Trajectories and personality correlates of change in perceptions of physical and mental health across adulthood and old age

Abstract

Subjective health is known to predict later outcomes, including survival. However, less is known about subjective health changes across adulthood, how personality moderates those changes, and whether such associations differ with age. We applied growth models to 10 waves of data from the Household, Income and Labour Dynamics in Australia Survey (HILDA, N = 7,172; median ages 20–93) to examine age-related differences in trajectories of subjective physical and mental health. On average, perceptions of physical health declined with increasing steepness in old age, whereas self-rated mental health remained relatively stable across all ages. Higher neuroticism and lower extraversion and conscientiousness were each related to less successful aging. The health implications of personality did not differ by age for physical health, but were weaker for mental health in old age. We discuss implications of our results for accelerated longitudinal designs and consider avenues for future more mechanism-oriented research.

Keywords

adulthood and old age growth curve model longitudinal self-rated health well-being

Maintenance of physical and mental health is among the central ingredients of successful aging and quality of life (Baltes & Baltes, 1990; Ryff & Singer, 1998). For example, Rowe and Kahn (1997) define successful aging in part as the maintenance of health and low disease and disability. A large body of evidence has accumulated that between-person differences in self-perceptions of health predict key outcomes such as mortality (for review, see Idler & Benyamini, 1997). However, less is known about how perceptions of health change across adulthood, what moderates those changes, and whether such moderator associations differ with age. In this report, we make use of a large, nation-wide sample assessed annually for 10 years to track how perceptions of physical and mental health develop across adulthood and old age (for people in a First World nation), to examine the role personality plays in health change, and to investigate whether that role differs with age. Precisely understanding how people perceive their health to change can inform public health and policy aimed at promoting the well-being and health of our society. In the present study, we applied growth curve models to 10 waves of annual longitudinal data from the Household, Income, and Labour Dynamics in Australia Survey (HILDA). We first modeled age-related differences in change trajectories of perceived physical and mental health for individuals aged 20 to 93 years (the median age of each participant across all of his or her waves). We then investigated whether and how personality factors were related to between-person differences in perceived health trajectories and whether the direction and strength of these associations differed with age.

Typical age-related changes in physical and mental health

Physical health can be construed as a multi-dimensional construct encompassing physician-diagnosed health symptoms, limitations in completing everyday tasks of living, and self-ratings of health (Lawton & Lawrence, 1994). Age-related differences in each facet are well documented, with older adults (e.g. age 50 and over) often reporting poorer physical health and suffering from more frequent and severe limitations than younger adults (McCullough & Laurenceau, 2004; Rook et al., 2007; Sargent-Cox et al., 2010; Smith et al., 2002; Suzman et al., 1992). On average, the older someone is then the worse his or her health will be compared to a younger person. Mental health can be construed as a multi-dimensional construct that includes the absence of mental disorders, the ability to deal with life’s problems effectively, the cognitive evaluation of life, emotional well-being, and social functioning including feelings of social integration (Gatz & Zarit, 1999; Keyes, 2005). In contrast to physical health, aspects of mental health usually remain relatively stable across adulthood and old age, with larger age-related decrements being observed only in very old age (e.g. age 85+; Aldwin et al., 1989; Baltes & Smith, 2003; Zarit, 2010) and at the end of life (e.g. approaching death; Gerstorf, Ram et al., 2010).

Personality correlates of age-related changes in physical and mental health

There are substantial between-person differences in how physical and mental health change with age. Some people maintain good health until late in life, and others experience precipitous declines early on (Spiro, 2007). Here, we are interested in how personality and age-based selection contribute to heterogeneity in the trajectories of individuals’ self-reported health. Personality traits have been linked with physical and mental health. Most prominently, higher neuroticism is often related to poorer subjective physical and mental health, primarily because of heightened stress in anticipating and reacting to challenging situations (Aldwin et al., 2001; Charles et al., 2008; Lahey, 2009). In contrast, people with higher conscientiousness tend to report better physical and mental health, perhaps because they tend to participate in health preventative (e.g. colorectal exams, mammography) and promoting behaviours (e.g. Boggs & Roberts, 2004; Friedman, Kern, Hampson, & Duckworth, 2012). Less is known, however, about how extraversion, openness, and agreeableness are associated with health (Lockenhoff, Sutin, Ferrucci, & Costa, 2008; Turiano et al., 2012). Initial evidence suggests that extraverted individuals tend to report better health and live longer (Turiano et al., 2012; Wilson, Mendes de Leon, Bienas, Evans, & Bennett, 2004), a trajectory that might be linked to the frequent experience of positive emotions and associated feelings of vitality and energy (Duberstein et al., 2003). For openness, several studies suggest a positive relationship with subjective health (e.g. Duberstein et al., 2003; Hampson & Friedman, 2008; Jerram & Coleman, 1999), perhaps because more open individuals lead more active lives, have greater mental flexibility, and deal with changes in life circumstances better than less open individuals. In turn, such engagement may have positive implications for both mental and physical health (Duberstein et al., 2003; Jerram & Coleman, 1999). Agreeableness has also been linked with better health behaviours and subjective health (Booth-Kewley & Vickers, 1994; Turiano et al., 2012). It is possible that agreeable individuals’ conflict avoidant and pro-social natures make it easier for them to seek out and utilize social support, which in turn is linked with health in a multitude of different ways (e.g. discouraging risky behaviours or buffering stress; Berkman, Glass, Brisette, & Seeman, 2001; Cohen, 2004).

