Effects of Timing and Level of Degree Attained on Depressive Symptoms and Self-Rated Health at Midlife

Abstract

Objectives. We examined whether attaining a higher educational degree after 25 years of age was associated with fewer depressive symptoms and better self-rated health at midlife than was not attaining a higher educational degree.

Methods. We analyzed data from National Longitudinal Survey of Youth, restricting our sample to respondents who had not attained a bachelor's degree by 25 years of age (n = 7179). We stratified all regression models by highest degree attained by 25 years of age.

Results. Among respondents with no degree, a high school diploma, or a post–high school certificate at 25 years of age, attaining at least a bachelor's degree by midlife was associated with fewer depressive symptoms and better self-rated health at midlife compared with respondents who did not attain a higher degree by midlife. Those with an associate's degree at 25 years of age who later attained a bachelor's degree or higher reported better health at midlife.

Conclusions. Attaining at least a bachelor's degree after 25 years of age is associated with better midlife health. Other specifications of educational timing and its health effects across the life course should be studied.

With increasing education, individuals experience more favorable health^1–5 and greater longevity.^6–9 Yet, much of the research implicitly, and sometimes explicitly, assumes that education is completed at or around 25 years of age and is stable thereafter.^8–12 As such, education–health research often ignores issues of educational timing, even though US educational pathways became much more differentiated after the 1970s as women entered the labor force in greater numbers, the economy shifted from industrial to service-based, higher education became more affordable, and a college degree became necessary to attain a middle-class lifestyle.^13–15 For example, in 1970, 28% of US undergraduate students were 25 years or older.¹⁶ By 2008, the percentage had increased to 38%.¹⁷ The number of individuals returning to school when older than 25 years is projected to increase from 6.9 million in 2006 to 8.1 million by 2017.¹⁸

Given that a significant proportion of US adults earn a college degree after their mid-20s, it is surprising that few studies have examined the health effects of returning to school later in life. Attaining a higher degree after the mid-20s may result in positive mental and physical health benefits for US adults. Regardless of the age at which one returns to school, educational activities can enhance social networks, provide individuals with fulfilling life experiences, and foster adult development and intellectual enrichment.¹⁹ Such experiences may be particularly important for mental health. For example, 1 study found that individuals who attained their general educational development (GED) certificate in adulthood had lower levels of depressive symptoms than did permanent high school dropouts.²⁰ Additionally, individuals who attain a higher degree after their mid-20s can experience increases in their earnings and more stable employment,²¹ which in turn may promote better mental and physical health. For example, Clark and Jaeger²¹ found that high school dropouts who attained their GED certificate in adulthood experienced wages closer to individuals who attained a high school diploma at the end of 12th grade than to permanent dropouts.

Alternatively, returning to school later in life may have little or no effect on health. Because individuals who attain their college degree later in adulthood have fewer years in the labor market to recoup their investment in education, they may be less likely to accumulate the economic and social (e.g., personal control, cognitive ability, and moral reasoning)^4,22 benefits associated with a college degree that may lead to better psychological and physical functioning. For example, Taniguchi reported that individuals who completed a bachelor's degree when 25 years or older received significantly lower wages than did those who received a bachelor's degree before 25 years of age.²³ Importantly, however, Taniguchi did not consider whether those who attained a college degree when 25 years or older received any economic benefit to their delayed degree completion. That is, attaining a bachelor's degree later in life may still yield economic benefits in comparison with never receiving a bachelor's degree. Given that 2 of the most important mechanisms linking education to health are employment and wages,^24–26 it is important to consider whether the timing of education contributes to health at midlife among US adults, as it seems to have some effect on wages. This question is especially timely as US policymakers tout retraining unemployed, often older or less educated workers as a viable strategy to help affected workers find new work.

Our study advances current research on the education–health relationship by exploring the extent to which attaining a higher degree after the mid-20s is associated with depressive symptoms and self-rated health at midlife. We chose depressive symptoms and self-rated health for 2 reasons. First, the few studies to have examined the effect of the timing of education on health have focused on depressive symptoms or psychological well-being. Second, self-rated health is commonly used in education–health research^1,4,24,27 and is a valid and reliable measure of subjective well-being that correlates strongly with physician assessments of morbidity and subsequent mortality.^28–30

