Longitudinal change in daily stress across 20 years of adulthood: Results from the National Study of Daily Experiences

Abstract

This study examined age-related patterns in exposure and affective reactivity to daily stressors across a 20-year time span among adults who were between 22 and 77 year-old at their baseline interview. Longitudinal data from the National Study of Daily Experiences (NSDE) consisted of three bursts of eight consecutive nightly interviews of stress and affect. Analyses made use of all available data from a U.S National sample of respondents who participated in any of the three NSDE bursts (N = 2,845; # daily assessments = 33,688). Findings revealed increasing age-related benefits. Younger adults (< 30 years) reported the highest levels of stressor exposure and reactivity, but their stress profile improved with age. Over time, adults averaged an 11% reduction in the occurrence of stressor days, and the younger adults exhibited an even steeper decline (a 47% reduction) in their levels of stressor reactivity. For people in midlife and old age, stressor occurrence continued to decrease over time, yet among adults aged 54 years or older at baseline, stress reactivity remained stable across time.

Every day, the news is filled with emotional accounts of life in the United States, from protests in the streets over political policy and social issues to economic uncertainty and concerns over the health and well-being of our families and the larger world. According to the Stress in America poll, which began documenting stress in the U.S. in 2007, levels of stress in America have reached an all-time high, with more than 80% of surveyed adults reporting emotional effects from prolonged stress (American Psychological Association, 2021). This uptick in stress is concerning; when considering psychological factors related to mental and physical health, few are as pervasive as stress. Decades of research have linked the stress response to multiple indices of physical and emotional health, ranging from depression, anxiety, and sleep disorders to cardiovascular disease, diminished recovery from cancer, higher rates of chronic disease, and earlier mortality (Charles et al., 2013; Chiang et al., 2018; Cohen, Gianaros, & Manuck, 2016; Piazza et al., 2010). Stress exacerbates almost every chronic health condition, is fundamental to emotional well-being and overall quality of life, and is experienced at high rates by a majority of adults. Moreover, stress influences our biological aging, acting as a speedometer of our life span (for review, see Almeida et al., 2011; Harvanek et al., 2021).

Stress affects health, and these effects unfold every day of our lives. It has been proposed that chronic stress can best be understood and measured in patterns of acute stressors and responses (Smyth, Zawadzki, & Gerin, 2013). For this reason, research has increasingly focused on the dynamics of stress in daily life, studying the naturally occurring events that trigger stress (i.e., stressors) and the emotional and behavioral sequalae of these events (i.e., stress responses). For the past two decades, we have studied how daily stressors – those minor but frequent occurrences arising out of normal day-to-day living, such as arguments with a spouse, pressing work deadlines, or one’s child spiking a fever – disrupt daily life. Daily stresors have potent immediate effects on emotional (Stawski et al., 2008), physical (Hoffman & Stawski, 2009), neuroendocrine (Stawski et al., 2013), and cognitive (Sliwinski et al., 2006; Stawski et al., 2010) health, and accumulate to have long-term effects on mental (Charles et al., 2013) and physical (Piazza et al., 2013) health, as well as longevity (Mroczek et al., 2015; Chaing et al., 2018). This work distinguishes the occurrence of stressors (i.e., stressor exposure) from the emotions that they engender (i.e,. stressor reactivity; Almeida, 2005; Bolger, et al., 1989). Exposure is often assessed as the number of stressors encountered or the proportion of days individuals report having a stressful event, whereas reactivity is often characterized as the difference in negative or positive affect on stressor days versus non-stressor days. Both stressor exposure and reactivity are important features of health and well-being across the life course (Almeida et al., 2020), and a main feature of previous work has been to examine the extent to which daily stress is age-graded. Given that stress is highly correlated with health outcomes, understanding how the stress process relates to age and changes over time is necessary to identify the age groups most at risk for stress-related emotional and physical disorders.

Across successively older age groups, stressor frequency decreases with age (Aldwin et al., 2014; Almeida & Horn, 2004; Stawski et al., 2008), a pattern that can be partially explained by differing goals and life tasks common at each life stage. In young adulthood, common developmental tasks include gaining educational and occupational skills, achieving financial independence, and cultivating adult social networks. These tasks entail challenges and struggles that are often accompanied with daily stressors. Over time, career trajectories may become more stable. Family and work responsibilities may give rise to daily stressors, although generally not at the level observed by younger adults. In older adulthood, decreased social roles and expectations associated with retirement and empty nests, coupled with greater leisure time, may explain lower levels of daily stressors (Bossé et al., 1991).