Research has also suggested that the relationship between personality and health is moderated by age. Although few studies have examined this, there is some evidence that the relationship of personality and health is different for older and younger individuals (Duberstein et al., 2003; Lockenhoff et al., 2008; Quinn, Johnson, Poon, & Martin, 1999). One hypothesis is that aging-related losses may magnify the effects of personality on health. For example, individuals with high levels of neuroticism may become even more likely to report poor health later in life because of the accumulation of age-related losses, which could accentuate the anxious and worrisome tendencies of these individuals. Alternatively, age-related declines may become too pervasive for personality to make a difference later in life (e.g. in the fourth age, 85+; Baltes & Smith, 2003). Conflicting evidence has been reported in the few studies examining such age moderation, with some reports supporting the notion that personality–health associations were stronger at older ages (Duberstein et al., 2003; Quinn et al., 1999), whereas others do not (Lockenhoff et al., 2008).

The present study

Taken together, our study aimed to describe age differences in trajectories of perceived physical and mental health, identify personality moderators of successful development/aging (i.e. high levels and maintenance of physical and mental health), and explore whether personality associations differ with age. We note that the observation interval (10 years) was relatively sparse compared with the broad age range across the adult life span that our model parameters generalize to (20 to 93 years). As a consequence, we used the accelerated longitudinal study design to model health trajectories over time and examine age-related differences therein as opposed to modeling change in health directly over age. Going this route allowed us to test rather than assume age convergence (i.e. that the younger individuals would turn into the older individuals in the sample; for discussion, see Sliwinski, Hoffman, & Hofer, 2010). We hypothesized that perceptions of physical health are characterized by decline across much of adulthood, with some concave curvature that manifests as pronounced decline in old age. In contrast, mental health is expected to remain relatively stable across the adult life span. Considering personality differences, we expected that higher neuroticism and lower conscientiousness each relate to less favourable trajectories, as characterized by lower levels of and steeper declines in self-rated health. We also explored the roles of extraversion, openness, and agreeableness and tested whether or not the direction and strength of associations differ with age.

To control for important correlates of self-rated health, our models also included gender, education, disability, and negative life events as covariates. Specifically, accumulated opportunity disadvantages (in most countries) along with limited financial resources to access healthcare and greater chronic strain often contribute to women having poorer physical and mental health, a greater number of incapacitating conditions and acute health issues, as well as a higher risk for depression than men (Case & Paxson, 2005; Crimmins et al., 2010; McCullough & Laurenceau, 2004; Moen, 1996; Nolen-Hoeksema, 1999; Smith & Baltes, 1998). Higher education and the accompanying resources such as financial assets and health-promoting behaviours are often associated with better health and may also serve as protective factors against health declines (Adler et al., 1994). Finally, poorer health is typically reported by individuals with a disability (e.g. Lucas, 2007; Zarit, Johansson, & Berg, 1993) and those who experience negative life events, presumably because of the stress and challenges involved (e.g. Cohen, Tyrrell, & Smith, 1993).

Method

Participants and procedure

The HILDA is a nationally representative, longitudinal panel survey of randomly selected households from a set of randomly selected census districts across Australia (for details, see Dyrenforth et al., 2010; Lucas & Donnellan, 2009; Watson, 2010). Within a household, all persons aged 15 and over were invited to participate. As new members have entered households, they have also been invited to participate. Data are collected annually, since 2001, via a combination of face-to-face and telephone interviews and self-completed questionnaires. We utilized data from the first ten waves of assessment (2001–2010), including in our analyses all participants with valid data on physical and mental health and all the correlates. Of particular note, personality was only measured at Wave 5 and was the main constraint on sample size, so that the study subsample consisted of n = 7,172 of the total N = 19,914 sample. Compared to participants who were not included in our sample, our “personality-rich” subsample was older (M = 44.8, SD = 15.9 vs. M = 42.4, SD = 19.3; F[1, 13,967] = 65.62, p < .001), more highly educated (57% vs. 42% with post-high school education; χ² [1, N = 12,783] = 261.70, p < .001), consisted of more women (54% vs. 49%; χ² [1, N = 19,914] = 50.61, p < .001), and reported greater physical health (M = 49.4, SD = 9.9 vs. M = 48.8, SD = 10.8; F[1, 12,321] = 9.97, p = .002) and mental health (M = 50.5, SD = 9.7 vs. M = 49.2, SD = 10.2; F[1, 12,321] = 48.90, p < .001) at Wave 1. Effect sizes for selectivity differences were in the small range (R² < .005 for all comparisons).