We hypothesized that among US adults who had attained less than a bachelor's degree by their mid-20s, attaining a higher degree by midlife would be associated with fewer depressive symptoms and better self-rated health compared with those who maintain their level of education after their mid-20s and that attaining at least a bachelor's degree after the mid-20s would provide greater health benefits than would attaining a lesser degree after the mid-20s. It is likely, however, that attaining at least a bachelor's degree after the mid-20s will provide greater health benefits than will attaining a lesser degree after the mid-20s, given the strong association of a college or advanced degree with employment options, occupational prestige, and wages.³¹

METHODS

We analyzed data from the National Longitudinal Survey of Youth 1979 (NLSY79), a nationally representative sample of individuals who were aged 14 to 21 years in 1979.³² Respondents were interviewed annually from 1979 to 1994 and interviewed biennially after 1994, with data collection ongoing. We restricted our analyses to respondents who (1) were interviewed during a window between 24 and 26 years of age, (2) were interviewed at least once when 40 years old or older, and (3) had not completed a bachelor's degree when interviewed at 24 to 26 years of age (n = 7251). We further excluded 72 respondents because of missing data. After exclusions, our final analytic sample consisted of 7179 respondents.

An important feature of our analytic plan was to measure educational attainment at 25 years of age because that is the age at which many studies assume that education is complete. However, some people were not interviewed at 25 years of age because of differences in the timing of NLSY79 interviews or because respondents skipped a year and were interviewed the next year. Of those interviewed between 24 and 26 years of age, 89% were interviewed at 25 years of age, 10% were interviewed at 26 years of age, and 1% were interviewed at 24 years of age.

Measures

We measured depressive symptoms using a 7-item Center for Epidemiological Studies Depression Scale (CES-D). Using a 4-point Likert scale, respondents were asked how often in the past week they (1) did not feel like eating or had poor appetite, (2) had trouble keeping their mind on what they were doing, (3) felt depressed, (4) felt that everything was an effort, (5) had restless sleep, (6) felt sad, and (7) could not get going. Per convention, the 7 items were summed (Cronbach α = 0.83), with higher values representing greater depressive symptomatology. The distribution was skewed, so we used a square root transformation to normalize the distribution. This type of transformation has been used in previous research using the CES-D scale to normalize the distribution.^5,33

The standard 20-item CES-D discriminates between clinically depressed and non–clinically depressed individuals.³⁴ Prior studies have demonstrated that the 7-item CES-D is highly correlated with the 20-item CES-D (r = 0.92).³⁵ Moreover, because the NLSY79 administered the 20-item CES-D in 1992, for this study, we were able to use the 1992 administration to correlate the 20-item CES-D and the 7-item CES-D. The correlation was 0.90. Thus, the 7-item CES-D appears to achieve high criterion validity with the 20-item CES-D in previous studies³⁵ and within the NLSY79 sample.

We measured self-rated health using an ordinal scale in which 1 = excellent, 2 = very good, 3 = good, 4 = fair, and 5 = poor. Thus, higher values on both dependent variables reflect worse health. Although the NLSY79 is a longitudinal study, self-rated health was only assessed once after 40 years of age, whereas the CES-D was assessed on 3 occasions after 26 years of age, including once after 40 years of age. We therefore chose to examine depressive symptoms and self-rated health at midlife to maintain consistency across analytic approaches.

We measured educational attainment as the highest degree the respondent had attained by 24 to 26 years of age, categorized as no degree, GED, high school diploma, post–high school academic degree or certificate (e.g., apprenticeship), and associate's degree. We included a set of dummy variables to measure the highest degree respondents had attained after 24 to 26 years of age but before approximately 40 years of age (henceforth referred to as midlife). We chose this cutpoint to coincide with our measures of depressive symptoms and self-rated health. Approximately 95% of respondents provided education and health data when they were 40 to 42 years of age. Highest degree attained at midlife was measured simultaneously with or before the health measures. We classified respondents as those who by midlife had attained (1) no degree, (2) a GED, (3) a high school diploma, (4) a post–high school academic degree or certificate, (5) an associate's degree, or (6) a bachelor's or graduate degree. We combined bachelor's and graduate degrees into 1 category owing to issues of data sparseness at the graduate level for respondents who had attained no degree, a GED, or a post–high school degree or certificate by 24 to 26 years of age.