Social roles and life tasks, however, may not be sufficient to explain age-related decreases in daily stressors. Even among older adults ranging from their mid-60s to their 90s, age is still related to lower levels of daily stressors (Charles et al., 2010). Another contributing factor may be related to a development shift where older age is associated with an increased preference for positive over negative stimuli, referred to as the positivity effect (Carstensen & Mikels, 2005). The positivity effect describes why older adults often perceive and remember situations less negatively than their younger counterparts, which may perhaps decrease the likelihood of them remembering a daily event as something that warrants the label of a stressor (e.g., Aldwin et al., 1996).

The finding that stressors decrease in frequency with age in cross-sectional studies is robust and replicated in studies examining groups of adults representing most of the entire adult life span, or just among samples of middle-aged and/or older adults (Aldwin et al., 2014; Charles et al., 2010). With respect to stressor reactivity, however, the pattern of cross-sectional age differences has been mixed (Schilling & Diehl, 2015), at times showing the expected age-related decreases (Birditt, 2014; Charles et al., 2009; Scott et al., 2013, 2017; Uchino et al., 2006), but also no significant age-related differences (Diehl & Hay, 2010; Schilling & Diehl, 2014; Stawski et al., 2008) and even the opposite pattern of age-related increases in stressor reactivity (Sliwinski et al., 2009; Wrzus et al., 2013, 2015). For example, Sliwinski and colleagues (2009) found that affective reactivity to daily stressors in samples of midlife (25-74) and older (65-95) adults increased longitudinally over 2.5- and 10-year periods. Meanwhile, Shilling and Diehl (2014) found that increased age was associated with reduced impact of accumulated stressor exposure (i.e., pile-up). A recent coordinated analysis of 7 daily studies reflects this varied pattern, showing some, but not universal, evidence for age related decrements in stressor reactivity (Stawski et al., 2019).

Given the importance of stress on our health and well-being, the current study followed the 20-year trajectory of two aspects of stress – the occurrence/exposure of stressors and emotional reactivity to these stressors – among a large national sample of adults across three time points. Many of the previous studies on age patterns of daily stress have been limited by age ranges and limited longitudinal time spans (Almeida & Horn, 2004; Diehl & Hay, 2010; Stawski et al., 2008, 2019), which may be part of the reason for the discrepant findings. The current study is the first look into age patterns in exposure and affective reactivity to daily stressors across a long (20-year) time span on a wide age range of adults who were between 22 and 77 at their baseline interview. Given the descriptive nature of the analyses, we did not preregester the hpotheses.We use cross-sectional and longitudinal data from the National Study of Daily Experiences (NSDE) to examine changes in exposure and reactivity to daily stress across 20 years, and how that may differ across the life span.

Method

Participants and Procedure

Participants were from the National Study on Daily Experiences (NSDE), a random subset of participants invited from the larger Midlife in the United States (MIDUS) project (https://www.icpsr.umich.edu/web/ICPSR/series/203). The study was approved by the IRB of the institution responsible for data collection and all respondents consented to their participation. Individuals who consented to participate responded to end-of-day telephone interviews for 8 consecutive days that assessed daily levels of stress and affect (for a detailed description of data collection, see Almeida, 2005; Almeida, McGonagle, & King, 2009). The NSDE data collection consisted of three bursts of daily assessments repeated at approximately 9-year intervals, providing longitudinal daily diary data across 20 years of adulthood (NSDE 1: ~1996; NSDE 2: ~2005; NSDE 3: ~2017). Daily diary data was collected on a total of 33,688 days out of 37,576 possible days (completion rate = 90%). The current research made use of all available data from respondents who participated in any of the three NSDE bursts (N = 2,845; # daily assessments = 33,688; 1,429 participants completed 2 or more bursts).

NSDE Daily Diary Measures

Daily stressors.

