Hispanic white (reference group), African American, Cuban, and other Hispanic. Socioeconomic status is captured using two variables: (1) a SES scale, a standardized and summed score using household income, educational attainment, and occupational category, divided by the number of available components (Hollingshead 1965) and (2) homeownership (coded 1 for homeowners and 0 for others). Married/cohabiting status is coded 1 for partnered respondents and 0 for all others. Social support (Wave 1, α = .86; Wave 2, α = .87) is the average response to 12 items from the Provisions of Social Relations Scale (Turner and Marino 1994), with half referring to friends and half to family. Responses range from 1 (not at all true for you) to 4 (true for you). We control for social support using two measures: support at Wave 1 and change in support (i.e., Wave 2 – Wave 1 mean scale). We include social support because it not only strongly reduces depressive symptoms but also plays a role, albeit a more limited one than that of self-conceptions, in explaining disabled individuals’ elevated depressive symptoms (Yang 2006). Our approach permits the assessment of self-conceptions’ contribution to explaining the higher depressive symptoms of those with physical impairments, over and above that of social support.

Methods

We use ordinary least squares (OLS) regression to address our two research questions, one centering on the impact of functional limitations on self-conceptions and the other on the role of this process in explaining the higher symptoms of individuals with more limitations, compared with their more able-bodied peers. Both are addressed using regression models assessing residual change (Kessler and Greenberg 1981); Wave 2 levels of the dependent variable (self-conceptions and mental health) regressed on corresponding Wave 1 levels of these measures. To address the first question, we construct models estimating the influence of baseline levels of and changes in functional limitations on change in self-conceptions. To address the second question, we construct models assessing how baseline levels of and changes in functional limitations influence changes in depressive symptoms. We address how self-conceptions contribute to explaining the impact of physical impairment on changes in depressive symptoms by creating a series of seven models. Model 1 regresses depressive symptoms at Wave 2 on symptoms at Wave 1, functional limitations, change in limitations, and control variables. Each of the next five models enters a component of self-conceptions, to assess the contribution each makes to explaining the higher symptoms of more functionally impaired individuals. A final model (Model 7) includes all five components, to assess the extent to which they collectively account for these differences. To assess potential mediators, we conduct Sobel tests and examine changes in the magnitude and level of significance of the functional limitation variables with the addition of each dimension of the self. In all models, we apply sampling weights to the data, which has the added benefit of estimating robust standard errors.

Results

Table 1 reports descriptive statistics. Mean levels of functional limitations were relatively low and changed little between waves, with the average respondent reporting mild to some physical difficulties. Levels of social support, self-esteem, and mattering were high and showed a significant, but modest, decline between waves. Mastery was high at both waves and did not significantly change. Average levels of introspection and emotional reliance were moderate and showed significant declines between waves. Average depressive symptoms were relatively low and changed little.

Table 1.

Descriptive Statistics.

Variable	Wave 1	Wave 2
Age (in years)	57.03 (16.81)
Female (0, 1)	0.53
Non-Hispanic white (0, 1)	0.24
African American (0, 1)	0.32
Cuban (0, 1)	0.24
Other Hispanic (0, 1)	0.20
Socioeconomic status	0.06 (1.01)
Homeowner (0, 1)	0.62
Married/cohabiting (0, 1)	0.60
Social support^a	3.49 (0.57)	3.44 (0.59)**
Functional limitations^a	1.43 (0.63)	1.42 (0.61)
Self-esteem^b	4.64 (0.56)	4.59 (0.59)*
Mastery^b	3.49 (1.10)	3.48 (1.05)
Mattering^a	3.39 (0.55)	3.35 (0.58)*
Introspection^b	2.94 (1.18)	2.82 (1.26)**
Emotional reliance^b	3.57 (1.05)	3.41 (1.03)***
Depressive symptoms^a	1.43 (0.53)	1.42 (0.48)

Source: Physical Disability, Mental Health, and Drug Use Study, 2001 and 2004 (Turner, Lloyd, and Taylor 2006).

Note: N = 1,362. Mean (standard deviation).

Range = 1 (low) to 4 (high). ^bRange = 1 (low) to 5 (high).

Significant differences between Wave 1 and Wave 2. * p < .05. **p < .01. ***p < .001 (two-tailed tests).

Table 2 presents the OLS regression results assessing residual change for the five self-concept dimensions. Across the dimensions, we find similar effects for functional limitations.

