A Change for the Worse: Negative Social Exchanges are Associated with an Accelerated Decline in Self-Rated Health Over Time

Abstract

Objectives

Age-associated accelerated declines in physical health vary across individuals, and researchers have suggested that individual differences in decline may vary as a function of stressors. The relation of one such stressor, negative social exchanges, to accelerated declines in self-rated health is investigated.

Method

Participants are from a 2-year, 5-wave, national, longitudinal study of social relationships among older adults. Growth curve analyses are used to examine the relation of negative and positive social exchanges to accelerated changes in self-rated health, controlling for age, sex, race/ethnicity, education, and depressive symptoms.

Results:

Individuals reporting more frequent negative social exchanges showed significantly accelerated declines in physical health. Positive social exchanges were not related to linear or accelerated declines in self-rated health over time.

Discussion:

The association between negative social exchanges and accelerated deterioration in self-rated health provides general support for hypotheses that interpersonal stressors play an important role age-related physical health decline.

Keywords

negative social interactions aging self-rated health

Aging of the human body is evident in the increasing risk of morbidity and mortality in older life. Although, on average, the probability of disease or death increases with advanced age (Arias, 2002), the rate of decline in physical health varies considerably across individuals (Fries, 1980; Strehler & Mildvan, 1960; Svedberg et al., 2005). Even at older ages, physical health may remain constant or improve for some individuals (Holt-Lunstad et al., 2017; Vogelsang, 2018). Factors such as genetic differences (Shadyab & LaCroix, 2015) are likely to explain variation in physical health trajectories across individuals, but individual differences in the accumulation of stress, or allostatic load, also may explain why physical health declines more quickly in some individuals than in others (Beckie, 2012; McEwen & Stellar, 1993). Exposure to stressors, particularly over prolonged periods, is known to be important in the development of serious physical health conditions, such as hypertension and heart disease (Kivimäki & Steptoe, 2018). Stress most likely leads to disease by encumbering neuroendocrine, immune, and cardiovascular system functioning (Kiecolt-Glaser et al., 2010), which compromises the ability of people to cope with stress in late life (Spence, 1995). Accordingly, individual differences in rates of decline in physical health associated with aging may be at least partially a function of exposure to different levels of stress.

With increasing age, the risk of failure of one or more of the body’s physiological subsystems increases. The combined risk of multiple subsystems results in accelerated increases in morbidity or mortality that are observed in life expectancy studies (Arias, 2002; Gavrilov & Gavrilova, 2015a; Strehler & Mildvan, 1960). Thus, a process in which stress impacts neuroendocrine, immune, and cardiovascular system functioning implies an exponential decline in overall physical health associated with aging. Although an accelerated decline in physical health in late life is observed on average, Fries (1980) emphasized that the modifiability of this trajectory accounts for the dramatic individual differences in physical health changes that accompany advancing age, and suggested that “social interaction, promotion of health, and personal autonomy may postpone many of the phenomena usually associated with aging” (p. 135).

Social Exchanges and Physical Health

Humans, like all primates, have strong needs for well-functioning social relationships (Baumeister & Leary, 1995; Ellis et al., 2019). These social needs have an important place in hierarchical structure of human motivations necessary for survival (Kenrick et al., 2010), explaining why social relationships have an integral function in later life (Antonucci et al., 2014). Indeed, a wealth of research has documented that social exchanges are an important factor accounting for individual differences in physical health (Hawkley & Capitanio, 2015; Holt-Lundstad et al., 2017; Kiecolt-Glaser & Wilson, 2017; Newsom et al., 2008; Pietromonaco & Collins, 2017; Rook et al., 2011). We define social exchanges as interpersonal contacts with family, friends, or others in one’s social network that can be of a favorable, affirming nature, or an unfavorable, upsetting nature.

Positive social exchanges primarily consist of what is commonly referred to as social support but can also include other positive interactions, such as companionship. Positive social exchanges have long been believed to play a preventive role in physical health (Cassel, 1976), and studies have borne out their connection to lower risk of chronic health conditions, functional limitations, and mortality (de Brito et al., 2017; Holt-Lunstad et al., 2017; Rook et al., 2012; Shor et al., 2013).

Social exchanges are not always beneficial to physical health, however, as they can also be a significant source of stress in daily life (Beckie, 2012; McEwen & Stellar, 1993; Wheaton et al., 2013). Negative social exchanges, which include aversive behaviors by social network members such as criticisms, intrusiveness, and insensitivities, can be chronic in nature (Krause & Rook, 2003), leading to important cumulative impacts on health (McEwen, 2017). The association between negative social exchanges and poorer physical health has been documented across a variety of different physical health indicators, including more arthritis symptoms (Pow et al., 2018), greater risk of cardiovascular disease (Orth-Gomer et al., 2000), poorer physical functioning (Hakulinen et al., 2016; Krause & Shaw, 2002; Mavandadi et al., 2007), lower self-reported health (Heymann et al., 2016; Krause, 1996; Newsom et al., 2008), and impaired immune response (Kiecolt-Glaser & Wilson, 2017). Although many studies documenting this association have been cross-sectional, leaving ambiguity about the causal direction of the relationship, prospective studies have lent support to the notion that negative social exchanges may lead to poorer physical health (Hakulinen et al., 2016; Newsom et al., 2008).

Positive and negative social exchanges may not have equivalent impacts on physical health, however. Findings indicate greater overall impact of negative social exchanges on mental health and physical health as compared to the effects of positive social exchanges (for a review, see Rook, 2015). The differences in impacts of positive and negative social exchanges may be rooted in the centrality of social relationships for survival, in which the adaptive benefit of avoiding threatening individuals outweighs the adaptive benefit of seeking supportive individuals (Haidt, 2006; Kemeny, 2009). Negative social exchanges tend to be stable over time (Newsom et al., 2008), constituting a chronic source of stress, and tend to involve closer network members (Sorkin & Rook, 2004). Given the greater overall impact of negative social exchanges, individuals who experience more negative social exchanges should show a more rapid decline in physical health, that would account for some of the well-known individual variation in physical health declines associated with aging (Gavrilov & Gavrilova, 2015a) that are predicted by major theoretical perspectives of biological aging, such as damage or error theories (Jin, 2010). Understanding the relative impacts of negative and positive social exchanges on declines in physical health has potential implications for interventions and public health initiatives. If impact of negative social exchanges outweighs the impact of positive social exchanges, efforts could be better tailored to be more effective.