Daily stressors were assessed using the Daily Inventory of Stressful Events (DISE; Almeida, Wethington, & Kessler, 2002). The inventory consisted of six questions inquiring whether certain types of stressors had been experienced in the last 24 hours (i.e., arguments, avoided arguments, work overloads, home overloads, network stressors, other). A dichotomous variable was used to characterize days as either stress days (at least one stressor was reported) or non-stress days (no stressor reported). A daily stressor was reported on 39% of all available days. Two-thirds of the participants reported that the daily assessment period was ‘about the same as a typical week’ in terms of stressfulness, whereas 15% indicated that the week was more stressful than a typical week, and 18% indicated that it was less stressful. Overall, the 8-day daily assessment period appeared to adequately capture a typical week for participants.

Negative Affect.

Daily negative affect was assessed during each burst of the NSDE data collections. Participants were presented with a list of six emotions (fidgety, nervous, worthless, so sad that nothing could cheer you up, everything was an effort, and hopeless; Mrazek & Kolarz, 1998) and asked to indicate how frequently they felt each emotion in the past 24 hours. Responses ranged from 0 (none of the time) to 4 (all of the time). Daily negative affect scores were computed by averaging across the items. Multilevel omega was used to estimate within- and between-person reliability (see Geldhof, Preacher, & Zyphur, 2014). Within-person reliability estimates were .60, .58, and .54 for bursts 1, 2 and 3, respectively. Between-person reliability was .81, .82, and .82 for bursts 1, 2, and 3, respectively.

Covariates.

Participant age at NSDE baseline, sex, education, and race were included as covariates to adjust for sample heterogeneity. Age at baseline was grouped into 5-year bins (<30; 31-35; 36-40; 41-45; 46-50; 51-55; 56-60; 61-65; 66-70; >70). Sex, education, and race were coded as dichotomous variables (0 = male, 1 = female; 0 = high school or less, 1 = some college or more, and 0 = white, 1 = not white, respectively). Each of the covariates were centered at the grand mean in all statistical models.

Data Analytic Strategy

Separate multilevel modeling analyses were used to examine changes in daily stressor exposure and stressor reactivity across 20 years of longitudinal data. Daily measurement occasions were nested within measurement bursts and measurement bursts were nested within people, resulting in three levels of analysis.

Stressor exposure change.

Stressor exposure was defined as the proportion of days within each burst that a stressor was reported. Changes in stressor exposure was estimated with the following equation:

$$\text{Level}1:{\text{Y}}_{ijk}={\pi }_{0ij}+e_{ijk}$$ (1a)

$$\text{Level}2:{\pi }_{0ij}={\beta }_{00\text{i}}+{\beta }_{\text{01i}}\left(Burs{t}_{ij}\right)+{\text{r}}_{0ij}$$ (1b)

$$\text{Level}3:{\beta }_{00\text{i}}={\gamma }_{000}+{\gamma }_{001}\left(Age.B{L}_i\right)+{\gamma }_{002}\left(Se{x}_i\right)+{\gamma }_{003}\left(Colleg{e}_i\right)+{\gamma }_{004}\left(Rac{e}_i\right)+{\text{u}}_{00i}$$ (1c)

$${\beta }_{01\text{i}}={\gamma }_{010}+{\gamma }_{011}\left(Age.B{L}_i\right)+{\text{u}}_{01i}$$ (1d)

The within-burst stress exposure estimate (π_0ij) was regressed on Burst_ij (coded 0, 1, or 2) to provide an estimate of macro-longitudinal change in stressor exposure across bursts, β_01i (between-burst, Level 2). Age at baseline (Age.BL_i), Sex_i, College_i, and Race_i were included as between-person (Level 3) covariates to adjust for differences in baseline levels of stress exposure. Age at baseline was also included as a between-person moderator of changes in stressor exposure (i.e., γ₀₁₁).

Stressor reactivity change.