On average, these individuals provided 9.1 (of 10 possible) repeated reports (SD = 1.4; 75% provided data on 9 or more occasions). Relative to participants who provided fewer waves of data (5 waves or less; n = 236), participants who provided six or more waves of data (n = 6,936) were younger (M = 44.6, SD = 15.7 vs. M = 51.2, SD = 20.2; F[1, 7,170] = 39.46, p < .001), more highly educated (57% vs. 39% with post-high school education; χ² [1, N = 7,172] = 31.69, p < .001), and reported better physical health (M = 49.5, SD = 9.8 vs. M = 45.4, SD = 12.5; F[1, 6,689] = 25.84, p < .001) at Wave 1. No gender or mental health differences were found. Effect sizes for selectivity differences were in the small range (R² < .005 for all comparisons). Descriptive statistics for all measures are given in Table 1.

Table 1.

Means, standard deviations, and interrcorrelations among study constructs.

Construct	M	SD	1	2	3	4	5	6	7	8	9	10	11	12
Outcomes
1. Physical health	49.43	9.91	−
2. Mental health	50.50	9.68	−.03	−
Correlates
3. Age	49.13	15.91	−.35*	.18*	−
4. Gender	0.54	0.50	−.01	−.07*	−.02	−
5. Education	0.57	0.50	.12*	−.02	−.16*	−.14*	−
6. Neuroticism	1.75	1.08	.00	−.31*	−.26*	−.06*	.01	−
7. Conscientiousness	5.17	1.02	.04*	.17*	.15*	.10*	.03*	−.30*	−
8. Extraversion	4.39	1.07	.05*	.11*	−.06*	.13*	.01	−.20*	.14*	−
9. Openness	4.19	1.05	.04*	−.10*	−.13*	−.06*	.22*	.20*	.08*	.06*	−
10. Agreeableness	5.41	0.91	−.00	.05*	.04*	.25*	−.00	−.16*	.30*	.17*	.26*	−
11. Disability	0.22	0.41	−.54*	−.13*	.22*	−.05*	−.08	.01	−.04*	−.04*	−.00	−.00	−
12. Negative major events	0.69	0.92	−.13*	−.13*	.02	.01	.01	.05*	−.01	.02	.06*	.05*	.12*	−

Note. Mean scores and standard deviations taken from T1 (except for age which was measured as the median age of participants across time in study, health conditions status from T2, and personality measures which were only collected at T5). *p < .01.

Measures

Physical and mental health

Annually, participants completed the SF-36 short form health survey, a tool validated for examination of health differences and population health changes (Butterworth & Crosier, 2004; Hemingway, Stafford, Stansfeld, Shipley, & Marmot, 1997; Hopman et al., 2006; Mishra, Ball, Dobson, & Byles, 2004). Extensive details on scale construction can be found in Ware and colleagues (1993). The scale consists of eight subscales: physical functioning (10 items), role limitations due to physical problems (4 items), bodily pain (2 items), general health (5 items), vitality (4 items), social functioning (2 items), role limitations due to emotional problems (3 items), and mental health (5 items). The eight subscales are typically grouped into two higher-order factors representing physical health and mental health summary scores (e.g. Butterworth & Crosier, 2004). All subscales load on and contribute to the higher order physical health and mental health factors, but the subscales physical functioning, role limitations due to physical problems, bodily pain, and general health contribute the most to the physical health factor. In contrast, the subscales vitality, social functioning, mental health, and role limitations due to emotional problems contribute primarily to the mental health factor. Following standard scoring procedures (see Ware, Kosinski, & Keller, 1994), we standardized each of the eight subscale scores using US normed population averages and standard deviations, calculated the weighted (factor score coefficients from the scoring manual) sums of those z-scores to obtain the physical and mental health factors scores, and transformed to t-scores (Note: scores are in US normed population units with Mean = 50, SD = 10). Higher scores indicate better health. Internal consistencies for both scales were good (α ≥ .82 for each subscale; see also Ware et al., 1995).

Age

An individual’s age was indexed as the median age across all waves that he or she participated. As a time-invariant, between-person characteristic, median age (rather than age at T1) facilitated more precise placement of the individual’s time series within the 20–93 year age span.

