Education and Psychosocial Functioning Among Older Adults: 4-Year Change in Sense of Control and Hopelessness

Abstract

Objectives

This study investigates education differences in levels and change in sense of control and hopelessness among older adults.

Method

We used data from the Health and Retirement Study, an ongoing biennial survey of a nationally representative sample of older Americans, to examine education differences in sense of control (e.g., mastery and perceived constraints) and hopelessness. Our sample included 8,495 adults aged 52 and older who were interviewed in 2006/2008 and 2010/2012. We assessed separate models for change in sense of control and hopelessness, accounting for recent changes in social circumstances and health status.

Results

Low mastery, perceived constraints, and hopelessness were highest among individuals with less than a high school education. Over a 4-year period, this group experienced the greatest declines in psychosocial functioning, as indicated by greater increases in low mastery, perceived constraints, and hopelessness. Education differences existed net of recent negative experiences, specifically the loss of intimate social relationships and social support and increases in disease and disability.

Discussion

These findings highlight the importance of education for sense of control and hopelessness in older adulthood and demonstrate the cumulative advantage of higher levels of education for psychosocial functioning.

Health and Retirement Study, Longitudinal analysis, Mastery, Perceived constraints

Maintaining psychosocial functioning in the transition to old age is critical. Psychosocial resources are important for coping with the declines and losses that often accompany aging (Jopp & Rott, 2006). Moreover, maintaining high levels of psychosocial functioning may increase older adults’ ability to take advantage of opportunities for positivity and productivity as they age (Rowe & Kahn, 1997). Prior research suggests that socioeconomically disadvantaged adults, such as those with low educational attainment, experience poorer psychosocial functioning (Harper et al., 2002). The advantages that accompany increased education, including greater access to social and material resources and fewer adverse experiences over the life course (Mirowsky & Ross, 2003), may increase one’s ability to maintain psychosocial functioning. It is unclear, however, whether more educated adults are also more likely to maintain greater functioning as they age. The purpose of this study is to determine whether there are education differences in change in psychosocial functioning over 4 years. We focus on two aspects of psychosocial functioning, control and hopelessness, because they capture an older adult’s ability to maintain self-efficacy and avoid losing hope as they confront the challenges and opportunities of growing older.

Background

Psychosocial functioning is a multidimensional concept that encompasses the adaptive psychosocial resources older adults call upon to cope with the challenges of aging. High psychosocial functioning is reflected in an older adult’s capacity to maintain a positive outlook and achieve positive outcomes despite experiencing the adverse, and often irreversible, losses that tend to occur in the transition to old age. This capacity is captured by two related measures of psychosocial functioning—sense of control and hopelessness—that assess the extent to which older adults feel like they can control what happens to them, rather than give up and lose hope as they age. Efficacy in managing current life conditions and the ability to maintain positivity about future events or conditions are important psychosocial resources for older adults (Jopp & Rott, 2006). A joint consideration of sense of control and hopelessness is, therefore, useful for obtaining a comprehensive assessment of psychosocial functioning in older adulthood.

Sense of control refers to a person’s belief in their ability to influence life events and circumstances. It captures the degree to which the conditions of life are attributed to an individual’s own actions as opposed to the actions of others (Ross & Mirowsky, 2013) or external forces (Lachman, 2006). Prior research suggests that control declines at older ages (Mirowsky & Ross, 2007; Wolinsky, Wyrwich, Babu, Kroenke, & Tierney, 2003), which is a period of the life course when it may be especially important for maintaining good health. Among older adults, control buffers the negative effects of stress on health (Pudrovska, Schieman, Pearlin, & Nguyen, 2005), protects against declines in health and physical functioning (Infurna & Mayer, 2015; Lachman & Agrigoroaei, 2010), and is associated with lower mortality risk (Infurna, Ram, & Gerstorf, 2013). Additionally, sense of control is associated with greater happiness during the transition into retirement (Calvo, Haverstick, & Sass, 2009) and better cognitive functioning (Lachman & Agrigoroaei, 2012).

Sense of control is also influenced by the aging experience. As older adults age they must increasingly deal with personal and social losses, including, for example, the loss of independence and social ties (Carr, 2004; Rowe & Kahn, 1997). The accumulation of these losses over the life course may tax feelings of efficacy and lead to a diminished sense of control at older ages. Older adults who, on the other hand, are able to maintain high levels of control despite these challenges are better equipped to confront and cope with additional challenges while maintaining a high quality of life. Thus, a strong sense of control is adaptive in advanced age; it indicates how well older adults are coping with changes in their life circumstances and whether they continue to feel efficacious in achieving their goals.