Stressor reactivity was defined as the difference in NA on days when a stressor was reported compared to days when a stressor was not reported, and specified at the within-burst level (Level 1). Changes in stressor reactivity was estimated with the following equation:

$$\text{Level}1:{\text{NA}}_{ijk}={\pi }_{0ij}+{\pi }_{1ij}\left(Stress.da{y}_{ijk}\right)+e_{ijk}$$ (2a)

$$\text{Level}2:{\pi }_{0ij}={\beta }_{00i}+{\beta }_{01i}\left(Burs{t}_{ij}\right)+{\beta }_{02i}\left(\overline{Stress.da{y}_{IJ}}\right)+{\text{r}}_{0ij}$$ (2b)

$${\pi }_{1ij}={\beta }_{10\text{i}}+{\beta }_{11\text{i}}\left(Burs{t}_{ij}\right)+{\text{r}}_{1ij}$$ (2c)

$$\text{Level}3:{\beta }_{00\text{i}}={\gamma }_{000}+{\gamma }_{001}\left(Age.B{L}_i\right)+{\gamma }_{002}\left(Se{x}_i\right)+{\gamma }_{003}\left(Colleg{e}_i\right)+{\gamma }_{004}\left(Rac{e}_i\right)+{\text{u}}_{00i}$$ (2d)

$${\beta }_{01\text{i}}={\gamma }_{010}$$ (2e)

$${\beta }_{02\text{i}}={\gamma }_{020}$$ (2f)

$${\beta }_{\text{10i}}={\gamma }_{100}+{\gamma }_{101}\left(Age.B{L}_i\right)+{\gamma }_{102}\left(Se{x}_i\right)+{\gamma }_{103}\left(Colleg{e}_i\right)+{\gamma }_{104}\left(Rac{e}_i\right)+{\text{u}}_{10i}$$ (2g)

$${\beta }_{\text{11i}}={\gamma }_{110}+{\gamma }_{111}\left(Age.B{L}_i\right)+{\text{u}}_{11i}$$ (2h)

The within-burst stress reactivity estimate (π_1ij) was regressed on Burst_ij (coded 0, 1, or 2) to provide an estimate of macro-longitudinal change in stressor reactivity across bursts, β_11i (between-burst, Level 2). Age at baseline (Age.BL), Sex, College, and Race were included as between-person (Level 3) covariates predicting baseline levels of negative affect (i.e., γ₀₀₁, γ₀₀₂, γ₀₀₃, and γ₀₀₄, respectively) and stress reactivity (i.e., γ₁₀₁, γ₁₀₂, γ₁₀₃, and γ₁₀₄, respectively). Age at baseline was also included as a between-person moderator of changes in stressor reactivity (i.e., γ₁₁₁). All models were estimated in Mplus v8.5 (Muthén & Muthén, 2017) using full information maximum likelihood with robust standard errors (MLR).

Results

Descriptive statistics at each burst of daily assessments are included in Table 1. Participants ranged in age from 22-77 years (Mean = 48.2, SD = 12.8) at baseline and 47-95 years (Mean = 67.7, SD = 10.3) at the third burst of assessment. Just over half of the participants were women (56%) and a majority (67%) had completed at least some college education. Across the bursts of assessment, participants reported relatively low levels of daily negative affect and experienced at least one stressor on approximately 40% of the study days.

Table 1.

Means, Standard Deviations, and Range among Study Variables

	Burst 1 (N=1,499)			Burst 2 (N=2,022)			Burst 3 (N=1,176)
Variable	M	SD	Range	M	SD	Range	M	SD	Range
Age	48.20	12.84	22-77	58.61	12.12	35-86	67.67	10.34	47-95
Sexa	.53	0.50	0-1	.57	0.49	0-1	.57	0.49	0-1
Collegeb	.65	0.48	0-1	.69	0.46	0-1	.75	0.43	0-1
Racec	.10	0.29	0-1	.15	0.36	0-1	.14	0.34	0-1
Daily negative affectd,f	0.18	0.29	0-3.39	0.20	0.31	0-3.08	0.17	0.25	0-2.92
Daily stressorse,f	.39	0.26	0-1	.40	0.27	0-1	.39	0.28	0-1

Note.

^aProportion of female participants.

^bProportion of participants with some college.

^cProportion of non-white participants.

^dAggregated across persons.

^eProportion of stress days across persons.

^fDaily variables were initially aggregated across days and then again across persons to create between-person averages of stressor exposure and negative affect. 1,429 participants completed 2+ bursts.