Self-Reported Health

One reflection of declining physical health can be found in changes in an individual’s own assessments of their general state of physical health. Self-ratings of health provide a useful global assessment of one’s overall, current physical health status (Benyamini, 2011; Latham & Peek, 2013). These simple but direct assessments of physical health have gained increasing credence as an important gauge of overall physical health status because studies have repeatedly demonstrated an association with mortality that is independent of other physical health measures and psychosocial predictors (Gupta et al., 2020; Idler & Cartwright, 2018; Latham & Peek, 2013; Osoba, 1999; Szybalska et al., 2018). Self-rated health is an indicator of overall physical health that may take many aspects of physical health into account, including the presence of disease (Falconer & Quesnel-Vallée, 2017), functional abilities (Latham & Peek, 2013; Takahashi et al., 2020), subclinical symptoms, and mental health. Self-rated health also exhibits exponential declines in physical health that occur with advancing age that would be expected to result from the failure of individual physiological subsystems (Bunda & Busseri, 2019; Ellaway et al., 2012). Self-rated health, therefore, is a measure that is well-suited for examining role of social exchanges in accounting for individual differences in physical health decline over time. Two studies have found that declines in self-rated health vary as a function of marital strain (Miller et al., 2013; Umberson et al., 2006). Results from these studies may not generalize to all individuals because they were limited to married individuals. Additionally, both studies modeled change over time linearly, which may be limiting if physical health in fact declines exponentially, rather than steadily over time.

Purpose of the Current Study

To date, studies that have compared the relative effects of positive and negative social exchanges on physical health have usually been cross-sectional. Cross-sectional studies are unable to demonstrate whether physical health declines over time are precipitated by social exchanges that would be expected through the supportive functions of social exchanges and deteriorating functions of negative social exchanges. The current study examines whether social exchanges explain variation in exponential declines in self-rated health associated with aging and whether positive or negative social exchanges may account for greater variability in this decline. We hypothesize that those with more frequent negative exchanges with network members will later exhibit more rapid declines in self-rated health than those with less frequent negative social exchanges. Further, we hypothesized that negative social exchanges would have a stronger relationship to declines in self-rate health than positive social exchanges. Physical health declines are expected to accelerate at an exponential rate because the increasing burden on physiological subsystems that is caused by the accumulation of greater interpersonal stress and will be manifested in accelerated risk of poor physical health. Moreover, because the impact of negative social exchanges on psychological distress has been found to outweigh the impact of positive social exchanges in cross-sectional studies, we hypothesize that negative social exchanges will have a greater effect than positive social exchanges in predicting declines in self-rated health over time as well. These hypotheses are investigated in a longitudinal study of a nationally representative sample of older adults that makes use of extensive and comparable measures of positive and negative social exchanges.

Method

Sample Recruitment and Participant Characteristics

The Late Life Study of Social Exchanges (LLSSE) is a national study of the physical and psychological health consequences positive and negative social exchanges. Participants were 916 community-dwelling older adults, ages 65–90. Only individuals who were English-speaking, cognitively functional, and living in the contiguous United States were eligible to participate. Survey responses were obtained over 2 years from the years 2000 to 2002, with interviews spaced 6 months apart (five waves). The first, third, and fifth waves were face-to-face interviews, whereas the second and fourth waves were telephone interviews. The sample was recruited from a 5% random sample of the Medicare Beneficiary Eligibility List, provided by the Centers for Medicare and Medicaid Services (CMS). CMS restricts access to the entire eligibility list, which contains tens of millions of adults. Thus, the random subsample process was used to help protect privacy and ensure that the study sample could be as representative of all Medicare-eligible older adults as practical. CMS, however, does not release information for beneficiaries who are 100 years of age and older or are non-legal residents. An initial contact was made by letter, followed by a phone call, resulting in a response rate of 53% of those who met eligibility criteria, who could be contacted, and who agreed to participate. All participants provided informed consent prior to participation in the study. This research was approved by the Portland State University (04254) and the University of California, Irvine (96-257) Institutional Review Boards.

Participants in the LLSSE were 74.16 years old on average (SD = 6.63), 62% were women, and 63% had a high school degree or less education. Eighty-three percent were White, 11% were African American, 5% were Latino, and approximately 1% were of other ethnicities (e.g., Asian, Native American). As a general check on the representativeness of the sample, an examination of U.S. Census data (U.S. Bureau of the Census: Current Population Survey, 2000) suggested the characteristics of the sample at wave 1 closely mirrored the characteristics of the older adult population at that time.

Sixty-seven percent (N = 666) completed the final wave of the study, and the primary reasons for attrition included death, physical health, change of residence, and refusal. Because of missing data on predictors in the growth curve analyses (including 30 participants who reported no positive or negative exchanges at wave 1), the final sample size used in all analyses was 608. Growth curve analysis does not require complete data on the dependent variable (self-rated health) at all waves, using expectation maximization estimation to obtain estimates where data are missing. We examined the extent to which missing data on the dependent variable was associated on model-related variables to gauge the extent to which analyses might be substantially biased by non-ignorable missing data. There were several significant differences in baseline characteristics between those who completed the final wave and those who did not, including lower self-reported health, greater difficulties in functioning, and lower education. Though significant, these differences were not large (r = −.07, r =. 08, and r = .07, respectively) and were not expected to lead to substantial biases in the overall findings.

Procedure

The survey, lasting approximately 70 minutes for in-person interviews and 15–30 minutes for telephone interviews, covered a wide range of content, including questions about sociodemographic characteristics, social relationships, physical health, and psychological distress. A major survey research firm with experience conducting surveys with older adults, Harris Interactive, Inc., conducted all interviews.

Measures

Social exchanges . A measure of negative social exchanges was designed to assess four domains of negative social exchanges: unwanted advice/intrusion, failure to provide help, unsympathetic/insensitive behavior, and rejection/neglect (Newsom et al., 2005). This measure, developed through extensive work combining qualitative methods (focus groups, card sorting tasks) and confirmatory factor analyses (Morgan, 2000; Newsom et al., 2003), was intended to provide a broad assessment of negative social exchanges with strong content validity and reliability and has been used in previously published studies (August et al., 2007; Mavandadi et al., 2007; Newsom et al., 2008; Palmer et al., 2016). Twelve items (3 per domain) assessed the frequency (ranging from 0 “never” to 4 “very often”) of negative social exchanges with a “spouse, family members, friends, neighbors, in-laws or others.” Participants were asked “In the past month, how often did the people you know…” about exchanges such as “… interfere or meddle in your personal matters?” (unwanted advice/intrusion), “…let you down when you needed help?” (failure to provide needed help), “… act unsympathetic or critical about your personal concerns?” (unsympathetic/insensitive behavior), and “…leave you out of activities you would have enjoyed?” (rejection/neglect). An average of the 12 items was used in the analyses, with higher scores indicating more frequent negative social exchanges. Cronbach’s alpha was .90 for the total negative exchanges scale.