Another indicator of positive adaptation to the negative aspects of aging is an older adult’s ability to avoid feelings of hopelessness. Hopelessness refers to a set of negative beliefs and expectations about one’s future that lead to a sense that challenges are insurmountable and opportunities for improving life conditions are minimal. Aging may increase the risk of hopelessness among older adults, eroding their sense of positivity about the future. For example, a study of German adults aged 70 and older found that a decline in positive orientation toward the future over a 14-year period was associated with decreased well-being over time (Kotter-Grühn & Smith, 2011). Aging-related declines in positive future orientation suggest that hopelessness may increase with age because the absence of a positive outlook is characteristic of hopelessness.

Hopelessness can have a profound impact on the lives of older adults. Greater levels of hopelessness have been linked to increased risk and accelerated progression of atherosclerosis (Everson, Kaplan, Goldberg, Salonen, & Salonen, 1997). Older adults who feel hopeless are also more likely to have poor memory (Crane, Bogner, Brown, & Gallo, 2007) and cognitive function (Håkansson, Soininen, Winblad, & Kivipelto, 2015). Hopelessness can also impact an older adult’s will to live. For instance, medically ill older adults who report feeling hopeless are less likely to want life-sustaining medical interventions (Menon, Campbell, Ruskin, & Hebel, 2000; Rodin et al., 2009). In addition, hopelessness has been identified as a risk factor for suicide in older adults (Britton et al., 2008; Stanley, Hom, Rogers, Hagan, & Joiner Jr, 2015). Maintaining hope in older adulthood is therefore a critical aspect of psychosocial functioning.

Greater educational attainment may help older adults maintain their sense of control and positive outlook on life. Education develops personal resources that are important for achieving personal goals and managing changes in life circumstances. Educational attainment also creates access to social and economic opportunities, such as higher earnings (Mirowsky & Ross, 2003), that can be used to minimize exposure to stressful experiences and adverse living conditions. These resources can be mobilized to help cope with aging-related declines in health and physical functioning and adverse live events including the loss of a spouse. For example, centenarians are a group of very long-lived adults who have experienced decades of aging-related changes and losses, yet those with more social and economic resources are able to maintain their well-being partly due to their greater sense of control and more positive outlook toward the future (Jopp & Rott, 2006). Highly educated older adults are more likely to have accumulated resources over the life course that they can draw on to overcome life’s challenges. Over time, success in overcoming challenges can enhance sense of control and bolster positivity and productivity (Rodin, 1986). More educated older adults may, therefore, be less vulnerable to declines in control and increases in hopelessness that can accompany aging because they are less likely to face the adverse experiences that erode control and hope and are more likely to experience positive outcomes that develop and sustain them.

Several studies have documented a positive association between education and control (House, 2002; Lachman & Weaver, 1998b; Mirowsky & Ross, 1998; Schieman & Plickert, 2008). We know very little, however, about education differences in change in control among older adults. One study examined longitudinal associations between age and sense of control and showed that more years of education was associated with a greater increase in control over a 1- to 2-year period (Wolinsky et al., 2003). This study, however, was based on a clinical sample of adults aged 18 and older living in a Midwestern state, and it remains unclear whether there is a relationship between education and change in sense of control in the older U.S. adult population.

There is very little research on the relationship between hopelessness and education. A few studies have found greater levels of hopelessness among those with less education (Anda et al., 1993; Harper et al., 2002; Lynch, Kaplan, & Salonen, 1997), but other studies have found no such association (Dong & Chang, 2014; Haatainen, Tanskanen, Kylmä, Honkalampi, et al., 2003). The majority of these studies are based on cross-sectional data and non-U.S. samples, so we have limited insight into whether hopelessness varies by education among older U.S. adults. Additionally, there are no studies, to our knowledge, that have examined whether change in hopelessness over time differs by education among older adults.

In the current study, we examine education differences in 4-year change in control and hopelessness in a large, nationally representative sample of U.S. adults aged 52 and older. Prior research has shown that these measures of psychosocial functioning differ by education, but it is unclear whether changes in control and hopelessness also differ by education. Our study addresses this gap in the literature and contributes to our understanding of the relationships between education, aging, and psychosocial functioning in older adulthood. Examining psychosocial functioning among older adults is important because it is a critical element in positive adaptation to the often negative experiences associated with aging. Examining change in psychosocial functioning is also important because the advantages associated with educational attainment can accumulate over time, exerting a dynamic rather than static influence on psychosocial functioning, and potentially leading to widening education differences in functioning. Prior theory and evidence suggest that individuals with more education are better able to maintain their health and physical functioning into old age (Dannefer, 2003; Ross & Wu, 1996). This educational advantage may extend to maintaining psychosocial functioning over time. We therefore expect older adults with lower educational attainment to experience greater declines in their sense of control and increases in hopelessness compared with the most educated.