Personality

Personality was measured in 2005 using Saucier’s Mini-Marker Inventory (1994). Respondents were asked to indicate how well a set of adjectives described them, using a scale from 1 (does not describe me at all) to 7 (describes me very well). For each trait, the average of six items (except for agreeableness with four items) indicated individuals’ level of neuroticism (e.g. moody, fretful), conscientiousness (e.g. systematic, disorganized), extraversion (e.g. talkative, lively), openness (e.g. intellectual, creative), and agreeableness (e.g. cooperative, sympathetic). Internal consistency for each scale was adequate (Cronbach’s α ≥ .74).

Covariates

Gender was a dichotomous variable with 54% of our sample being women. Education was indexed as either high school graduate or less (43%) vs. post high-school education (57%). Disability was measured by a single item from Wave 1 asking participants to indicate whether or not they had a health condition, which limited their everyday activity and lasted for six months or more. 22% of the sample reported a health condition (0 = no health condition, 1 = health condition present). Major negative life events were indexed by how many negative events (separation from spouse or long-term partner, serious injury or illness to self, injury or illness to close relative, death of spouse or child, death of a close relative, death of close friend, victim of physical violence, victim of a property crime, detained in jail, close family member detained in jail, fired or made redundant by employer, major worsening in financial situation) participants had experienced in the past 12 months (adapted from Holmes & Rahe, 1967). We used the measure from Wave 2 to stay consistent with using data from the first wave available for each covariate.

Data analysis

We estimated separate growth curve (i.e. multilevel) models (e.g. McArdle & Nesselroade, 2003; Ram & Grimm, 2007; Singer & Willett, 2003) to examine age-related differences in self-rated physical and mental health changes over time and to examine how personality variables were related to those changes. Models were specified as

{H e a l t h}_{t i} {= β}_{0 i} {+ β}_{1 i} {(t i m e}_{t i} {) + β}_{2 i} {(t i m e}_{t i}^{2} {) + e}_{t i},

where person i’s physical or mental health at time t, Health_ti , is a function of an individual-specific intercept parameter, β_0i _, individual-specific linear and quadratic slope parameters, β_1i and β_2i , that capture the linear rate and acceleration of change per year of time in study, and residual error, e_ti . Following standard growth curve modeling procedures, individual-specific intercepts, β_0i , linear slopes, β_1i , and quadratic slopes β_2i , were modeled as

\begin{aligned} β_{0 i} & = γ_{00} + γ_{01} {(a g e}_{i}) + γ_{02} {(a g e}_{i}^{2}) + γ_{03} {(g e n d e r}_{i}) \\ + γ_{04} {(e d u c a t i o n}_{i}) + γ_{05} {(n e u r o t i c i s m}_{i}) \\ + γ_{06} {(c o n s c i e n t i o u s n e s s}_{i}) + γ_{07} {(e x t r a v e r s i o n}_{i}) & (2) \\ + γ_{08} {(o p e n n e s s}_{i}) + γ_{09} {(a g r e e a b l e n e s s}_{i}) + γ_{010} {(d i s a b i l i t y}_{i}) \\ + γ_{011} {(n e g a t i v e e v e n t s}_{i}) + u_{0 i} \end{aligned}

\begin{aligned} β_{1 i} & = γ_{10} + γ_{11} {(a g e}_{i}) + γ_{12} {(a g e}_{i}^{2}) + γ_{13} {(g e n d e r}_{i}) \\ + γ_{14} {(e d u c a t i o n}_{i}) + γ_{15} {(n e u r o t i c i s m}_{i}) \\ + γ_{16} {(c o n s c i e n t i o u s n e s s}_{i}) + γ_{17} {(e x t r a v e r s i o n}_{i}) \\ + γ_{18} {(o p e n n e s s}_{i}) + γ_{19} {(a g r e e a b l e n e s s}_{i}) + γ_{110} {(d i s a b i l i t y}_{i}) \\ + γ_{111} {(n e g a t i v e e v e n t s}_{i}) + u_{1 i} \\ β_{2 i} = γ_{20} + u_{2 i}, \end{aligned}

where the γs are the sample-level associations between socio-demographic and personality factors and the 10-year trajectories (level and slope), and u_0i , u_1i , and u_2i are unexplained differences. We note that between-person differences in curvature, u_2i , were left un-modeled. Time_ti was centred at the middle of each person’s repeated measures series, and all person-level predictors were centred at sample means. Models were estimated in SAS 9.2 using PROC MIXED (Littell et al., 2006) using full information maximum likelihood estimation, thereby treating incomplete data as missing at random (Little & Rubin, 1987) and adjusting for unbalanced data (Singer & Willett, 2003). Personality interactions with age were also tested in the models and trimmed iteratively, so that the final model included only those age interaction effects that were significant at the α = .01 level.