Table 2.

Ordinary Least Squares Regression of Self-conceptions at Wave 2 on Functional Limitations.

Variable	Self-esteem	Mastery	Mattering	Introspection	Emotional reliance
Age^a (in years)	−.00	−.00	−.00	−.01***	−.00
	(.00)	(.00)	(.00)	(.00)	(.00)
Age-squared^a	.00	.00	.00	−.00	−.00
	(.00)	(.00)	(.00)	(.00)	(.00)
Female	.03	.00	.04	−.00	−.09
	(.03)	(.06)	(.03)	(.07)	(.06)
African American^b	.08	−.07	.15***	.00	−.10
	(.04)	(.08)	(.04)	(.09)	(.08)
Cuban^b	−.01	−.20*	−.03	.37***	.12
	(.05)	(.08)	(.04)	(.10)	(.08)
Other Hispanic^b	−.06	−.12	.03	.33**	.10
	(.06)	(.09)	(.05)	(.11)	(.09)
Socioeconomic status	.04**	.06	.03	−.02	−.07*
	(.02)	(.03)	(.02)	(.04)	(.03)
Homeowner	.09*	.18**	.10**	−.13	−.03
	(.04)	(.07)	(.04)	(.08)	(.07)
Married/cohabiting	−.04	−.03	.04	.10	.09
	(.03)	(.06)	(.03)	(.07)	(.06)
Social support	.21***	.30***	.36***	−.09	−.04
	(.04)	(.06)	(.04)	(.08)	(.06)
ΔSocial support	.34***	.34***	.38***	.06	.05
	(.04)	(.05)	(.03)	(.06)	(.05)
Functional limitations	−.16***	−.28***	−.01	.25***	.10*
	(.03)	(.05)	(.03)	(.06)	(.05)
ΔFunctional limitations	−.16***	−.28***	−.04	.23***	.24***
	(.03)	(.06)	(.03)	(.07)	(.06)
Self-esteem, Wave 1	.14***
	(.04)
Mastery, Wave 1		.23***
		(.03)
Mattering, Wave 1			.21***
			(.03)
Introspection, Wave 1				.29***
				(.03)
Emotional reliance, Wave 1					.33***
					(.03)
Adjusted R²	.23	.22	.31	.17	.17

Source: Physical Disability, Mental Health, and Drug Use Study, 2001 and 2004 (Turner, Lloyd, and Taylor 2006).

Note: N =1,362. Unstandardized coefficient (robust standard error).

Centered at the mean. ^bNon-Hispanic whites = reference group.

p < .05. **p < .01. ***p < .001.

Results for self-esteem reveal that higher baseline level and greater increases in functional limitations predict greater declines in self-esteem over time. Having higher socioeconomic status and being a homeowner also predict higher self-esteem at Wave 2, controlling on Wave 1 levels.

We find similar results for mastery. Higher baseline level and greater increases in functional limitations predict greater declines in mastery. Having more social support at baseline and more improvement in support also predicts higher mastery at Wave 2, controlling on Wave 1 levels. Race-ethnicity patterns indicate that, controlling on Wave 1 levels of mastery, Cubans have higher mastery than do non-Hispanic whites. Homeowners have greater mastery at Wave 2 than do nonhomeowners.

In contrast, the results for mattering reveal no significant effect of functional limitations. However, change in limitations between waves significantly predicts mattering in a model omitting from the equation change in social support, raising the possibility that change in social support mediates the association between change in functional limitations and mattering. Consistent with this causal chain, results indicate that higher levels of social support and more change in support between waves predict greater increases mattering. To further explore this possibility, we constructed a model assessing whether change in functional limitations significantly predicts change in social support. We regressed social support at Wave 2 on social support at Wave 1, functional limitations at Wave 1, change in functional limitations between waves, and our controls. Consistent with our post hoc hypothesis, results revealed that greater change in functional limitations is associated with lower levels of social support at Wave 2, controlling on Wave 1 support (b = −.09, p < .05). In addition to social support, race-ethnicity and homeownership are significant, with African Americans and homeowners having higher mattering than their respective peers.

Results for introspection contrast with those for mattering and are consistent with those for self-esteem and mastery. We find that having higher levels of limitations at Wave 1 and greater increases in limitations predict greater increases in introspection between waves. Other significant predictors are age and race-ethnicity. Being older predicts lower levels of introspection, while Hispanics have higher introspection scores, compared with non-Hispanic whites.