Education may protect older adults from declines in psychosocial functioning because it helps them avoid adverse events and changes associated with a low sense of control (Pudrovska et al., 2005) and hopelessness (Haatainen, Tanskanen, Kylmä, Honkalampi, et al., 2003). Changes in psychosocial functioning may be influenced by concurrent changes in life circumstances. In order to isolate the role of education in change in sense of control and hopelessness from that of recent changes in social circumstances and health status, we account for the loss of a spouse or partner, the perceived loss of support from family and/or friends, and the presence or onset of disease or disability during the same 4-year period. We selected these particular experiences because they represent aversive, and largely irreversible, challenges encountered in older age, often requiring substantial social and psychological adjustment. They also are likely to have effects on psychosocial functioning over a relatively short time period, such as the 4 years covered in this study. We expect education to have an enduring effect on psychosocial functioning in older adulthood that cannot be explained by differences in recent negative changes in social circumstances and health status.

Method

Data and Sample

The data come from the Health and Retirement Study (HRS), an ongoing nationally representative study of older Americans. In 2006, the HRS began collecting data on psychosocial characteristics using a self-administered questionnaire (SAQ). A random half-sample of households, excluding nursing homes and other institutions, received the SAQ in 2006 and the second half-sample received it in 2008. Follow-up assessments occurred 4 years later in 2010 and 2012, respectively.

In the combined 2006/2008 sample, 14,866 eligible individuals received the SAQ and 13,079 responded by mail or phone, for a completion rate of 88%. At follow-up, 11,091 respondents received the SAQ and 9,561 completed it for a completion rate of 86%. Sample weights for the SAQ subsample were not available for 228 respondents. To retain these respondents in the analytic sample, we substituted the respondent-level sampling weights from the core interview. We conducted all analyses with and without these individuals and found no differences in the results. We excluded 616 (6.4%) respondents with missing values on variables included in our analysis. The final analytic sample included 8,945 adults aged 52 and older (i.e., born in 1953 or earlier) with complete data on all demographic, socioeconomic, psychosocial, and health measures.

Study Measures

Psychosocial functioning

We assessed three indicators of psychosocial functioning: mastery, perceived constraints, and hopelessness. Mastery and perceived constraints represent two distinct aspects of control: mastery refers to beliefs about personal efficacy, whereas perceived constraints refer to beliefs about the presence of external barriers or obstacles to achieving one’s goals (Lachman & Weaver, 1998a). Because mastery and perceived constraints have been shown to have different implications for health and aging (Infurna & Mayer, 2015), we examine both constructs separately to get a comprehensive assessment of sense of control among older adults. Our scales for mastery and perceived constraints have been used in prior studies and are standard measures of a general sense of control over one’s life (Lachman & Weaver, 1998a; Pearlin & Schooler, 1978; as cited in Smith et al., 2013).

We created measures of low mastery by averaging values from five items (e.g., “I can do just about anything I really set my mind to.”) that were reverse coded from 1 (strongly agree) to 6 (strongly disagree); higher values represent less mastery. For perceived constraints, five items (e.g., “Other people determine most of what I can and cannot do.”) were also averaged with values ranging from 1 (strongly disagree) to 6 (strongly agree); higher values indicate greater perceptions of constraints.

Hopelessness was measured with a four-item scale, consisting of two items each from two previously validated scales (Beck, Weissman, Lester, & Trexler, 1974; Everson et al., 1997; as cited in Smith et al., 2013). The items assessed the extent to which respondents believed that they cannot reach their goals and that their lives are not changing for the better. Responses for hopelessness ranged from 1 (strongly disagree) to 6 (strongly agree), with higher values representing more hopelessness.

Scale scores were not constructed if more than two items were missing data. The scales have high reliability (e.g., Cronbach’s α > .80) for each year of data collection. We construct change scores for each measure that take the difference between levels at baseline and follow-up. Positive change scores represent an overall average increase in low mastery, perceived constraints, and hopelessness. Details on scale reliability, item-wording, response options, and nonresponse are provided in Supplementary Table 1. Correlations among our measures at baseline ranged from .33 to .60 (Supplementary Table 2).