Stressor Exposure

Changes in stressor exposure are presented in Table 2. On average individuals declined in the proportion of stress days reported across the 20-year period (γ₀₁₀ = −0.022, p < .001). By the third burst, stressor exposures had declined by 11% compared to baseline levels. Furthermore, cross-sectional age differences indicated that older adults at baseline reported fewer stress days than younger adults (γ₀₀₁ = −0.020, p < .001). For example, the proportion of stress days reported by individuals >70 years old at baseline was estimated to be 25% fewer than individuals 46-50 years old at baseline. Both the longitudinal age changes and the cross-sectional age differences in stress exposure are illustrated in Figure 1A.

Table 2.

Multilevel Modeling Analyses of the Longitudinal Changes in Daily Stressor Exposure across 20 years

Variable	Estimate (SE)	p-value	95% CI
Fixed Effects
Within-burst variables (Level 1)
Stress exposure intercept (γ000)	0.400 (.006)	<.001	[0.389, 0.411]
Between-burst variables (Level 2)
Stress exposure change (γ010)	−0.022 (.004)	<.001	[−0.030, −0.013]
Between-person variables (Level 3)
Sex (γ002)	0.048 (.009)	<.001	[0.031, 0.065]
College (γ003)	0.099 (.009)	<.001	[0.081, 0.116]
Race (γ004)	−0.045 (.012)	<.001	[−0.070, −0.021]
Age at baseline (γ001)	−0.020 (.002)	<.001	[−0.025, −0.016]
Stress exposure change * Age at baseline (γ011)	−0.002 (.002)	.400	[−0.005, 0.002]
Random effects
Within-person residual (σe2)	0.195 (.002)	<.001	[0.192, 0.198]
Between-burst
Stress exposure (σr02)	0.005 (.001)	<.001	[0.003, 0.008]
Between-person
Stress exposure (σu002)	0.028 (.002)	<.001	[0.025, 0.031]
Stress exposure change (σu012)	0.003 (.001)	<.001	[0.001, 0.004]

Note. Results are based on 33,688 daily assessments (N = 2,845). CI = confidence interval. Sex (0=male; 1=female). College (high school or less = 0; some college or more = 1). Race (white = 0; not white = 1). Estimates of fixed effects are reported as unstandardized regression coefficients. Estimates of random effects are reported as variances.

Figure 1.

A. Longitudinal age changes and cross-sectional age differences in stressor exposure across adulthood. B. Johnson-Neyman plot to identify regions of significance. The simple slope of stress exposure change is shown across varying Ages at baseline (thick black line). The gray bands represent the 95% confidence interval that can be used to infer statistical significance. When the horizontal zero line is included in the confidence bands, the simple slope is not statistically significant at that age. The vertical hatched line denotes the boundary age where longitudinal change in stress exposure is statistically significant.

Age at baseline did not moderate the longitudinal changes in stressor exposure (γ₀₁₁ = −0.002, p = .400). That is, older adults did not significantly differ from younger adults in the rate of decline in stressor exposure. Figure 1B displays the simple slope of longitudinal stress exposure changes at varying baseline ages. For individuals over 30 years old at baseline, stress exposure significantly declined across the 20-year follow-up. Differences in baseline levels of stress exposure were also detected for sex, education, and race. Females reported more stress days than males; individuals with some college reported more stress days than those who completed high school or less; and white participants reported more stress days than non-white participants.

Stressor Reactivity

Table 3 presents the results of the multilevel model examining changes in stressor reactivity. Consistent with much previous research (Stawski et al., 2019), a significant stress reactivity estimate was established. On days when a stressor was reported, participants reported higher NA compared to days when a stressor was not reported (γ₁₀₀ = 0.139, p < .001). Longitudinal changes in stressor reactivity indicated that on average individuals declined in their reactivity to stressors across 20 years of follow-up (γ₁₁₀ = −0.017, p = .002). Individuals at the sample mean age (i.e., 46-50 years old) at baseline were estimated to be 24% less reactive to daily stressors at the 20-year follow-up relative to their baseline levels. Cross-sectional age differences also emerged, such that older adults at baseline were less reactive to daily stressors compared to younger adults (γ₁₀₁ = −0.010, p < .001). Individuals 66-70 years old at baseline were estimated to be 29% less reactive to daily stressors than individuals 46-50 years old at baseline.

Table 3.