Results

Using intercept-only models, we first estimated the proportion of between-person and within-person variation (intraclass correlation) in the repeated measures of health. Although the majority of the total variance was between-persons (physical: 70%; mental: 53%), there was substantial within-person variance as well (physical health: 30%; mental health: 47%).With both within-person and between-person variability to model, we used growth models to describe and evaluate how this variation was structured over time and across age. Results from the final models are given in Table 2. We present results from each portion of the model in turn.

Table 2.

Growth models of physical and mental health over time, as applied to ten waves of annual (2001–2010) longitudinal data from the Household, Income and Labour Dynamics in Australia Survey (HILDA).

	SF-36 Component Measures
Effect (se)	Physical health		Mental health
Fixed effects
Intercept (age 49), γ₀₀	48.954*	(0.112)	50.832*	(0.106)
Time (age 49), γ₁₀	– 0.252*	(0.015)	0.124*	(0.015)
Time² (age 49), γ₂₀	– 0.005	(0.003)	– 0.004	(0.004)
Age, γ₀₁	–0.216*	(0.005)	0.061*	(0.005)
Age², γ₀₂	–0.003*	(0.000)	0.000	(0.000)
Gender (women)	– 0.816*	(0.165)	– 1.813*	(0.157)
Education	0.757*	(0.165)	0.034	(0.156)
Neuroticism	–0.392*	(0.082)	–2.239*	(0.078)
Conscientiousness	0.535*	(0.084)	0.653*	(0.080)
Extraversion	0.276*	(0.076)	0.698*	(0.072)
Openness	0.032	(0.082)	–0.393*	(0.078)
Agreeableness	– 0.095	(0.097)	0.270*	(0.092)
Disability	–8.927*	(0.195)	-3.423*	(0.186)
Negative life events	–0.909*	(0.085)	–0.977*	(0.081)
Age X time, γ₁₁	– 0.013*	(0.001)	–0.004*	(0.001)
Age²X time, γ₁₂	–0.000	(0.000)	–0.000*	(0.000)
Gender X linear slope	– 0.074*	(0.023)	0.042	(0.024)
Education X linear slope	0.018	(0.023)	0.013	(0.024)
Neuroticism X linear slope	– 0.008	(0.012)	– 0.010	(0.012)
Conscientiousness X linear slope	0.023	(0.012)	–0.014	(0.012)
Extraversion X linear slope	0.024	(0.011)	0.003	(0.011)
Openness X linear slope	–0.016	(0.012)	0.027	(0.012)
Agreeableness X linear slope	0.008	(0.014)	0.020	(0.014)
Disability X linear slope	0.287*	(0.028)	–0.003	(0.029)
Negative life events X linear slope	–0.002	(0.012)	0.054*	(0.013)
Age X openness	–		0.018*	(0.004)
Random effects
Variance of intercept, σ² _u0	47.873*	(0.917)	38.818*	(0.814)
Variance of linear slope, σ² _u1	0.470*	(0.015)	0.398*	(0.016)
Variance of quadratic slope, σ² _u2	0.026*	(0.001)	0.029*	(0.002)
Cov. intercept, linear slope, σ_u1u0	1.096*	(0.083)	0.570*	(0.080)
Cov. intercept, quadratic slope, σ_u2u0	–0.011*	(0.003)	– 0.365*	(0.030)
Cov. slopes, σ_u2u1	– 0.011*	(0.003)	– 0.026*	(0.004)
Residual variance, σ² _e	26.642*	(0.180)	37.583*	(0.253)
–2LL	425,518		442,568

Note. Age refers to the median age of participants and is centred at age 49. Time centred at T4. Numbers in the parentheses are the standard errors of the estimates. N = 7,172. *p < .01.

Age-related differences in physical and mental health trajectories

The prototypical Australian (49-year-old) participant’s self-rated physical health declined a little over 2 ½ T-score units over the 10 years of time in study (γ₁₀ = – 0.252, p < .01), with negligible curvature (γ₂₀ = –.005, p > .01), at comparable average levels to the US (γ₀₀ = 48.954, p < .01). Because the SF-36 data was standardized using US normed population averages and standard deviations, we could compare the average level of our Australian sample to the US average level (e.g. 50). Trajectories of change were significantly related to age and age², such that older age was associated with both lower levels of physical health (γ₀₁ = –0.216 and γ₀₂ = –0.003, ps < .01) and steeper rates of decline (γ₁₁ = –0.013, p < .01). The pattern of age-related moderation of 10-year changes is shown in the top panel of Figure 1. Age convergence in the sample would be evident if the 10 year trajectories fit well to one continuous trajectory over all ages. However, non-convergence of 10-year age differences and 10-year changes is readily discernible by the mainly non-overlapping trajectories.