Education

We classified education into four categories: less than high school, high school, some college, and college degree or higher (reference). Education is a preferred indicator of socioeconomic status among older adults because people tend to report their educational history more accurately, which means there is less reporting bias than other socioeconomic indictors such as income (Mirowsky & Ross, 2003). Additionally, education is typically established earlier in life and rarely changes after the age of 50 years. This stability is particularly useful when studying the psychosocial functioning of older adults because changes in functioning occur after education has been established, which offers reasonable assurance that the changes are a result, rather than a driver, of education differences.

Recent negative events and changes

To account for adverse changes in circumstances that may influence psychosocial functioning, we constructed measures reflecting recent negative changes in social and health-related experiences. Change in marital status represents the loss or absence of support from an intimate social relationship. We classified respondents into one of three categories—those who were married or partnered at baseline and follow-up (reference); those who were divorced, separated, or widowed at baseline and remained unmarried/unpartnered at follow-up; and those who were married or partnered at baseline but experienced a divorce, separation, or death of a spouse/partner by follow-up.

We also measured the loss or absence of perceived social support from family and friends using three items assessing how much respondents feel family/friends understand their feelings, can be relied upon for serious problems, and can be confided in about worries. Respondents answered this set of questions in reference to their spouse/partner, children, family, and friends. Response options were reverse coded toward higher levels of support: 1 = “not at all”, 2 = “a little”, 3 = “some,” and 4 = “a lot”. We created separate scales for each relationship type and then averaged across the relationship-specific scales to derive a summary measure. We then categorized high support as a value of greater than 3 (i.e., above “some” support), which is close to the scale median score of 3.2. We compare individuals who either gained support or had consistently high levels of support at baseline and follow-up (reference) with individuals who perceived a loss of support and individuals who consistently reported low levels of support (i.e., scores ≤ 3 at baseline and follow-up).

The health measures include changes in the number of limitations in activities of daily living (ADLs) and the number of health conditions. ADL limitations include difficulties walking across a room, dressing, bathing, eating, getting into bed, and toileting. We categorize respondents into three groups: those with no ADL limitations at baseline or follow-up (reference); those who had no ADL limitations at baseline but experienced the onset of at least one limitation by follow-up; and those with at least one limitation at baseline and follow-up. We counted the number of self-reported physician-diagnosed health conditions, including stroke, diabetes, lung disease, heart problems, or cancer (other than skin), and assessed change in the number of conditions by comparing individuals who had no conditions at baseline or follow-up (reference) with those who had no conditions at baseline but experienced the onset of at least one condition by follow-up and those who had at least one condition at both baseline and follow-up.

Other covariates

Because the presence of depressive symptoms can negatively bias appraisals of one’s current and future situation (Beck, 2008), we included an indicator for depressed affect. The HRS includes eight items from the validated Center for Epidemiologic Studies Depression Scale (CES-D; Radloff, 1977): felt depressed, everything is an effort, trouble sleeping, felt lonely, felt sad, can’t get going, enjoyed life, and was happy. We reverse coded the last two items and created a count variable. We classified individuals reporting three or more symptoms at baseline as depressed and individuals reporting less than three symptoms as not depressed; on this shortened CES-D scale, using a cutpoint of three or more symptoms as an indicator of depressed affect has been shown to produce prevalence rates similar to those estimated using the full CES-D scale (Wallace et al, 2000).

Our sociodemographic controls include age at baseline (divided by 10), gender, and race/ethnicity. We include four racial/ethnic groups: non-Hispanic white (reference), non-Hispanic black, Hispanic, and non-Hispanic other. Foreign-born respondents included individuals born outside of the United States; we compared them with U.S. natives.

Statistical Analyses

We used an adjusted Wald test to compare differences in the distribution of sample characteristics over time and differences by education in low mastery, perceived constraints, and hopelessness at baseline. We used ordinary least squares (OLS) regression to assess education differences in levels and change in each outcome. All three outcomes are normally distributed, and inspection of normal quantile plots of regression residuals suggested that OLS regression was an appropriate modeling strategy. In order to improve interpretation of findings across the outcomes, we standardized the scales for mastery, constraints, and hopelessness to a mean of zero and standard deviation of one. We first examined education differences in psychosocial functioning at baseline and then examined differences in change in psychosocial functioning over 4 years. We estimated two models for the multivariate analyses: Model 1 included education and the covariates; Model 2 added indicators for baseline social relationship characteristics and health status for the baseline analysis and indicators of stability and change in these characteristics from baseline to follow-up for the change analysis. For all analyses, we applied sample weights provided by the HRS to make our findings generalizable to the older U.S. population and used the survey (svy) suite of commands in Stata 13.1 to account for the complex sample design.