The measure of positive social exchanges, designed to have comparable content validity and reliability, assessed four parallel domains of social exchanges: informational support, instrumental support, emotional support, and companionship. This measure included 12 items (3 per domain) assessed the frequency (ranging from 0 “never” to 4 “very often”) of positive social exchanges with a “spouse, family members, friends, neighbors, in-laws or others.” Participants were asked “In the past month, how often did the people you know…” about exchanges such as “do favors and other nice things for you” or “say things that were kind or considerate of you.” An average of the 12 items was used in the analyses, with higher scores indicating more frequent positive social exchanges. Cronbach’s alpha was .90 for the total positive exchanges measure.

Self-rated health . Self-rated health was measured by the commonly used single-item “How would you describe your health at the present time? Would you say it is excellent, very good, good, fair, or poor?” (ranging from 0 “poor” to 4 “excellent,” Ware & Sherbourne, 1992). This measure is identical to the measure of self-rated health widely used in population health studies. It has shown a remarkable consistency as an independent predictor of mortality in major health studies in the U.S. and abroad (Benyamini, 2011).

Depressive symptoms . Because some researchers have hypothesized that physical health ratings may be partially based on psychological health (Bailis et al., 2003; Fayers & Sprangers, 2002; Krause & Jay, 1994), it is important to include a measure of psychological health among covariates, thus depressive symptoms were included as a covariate in the analyses. A 9-item version of the Center for Epidemiologic Studies-Depression scale (CES-D; Radloff, 1977) adapted by Santor and Coyne (1997) assessed psychological health. This version of the measure does not include the CES-D item regarding loneliness. The standard 4-point response scale was used for the items (ranging from 0 “none of the time (<1 day)” to 3 “most of the time (5-7 days)”). Items were summed to create the scale score. Cronbach’s alpha for the resulting 8-item measure of depressive symptomatology is .84 in this sample.

Demographic characteristics . Demographic characteristics such as age, sex (0 = male, 1 = female), education (ranging from 1 “less than eighth grade” to 9 “completed graduate school or professional training”), and race/ethnicity (0 = white, 1 = racial/ethnic minority) were used as time-invariant covariates in the analyses. Depressive symptoms at baseline and demographic characteristics were used as covariates because some prior research has indicated a relationship between these factors and self-rated health (Whitmore et al., 2021).

Overview of Analyses

Our hypotheses about change in self-rated health over time and negative social exchanges were tested using growth curve modeling in HLM version 8.0 (Raudenbush et al., 2019). We first examined a growth model of linear and exponential change before testing whether changes in self-rated health were dependent on negative social exchanges and a set of time-invariant covariates. Exponential change was tested by transformation of the time scores using the following formula

X_{i t_{\exp}} = α - e^{γ X_{i t}}

where

X_{i t_{\exp}}

is the time score used to test the exponential effect, with i and t subscripts for individual and time, respectively, α is a constant representing an asymptotic maximum value for self-rated health, e is the exponential constant (i.e., approximately 2.72), γ is a constant acceleration parameter, and

X_{i t}

represents the linear time scores (0, 1, 2, 3, 4). The value of the constant, α, was 5 to obtain a maximum asymptotic value for self-rated health of 4, and the value of the constant, γ, was chosen a priori as .3 to reflect a slow deceleration in self-rated health that would approximate the deceleration found in theories of biology of aging (Kirkwood, 2015). All covariates (except the time variables) were grand mean centered, so that coefficients represented effects when all variables were at their mean values. Unstandardized coefficients and robust standard errors from the models are reported.

Results

Descriptive Analyses

Means, standard deviations, and correlations among all variables are presented in Table 1. The means for self-rated health show a slight decline over time, and the correlations among self-rated health measurements over time indicated that perceptions of physical health were relatively stable across the five waves (ranging from r = .530 to r = .617). Among the control variables, minority status (r = −.122, p < .01), education (r = −.218, p < .001), and depressive symptoms (r = −.283, p <.001) were significantly correlated with Wave 1 self-rated health, such that minority participants had lower self-rated health, participants with higher education had higher self-rated health, those with higher depressive symptom scale scores had lower self-rated health, and those with more frequent negative social exchanges had lower self-rated health. Positive social exchanges were unrelated to self-rated health (r = .026, ns), whereas negative social exchanges were related to worse self-rated health (r = −.101, p < .05).

Table 1.

Correlations Among Variables at Wave 1 and Self-Rated Health at All Waves (N = 608).

S. No		1	2	3	4	5	6	7	8	9	10	11	12
1	Age
2	Sex	.026
3	Minority	.007	.001
4	Education	−.008	−.123**	−.219***
5	Depressive symptoms	−.035	.124**	.033	−.109**
6	Positive social exchanges	−.039	.098*	.027	−.056	.053
7	Negative social exchanges	−.161***	.093*	.106**	−.086*	.346***	.113**
8	Wave 1 self-rated health	−.051	.045	−.122**	.218***	−.283***	.026	−.101*
9	Wave 2 self-rated health	−.083*	-.017	−.132***	.252***	−.237***	.057	−.054	.614***
10	Wave 3 self-rated health	−.117**	.023	−.058	.220***	−.268***	.028	−.050	.587***	.563***
11	Wave 4 self-rated health	−.128**	-.020	−.036	.191***	−.250***	.001	−.035	.530***	.569***	.581***
12	Wave 5 self-rated health	−.097*	.015	−.076	.182***	−.256***	.019	−.093*	.584***	.542***	.601***	.617***
Mean		73.67	.63	.16	4.69	4.74	2.35	.42	2.17	2.19	2.15	2.08	2.11
SD		6.40	.48	.37	1.98	4.55	.86	.55	1.11	1.08	1.05	1.05	1.05

Note. * p < .05, ** p < .01, *** p < .001. The following codes were used for binary variables: sex (0 = male, 1 = female) and minority status (0=white, 1=racial/ethnic minority).

Change in Self-Rated Health

We first examined overall change in self-rated health with an unconditional model that included linear and exponential terms but no covariates. In accordance with Singer and Willett’s (2003) recommendations, we tested the assumption of constant residual variance over time (i.e., the homoscedasticity assumption). The chi-square test of this assumption indicated a significantly better fitting model if the residuals were allowed to vary over time (χ²(2) = 6.12, p = .045), and, thus, all subsequent results are derived from models that relax this assumption. Results from the unconditional model show that, although neither linear nor exponential change in self-rated health were significant (b = −.07, SE_b = .06, p = .232; b = −.11, SE_b = .14, p = .422, respectively), both linear and exponential trajectories varied significantly across individuals (τ₁₁ = .32, χ²(606) = 707.32, p = .003; τ₂₂ = 1.26, χ²(606) = 701.89, p = .004, respectively). The next analyses examined whether social exchanges account for variability in linear or exponential changes in self-rated health over time.