We included covariates that may be associated with attaining a higher degree after 24 to 26 years of age as our dependent variables. Respondents self-reported their race/ethnicity, which we categorized as non-Hispanic White, non-Hispanic Black, Hispanic (any race), or other. We classified respondents’ nativity as foreign-born versus US-born. We categorized the nativity of respondents’ parents as both or known parent(s) US-born, 1 of 2 parents foreign-born, both or known parent(s) foreign-born, or unknown nativity status. If the respondent indicated never knowing 1 of his or her parents, the nativity of the known parent was used to categorize parents’ nativity (i.e., if the known parent was US-born, we categorized this as “both or known parents US-born”). We measured parents’ education using the education of the parent who completed the most years of schooling and categorized it as 0 to 11 years, 12 years, 13 to 15 years, 16 or more years, or unknown. For those who reported the education of only 1 parent, we used that parent's education. We categorized respondent's poverty status at 24 to 26 years of age as in poverty (0%–100% of poverty), not in poverty, or unknown. We included a measure of marital status (married; divorced, separated, or widowed; or never married) when respondents were 24 to 26 years of age. Other covariates included gender, birth cohort (1957–1960 vs 1961–1964), whether respondents had a health condition at 24 to 26 years of age that limited their ability to work or the amount or type of work they could perform, and whether respondents were in the active military forces when they were 24 to 26 years of age.

We considered additional covariates, including family structure at 14 years of age, parental occupation, and attitudes toward women working, but the inclusion of these covariates did not alter our results. Thus, for parsimony and to retain sample size, we did not include these variables in our final models.

Analytic Approach

We began with descriptive statistics to understand the data distribution. We employed multivariate linear regression (for depressive symptoms) and ordered logistic regression (for self-rated health) to examine the association of attaining a higher degree when older than 24 to 26 years of age with the health outcomes. We stratified multivariate analyses by the highest degree respondents had attained by the time they were 24 to 26 years of age to allow us to examine the effect of attaining a higher degree when they were older than 24 to 26 years of age, conditional on the degree they had attained by the time they were 24 to 26 years of age. We weighted all analyses to adjust for the complex sampling design and respondent attrition using the custom weights developed by NLSY79 and the SVY commands in Stata version 11.0 (StataCorp LP, College Station, TX).³⁶

Given potential gender differences in not only the timing of education but also health, we also ran gender-stratified models. These results were generally comparable to the estimates we have presented for the full sample; however, issues with data sparseness for some of the cell sizes resulted in issues of stability for the gender-stratified models. Thus, we present results from our full sample only.

RESULTS

NLSY79 respondents were primarily White (61.6%), US-born (95.7%), not in poverty when 24 to 26 years of age (83.2%), and born into families in which both or known parent(s) were US-born (91.0%; Table 1). Approximately 48.0% of respondents were born between 1961 and 1964, 51.6% were female, 44.6% had a parent who completed 12 years of schooling, 47.3% were married when 24 to 26 years of age, and 5.5% reported a health limitation when 24 to 26 years of age.

TABLE 1—

Sample Characteristics of Respondents Without a Bachelor's Degree by 24 to 26 Years of Age: National Longitudinal Study of Youth, 1979