Results

Table 1 presents weighted characteristics of the sample at baseline and follow-up. The majority of respondents were women, white, and born in the U.S, and most respondents had at least a high school degree or equivalent. The average age was approximately 65 years. More than two thirds were married or partnered and slightly more than 40% reported low levels of social support. One in eight had at least one ADL limitation at baseline, nearly 45% had at least one health condition, and about 20% were considered depressed.

Table 1.

Weighted Sample Characteristics: Health and Retirement Study at Baseline (2006/2008) and Follow-Up (2010/2012) (n = 8,945)

	Baseline (2006/2008)	Follow-up (2010/2012)
	% or Mean (SE)	% or Mean (SE)
Female	55.5
Non-Hispanic white	83.7
Non-Hispanic black	7.8
Hispanic	6.3
Non-Hispanic other	2.2
Foreign-born	7.7
Less than high school	13.1
High school degree	36.0
Some college	24.5
College degree	26.5
Age (years)	64.7 (0.17)	68.8 (0.17)
Married/partnered	70.2	66.2
Separated/divorced	12.9	12.7
Widowed	13.4	17.5
Never married	3.6	3.7
Low support	42.2	40.9
ADL limitations (one or more)	12.2	15.8
Chronic conditions (one or more)	44.9	56.0
Depresseda	18.8
Low mastery	2.17 (0.02)	2.28 (0.02)
Perceived constraints	2.10 (0.02)	2.16 (0.02)
Hopelessness	2.26 (0.02)	2.33 (0.02)

Notes. ADL = activity of daily living; SE = standard error.

^aWe only assess depression at baseline; the differences in the distribution of each characteristic at baseline and follow-up are statistically significant (p < .001).

The mean unstandardized scores for low mastery, perceived constraints, and hopelessness are presented at the bottom of Table 1. For low mastery, the mean value indicates that, on average, respondents “somewhat” and “slightly” agreed with affirmative statements about mastery. Mean values for perceived constraints and hopelessness indicate that on average respondents “somewhat” and “slightly” disagreed with affirmative statements about constraints and hopelessness. Values of each scale were higher at follow-up.

Figure 1 shows baseline mean values of low mastery, perceived constraints, and hopelessness by education. There was a graded decrease in each psychosocial measure as education increased. The highest scores of poor psychosocial functioning were among respondents with less than a high school education and the lowest were among those with a college degree or higher. The difference between these two groups was most pronounced for hopelessness: The average level of hopelessness among respondents with less than a high school degree was nearly 64% higher than the hopelessness among respondents with a college degree or higher.

Figure 1.

Weighted mean levels of psychosocial functioning by education at baseline (2006/2008).

Table 2 presents education differences in low mastery, perceived constraints, and hopelessness at baseline. For each measure of psychosocial functioning, lower levels of education were associated with worse psychosocial functioning at baseline net of sociodemographic characteristics (Model 1). Older adults with less than a high school degree had the lowest levels of functioning compared with those with a college degree or higher. The inclusion of baseline social characteristics and health indicators (Model 2) had a minimal effect on education differences in psychosocial functioning, with the exception that the difference in low mastery between those with some college and those with a college degree was no longer statistically significant.

Table 2.

OLS Regression Models of Low Mastery, Constraints, and Hopelessness: Health and Retirement Study at Baseline (2006/2008; n = 8,945)