A subsequent growth curve model, therefore, examined whether negative or positive social exchanges were predictors of growth parameters (baseline self-rated health and rate of change in self-rated health over time) after controlling for age, sex, education, minority status, and depressive symptoms. Results are presented in Table 2. Older age (b = −.02, SE_b = .01, p = .001), being male (b = .13, SE_b = .07, p = .048), and lower education (b = .10, SE_b = .02, p < .001) were all significant predictors of poorer self-rated health at baseline. Depressive symptoms were unrelated to baseline self-rated health (b = −.06, SE_b = .13, p = .616). Negative social exchanges (b = −2.39, SE_b = .95, p = .013) significantly predicted baseline self-rated health, but positive social exchanges did not (b = .33, SE_b = .75, p = .659). The cross-level interactions between positive or negative social exchanges and linear or exponential time test whether changes in self-rated health vary as a function of baseline social exchanges (Preacher et al., 2006). Results from the model indicate that negative, but not positive, social exchanges were significantly associated with both linear (b = .26, SE_b = .10, p = .012) and exponential change (b = .59, SE_b = .23, p = .011) in self-rated health over time. Variance components for linear and exponential time were also significant in these models (τ₁₁ = .30, χ²(603) = 702.08, p = .003 and τ₂₂ = 1.50, χ²(603) = 697.37, p = .005, respectively).

Table 2.

Growth curve analyses examining the effects of negative social exchanges on exponential changes in self-rated health (N = 608).

Predictor	B	SE_b	p
Intercept	2.66	0.58	.000
Age	−0.02	0.01	.001
Sex	0.13	0.07	.048
Minority	−0.12	0.08	.158
Education	0.10	0.02	.000
Depressive symptoms	−0.06	0.13	.616
Positive social exchanges	0.33	0.76	.659
Negative social exchanges	−2.39	0.96	.013
Linear SRH	−0.07	0.06	.225
Exponential SRH	−0.11	0.14	.413
Positive social exchanges X linear SRH	−0.04	0.08	.646
Negative social exchanges X linear SRH	0.26	0.10	.012
Positive social exchanges X exponential SRH	−0.07	0.18	.717
Negative social exchanges X exponential SRH	0.59	0.23	.011

Note. Time was coded 0, 1, 2, 3, and 4, and, thus, the intercept represents average self-rated health (SRH) at baseline. Linear SRH represents linear growth in self-rated health over time, using time coefficients 0, 1, 2, 3, and 4. Exponential SRH represents nonlinear growth using an exponential transformation of the linear time coefficients. All predictors, including covariates, were measured at baseline. The following codes were used for binary variables: sex (0 = male, 1 = female) and minority status (0=white, 1=racial/ethnic minority).

The significant interaction between negative social exchanges and exponential time indicates that self-rated health changes at different rates for individuals reporting more frequent or less frequent negative social exchanges. To explore the nature of this interaction, we used the resulting growth curve equation to plot simple slopes for exponential change when the value of negative social exchanges was equal to 0, the mean, and 1 SD above the mean. Although 1 SD below the mean is often conventionally used for the lowest value at which the simple slope is plotted (Aiken & West, 1991), we used 0 (i.e., negative exchanges experienced “not at all” over the past month), as our lower bound plot point, because using −1 SD below the mean would have resulted in a slightly negative value on the scale. As shown in Figure 1, these simple slopes suggest that individuals with the highest negative social exchanges experienced more rapid exponential declines in self-rated health.

Figure 1.

Simple slopes for changes in self-rated health at no negative exchanges, average frequency of negative exchanges, and negative exchanges 1 SD above the mean.

Discussion

This study goes beyond prior work by investigating whether social exchanges predict differing rates of decline in self-rated health over a 2-year period, and, furthermore, whether negative social exchanges might be a stronger predictor of self-rated health decline than positive exchanges. Negative, but not positive, social exchanges significantly predicted declining self-rated health. Moreover, the results indicated that those who reported more frequent negative social exchanges were significantly more likely to show an exponential decay in self-rated health over the course of the study, reflecting a more rapid rate of deteriorating physical health for these individuals.

The results of this study have important implications for the role of stress in the aging process and add to several studies that have shown a relation between interpersonal conflicts and changes in self-rated health (Newsom et al., 2008; Umberson et al., 2006; Wickrama et al., 1997). Exposure to stressors is expected to accelerate physical health decline in older age through a weakening of the immune system’s capacity to recover from stress (McEwen & Lasley, 2002). Results of this study identify interpersonal interactions as one important source of stress associated with an accelerated decline in physical health. The potency of negative social exchanges as a stressor is likely derived from its interference with a marked need for affiliation among humans and other primates (Baumeister & Leary, 1995). Experimental studies of social disruption among non-human primates, in fact, have illustrated that social stress is noxious enough to have appreciable effects on longevity (Lund et al., 2014). This study forges new ground, however, in documenting the association of negative social exchanges and accelerated changes in self-reported health in a nationally representative sample of older adults.

The negative social exchanges measured in our study may have their impacts on physical health through cumulative process that occurs when interpersonal stressors are chronic in nature (McEwen, 2017; Pearlin & Bierman, 2013; Wheaton et al., 2013). We did not examine cumulative or chronically occurring negative social exchanges in these analyses, and, therefore, the results may represent an underestimation of the association with physical health (Newsom et al., 2008). Our measurement of negative social exchanges was global in nature and was not able to distinguish between less serious negative social exchanges, such as minor criticisms and more serious negative social exchanges, such as abuse. Moreover, the point during the life course that negative socials exchanges occur also may matter. Some authors have speculated that early damage due to more extreme stressors may have an important impact on physical health decline in later life (Gavrilov & Gavrilova, 2015b). Future research may help uncover how early life interpersonal stressors may interact with those experienced in middle and later adulthood to affect the rate of deterioration in physical health in later life (Suglia et al., 2021).

One question raised by the study is why positive exchanges were not protective against declines in self-rated health. Evidence suggests that receiving social support can trigger feelings of low self-efficacy or indebtedness in recipients (Zee & Bolger, 2019). Support can also be mismatched with recipients’ needs, as is sometimes the case with unwanted or insensitive advice (J. Smith & Goodnow, 1999). Thus, even well-intentioned support can have unintended negative effects. Compared with instrumental support, companionship should have fewer negative effects, as it is usually sought for the purpose of mutual enjoyment and, therefore, is less likely to be unwanted. Companionship also occurs between equals, rather than a helper and help-receipt, and, as a result, is less likely to arouse feelings of low self-efficacy or indebtedness (Rook et al., 2011). Comparing the links between self-rated health and support versus companionship was beyond the scope of the current study, but it warrants investigation in future research.