Similar patterns are found for emotional reliance, with functional limitations emerging as one of the few significant predictors. Having higher baseline levels and greater increases in limitations predicts higher levels of emotional reliance at Wave 2, controlling on Wave 1 levels. Lower socioeconomic status also predicts greater emotional reliance at Wave 2.

Table 3 reports results of the regression of depressive symptoms on functional limitations and self-conceptions. Model 1 reveals that higher initial levels and greater increases in functional limitations predict higher levels of symptoms at Wave 2, controlling on Wave 1 levels. More symptoms also are associated with being younger and a woman and having lower initial levels and decreases in support. Other significant predictors of symptoms are race-ethnicity and homeownership, with African Americans and homeowners having fewer symptoms at Wave 2 than their respective counterparts, controlling for Wave 1 symptoms.

Table 3.

Ordinary Least Squares Regression of Depressive Symptoms at Wave 2 on Functional Limitations and Self-conceptions.

Variable	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6	Model 7
Depressive symptoms, Wave 1	.19***	.18***	.14***	.19***	.18***	.17***	.12***
	(.03)	(.03)	(.03)	(.03)	(.03)	(.03)	(.03)
Age^a (years)	−.00***	−.00***	−.00***	−.00***	−.00**	−.00**	−.00***
	(.00)	(.00)	(.00)	(.00)	(.00)	(.00)	(.00)
Age-squared^a	−.00	−.00	−.00	−.00	−.00	−.00*	−.00
	(.00)	(.00)	(.00)	(.00)	(.00)	(.00)	(.00)
Female	.04*	.04**	.03*	.04**	.04*	.04*	.04**
	(.01)	(.01)	(.01)	(.01)	(.01)	(.01)	(.01)
African American^b	−.04*	−.03	−.05**	−.03	−.04*	−.04*	−.03
	(.02)	(.02)	(.02)	(.02)	(.02)	(.02)	(.02)
Cuban^b	−.00	−.01	−.03	−.01	−.02	−.01	−.03
	(.02)	(.02)	(.02)	(.02)	(.02)	(.02)	(.02)
Other Hispanic^b	−.03	−.04*	−.05*	−.03	−.05	−.04	−.06**
	(.02)	(.02)	(.02)	(.02)	(.02)	(.02)	(.02)
Socioeconomic status	−.01	−.00	.00	−.01	−.01	−.00	.01
	(.01)	(.01)	(.01)	(.01)	(.01)	(.01)	(.01)
Homeowner	−.04*	−.03	−.03	−.03	−.04*	−.04*	−.01
	(.02)	(.02)	(.02)	(.02)	(.02)	(.02)	(.02)
Married/cohabiting	−.03	−.03*	−.03	−.02	−.03	−.04*	−.04**
	(.02)	(.02)	(.02)	(.02)	(.02)	(.02)	(.01)
Social support	−.09***	−.05**	−.06***	−.05**	−.09***	−.09***	−.02
	(.02)	(.02)	(.02)	(.02)	(.02)	(.02)	(.02)
ΔSocial support	−.13***	−.08***	−.10***	−.09***	−.13***	−.13***	−.05***
	(.02)	(.01)	(.01)	(.02)	(.02)	(.01)	(.01)
Functional limitations	.12***	.09***	.10***	.12***	.11***	.11***	.08***
	(.02)	(.01)	(.01)	(.01)	(.01)	(.01)	(.01)
ΔFunctional limitations	.11***	.09***	.09***	.11***	.10***	.10***	.07***
	(.02)	(.02)	(.02)	(.02)	(.02)	(.02)	(.01)
Self-esteem		−.15***					−.12***
		(.02)					(.02)
ΔSelf-esteem		−.15***					−.10***
		(.02)					(.02)
Mastery			−.09***				−.06***
			(.01)				(.01)
ΔMastery			−.07***				−.05***
			(.01)				(.01)
Mattering				−.08***			−.03
				(.02)			(.02)
ΔMattering				−.11***			−.06***
				(.02)			(.02)
Introspection					.03***		.02*
					(.01)		(.01)
ΔIntrospection					.03***		.02***
					(.01)		(.01)
Emotional reliance						.05***	.03**
						(.01)	(.01)
ΔEmotional reliance						.03***	.02**
						(.01)	(.01)
Adjusted R²	.26	.33	.32	.29	.27	.28	.39

Source: Physical Disability, Mental Health, and Drug Use Study, 2001 and 2004 (Turner, Lloyd, and Taylor 2006).