	Low mastery				Perceived constraints				Hopelessness
	Model 1		Model 2		Model 1		Model 2		Model 1		Model 2
	b	SE	b	SE	b	SE	b	SE	b	SE	b	SE
Education
Less than high schoola	0.15	0.05**	0.11	0.05*	0.50	0.04***	0.45	0.04***	0.73	0.04***	0.68	0.04***
High school degreea	0.12	0.04**	0.11	0.04**	0.25	0.03***	0.23	0.03***	0.41	0.03***	0.38	0.03***
Some collegea	0.08	0.03*	0.05	0.03	0.14	0.04***	0.11	0.04**	0.21	0.03***	0.17	0.03***
Baseline social circumstances and health
Separated/divorcedb			−0.04	0.05			−0.02	0.04			0.12	0.04**
Widowedb			−0.04	0.04			0.01	0.04			0.10	0.03**
Never marriedb			0.11	0.07			0.23	0.09*			0.26	0.08**
Low social supportc			0.27	0.03***			0.32	0.02***			0.34	0.02***
Any ADL limitationd			0.29	0.04			0.31	0.04***			0.18	0.03***
Any health conditionse			0.05	0.03			0.08	0.03**			0.10	0.03***
Covariates
Age/10	0.02	0.01	0.02	0.01	0.03	0.01	0.02	0.01	0.02	0.01	0.01	0.01
Femalef	0.01	0.02	0.05	0.02	−0.01	0.02	0.03	0.03	−0.11	0.02***	−0.08	0.02***
Non-Hispanic blackg	−0.03	0.05	−0.05	0.05	−0.00	0.06	−0.03	0.06	−0.04	0.05	−0.07	0.05
Hispanicg	−0.13	0.06*	−0.12	0.06*	−0.01	0.08	−0.00	0.06	0.06	0.06	0.07	0.05
Non-Hispanic otherg	−0.01	0.08	−0.05	0.08	0.22	0.09*	0.17	0.09	0.22	0.08**	0.18	0.08*
Foreign-bornh	0.12	0.07	0.13	0.06*	0.04	0.05	0.05	0.05	−0.01	0.06	0.03	0.06
Depressedi	0.55	0.04***	0.43	0.04***	0.81	0.03***	0.67	0.03***	0.75	0.04***	0.62	0.04***
Constant	−0.35	0.08***	−0.22	0.08*	−0.53	0.09***	−0.35	0.09***	−0.56	0.09***	−0.35	0.09***
F statistic	F(10, 47) = 26.94***		F(16, 41) = 36.25***		F (10, 47) = 129.04***		F(16, 41) = 102.63***		F(10, 47) = 112.58***		F(16, 41) = 111.10***
Pseudo R2	.05		.08		.15		.18		.17		.21

Notes. ADL = activity of daily living; b = unstandardized regression coefficient; OLS = ordinary least squares; SE = standard error.

^aReference group is college education, ^bReference group is married/partnered, ^cReference group is high social support, ^dReference is no ADL limitations, ^eReference is no health conditions, ^fReference is male, ^gReference group is non-Hispanic white, ^hReference group is U.S. born, ⁱReference is not depressed at baseline.

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

Low mastery was positively associated with low social support and the presence of at least one ADL at baseline. Perceived constraints were also higher among those with low support and ADL limitations, and respondents who were never married or had at least one health condition perceived greater constraints in their environment compared with their married/partnered and disease-free counterparts. For hopelessness, being married or partnered was protective, and low support, the presence of any ADL limitation, and the presence of a health condition were associated with more hopelessness.

Table 3 presents results from multivariate regression models examining 4-year change in low mastery, perceived constraints, and hopelessness. Two models are again presented for each outcome. Model 1 includes controls for baseline levels of the outcome, age, gender, race/ethnicity, foreign-born status, and depressive affect. In Model 2, we added variables reflecting change in marital status, perceived social support, ADL limitations, and chronic conditions and disease.

Table 3.

Models Predicting 4-Year Change in Psychosocial Functioning: Health and Retirement Study (n = 8,945)

	Low mastery				Perceived constraints				Hopelessness
	Model 1		Model 2		Model 1		Model 2		Model 1		Model 2
	b	SE	b	SE	b	SE	b	SE	b	SE	b	SE
Education
Less than high schoola	0.17	0.04***	0.12	0.04**	0.26	0.04***	0.22	0.04***	0.37	0.04***	0.35	0.04***
High school degreea	0.13	0.03***	0.11	0.03***	0.16	0.02***	0.14	0.03***	0.22	0.02***	0.21	0.02***
Some collegea	0.10	0.03**	0.07	0.03*	0.13	0.03***	0.11	0.02***	0.12	0.03***	0.10	0.03***
Negative changes and events
Lost spouse/partner (by follow-up)b			−0.03	0.05			−0.01	0.04			0.02	0.04
Not married/partnered (baseline & follow-up)b			0.00	0.03			−0.02	0.02			0.02	0.02
Loss of support (by follow-up)c			0.19	0.04***			0.19	0.03***			0.22	0.03***
Chronic low support (baseline & follow-up)c			0.19	0.02***			0.23	0.03***			0.22	0.02***
Onset of ADL Limitations (by follow-up)d			0.31	0.05***			0.36	0.05***			0.24	0.05***
Chronic ADL limitations (baseline & follow-up)d			0.41	0.06***			0.33	0.05**			0.21	0.04***
Onset of health condition(s) (by follow-up)e			0.05	0.04			0.12	0.04**			0.08	0.03*
Chronic health condition (baseline & follow-up)e			0.04	0.02			0.04	0.03			0.06	0.02**
Covariates
Baseline value	−0.59	0.01***	−0.61	0.01***	−0.54	0.01***	−0.57	0.01***	−0.46	0.01***	−0.49	0.01***
Age/10	0.08	0.01***	0.06	0.01***	0.07	0.01***	0.05	0.01***	0.05	0.01***	0.04	0.01**
Femalef	0.00	0.03	0.03	0.02	−0.02	0.02	0.02	0.02	−0.08	0.02***	−0.06	0.02**
Non-Hispanic blackg	−0.04	0.03	−0.06	0.03	−0.05	0.04	−0.07	0.04	−0.02	0.03	−0.04	0.03
Hispanicg	−0.06	0.05	−0.07	0.05	−0.05	0.05	−0.06	0.05	0.10	0.04*	0.10	0.04*
Non-Hispanic otherg	−0.13	0.08	−0.16	0.08*	0.02	0.09	−0.01	0.09	0.17	0.06**	0.14	0.06*
Foreign-bornh	0.00	0.04	0.01	0.05	0.10	0.05	0.10	0.05*	−0.00	0.04	0.00	0.04
Depressedi	0.26	0.03***	0.16	0.03***	0.34	0.03***	0.26	0.03***	0.26	0.04***	0.19	0.04***
Constant	−0.69	0.09***	−0.70	0.09***	−0.63	0.08***	−0.69	0.08***	−0.51	0.07***	−0.58	0.07***
F statistic	F(11, 46) = 138.18***		F(19, 38) = 83.19***		F(11, 46) = 153.67***		F(19, 38) = 118.91***		F(11, 46) = 93.01***		F(19, 38) = 53.54***
Pseudo R2	.31		.33		.27		.29		.23		.25