Limitations

Although an accelerated decline in self-rated health occurred in some individuals, we did not observe an accelerated decline in self-rated health on average. The relatively short 2-year time frame of our study, compared with other studies that have traced self-rated health changes over the life span (Svedberg et al., 2005), may explain why we did not find an average change in self-rated health in the sample. The short time frame of the current study represents only a brief cross-section of a longer developmental process, and, therefore, the nonlinear decline in health observed in our study should be far less evident than the decline observed over longer periods in other studies. Over longer time spans and at older ages, one might expect to see a more dramatic decay pattern on average. The role of negative social exchanges, to the extent that they are persistent, would also be expected to be stronger if observations were taken over longer periods, because development of more serious, chronic conditions takes place gradually over many years (Kaplan et al., 1999). Presumably those exposed to chronic stressors will experience greater impairment of the body’s restorative systems, resulting in greater risk of developing disease.

Our measure of global health is identical to the measure used in numerous national health surveys and epidemiological research, such as the Health and Retirement Study (HRS; Heeringa & Conner, 1995) or the National Health and Aging Trends Study (NHATS; Kasper & Freedman, 2014). Despite its widespread use, the single-item rating of overall health has been viewed with some skepticism by some researchers because the factors that individuals take into account in making ratings and the processes by which they weight those factors are largely unknown (Bombak, 2013). Respondents appear to take into account health conditions, functioning levels, and subclinical symptoms, and, thus, self-ratings of health do not appear to reflect a consistent or specific facet of health (Kaplan et al., 1999; Krause & Jay, 1994; Schnittker, 2003). A single-item rating may be subject to response bias, such as the tendency for some individuals to rate their health more negatively because they experience greater negative affect (Bailis et al., 2003), although the inclusion of a depression measure in this study at least partially addresses this possibility. Evidence supporting the predictive value of global self-ratings of health is irrefutable, however. Idler and Benyamini (1997) concluded from a review of 27 studies, that self-rated health represents an “irreplaceable dimension of health status” and that “an individual’s health status cannot be assessed without it” (p. 34).

Although the association of negative social exchanges and exponential declines in health is consistent with proposed notions of stress and aging, our study was unable to directly confirm these processes because we did not obtain data on immunologic or neuroendocrine markers involved in the hypothesized underlying mediational process. Cumulative effects of recurring or chronic stress are expected to increase the risk of disease and health decline through several physiologic systems (McEwen, 2017). Additional research that tests this proposed process and links it to older adults’ social exchanges will be valuable in advancing understanding of how some kinds of social ties and interactions may be implicated in declining health in later life.

Conclusion

The findings from our study provide further insight into two broad observations that frequently appear in the literature. The first observation is that changes in health in later life are nonlinear and that considerable variability in trajectories of health occurs with age (Gavrilov & Gavrilova, 2003; Nelson & Dannifer, 1992; Strehler & Mildvan, 1960). Evidence for this variability has generally been based on simple measures of dispersion taken cross-sectionally across cohorts. We add evidence by bringing more sophisticated individual growth curve analyses to bear on the issue, however. Our data suggest, as well, that there is significant individual variation in these nonlinear patterns of change in health over time.

The second broad observation in the literature that our study addresses concerns the differential effects of positive and negative social exchanges on health. The effects of negative social exchanges on well-being tend to be greater than the corresponding effects of positive social exchanges (Rook, 2015). This finding is consistent with views in the literature that “… bad is stronger than good. Responses to threats and unpleasantness are faster, stronger, and harder to inhibit than responses to opportunities and pleasures” (Haidt, 2006, p. 29; see also; Baumeister et al., 2001). Such “negativity bias” is not a new concept (A. Smith, 1759/2002), but is a well-documented phenomenon (Tierney & Baumeister, 2021). Given that the adverse health impact of negative social exchanges appears to outweigh the ameliorative impact of positive exchanges, these results have implications for interventions and policy. Specifically, the results suggest that greater focus is needed on reducing the impacts of harmful social interactions, in contrast to a more widely assumed focus of bolstering supportive interactions (Sitges-Maciá et al., 2021).

Footnotes

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported by National Institute on Aging Grants R01 AG022957 (Newsom), R01 AG14130 (Rook), and R01 AG009221 (Krause).

ORCID iD

Jason T. Newsom

References

Aiken

L. S.

West

S. G.

(1991). Multiple regression: Testing and interpreting interactions. Sage.

Antonucci

T. C.

Ajrouch

K. J.

Birditt

K. S.

(2014). The convoy model: Explaining social relations from a multidisciplinary perspective. The Gerontologist, 54(1), 82–92. https://doi.org/10.1093/geront/gnt118

Arias

(2002). United States life tables, 2000. National vital statistics reports (51). National Center for Health Statistics.

August

K. J.

Rook

K. S.

Newsom

J. T.

(2007). The joint effects of life stress and negative social exchanges on emotional distress. The Journals of Gerontology Series B: Psychological Sciences and Social Sciences, 62(5), S304–S314. https://doi.org/10.1093/geronb/62.5.S304

Bailis

D. S.

Segall

Chipperfield

J. G.

(2003). Two views of self-rated general health status. Social Science & Medicine, 56(2), 203–217. https://doi.org/10.1016/S0277-9536(02)00020-5

Baumeister

R. F.

Bratslavsky

Finkenauer

Vohs

K. D.

(2001). Bad is stronger than good. Review of General Psychology, 5(4), 323–370. https://doi.org/10.1037/1089-2680.5.4.323

Baumeister

R. F.

Leary

M. R.

(1995). The need to belong: Desire for interpersonal attachments as a fundamental human motivation. Psychological Bulletin, 117(3), 497–529. https://doi.org/10.1037/0033-2909.117.3.497

Beckie

T. M.

(2012). A systematic review of allostatic load, health, and health disparities. Biological Research for Nursing, 14(4), 311–346. https://doi.org/10.1177/1099800412455688

Benyamini

(2011). Why does self-rated health predict mortality? An update on current knowledge and a research agenda for psychologists. Psychology & Health, 26(11), 1407–1413. https://doi.org/10.1080/08870446.2011.621703

10.

Bombak

A. E.

(2013). Self-rated health and public health: A critical perspective. Frontiers in Public Health, 1(15), 1–4. https://doi.org/10.3389/fpubh.2013.00015.

11.

Bunda

Busseri

M. A.

(2019). Subjective trajectories for self-rated health as a predictor of change in physical health over time: Results from an 18-year longitudinal study. Social Cognition, 37(3), 206–228. https://doi.org/10.1521/soco.2019.37.3.206

12.