Note: N =1,362. Unstandardized coefficient (robust standard error).

Centered at the mean. ^bNon-Hispanic whites = reference group.

p < .05. **p < .01. ***p < .001.

The remaining models allow examinations of whether the effect of functional limitations on dimensions of the self-concept, individually (in Models 2 through 6) or collectively (in Model 7), accounts for the elevated depressive symptoms of more functionally limited individuals. Each dimension predicts depressive symptoms in the expected direction: Controlling on Wave 1 depressive symptoms, greater Wave 2 symptoms are predicted by lower initial levels of and greater declines in self-esteem, mastery, and mattering and higher initial levels of and greater increases in introspection and emotional reliance. However, analyses of change in the coefficients for functional limitations between Model 1 and each subsequent model (from Model 2 to Model 6) reveal that all but one of the five dimensions—mattering—contribute to explaining more impaired individuals’ higher depressive symptoms. These analyses also indicate that relative contributions vary across dimensions, with self-esteem and mastery exerting the largest mediating effects. With self-esteem included (Model 2), coefficients for initial levels and change in functional limitations are reduced by 25 and 18 percent, respectively. We find comparable effects for mastery. With mastery included (Model 3), coefficients for initial levels and change in limitations are reduced by 17 and 18 percent. We observe modest mediating effects for introspection (Model 5) and emotional reliance (Model 6), with similar declines in coefficients (i.e., 8 percent for initial level and 9 percent for change in functional limitations). With all self-conceptions included (Model 7), the coefficients for initial levels and change in limitations are reduced by 33 and 36 percent. However, functional impairments retain independent influences on depressive symptoms.

These conclusions are supported by results of Sobel tests, revealing significant mediating effects for self-esteem, mastery, introspection, and emotional reliance. Self-esteem at baseline mediates baseline limitations’ effect on depressive symptoms (z = 4.11, p = .000), while change in esteem mediates the effects of baseline (z = 4.47, p = .000) and change in limitations (z = 4.51, p = .000) on depressive symptoms. Mastery at baseline is a mediator of baseline limitations’ effect on depressive symptoms (z = 4.45, p = .000), while change in mastery mediates the effects of baseline (z = 4.33, p = .000) and change in limitations (z = 4.67, p = .000) on depressive symptoms. Similar to the self-esteem and mastery results, introspection at baseline mediates baseline limitations’ effect on depressive symptoms (z = 2.54, p = .01), while change in introspection mediates the effects of baseline (z = 2.76, p = .005) and change in limitations (z = 2.65, p = .008) on depressive symptoms. We find slightly different results for emotional reliance. Emotional reliance at baseline mediates baseline limitations’ effect on depressive symptoms (z = 3.49, p = .000), while change in emotional reliance mediates the effect of change in limitations on depressive symptoms (z = 3.21, p = .001). However, change in reliance does not mediate the effect of baseline functional limitations on depressive symptoms.

Discussion

This study responds to the call for sociologists of mental health to give greater attention to social processes related to the self, including their role in explaining stressful life circumstances’ erosion of mental health (Thoits 2013). Our study examines one potential source of stress, or ongoing strain, facing many older adults—functional limitations. It addresses how they affect five dimensions of the self-concept and whether these effects partially explain the elevated depressive symptoms of persons with disabilities compared with those who are more able-bodied. It extends the literature on functional limitations and mental health in two ways. First, it adds to the relatively small number of studies of these issues that use panel data. Our data permit an examination of the consequences for psychological distress of not only baseline levels but also change in functional limitations. Second, our study examines a wider range of dimensions of the self-concept, including several receiving scant attention in research on physical limitations, in particular sense of mattering, introspection, and emotional reliance. This feature of our study allows an examination of the relative effects of limitations on these dimensions, addressing the broader question of whether disability has pervasive or more limited effects on self-conceptions.