Multilevel Modeling Analyses of the Longitudinal Changes in Daily Stressor Reactivity across 20 years

Variable	Estimate (SE)	p-value	95% CI
Fixed Effects
Within-burst variables (Level 1)
NA Intercept (γ000)	0.064 (.007)	<.001	[0.050, 0.079]
Stress reactivity intercept (γ100)	0.139 (.007)	<.001	[0.125, 0.153]
Between-burst variables (Level 2)
Burst-mean stress (γ020)	0.172 (.017)	<.001	[0.139, 0.205]
NA change (γ010)	0.011 (.003)	.002	[0.004, 0.018]
Stress reactivity change (γ110)	−0.017 (.006)	.002	[−0.028, −0.006]
Between-person variables (Level 3)
Sex (γ002)	−0.002 (.008)	.798	[−0.018, 0.014]
College (γ003)	−0.066 (.009)	<.001	[−0.084, −0.048]
Race (γ004)	0.086 (.015)	<.001	[0.057, 0.116]
Age at baseline (γ001)	0.001 (.002)	.736	[−0.003, 0.004]
Stress reactivity * Sex (γ102)	0.024 (.009)	.005	[0.007, 0.041]
Stress reactivity * College (γ103)	−0.052 (.011)	<.001	[−0.074, −0.031]
Stress reactivity * Race (γ104)	0.016 (.015)	.292	[−0.014, 0.045]
Stress reactivity * Age at baseline (γ101)	−0.010 (.003)	<.001	[−0.016, −0.005]
Stress reactivity change * Age at baseline (γ111)	0.006 (.002)	.011	[0.001, 0.010]
Random effects
Within-person NA (σe2)	0.054 (.002)	<.001	[0.050, 0.058]
Between-burst
NA (σr02)	0.014 (.003)	<.001	[0.009, 0.019]
Stress reactivity (σr12)	0.023 (.004)	<.001	[0.016, 0.030]
Between-person
NA (σu002	0.032 (.006)	<.001	[0.021, 0.043]
Stress reactivity (σu102)	0.012 (.005)	.007	[0.003, 0.021]
Stress reactivity change (σu112)	0.000 (.004)	.976	[−0.008, 0.008]

Note. Results are based on 33,688 daily assessments (N = 2,845). CI = confidence interval. Sex (0=male; 1=female). College (high school or less = 0; some college or more = 1). Race (white = 0; not white = 1). Estimates of fixed effects are reported as unstandardized regression coefficients. Estimates of random effects are reported as variances.

Importantly, age at baseline significantly moderated changes in stress reactivity (γ₁₁₁ = 0.006, p = .011). Figure 2A illustrates how younger adults declined more rapidly longitudinally across time in their reactivity to daily stressors compared to older adults. The Johnson-Neyman technique was applied to identify the ages where individuals were significantly changing in stressor reactivity. As displayed in Figure 2B, individuals younger than age 54 at baseline were significantly declining in stressor reactivity, whereas individuals older than 54 did not significantly decline across time. Furthermore, around age 65 at baseline, individuals demonstrated slight (but non-significant) longitudinal increases in stressor reactivity.

Figure 2.

A. Longitudinal age changes and cross-sectional age differences in stressor reactivity across adulthood. B. Johnson-Neyman plot to identify regions of significance. The simple slope of stress reactivity change is shown across varying Age at baseline (thick black line). The gray bands represent the 95% confidence interval that can be used to infer statistical significance. When the horizontal zero line is included in the confidence bands, the simple slope is not statistically significant at that age. The vertical hatched line denotes the boundary age where longitudinal change in stress reactivity is no longer statistically significant.

In addition to the primary age-related effects, sex and education also accounted for differences in baseline levels of stress reactivity. Females were more reactive to daily stressors compared to males, whereas individuals with some college were less reactive to daily stressors compared to those who completed high school or less.

Discussion

Stress is a contributing factor to many mental and physical health conditions (Yaribeygi et al., 2017). Given how important it is to lifespan health and well-being, the current study examined how two spects of stress—exposure to stressors and affective reactivity to these stressors—change over a 20-year time span among a large national sample of adults who were interviewed across three time points. Findings overall reveal a clear and robust benefit with age. Younger adults have the highest levels of stress – both in terms of stressor exposure and reactivity – compared to any other age group. Yet, their stress profile improves as they age, with adults reporting, on average, an 11% reduction in the occurrence of stressor days over time and younger adults exhibiting the steepest declines (47% reduction) in levels of stressor reactivity. For people in their mid fifties and older, however, stressor occurrence continues to decrease over time, yet stress reactivity remains stable across time.