Cassel

(1976). Contribution of the social environment to host resistance. American Journal of Epidemiology, 104(2), 107–123. https://doi.org/10.1093/oxfordjournals.aje.a112281

13.

de Brito

T. R. P.

Nunes

D. P.

Corona

L. P.

da Silva Alexandre

de Oliveira Duarte

Y. A.

(2017). Low supply of social support as risk factor for mortality in the older adults. Archives of Gerontology and Geriatrics, 73(2017), 77–81. https://doi.org/10.1016/j.archger.2017.07.016

14.

Ellaway

Benzeval

Green

Leyland

Macintyre

(2012). Getting sicker quicker”: Does living in a more deprived neighbourhood mean your health deteriorates faster? Health & Place, 18(2), 132–137. https://doi.org/10.1016/j.healthplace.2011.08.005

15.

Ellis

Snyder-Mackler

Ruiz-Lambides

Platt

M. L.

Brent

L. J.

(2019). Deconstructing sociality: The types of social connections that predict longevity in a group-living primate. Proceedings of the Royal Society B, 286(1917), Article 20191991. https://doi.org/10.1098/rspb.2019.1991

16.

Falconer

Quesnel-Vallée

(2017). Pathway from poor self-rated health to mortality: Explanatory power of disease diagnosis. Social Science & Medicine, 190(2017), 227–236. https://doi.org/10.1016/j.socscimed.2017.08.008

17.

Fayers

P. M.

Sprangers

M. A. G.

(2002). Understanding self-rated health. The Lancet, 359(9302), 187–188. https://doi.org/10.1016/s0140-6736(02)07466-4

18.

Fries

J. F.

(1980). Aging, natural death, and the compression of morbidity. The New England Journal of Medicine, 303(3), 130–135. https://doi.org/10.2307/3349864

19.

Gavrilov

L. A.

Gavrilova

N. S.

(2003). The quest for a general theory of aging and longevity. Science of Aging Knowledge and Environment, 2003(28), 1–10. https://doi.org/10.1126/sageke.2003.28.re5

20.

Gavrilov

L. A.

Gavrilova

N. S.

(2015a). New developments in the biodemography of aging and longevity. Gerontology, 61(4), 364–371. https://doi.org/10.1159/000369011

21.

Gavrilov

L. A.

Gavrilova

N. S.

(2015b). Predictors of exceptional longevity: Effects of early-life and midlife conditions, and familial longevity. North American Actuarial Journal, 19(3), 174–186. https://doi.org/10.1080/10920277.2015.1018390

22.

Gupta

Montgomery

(2020). The role of inflammation in the relationship of self-rated health with mortality and implications for public health: Data from the English Longitudinal Study of Aging (ELSA). Brain, Behavior, & Immunity-Health, 8(2020), Article 100139. https://doi.org/10.1016/j.bbih.2020.100139

23.

Haidt

(2006). The happiness hypothesis: Finding modern truth in ancient wisdom. Basic Books.

24.

Hakulinen

Pulkki-Råback

Jokela

Ferrie

J. E.

Aalto

A. M.

Virtanen

Kivimäki

Vahtera

Elovainio

(2016). Structural and functional aspects of social support as predictors of mental and physical health trajectories: Whitehall II cohort study. Journal of Epidemiology and Community Health, 70(7), 710–715. https://doi.org/10.1136/jech-2015-206165

25.

Hawkley

L. C.

Capitanio

J. P.

(2015). Perceived social isolation, evolutionary fitness and health outcomes: A lifespan approach. Philosophical Transactions of the Royal Society B: Biological Sciences, 370(1669), Article 20140114. https://dx-doi-org.web.bisu.edu.cn/10.1098/rstb.2014.0114

26.

Heeringa

S. G.

Connor

(1995). Technical description of the health and retirement study sample design. Institute for Social Research, University of Michigan. https://doi.org/10.7826/isr-um.06.585031.001.05.0001.1995

27.

Heymann

A. D.

Kaplan

Freidman

Baron-Epel

(2016). Unsupportive social interactions are associated with poorer self-reported health in Israeli patients with diabetes. American Journal of Health Behavior, 40(5), 645–651. https://doi.org/10.5993/AJHB.40.5.11

28.

Holt-Lunstad

Robles

T. F.

Sbarra

D. A.

(2017). Advancing social connection as a public health priority in the United States. American Psychologist, 72(6), 517–530. https://doi.or/10.1037/amp0000103

29.

Idler

E. L.

Benyamini

(1997). Self-rated health and mortality: A review of twenty-seven community studies. Journal of Health and Social Behavior, 38(1), 21–37. https://doi.org/10.2307/2955359

30.

Idler

E. L.

Cartwright

(2018). What do we rate when we rate our health? Decomposing age-related contributions to self-rated health. Journal of Health and Social Behavior, 59(1), 74–93. https://doi.org/10.1177/0022146517750137

31.

Jin

(2010). Modern biological theories of aging. Aging and Disease, 1(2), 72–74.

32.

Kaplan

G. A.

Haan

M. N.

Wallace

B. R.

(1999). Understanding the changing risk factor associations with increasing age in adults. Annual Review of Public Health, 20, 89–108. https://doi.org/10.1146/annurev.publhealth.20.1.89

33.

Kasper

Freedman

(2014). National health and aging trends study user guide: Rounds 1, 2 & 3 beta release. Johns Hopkins University School of Public Health. www.NHATS.org

34.

Kemeny

M. E.

(2009). Psychobiological responses to social threat: Evolution of a psychological model in psychoneuroimmunology. Brain, Behavior, and Immunity, 23(1), 1–9. https://doi.org/10.1016/j.bbi.2008.08.008

35.

Kenrick

D. T.

Griskevicius

Neuberg

S. L.

Schaller

(2010). Renovating the pyramid of needs: Contemporary extensions built upon ancient foundations. Perspectives on Psychological Science, 5(3), 292–314. https://doi.org/10.1177/1745691610369469

36.

Kiecolt-Glaser

J. K.

Gouin

J. P.

Hantsoo

(2010). Close relationships, inflammation, and health. Neuroscience and Biobehavioral Reviews, 35(1), 33–38. https://doi.org/10.1016/j.neubiorev.2009.09.003

37.

Kiecolt-Glaser

J. K.

Wilson

S. J.

(2017). Lovesick: How couples’ relationships influence health. Annual Review of Clinical Psychology, 13, 421–443. https:doi.org/10.1146/annurev-clinpsy-032816-045111

38.