Notes. ADL = activity of daily living; b = unstandardized regression coefficient; SE = standard error.

^aReference group is college education, ^bReference group is stable partnership of became married/partnered, ^cReference group is consistently high social support or gained support, ^dReference is no ADL limitations or limitations improved, ^eReference is no health conditions, ^fReference is male, ^gReference group is non-Hispanic white, ^hReference group is U.S. born, ⁱReference is not depressed at baseline.

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

Lower levels of education were associated with worse psychosocial functioning over time. Compared with respondents with a college degree or higher, respondents with lower levels of education experienced greater increases in low mastery, perceived constraints, and hopelessness as they aged, even after accounting for changes in life experiences. The education differences for hopelessness were considerably larger than those for low mastery and perceived constraints, which suggest that hopelessness may be more closely tied to educational attainment than sense of control. For all three measures, a loss of perceived support, chronically low levels of support, and the onset or chronic presence of any ADL limitation was associated with a significant worsening of psychosocial functioning. The onset of one or more health conditions was associated with increased perceived constraints, and the onset or chronic presence of one or more health conditions was associated with increased hopelessness.

Because there are known gender differences in education, control, hopelessness, and life experiences, we also assessed gender differences in the relationship between education and low mastery, perceived constraints, and hopelessness. We did not find gender difference in the association between education and psychosocial functioning. Of the 57 interactions we test, only three were statistically significant (p < .05) and none involving education.

Discussion

This study used data from a nationally representative sample of older U.S. adults to examine education differences in levels and 4-year change in three indicators of psychosocial functioning—mastery, perceived constraints, and hopelessness. We found that older adults with less than a high school education had the lowest mastery, the highest perceived constraints, and the greatest feelings of hopelessness. Compared with older adults with a college degree or higher, those with lower levels of education also experienced greater declines in control and increases in hopelessness over 4 years. Education differences in change in psychosocial functioning remained after accounting for changes in marital status, social support, and physical health and functioning that occurred in the same time period. The results suggest that more highly educated older adults not only have better psychosocial functioning, on average, but also maintain more of their functioning over time compared with those with less education.

Two explanations have been proposed for why education is related to a greater sense of control: education increases knowledge and problem-solving skills that help people feel efficacious in managing their lives, and education provides more opportunities for social and economic advancement that lead to more frequent experiences of success and fewer experiences of adversity, which, in turn, increase sense of control and protect against declines with age (Mirowsky & Ross, 2007).

There has been relatively less theoretical and empirical research on education differences in hopelessness. Our study provides evidence that hopelessness is greatest among the least educated and increases over time. Prior theory and research suggest that hopelessness results from repeated exposure to negative events and losses (Abramson, Metalsky, & Alloy, 1989; Bolland, Lian, & Formichella, 2005; Haatainen, Tanskanen, Klymä, Antikainen, et al., 2003), and we found that recent negative events and changes, such as the onset of disease and disability, were associated with increases in hopelessness.