Our findings point to more pervasive effects. Consistent with other studies (Orth et al. 2010; Pudrovska 2010; Reitzes and Mutran 2006), our results indicate that greater limitations diminish self-esteem and mastery. Not only higher initial levels but also greater increases in limitations predict worse self-esteem and mastery, with the magnitude of these effects similar. These results may suggest that self-protective strategies employed in the face of functional impairment—even over a period of three years—are inadequate to protect self-conceptions, a possibility that could be further explored in data spanning a longer time. Similar results are found for introspection and emotional reliance, dimensions not examined previously using panel data. These results suggest that impairment can, somewhat paradoxically, produce both inwardly and outwardly focused orientations. Experiencing greater limitations may signal a divergence from normative physical health standards, resulting in intensified cognitive engagement with the self (e.g., Charmaz 1991), but it also can require increased reliance on others for physical assistance, which may require a level of intimacy that can produce emotional reliance. While our hypothesis regarding the impact of functional limitations on self-conceptions was supported for self-esteem, mastery, introspection, and emotional reliance, it was not for mattering. Neither initial levels nor changes in limitations predicted mattering. However, further analyses suggested that the association between limitations and mattering is mediated by social support. In other words, having greater impairments predicts lower levels of support from family and friends, accounting for the lower sense of mattering reported by more impaired individuals. These results suggest that, rather than being inconsequential for sense of mattering, functional limitations may be linked to mattering through different processes from those of other dimensions of the self, including those centering less on relationships with others—like mastery and self-esteem.

We also examined whether self-conceptions partially explain the elevated psychological distress among people with greater levels of impairment. We find support for our hypothesis, with the effect of self-conceptions explaining approximately one-third of the relationship between limitations and depressive symptoms. We note that these patterns were found in models including social support, another factor found to partially explain the link between limitations and mental health (Thoits 2011; Yang 2006). However, mattering is not found to be a mediator, and the explanatory roles vary across other dimensions of the self. Self-esteem and mastery play the largest mediating roles. Introspection plays a smaller role, a finding consistent with that of the only other study examining the mediating role of introspection (van Gundy and Schieman 2001). A similarly small role is found for emotional reliance, though this finding does reveal another mechanism through which limitations impact psychological distress.

Our study is limited in several ways, each raising questions for future research. Our use of a community sample limits the generalizability of results, pointing to the need for replication with national data. Other limitations center on measurement issues. This study addresses the role of self-conceptions in explaining the effect of limitations on depressive symptoms, not diagnosable mental disorders. Although a robust relationship with disability is found across different measures of depression, subtle variation may be found in the underlying processes (Bruce 2001). For example, while the effect of physical limitations on the self-concept may be sufficient to elevate depressive symptoms, they may not reach a diagnostic threshold unless impairments substantially diminish other protective factors, like social support. In addition, our study examines the full range of neither impairments nor dimensions of self-conceptions. We note our study’s neglect of the impact of functional limitations on identities, including those hinging on sociodemographic characteristics (e.g., age) and social roles (e.g., spouse). Our study also does not examine perceived disability status, which can differ across individuals with similar levels of impairment (Kelley-Moore et al. 2006) and produce variation in limitations’ effects on self-conceptions and well-being. Also suggesting avenues for future research, our study examines change between two points separated by three years, limiting our ability to examine either the short-term, perhaps more dramatic but temporary, effects of increasing physical limitations on self-conceptions and mental health or their long-term implications. Questions for future studies include whether self-conceptions follow a continued downward trajectory in tandem with increasing physical limitations or if, over time, individuals adapt to changes in their physical abilities and find ways to bolster their self-conceptions, thus protecting their psychological well-being.

Footnotes

Appendix

References

American Psychiatric Association. 2013. Diagnostic and Statistical Manual of Mental Disorders (5th ed.). Arlington, VA: Author.

Bierman

Alex

Pearlin

Leonard I.

2011. “SES, Trajectories of Physical Limitations, and Change in Depression in Late Life.” Society and Mental Health 1:139-52.

Boeije

Hennie R.

Duijnstee

Mia S. H.

Grypdonck

Mieke H. F.

Pool

Aart

. 2002. “Encountering the Downward Phase: Biographical Work in People with Multiple Sclerosis Living at Home.” Social Science & Medicine 55:881-93.

Bruce

Martha L.

2001. “Depression and Disability in Late Life: Directions for Future Research.” American Journal of Geriatric Psychiatry 9:102-12.

Bury

Michael

. 1982. “Chronic Illness as Biographical Disruption.” Sociology of Health & Illness 4:167-82.

Cahill

Spencer E.

Eggleston

Robin

. 1995. “Reconsidering the Stigma of Physical Disability: Wheelchair Use and Public Kindness.” Sociological Quarterly 36:681-98.