Kirkwood

T. B.

(2015). Deciphering death: A commentary on Gompertz (1825)‘On the nature of the function expressive of the law of human mortality, and on a new mode of determining the value of life contingencies’. Philosophical Transactions of the Royal Society B: Biological Sciences, 370(1666), Article 20140379. https://doi.org/10.1098/rstb.2014.0379

39.

Kivimäki

Steptoe

(2018). Effects of stress on the development and progression of cardiovascular disease. Nature Reviews Cardiology, 15(4), 215–229. https://doi.org/10.1038/nrcardio.2017.189

40.

Krause

(1996). Neighborhood deterioration and self-rated health in later life. Psychology and Aging, 11(2), 342–352. https://doi.org/10.1037/0882-7974.11.2.342

41.

Krause

Jay

G. M.

(1994). What do global self-rated health items measure? Medical Care, 32(9), 930–942. https://doi.org/10.1097/00005650-199409000-00004

42.

Krause

Rook

K. S.

(2003). Negative interaction in late life: Issues in the stability and generalizability of conflict across relationships. Journal of Gerontology: Psychological Science, 58(2), P88–P99. https://doi.org/10.1093/geronb/58.2.P88

43.

Krause

Shaw

B. A.

(2002). Negative interaction and changes in functional disability during late life. Journal of Social and Personal Relationships, 19(3), 339–359. https://doi.org/10.1177/0265407502193003

44.

Latham

Peek

C. W.

(2013). Self-rated health and morbidity onset among late midlife US adults. Journals of Gerontology Series B: Psychological Sciences and Social Sciences, 68(1), 107–116. https://doi.org/10.1093/geronb/gbs104

45.

Lund

Christensen

Nilsson

C. J.

Kriegbaum

Rod

N. H.

(2014). Stressful social relations and mortality: A prospective cohort study. Journal of Epidemiology and Community Health, 68(8), 720–727. https://doi:10.1136/jech-2013-203675

46.

Mavandadi

Rook

K. S.

Newsom

J. T.

(2007). Positive and negative social exchanges and disability in later life: An investigation of trajectories of change. The Journals of Gerontology Series B: Psychological Sciences and Social Sciences, 62(6), S361–S370. https://doi.org/10.1093/geronb/62.6.S361

47.

McEwen

B. S.

(2017). Neurobiological and systemic effects of chronic stress. Chronic Stress, 1, 2470547017692328. https://doi.org/10.1177/2470547017692328

48.

McEwen

B. S.

Lasley

E. N.

(2002). The end of stress as we know it. Joseph Henry Publishing.

49.

McEwen

B. S.

Stellar

(1993). Stress and the individual: Mechanisms leading to disease. Archives of Internal Medicine, 153(18), 2093–2101. https://doi.org/10.1001/archinte.153.18.2093

50.

Miller

R. B.

Hollist

C. S.

Olsen

Law

(2013). Marital quality and health over 20 years: A growth curve analysis. Journal of Marriage and Family, 75(3), 667–680. https://doi.org/10.1111/jomf.12025

51.

Morgan

D. L.

(2000, November). Creating comparable measures of positive and negative social exchanges. Paper presented at the annual meeting of the Gerontological Society of America. Gerontological Society of America

52.

Nelson

E. A.

Dannifer

(1992). Aged heterogeneity: Fact or fiction? The fate of diversity in gerontological research. The Gerontologist, 32(1), 17–23. https://doi.org/10.1093/geront/32.1.17

53.

Newsom

J. T.

Rook

K. S.

Nishishiba

Sorkin

Mahan

T. L.

(2005). Understanding the relative importance of positive and negative social exchanges: Examining specific domains and appraisals. Journals of Gerontology: Psychological Sciences, 60(6), P304–P312. https://doi.org/10.1093/geronb/60.6.P304

54.

Newsom

J. T.

Nishishiba

Morgan

D. L.

Rook

K. S.

(2003). The relative importance of three domains of positive and negative social exchanges: A longitudinal measure with comparable measures. Psychology and Aging, 18(4), 746–754. https://doi.org/10.1037/0882-7974.18.4.746

55.

Newsom

J. T.

Mahan

T. L.

Rook

K. S.

Krause

(2008). Stable negative social exchanges and health. Health Psychology, 27(1), 78–86. https://doi.org/10.1037/0278-6133.27.1.78

56.

Orth-Gomer

Wamala

S. P.

Horsten

Schenk-Gustafsson

Schneiderman

Mittleman

M. A.

(2000). Marital stress worsens prognosis for women with coronary artery disease: The Stockholm female coronary risk study. Journal of the American Medical Association, 284(23), 3008–3014. https://doi.org/10.1001/jama.284.23.3008

57.

Osoba

(1999). What has been learned from measuring health-related quality of life in clinical oncology. Oral Oncology, 35(11), 1565–1570. https://doi.org/10.1016/S0959-8049(99)00192-6

58.

Palmer

A. D.

Newsom

J. T.

Rook

K. S.

(2016). How does difficulty communicating affect the social relationships of older adults? An exploration using data from a national survey. Journal of Communication Disorders, 62(2016), 131–146. https://doi.org/10.1016/j.jcomdis.2016.06.002

59.

Pearlin

L. I.

Bierman

(2013). Current issues and future directions in research into the stress process. In Aneshensel

C. S.

Phelan

J. C.

Bierman

(Eds.), Handbook of the sociology of mental health (2nd ed., pp. 325–340). Springer. https://doi.org/10.1007/978-94-007-4276-5_16

60.

Pietromonaco

P. R.

Collins

N. L.

(2017). Interpersonal mechanisms linking close relationships to health. American Psychologist, 72(6), 531–542. https://doi.org/10.1037%2Famp0000129

61.

Pow

Stephenson

Hagedoorn

DeLongis

(2018). Spousal support for patients with rheumatoid arthritis: Getting the wrong kind is a pain. Frontiers in Psychology, 9(1760), 1–11. https://doi.org/10.3389/fpsyg.2018.01760

62.

Preacher

K. J.

Curran

P. J.

Bauer

D. J.

(2006). Computational tools for probing interactions in multiple linear regression, multilevel modeling, and latent curve analysis. Journal of Educational and Behavioral Statistics, 31(4), 437–448. https://doi.org/10.3102/10769986031004437

63.

Radloff

L. S.

(1977). The CES-D scale: A self-report depression scale for research in the general population. Applied Psychology and Measurement, 1(3), 385–401. https://doi.org/10.1177/014662167700100306

64.

Raudenbush

S. W.

Bryk

A. S.

Cheong

Y. F.

Congdon

du Toit

(2019). HLM 8: Hierarchical linear & non-linear modeling. Scientific Software International.

65.