Daily stressor occurrence.

Both cross-sectional and longitudinal results reveal the same age-related pattern: daily stressors are most prevalent in young adulthood and steadily decrease in number with age. The proportion of stressor days reported by older adults are 25% fewer than middle-aged adults and 38% fewer than the youngest adults. Although this study did not test any underlying mechanisms to explain these age benefits, findings are consistent with a number of sociological and psychological theories. Life stage models of development describe the patterns of behavior typical of each life stage. Young adulthood is a time of constant role changes, with people starting careers, navigating financial independence, and often establishing families. Younger adults report more major life events than middle-aged and older adults (Hughes et al., 1988), and daily stressors generally accompany these major life transitions (Almeida & Wong, 2009). Perhaps for older adults, retirement, independent children, and less active lifestyles (e.g. Lee et al., 2018) may partially account for the fewer stressors they experience compared to middle-aged and younger adults.

In addition to this life course model, the socioemotional selectivity theory posits that time perspective is linearly associated with age, and the greater awareness of the finitude of life leads people to place increasingly greater emphasis on emotionally meaningful goals. As a result, older adults often appraise their lives less negatively and more positively, a phenomenon known as the positivity effect (Carstensen & Mikels, 2005; Reed & Carstensen, 2012). With age, then, people may be more likely to seek positive situations and avoid unnecessary sources of stressors.

Stressor reactivity.

Changes in life perspectives may also play a role in the age-related reduction of affective reactivity to stressors we observed across younger adulthood and into midlife. Appraisals are strongly related to stress responses (Lazurus, 2000), and older adults perceive equally objectively severe stressors as less threatening than do younger adults (Aldwin, 2007). With age, researchers posit that people grow increasingly better at regulating their emotions by proactively avoiding stressors (Charles, 2010), and this decreased reactivity may reflect increases in antecedent emotion regulatory skills.

Yet, the age-related declines in stressor reactivity are not as linear as the findings for stressor occurrence. Starting in the midst of midlife (estimated around 54 years old), people were stable in their stressor reactivity across 20 years. Again, we did not test mechanisms explaining why the decline in reactivity ceases at this time. One possibility is that a stressor, by definition, is usually associated with some degree of reactivity, and by midlife people have reached a low level of reactivity that cannot decline further (i.e., the floor effect). Another possibility is that although older adults may be more adept at avoiding the presence of stressors, they may not be more adept at regulating the high levels of distress that stressors elicit. Strength and Vulnerability Integration (SAVI) states that age-related changes in physiology may create difficulties down-modulating higher levels of distress. As a result, this model posits that older adults will report fewer stressors, but may not exhibit age-related reductions in reactivity when stressors do occur.

Limitations and future directions.

Of course, the strengths of this study must be interpreted alongside its limitations. Stressors were assessed through self-reports at the end of the day, and not by collecting objective measures of these events. As such, we rely on how people appraise and remember their stressors, and these appraisals and memory requirements may be affected by age. These limitations were offset by the large number of people capturing naturalistic stress processes that occur in daily life. Another major limitation is the lack of diversity in racial and ethnic composition of the sample, and the lack of people who fall in the lowest socioeconomic stratum of America (who are arguably most vulnerable to the effects of daily stress). In addition, younger adults at the bottom of the adult age spectrum (ages 18-22) were not included in NSDE and only a small number of the oldest old (e.g., those who are 90 years and older) were represented. We acknowledge that our results may not generalize to understudied groups. With the population increasingly becoming socioeconomically, racially, and ethnically diverse, and living longer than ever, studying how changes vary by these critical demographic dimensions is an ethical and scientific necessity.

Conclusion.

Stress is a speedometer of our life course. Stress permeates our daily lives and plays an important role in psychological health, chronic conditions, and even mortality. Given its pernicious effects, an important question is how do individuals experience stress as they age? Using our large U.S. national sample with 20 years of longitudinal data, we show that age confers some benefits in terms of decreased exposure and emotional reactivity to daily stressors. Although stress might accelerate the aging processes, growing older may also allow us to lead less stressful lives.