Charmaz

Kathy

. 1983. “Loss of Self: A Fundamental Form of Suffering in the Chronically Ill.” Sociology of Health & Illness 5:168-95.

Charmaz

Kathy

. 1991. Good Days, Bad Days: The Self in Chronic Illness and Time. New Brunswick, NJ: Rutgers University Press.

Dixon

Andrea L.

2007. “Mattering in the Later Years: Older Adults’ Experiences of Mattering to Others, Purpose in Life, Depression, and Wellness.” Adultspan Journal 6:83-95.

10.

Fries

James F.

Patricia Spitz

Kraines

Guy

Holman

Halsted R.

1980. “Measurement of Patient Outcome in Arthritis.” Arthritis & Rheumatism 23:137-45.

11.

Gardner

D. K.

Helmes

2006. “Interpersonal Dependency in Older Adults and the Risks of Developing Mood and Mobility Problems when Receiving Care at Home.” Aging & Mental Health 10:63-68.

12.

Gecas

Viktor

. 1982. “The Self-concept.” Annual Review of Sociology 8:1-33.

13.

Goffman

Erving

. 1963. Stigma: Notes on the Management of Spoiled Identity. New York: Simon & Schuster.

14.

Hirschfeld

Robert M. A.

Klerman

Gerald L.

Gough

Harrison G.

Barrett

James

Korchin

Sheldon J.

Chodoff

Paul

. 1977. “A Measure of Interpersonal Dependency.” Journal of Personality Assessment 41:610-18.

15.

Hollingshead

August B.

1965. Two Factor Index of Social Position. New Haven, CT: Yale University.

16.

Jang

Yuri

Chiriboga

David A.

Lee

Jeongnam

Cho

Soyeon

. 2009. “Determinants of a Sense of Mastery in Korean American Elders: A Longitudinal Assessment.” Aging & Mental Health 13:99-105.

17.

Kelley-Moore

Jessica A.

Schumacher

John G.

Kahana

Eva

Kahana

Boaz

. 2006. “When Do Older Adults Become “Disabled”? Social and Health Antecedents of Perceived Disability in a Panel Study of the Oldest Old.” Journal of Health and Social Behavior 47:126-41.

18.

Kessler

Ronald C.

Greenberg

David F.

1981. Linear Panel Analysis: Models of Quantitative Change. New York, NY: Academic Press.

19.

Liddiard

Kirsty

. 2014. “The Work of Disabled Identities in Intimate Relationships.” Disability & Society 29:115-28.

20.

Long

J. Scott

Pavalko

Eliza

. 2004. “Comparing Alternative Measures of Functional Limitation.” Medical Care 42:19-27.

21.

Mead

George H.

1934. Mind, Self, and Society. Chicago, IL: University of Chicago Press.

22.

Mechanic

David

Hansell

Stephen

. 1987. “Introspection and Illness Behavior.” Psychiatric Medicine 5:5-14.

23.

Mirowsky

John

Ross

Catherine E.

1992. “Age and Depression.” Journal of Health and Social Behavior 33:187-205.

24.

Mirowsky

John

Ross

Catherine E.

2003. The Social Causes of Psychological Distress (2nd ed.). New Brunswick, NJ: Aldine Transaction.

25.

Orth

Ulrich

Trzesniewski

Kali H.

Robins

Richard W.

2010. “Self-esteem Development from Young Adulthood to Old Age: A Cohort-sequential Longitudinal Study.” Journal of Personality and Social Psychology 98:645-58.

26.

Overholser

James C.

1997. “Treatment of Excessive Interpersonal Dependency: A Cognitive-behavioral Model.” Journal of Contemporary Psychotherapy 27:283-301.

27.

Pearlin

Leonard I.

1999. “Stress and Mental Health: A Conceptual Overview.” Pp. 161-75 in A Handbook for the Study of Mental Health: Social Contexts, Theories, and Systems, edited by Horwitz

A. V.

Scheid

T. L.

Cambridge, UK: Cambridge University Press.

28.

Pearlin

Leonard I.

Schooler

Carmi

. 1978. “The Structure of Coping.” Journal of Health and Social Behavior 19:2-21.

29.

Porcerelli

John H.

Bornstein

Robert F.

Markova

Tsveti

Huprich

Steven K.

2009. “Physical Health Correlates of Pathological and Healthy Dependency in Urban Women.” Journal of Nervous and Mental Disease 197:761-65.