Figure 1.

Age-related differences (for 10-year slices of age) in trajectories of self-rated physical health (top panel) and mental health (bottom panel) scales of the SF-36. Self-rated physical health becomes lower throughout adulthood and shows steep declines in old age. In contrast, mental health remains relatively stable across most of the adult life with slightly steeper declines in later life.

In contrast to prototypical declines in physical health, the prototypical Australian (49-year-old) participant’s self-rated mental health increased about 1 T-score unit over the 10 years of study time (γ₁₀ = 0.124, p < .01), with negligible curvature (γ₂₀ = –.004, p > .01), at a level similar to US averages (γ₀₀ = 50.832, p < .01). Trajectories of change were again significantly related to age and age². In contrast to self-rated physical health, older age was associated with higher levels of mental health (γ₀₁ = 0.061, p < .01) and steeper rates of decline (γ₁₁ = –0.004 and γ₁₂ = –0.0003, ps < .01). The resulting pattern of age-related moderation of 10-year changes is shown in the bottom panel of Figure 1. Relative convergence of 10-year age-related differences across slices of 10-year change is apparent at younger ages by mostly overlapping 10-year trajectories, but becomes more and more non-convergent with increasing age. In particular, intercepts gradually get higher in older ages with older adults in the sample appearing to have greater levels of mental health at baseline.

Personality differences in physical and mental health trajectories

Table 2 also presents parameter estimates for the personality moderators of perceived health trajectories. Many of the personality traits were found to moderate levels and changes in perceived health. Specifically, higher neuroticism was associated with both poorer physical and mental health (γ₀₅ = – 0.392; – 2.239, respectively). Similarly, participants high in conscientiousness and extraversion each reported better physical (γ₀₆ = 0.535; γ₀₇ = 0.276) and mental health (γ₀₆ = 0.653; γ₀₇ = 0.698). Participants scoring high on openness to experience perceived their mental health less favourably than those low on openness (γ₀₈ = -0.393). Finally, more agreeable individuals reported better mental health (γ₀₉ = 0.270).

In addition, several of the covariates related to level and linear change in physical and mental health. Results indicated that being a woman related to poorer physical (γ₀₃ = – 0.816) and mental health (γ₀₃ = – 1.813) as well as to steeper 10-year declines in physical health (γ₁₃ = –0.074). In addition, lower education was associated with poorer physical health (γ₁₄ = 0.757). As can be expected, health conditions (γ₀₁₀ = –8.927 and γ₀₁₀ = –3.423, respectively) and negative major life events (γ₀₁₁ = –0.909 and γ₀₁₁ = –0.977, respectively) were both associated with poor physical and mental health. Over time, suffering from a health condition related to less decline in physical health (γ₁₁₀ = 0.287), and negative events related to somewhat stronger increases in mental health (γ₁₁₁ = 0.054).

Personality effects moderated by age

To test whether the effects of the personality traits (on level or change in health) differed by age, age interaction terms were tested for all personality traits, with only statistically significant terms retained in the final models (p < .01). Overall, results revealed that the effects of personality associated with physical and mental health were unrelated to age with the exception of openness, which had a weaker association to mental health with increasing age (γ₀₁₂ = 0.018). The Johnson-Neyman technique (Johnson & Neyman, 1936; Preacher, Curran, & Bauer, 2006) was used to probe the interaction. This technique identifies the values of the moderator variable for which the focal predictor variable and outcome variable have a significant association. These probes indicated that openness and mental health were significantly related for ages less than 57 years and greater than 99 years. However, we note that the upper boundary is beyond the age range (ages 20–93) of the sample. Thus we qualify the generalization to note that in this data, greater levels of openness were related to worse levels of mental health until age 57 and unrelated after. Figure 2 shows the pattern of age-related moderation of 10-year changes in mental health for individuals with high levels of openness (+1 SD) and low levels of openness (–1 SD). A vertical line extends from age 57 to indicate the age cut-off for a significant openness-mental health association (e.g. before age 57). The older people were, the less difference openness appeared to make.

Figure 2.

The pattern of age-related moderation of 10-year changes in mental health for individuals high (+1 SD) and low (-1SD) in openness. Younger adults with high levels of openness report worse mental health compared to younger adults with low levels of openness, but the difference becomes less pronounced with increasing age. In particular, the vertical line represents that the openness-mental health association becomes non-significant at age 57 and over.