		Degree Attained by 24-26 Years of Age
Characteristic	Full Sample (n = 7179), Mean or %	No Degree (n = 1428), Mean or %	GED (n = 666), Mean or %	High School Diploma (n = 4301), Mean or %	Other Degree or Certificatea (n = 334), Mean or %	Associate's Degree (n = 450), Mean or %
Square root depressive symptomsb*	1.90	2.15	2.09	1.83	1.87	1.71
Self-rated healthc*
Excellent	20.3	12.4	14.2	21.5	27.9	29.1
Very good	38.9	32.7	36.0	40.7	33.8	43.7
Good	28.3	31.6	33.2	27.7	26.7	22.4
Fair	9.8	18.0	13.5	8.1	7.6	4.3
Poor	2.6	5.4	3.2	2.0	4.1	0.5
Education at age 24–26 y
No degree	15.9
GED	8.3
High school diploma	63.6
Other degree or certificatea	5.1
Associate's degree	7.2
Education at midlifec*
No degree	10.9	68.8	0.0	0.0	0.0	0.0
GED	10.5	19.7	88.9	0.0	0.0	0.0
High school diploma	54.3	7.2	1.6	83.4	0.0	0.0
Other degree or certificatea	5.0	0.5	2.2	1.2	77.4	0.0
Associate's degree	10.4	1.9	3.6	6.3	10.5	73.0
Bachelor's or graduate degree	8.9	1.9	3.6	9.2	12.0	27.0
Race/ethnicityc*
White	61.6	47.7	54.8	64.3	69.8	70.3
Black	16.7	22.0	19.7	15.9	13.9	10.7
Hispanic	8.3	14.9	10.1	6.8	6.5	6.5
Other	13.4	15.5	15.4	13.0	9.7	12.6
Femalec*	51.6	48.4	47.0	52.2	49.2	60.2
Birth cohortc*
1957–1960	51.6	50.8	41.5	52.6	63.8	47.7
1961–1964	48.4	49.2	58.5	47.4	36.2	52.3
Nativity statusc*
US-born	95.7	93.0	95.3	96.3	95.2	96.8
Foreign-born	4.3	7.0	4.7	3.7	4.8	3.2
Parents’ nativity statusc*
Both parents foreign-born	4.1	7.3	4.5	3.3	3.7	4.2
1 parent foreign-born	4.2	4.0	5.4	4.1	3.9	4.7
Both parents US-born	91.0	87.0	89.6	92.1	91.6	90.7
Unknown	0.7	1.6	0.6	0.5	0.8	0.4
Parent's education, yc*
0–11	26.5	47.8	32.9	22.4	20.8	12.7
12	44.6	33.2	43.2	48.0	45.9	40.4
13–15	12.7	6.5	11.6	13.3	15.1	21.1
≥ 16	13.3	5.2	6.9	14.6	16.8	24.0
Unknown	2.9	7.3	5.4	1.7	1.5	1.9
Household poverty at age 24–26 yc*
In poverty	14.6	32.7	22.7	10.4	11.6	4.6
Not in poverty	83.2	64.9	73.8	87.5	87.0	93.5
Unknown	2.3	2.4	3.5	2.2	1.5	1.9
Marital status at age 24–26 yc*
Married	47.3	43.8	43.7	48.6	49.8	46.3
Divorced, separated, or widowed	11.9	16.7	20.8	10.3	11.1	6.6
Never married	40.8	39.5	35.5	41.2	39.1	47.1
Health limitations at age 24–26 yc*	5.5	8.3	5.3	5.0	7.1	3.3
In active military forces at age 24–26 yc*	2.8	1.0	3.1	3.3	2.3	1.8

Note. GED = general equivalency diploma. Percentages may not add to 100 as a result of rounding error. All variables except depressive symptoms are categorical and can be interpreted as percentages.

^aOther post–high school degree or certificate;

^bF-test;

^cχ².

*P < .05.

The mean of square root depressive symptoms was 1.90, and 12.4% of respondents reported their health as fair or poor at midlife. Approximately 63.6% of respondents had attained a high school diploma by 24 to 26 years of age, and 15.9% had attained no degree. By midlife, 54.3% had attained a high school diploma, 10.9% had attained no degree, and 8.9% had attained at least a bachelor's degree.

Most respondents had not attained a higher degree by midlife regardless of the degree attained by 24 to 26 years of age. For example, 68.8% of respondents with no degree at 24 to 26 years of age did not attain a higher degree by midlife. This percentage was 88.9%, 83.4%, 77.4%, and 73.0%, respectively, of those with a GED, high school diploma, post–high school degree or certificate, or associate's degree at 24 to 26 years of age. Even so, a nonnegligible percentage of respondents received at least a bachelor's degree by midlife, ranging from approximately 2.0% for those with no degree at 24 to 26 years of age to 27.0% for those with an associate's degree at 24 to 26 years of age.

We examined the extent to which attaining a higher degree after 24 to 26 years of age was associated with depressive symptoms at midlife, given the degree attained by 24 to 26 years of age (Table 2). We adjusted our models for a range of sociodemographic and health covariates assessed at 24 to 26 years of age. For respondents who had attained no degree, a high school diploma, or a post–high school degree or certificate by 24 to 26 years of age, attaining at least a bachelor's degree by midlife was associated with fewer depressive symptoms at midlife (−0.61, −0.12, and −0.48, respectively) compared with respondents who maintained their level of education. We did not find a similar association between attaining at least a bachelor's degree and depressive symptoms among respondents who had attained a GED or an associate's degree by 24 to 26 years of age. Moreover, attaining an associate's degree by midlife was not associated with depressive symptoms, regardless of degree attained at 24 to 26 years of age.