Rook

K. S.

(2015). Social networks in later life: Weighing positive and negative effects on health and well-being. Current Directions in Psychological Science, 24(1), 45–51. https://doi.org/10.1177/0963721414551364

66.

Rook

K. S.

August

K. J.

Sorkin

D. H.

(2011). Social network functions and health. In Contrada

R. J.

Baum

(Eds.), The handbook of stress science: Biology, psychology, and health (pp. 123–135). Springer.

67.

Rook

K. S.

Luong

Sorkin

D. H.

Newsom

J. T.

Krause

(2012). Ambivalent versus problematic social ties: Implications for psychological health, functional health, and interpersonal coping. Psychology and Aging, 27(4), 912–923. https://dx-doi-org-s.web.bisu.edu.cn/10.1037%2Fa0029246

68.

Santor

D. A.

Coyne

J. C.

(1997). Shortening the CES-D to improve its ability to detect cases of depression. Psychological Assessment, 9(3), 233–243. https://doi.org/10.1037/1040-3590.9.3.233

69.

Schnittker

(2003). The shifting signification of self-rated health. University of Pennsylvania Population Aging Research Center Working Paper Series.

70.

Shadyab

A. H.

LaCroix

A. Z.

(2015). Genetic factors associated with longevity: A review of recent findings. Ageing Research Reviews, 19(2), 1–7. https://doi.org/10.1016/j.arr.2014.10.005

71.

Shor

Roelfs

D. J.

Yogev

(2013). The strength of family ties: A meta-analysis and meta-regression of self-reported social support and mortality. Social Networks, 35(4), 626–638. https://doi.org/10.1016/j.socnet.2013.08.004

72.

Singer

J. D.

Willett

J. B.

(2003). Applied longitudinal data analysis: Modeling change and event occurrence. Oxford University Press.

73.

Sitges-Maciá

Bonete-López

Sánchez-Cabaco

Oltra-Cucarella

(2021). Effects of e-health training and social support interventions for informal caregivers of people with dementia—a narrative review. International Journal of Environmental Research and Public Health, 18(15), Article 7728. https://www.mdpi.com/1660-4601/18/15/7728

74.

Smith

(1759/2002) The theory of moral sentiments. edited by K. Haakonssen. Cambridge University Press.

75.

Smith

Goodnow

J. J.

(1999). Unasked-for support and unsolicited advice: Age and the quality of social experience. Psychology and Aging, 14(1), 108–121. https://doi.org/10.1037/0882-7974.14.1.108

76.

Sorkin

D. H.

Rook

K. S.

(2004). Interpersonal control strivings and vulnerability to negative social exchanges in later life. Psychology and Aging, 19(4), 555–564. https://10.1037/0882-7974.19.4.555

77.

Spence

A. P.

(1995). Biology of human aging. Prentice-Hall.

78.

Strehler

B. L.

Mildvan

A. S.

(1960). General theory of mortality and aging. Science, 132(3418), 14–21. https://doi.org/10.1126/science.132.3418.14

79.

Suglia

S. F.

Appleton

A. A.

Bleil

M. E.

Campo

R. A.

Dube

S. R.

Fagundes

Heard-Garris

N. J.

Johnson

S. B.

Slopen

Stoney

C. M.

Watamura

S. E.

(2021). Timing, duration, and differential susceptibility to early life adversities and cardiovascular disease risk across the lifespan: Implications for future research. Preventive Medicine, 153(106736), 1–8. https://doi.org/10.1016/j.ypmed.2021.106736

80.

Svedberg

Gatz

Lichtenstein

Sandin

Pedersen

N. L.

(2005). Self-rated health in a longitudinal perspective: A 9-year follow-up twin study. Journal of Gerontology: Social Sciences, 60(6), S331–S340. https://doi.org/10.1093/geronb/60.6.S331

81.

Szybalska

Broczek

Puzianowska-Kuznicka

Slusarczyk

Chudek

Skalska

Mossakowska

(2018). Self-rated health and its association with all-cause mortality of older adults in Poland: The PolSenior project. Archives of Gerontology and Geriatrics, 79(2018), 13–20. https://doi.org/10.1016/j.archger.2018.07.016

82.

Takahashi

Tanno

Yonekura

Ohsawa

Kuribayashi

Ishibashi

Omama

Tanaka

Sasaki

Tsubota-Utsugi

Takusari

Koshiyama

Onoda

Sakata

Itai

Okayama

(2020). Poor self-rated health predicts the incidence of functional disability in elderly community dwellers in Japan: A prospective cohort study. BMC Geriatrics, 20(1), 328. https://doi.org/10.1186/s12877-020-01743-0

83.

Tierney

Baumeister

R. F.

(2021). The power of bad: How the negativity effect rules us and how we can rule it. Penguin Books.

84.

Umberson

Williams

Powers

D. A.

Liu

Needham

(2006). You make me sick: Marital quality and health over the life course. Journal of Health and Social Behavior, 47(1), 1–16. https://doi.org/10.1177/002214650604700101

85.

U. S. Bureau of the Census: Current population survey (2000). America’s families and living arrangements: Population characteristics. Retrieved on September 9, 2003, from http://www.census.gov/prod/2001pubs/p20-537.pdf

86.

Vogelsang

E. M.

(2018). Feeling better at this age? Investigating three explanations for self-rated health improvements among the oldest-old. The Gerontologist, 58(5), 825–834. https://doi.org/10.1093/geront/gnx149

87.

Ware

J. E.

Sherbourne

C. D.

(1992). The MOS 36-item short form health survey (SF-36): Conceptual framework and item selection. Medical Care, 30(6), 473–483. https://doi.org/10.1097/00005650-199206000-00002

88.

Wheaton

Young

Montazer

Stuart-Lahman

(2013). Social stress in the twenty first century. In Aneshensel

C. S.

Phelan

J. C.

Bierman

(Eds.), Handbook of the sociology of mental health (pp. 299–323). Springer. https://doi.org/10.1007/978-94-007-4276-5_15

89.

Whitmore

Markle-Reid

Fisher

McAiney

Ploeg

(2021). What factors shape self-reported health among community-dwelling older adults? A scoping review. Canadian Journal on Aging, 40(2), 177–192. https://doi.org/10.1017/S0714980820000124

90.

Wickrama

K. A. S.

Lorenz

F. O.

Conger

R. D.

Elder

G. H.

Jr (1997). Marital quality and physical illness: A latent growth curve analysis. Journal of Marriage and the Family, 59(1), 143–155. https://doi.org/10.2307/353668

91.

Zee

K. S.

Bolger

(2019). Visible and invisible social support: How, why, and when. Current Directions in Psychological Science, 28(3), 314–320. https://doi.org/10.1177/0963721419835214