Discussion

We applied growth models to 10 waves of data from the HILDA study to describe age-related differences in self-rated physical and mental health trajectories over time, to investigate whether and how personality traits were associated with between-person differences in perceived health trajectories, and to test if these associations differ with age. Our findings suggest that physical health declines throughout adulthood and exhibits progressively steeper decrements through old age. In contrast, mental health appears to remain steady across adult life. We also found that higher neuroticism and lower extraversion and conscientiousness were each related to less favourable health trajectories. Surprisingly, participants high on openness reported lower mental health. Our results also revealed that the health implications of the personality correlates did not differ by age for physical health and were weaker for mental health in later life (e.g. openness). Methodologically, our modeling strategy allowed us to relax the often-made convergence assumption and provided clear evidence for selection effects. In particular, increasing intercepts for mental health with age indicated that younger participants (e.g. 50-year-olds) will not turn into older participants (e.g. 80-year-olds). With this accommodation, we discuss implications of our results for use of accelerated longitudinal designs and consider avenues for future more mechanism-oriented research to better understand the pathways by which these perceived health disparities emerge.

Age-related differences in physical and mental health trajectories

Our findings of steeper declines in self-rated physical health in older ages corroborate previous reports of normative physical health trajectories across adulthood (Aldwin et al., 2006; McCullough & Laurenceau, 2004; Rook et al., 2007; Sargent-Cox et al., 2010; Smith et al., 2002). Strains associated with increased risks for functional limitations and early precursors or the onset of (chronic) diseases experienced in midlife make it more and more difficult for people to perceive their physical health as stable. Compared to physical health, we found that mental health remained relatively unchanging throughout adulthood. This finding is in line with conceptual work and a myriad of empirical reports illustrating that the self-regulation system remains highly efficient in helping people adapt to a variety of (changes in) life circumstances (Baltes & Smith, 2003; Carstensen et al., 2000; Charles et al., 2001; Diener & Suh, 1998; Mroczek & Kolarz, 1998; Zarit, 2010). Moreover, some work has suggested that the stability of mental health may be due to its strong association with mostly static personality traits (e.g. Steel, Schmidt, & Shultz, 2008). Precipitous declines in mental health may only emerge at the very end of life when regulatory and/or other supporting structures break down (Gerstorf, Ram et al., 2010; Mroczek & Spiro, 2007). On the other hand, some studies have reported an increase in well-being throughout late life (Stone, Schwartz, Broderick, & Deaton, 2011). However, such results come from primarily cross-sectional studies which cannot address changes but rather only level differences. The difference in levels is likely due to selection effects with older adults who participate in the study having better mental health than those who do not participate.

Our findings corroborate previous studies examining longitudinal changes in the SF-36 that reported declines in physical and mental health with increasing age (e.g. Chandola, Ferie, Sacker, & Marmot, 2007; Hopman et al., 2006; Sacker, Head, Gimneo, & Bartley, 2009). For example, Chandola and colleagues (2007) used a sample of men and women aged 35 to 55 at baseline from the Whitehall II Study to examine long-term changes in the physical and mental health summary scores up to 18 years and found negative linear changes over age for both health dimensions. Our study adds to these earlier reports by placing their results in an adult lifespan perspective using data from a large and nationally representative sample.

In addition, our results illustrate that the assumption of age convergence in accelerated longitudinal designs may not always be accurate. Here, when the current 20- to 30-year-old participants reach old age, they will not necessarily develop into the 70- to 80-year-old participants in our study. For example, as depicted in the bottom panel of Figure 1, mental health appears to decline more in older age groups, yet older adults have greater levels of mental health at the start of the study than younger adults. If mental health takes a downturn in adulthood at some point, then it would be impossible for the younger adults in the study to be on the same trajectory in late life as the older adults in the study (i.e., the younger adults could not have greater levels of mental health in late life). Among the reasons contributing to non-convergence are birth cohort differences and selection effects (see Sliwinski, Hoffman, & Hofer, 2010). For the mental health example presented, it is plausible that the older adults who participated in the study are healthier and happier than those who did not participate. Simply by surviving to the last stage of life, these people may be a more select group of individuals than the younger people in the study.

Personality correlates of changes in physical and mental health

Our results also revealed that personality factors moderate changes in perceived health across adulthood. All of the Big Five personality dimensions were associated with between-person differences in age-related perceived health trajectories. Consistent with many reports (Bolger & Schilling, 1991; Costa & McCrae, 1987; Lahey, 2009; Neeleman et al., 2004), people higher in neuroticism tended to report poorer physical and mental health. These differences in health may reflect the effects of poor coping strategies in anticipation of or reaction to stressors, somatic complaints, and emotional instability that accompany neuroticism (Mroczek & Almeida, 2004; Watson & Hubbard, 1996). For example, when faced with a challenge, people high in neuroticism may use less effective coping strategies (e.g. rumination), which in turn results in poorer health outcomes relative to people low in neuroticism. In addition, it has been suggested that high levels of neuroticism are linked to reporting more health problems, and such a bias may contribute to the neuroticism-health association found in our study (e.g. Watson & Pennebaker, 1989). Participants higher in conscientiousness and extraversion each reported better physical and mental health, and more extraverted individuals perceived shallower physical health declines. Conscientious individuals tend to engage in health-promoting behaviours such as exercise and less in health-harming behaviours such as smoking (Boggs & Roberts, 2004), whereas the health implications of extraversion appear to be primarily mediated by social activities (e.g. utilizing resources from one’s social network in times of need; Wilson et al., 2005).