TABLE 2—

Weighted Estimates From Linear Regression Models Predicting Square Root Depressive Symptoms: National Longitudinal Study of Youth, 1979

	Highest Degree Attained by 24-26 Years of Age
Highest Degree Attained by Midlife	No Degree (n = 1428), b (SE)	GED (n = 666), b (SE)	High School Diploma (n = 4,301), b (SE)	Other Degreea (n = 334), b (SE)	Associate's Degree (n = 450), b (SE)
Intercept	2.04*** (0.09)	1.72*** (0.10)	1.69*** (0.04)	1.74*** (0.13)	1.63*** (0.10)
No degree	…
GED	−0.03 (0.09)	…
High school diploma	0.05 (0.13)	0.18 (0.30)	…
Other degreea	0.72 (0.63)	0.70* (0.31)	0.08 (0.16)	…
Associate's degree	−0.17 (0.20)	−0.34 (0.23)	0.08 (0.07)	−0.01 (0.20)	…
Bachelor's or graduate degree	−0.61** (0.24)	0.15 (0.23)	−0.12** (0.05)	−0.48*** (0.15)	−0.02 (0.09)

Note. GED = general equivalency diploma. Adjusted for race/ethnicity, gender, birth cohort, respondent's nativity, parents’ nativity, parent's education, poverty status at 24-26 years of age, marital status at 24-26 years of age, health limitations at 24-26 years of age, and in active forces at 24-26 years of age.

^aOther post–high school degree or certificate.

*P < .05; **P < .01; ***P < .001.

Next, we assessed the association between attaining a higher degree after 24 to 26 years of age and self-rated health at midlife (Table 3). Again, we stratified analyses by degree attained by 24 to 26 years of age. Except for respondents with a GED at 24 to 26 years of age, respondents who attained at least a bachelor's degree by midlife reported lower odds of poorer health at midlife compared with respondents who maintained their level of education. For example, compared with respondents who had a high school diploma at 24 to 26 years of age and at midlife, attaining at least a bachelor's degree by midlife was associated with lower odds of poorer health (odds ratio [OR] = 0.66).

TABLE 3—

Weighted Estimates From Ordered Logistic Regression Models Predicting Self-Rated Health: National Longitudinal Study of Youth, 1979

	Highest Degree Attained by 24–26 Years of Age
Highest Degree Attained by Midlife	No Degree (n = 1428), OR (95% CI)	GED (n = 666), OR (95% CI)	High School Diploma (n = 4301), OR (95% CI)	Other Degreea (n = 334), OR (95% CI)	Associate's Degree (n = 450), OR (95% CI)
No degree	1.00
GED	0.88 (0.65, 1.18)	1.00
High school diploma	0.77 (0.50, 1.17)	1.10 (0.18, 6.68)	1.00
Other degreea	1.78 (0.34, 9.47)	1.57 (0.57, 4.27)	0.46 (0.24, 0.89)	1.00
Associate's degree	0.54 (0.16, 1.78)	0.82 (0.35, 1.94)	0.80 (0.61, 1.06)	0.75 (0.31, 1.82)	1.00
Bachelor's or graduate degree	0.30 (0.11, 0.83)	0.94 (0.32, 2.75)	0.66 (0.52, 0.85)	0.28 (0.13, 0.63)	0.62 (0.41, 0.96)

Note. CI = confidence interval; GED = general equivalency diploma; OR = odds ratio. Adjusted for race/ethnicity, gender, birth cohort, respondent's nativity, parents’ nativity, parent's education, poverty status at 24–26 years of age, marital status at 24–26 years of age, health limitations at 24–26 years of age, and in active military forces at 24–26 years of age. Self-rated health coded 1 = excellent to 5 = poor.

^aOther post–high school degree or certificate.

DISCUSSION

Education is often viewed as a measure of socioeconomic status that is “generally stable over the course of an adult's lifetime.”^10(p38) Our results, however, suggest that a nontrivial percentage of US individuals attain their highest degree after their mid-20s. Our findings correspond to secular trends occurring in the 1970s and 1980s, which saw greater percentages of adults returning to school later in life.^16,17 Such correspondence is expected given that the NLSY79 cohort was making the transition to adulthood during this same period. The trend toward less stability in measures of education will likely increase even more among younger US birth cohorts, as an increasing number of people return to school later in life.¹⁸

We had 2 hypotheses: (1) attaining a higher degree after the mid-20s would be associated with fewer depressive symptoms and better self-rated health compared with never attaining a higher degree, and (2) attaining at least a bachelor's degree after the mid-20s would provide greater health benefits than would attaining a lesser degree after the mid-20s. Our results generally supported our hypotheses. Among respondents who attained less than a bachelor's degree after their mid-20s, going back to school and attaining at least a bachelor's degree by midlife was associated with fewer depressive symptoms and better self-rated health compared with respondents who did not attain a higher degree by midlife, with 2 exceptions: first, among respondents who had attained a GED by 24 to 26 years of age, attaining at least a bachelor's degree by midlife was unrelated to depressive symptoms and self-rated health; second, among respondents who had attained an associate's degree by 24 to 26 years of age, attaining at least a bachelor's degree by midlife was unrelated to depressive symptoms.