As expected, education had an enduring effect on sense of control and hopelessness. Demonstrating the net effects of education on sense of control and hopelessness above and beyond adverse changes in recent social circumstances and health is a distinguishing feature and contribution of this study. Among the recent negative changes we assessed, onset of disability was fundamentally important to declines in psychosocial functioning. The onset and presence of any ADL limitation were associated with the greatest increases in low mastery and perceived constraints, on the magnitude of two to three times those of the other events and changes. Declines in physical functioning and social support were associated with the largest increases in hopelessness over time.

We did not observe an association between spousal loss and any of our measures of psychosocial functioning. Additional analysis did not suggest that the effects of spousal loss were mediated by other negative changes (e.g., the loss of perceived social support). There was an association, however, between loss due to divorce/separation and perceived constraints (p < .05) and hopelessness (p < .05). The bereavement process is complex, and research suggests that although some older adults suffer psychologically from the death of a spouse (Pudrovska & Carr, 2008) others show psychological resilience (Bonanno, Wortman, & Nesse, 2004). This may explain the absence of an association between recent widowhood and psychosocial functioning in our study.

Our findings can be understood within the cumulative advantage/disadvantage framework (Dannefer, 2003), which argues that inequalities in aging result from the accumulation of social and economic advantages among those of higher status and adversity among those of lower status. Education advantages older adults by increasing access to resources and opportunities that help individuals accomplish their goals and cope with adverse experiences and by minimizing exposure to the events and experiences that erode control. Both processes likely operate simultaneously over the life course; however, the former explanation—that is, education’s association with higher status and greater availability of resources—seems particularly relevant for understanding education differences in change in psychosocial functioning in old age. It highlights how greater educational attainment leads to the accumulation of resources that help individuals successfully overcome aging-related challenges. Experiencing this success reinforces beliefs in personal control and a positive outlook toward the future. Lower educational attainment, on the other hand, is associated with fewer resources to cope with aging-related challenges but greater exposure to stressors that erode the already limited set of available resources. Because of this limitation in coping resources, less educated older adults experience fewer positive outcomes and therefore are at risk of declines in control and hope. Thus, the very resources needed to overcome the stressors they disproportionately face are more rapidly depleted (Aneshensel & Mitchell, 2014). The depletion of psychosocial functioning over time can have profound implications for the well-being of less educated older adults as they age: specifically that the least educated among them experience the compounding disadvantages of low education, a worsening sense of control, and greater hopelessness.

There are some limitations of the current study. First, the individuals in our sample are aged 52 and older when baseline differences in psychosocial functioning are assessed. The relationship between education and our measures of psychosocial functioning may vary over the life course because life-course stages are characterized by different living conditions, experiences, and resources. Moreover, levels of control and hopelessness in the transition to adulthood may influence educational attainment and thus confound the association between education and psychosocial functioning in later life. However, in our analysis of change, the temporal ordering is evident because we assessed psychosocial functioning after formal education has long been completed and we examined the role of experiences occurring after assessment of baseline levels of functioning. An additional limitation is that the set of recent events and changes examined in this study is incomplete; there are likely other social and health factors that influence psychosocial functioning. However, the recent experiences we did account for represent some of the most adverse challenges older adults encounter. Finally, our sample is limited to community-dwelling older adults who likely have a stronger sense of control and lower levels of hopelessness than those living in nursing homes or other institutionalized settings. Because institutionalized adults were not asked to report on their psychosocial functioning, we likely underestimate baseline levels of control and hopelessness and education differences in these factors in the broader older adult population.

The overall decline in psychosocial functioning observed among less educated older adults occurred over a relatively short period of time: approximately 4 years. This short-term change in functioning can compound over a longer period of time and eventually become large and pervasive disparities. Thus, our study demonstrates that education’s role as a fundamental cause of health inequalities (Phelan, Link, & Tehranifar, 2010) extends to inequalities in psychosocial functioning in older adulthood and that the depletion of psychosocial resources over time can have profound implications for the well-being of less educated older adults, namely, the compounding disadvantages of low education and declining psychosocial functioning over time.

Supplementary Material

Supplementary material is available at The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences online.

Funding

This work was supported by the National Institute on Aging at the National Institutes of Health (grant numbers T32-AG000037, R00-AG039528, P30-AG17265, P30-AG043073, and R24-AG045061).

Acknowledgments

U. A. Mitchell contributed to the conceptualization of the study, completed the data analysis, and wrote the paper. J. A. Ailshire, E. M. Crimmins, and L. L. Brown contributed to the conceptualization of the study and revision of the paper. M. E. Levine contributed to the data analysis.