30.

Pudrovska

Tetyana

. 2010. “Cancer and Mastery: Do Age and Cohort Matter?” Social Science & Medicine 71:1285-91.

31.

Radloff

Lenore S.

1977. “The CES-D Scale: A Self-report Depression Scale for Research in the General Population.” Applied Psychological Measurement 1:385-401.

32.

Raque-Bogdan

Trisha L.

Ericson

Sara K.

Jackson

John

Martin

Helena M.

Bryan

Nicole A.

2011. “Attachment and Mental and Physical Health: Self-compassion and Mattering as Mediators.” Journal of Counseling Psychology 58:272-78.

33.

Reitzes

Donald C.

Mutran

Elizabeth J.

2002. “Self-concept as the Organization of Roles: Importance, Centrality, and Balance.” Sociological Quarterly 43:647-67.

34.

Reitzes

Donald C.

Mutran

Elizabeth J.

2006. “Self and Health: Factors that Encourage Self-esteem and Functional Health.” Journal of Gerontology: Social Sciences 61B:S44-S51.

35.

Riolo

Stephanie A.

Nguyen

Tuan Anh

Greden

John F.

King

Cheryl A.

2005. “Prevalence of Depression by Race/ethnicity: Findings from the National Health and Nutrition Examination Survey III.” American Journal of Public Health 95:998-1000.

36.

Rosenberg

Morris

. 1965. Society and the Adolescent Self-image. Princeton, NJ: Princeton University Press.

37.

Rosenberg

Morris

McCullough

B. Claire

. 1981. “Mattering: Inferred Significance and Mental Health among Adolescents.” Research in Community & Mental Health 2:163-82.

38.

Russell

David

Turner

R. Jay

Joiner

Thomas E.

2009. “Physical Disability and Suicidal Ideation: A Community-based Study of Risk/protective Factors for Suicidal Thoughts.” Suicide and Life-threatening Behavior 39:440-51.

39.

Schieman

Scott

. 1999. “Age, Physical Impairment, and Symptoms of Anxiety: A Test of Mediating and Moderating Factors.” International Journal of Aging and Human Development 49:43-59.

40.

Schieman

Scott

Campbell

Janice E.

2001. “Age Variations in Personal Agency and Self-esteem: The Context of Physical Disability.” Journal of Aging and Health 13:155-85.

41.

Schieman

Scott

Turner

Heather A.

1998. “Age, Disability, and the Sense of Mastery. Journal of Health and Social Behavior 39:169-86.

42.

Sparkes

Andrew C.

Smith

Brett

. 2008. “Men, Spinal Cord Injury, Memories and the Narrative Performance of Pain.” Disability & Society 23:679-90.

43.

Taylor

John

Turner

R. Jay

. 2001. “A Longitudinal Study of the Role and Significance of Mattering to Others for Depressive Symptoms.” Journal of Health and Social Behavior 42:310-25.

44.

Thoits

Peggy A.

2011. “Mechanisms Linking Social Ties and Support to Physical and Mental Health.” Journal of Health and Social Behavior 52:145-61.

45.

Thoits

Peggy A.

2013. “Self, Identity, Stress, and Mental Health.” Pp. 357-77 in Handbook of the Sociology of Mental Health (2nd ed.), edited by Aneshensel

C. S.

Phelan

J. C.

Bierman

Dordrecht, Netherlands: Springer.

46.

Trzesniewski

Kali H.

Donnellan

M. Brent

Robins

Richard W.

2003. “Stability of Self-esteem across the Life Span.” Journal of Personality and Social Psychology 84:205-20.

47.

Turner

R. Jay

Lloyd

Donald A.

Taylor

John

. 2006. “Physical Disability and Mental Health: An Epidemiology of Psychiatric and Substance Disorders.” Rehabilitation Psychology 51:214-23.

48.

Turner

R. Jay

Marino

Franco

. 1994. “Social Support and Social Structure: A Descriptive Epidemiology.” Journal of Health and Social Behavior 35:193-212.

49.

van Gundy

Karen

Schieman

Scott

. 2001. “Looking Inward: Introspectiveness, Physical Disability, and Depression over the Life Course.” International Journal of Aging and Human Development 53:293-310.

50.

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.

51.

Zitzelsberger

Hilde

. 2010. “(In)visibility: Accounts of Embodiment of Women with Physical Disabilities and Differences.” Disability & Society 20:389-403.