Numerous factors may explain why attaining at least a bachelor's degree after the mid-20s is associated with fewer depressive symptoms and better self-rated health among midlife US adults as compared with those who maintain the level of education they had attained by their mid-20s. First, college education is associated with increased knowledge, cognitive development, and moral reasoning as well as greater social support and feelings of personal control.^4,22 All these things may result in a more optimistic outlook on life³⁷ and provide individuals with the means to respond effectively to stressful life experiences when they arise.⁴ Second, attaining a bachelor's or graduate degree is associated with higher earnings and more stable employment,³¹ which allows individuals a greater opportunity to engage in healthy behaviors that can prevent the onset of illness. Attaining a bachelor's or graduate degree is also associated with occupations that allow greater autonomy on the job and engagement in less routine work, both of which are associated with greater work fulfillment and better psychological and physical functioning.⁴ Examining the mechanisms linking the timing of education to depressive symptoms and self-rated health is an important next step for future investigation.

Beyond degree attained, completing additional years of schooling without attaining a degree might also be associated with health. Attending school takes money and time; expending resources to attain a degree and being unsuccessful in doing so could result in no or negative changes in health given that individuals may not receive significant economic gains if they do not attain a degree. We chose to focus on degree attainment because of its link to wages and employment; however, additional studies that investigate other specifications of attainment are needed.

A common practice in health research, especially in cross-sectional studies and studies of middle-aged or older adults, is to measure educational attainment at the time of the interview without regard to the timing of educational attainment. Measuring only 1 dimension of education—level—may be the norm because researchers are attempting to reduce respondent burden, or because they assume that education is generally stable over an adult's life course.^1,10–12 Given secular changes in the timing at which education is completed,^{14,16,17,38,39} such assumptions limit our ability to fully explicate the education–health relationship. In fact, our results suggest that studies that ignore issues of timing may underestimate education's effect on health. That is, in our study the effect of attaining at least a bachelor's degree after the mid-20s was not consistently associated with fewer depressive symptoms or better self-rated health. Had we only examined level of educational attainment at midlife, the effects of having at least a bachelor's degree on depressive symptoms or self-rated health would have been averaged across individuals who followed different educational pathways.

We measured educational attainment at 2 points in the life course (i.e., before and after 24 to 26 years of age) to correspond with the assumption in the current education–health literature that education is stable after the mid-20s. However, the timing of education may have a more complex association with health than we were able to demonstrate. Future studies should consider whether other specifications of timing (e.g., age at which education is completed) are also related to health.

Limitations

Our sample represents individuals who were born between 1957 and 1964; inferences can be made only about this population. However, our study is one of few to investigate whether the timing of education is related to depressive symptoms and self-rated health. Because the NLSY79 did not collect information on depressive symptoms or self-rated health before respondents were in their mid-20s, we were unable to fully adjust for prior depressive symptomatology or self-rated health. The only health-related measure collected consistently since the start of the study was a measure of health-induced work limitations, which we used to adjust for health in early adulthood. Although this measure likely underestimates health conditions in a young adult population, previous research has found that measures of health-induced work limitations are highly correlated with disability, functional limitations, health impairments, activities of daily living, instrumental activities of daily living, and self-reported health.^28,40,41 Thus, many of the health conditions associated with depressive symptoms and self-rated health were likely accounted for in our measure of health-induced work limitations. Additionally, depressive symptoms and self-rated health constitute only 2 dimensions of health. Other outcomes (e.g., clinical depression, disability) should be explored in future research. Finally, as the result of issues of data sparseness, we were unable to examine whether the timing of education differentially affects health by race/ethnicity. Future efforts should consider this possibility.

Conclusions

Education is often seen as the gateway to the American Dream.⁴² Yet, many individuals are not able to attain their degree of choice by their mid-20s. Our study provides preliminary evidence that the timing of education is associated with health and that attaining at least a bachelor's degree after the mid-20s may result in health benefits for US adults. Additional studies are needed to identify the extent to which the timing of educational attainment matters for population health and the mechanisms by which education in later life benefits health.