We also found that openness was associated with poorer mental health in young adulthood but not in old age. We note that this result is somewhat at odds with previously published work suggesting that openness relates to better mental health (e.g. Duberstein et al., 2003; Jerram & Coleman, 1999; for a null finding, see Turiano et al., 2012). We can only speculate about possible reasons underlying these discrepant findings. For example, we used an adjective checklist to measure openness, whereas other studies have used 60+ item scales such as the NEO-FFI. It is possible that the seven adjectives used in our study thus do not capture all the openness aspects the scale measures and potentially all the aspects of the other personality traits. Substantively, this finding may reflect a biased health perception in that individuals who live an active and engaged life may perceive a given health loss more negatively and be more detrimentally affected than people who live a less active life, especially in young adulthood when being active is more the norm. Further, openness has been discussed as a key predictor in theoretical accounts of personality growth, which, for example, includes the process of developing wisdom and individualized morals that transcend social norms (Staudinger & Kunzmann, 2005). Being more open to experience is related to moving forward on the difficult path of personality growth, and this journey is often filled with challenges and stress. To achieve personal growth and self-fulfilment, individuals must experience difficult situations which they cope with and ultimately learn from. Thus, being more open can expose individuals to stressful situations (though the outcome may be good), which in turn can relate to poorer levels of mental health. Further studies are needed to replicate and explore the underlying mechanisms of this openness–mental health association. Taken together, our longitudinal study adds to earlier cross-sectional reports of personality–health associations (e.g., using the SF-36: Duberstein et al., 2003; Jerram & Coleman, 1999; Lockenhoff et al., 2008) by demonstrating that a number of personality characteristics are important factors in shaping how perceptions of one’s health change across adult life.

Our results also revealed that the health implications of personality did not differ by age for physical health, whereas some personality effects for mental health were less pronounced in older ages (e.g. openness). Importantly, many of the examined personality correlates’ relationships with health were not moderated by age suggesting that the relationships do not differ based on age. Despite this, future work should further explore interactions of personality and other variables with age because doing so can provide information about which characteristics of individuals to target for health interventions. The characteristics of individuals that have greater health implications with age should be targeted to ideally lessen the health problems that will become exacerbated with increasing age. For example, in our study openness was found to have negative health implications in young adulthood but not in old age. Thus, targeting openness for mental health interventions may not be very worthwhile because of this tendency to improve with age regardless.

Limitations and outlook

We note several limitations of our study. First, as with most large-sample studies the specific measures available have some shortcomings. For example, health is a multidimensional construct including objective indicators such as medical diagnoses and subjective indices such as self-ratings. With only moderate overlap among these indices (Steinhagen-Thiessen & Borchelt, 1999), our findings based on changes in self-ratings may not generalize to more objective markers of health. Specifically, we note that the SF-36 may not be an ideal measure of mental and physical health because it is geared towards primarily assessing negative health aspects rather than positive ones such as well-being (e.g. emotional well-being, life satisfaction). In a similar vein, personality was measured only once in the HILDA, thereby not allowing us to examine how changes in personality and more broadly, psychological factors, are linked to changes in health (Infurna, Ram, & Gerstorf, in press; Mroczek & Spiro, 2007; Turiano et al., 2012). Following the same “trait” assumption underlying the single-occasion measurement of personality, we used a single-occasion measure of negative life events, a decision that precluded examining how fluctuations in experience of negative life events may affect health trajectories. In some cases, the response scale was also limited. For example, the measure of disability was dichotomous, and thus did not distinguish severity of disability or allow for examining how severity was related to health. Second, although our study describes how personality and age are related to adults’ health trajectories, we have not drilled into or explored the specific mechanisms through which these factors influence health. Further studies addressing such pathways are warranted, including examination of how health behaviours, physiological arousal, or social support mediate the effects noted here. Third, the relatively low number (only 10) and frequency (only yearly) of within-person assessments did not permit modeling of discrete shifts and change points in individual-level health trajectories. More closely spaced assessments would allow for a wider variety of change models (non-linear trajectories with many changes in direction), potential identification of individual-level timing of onset of health declines (e.g. second analysis in Gerstorf et al., 2008), and better opportunity to examine the interplay between physical and mental health, to identify time-varying predictors and to examine time-varying effects. Fourth, the data used here, obtained from a nationally representative sample of the population, certainly improves upon the predominantly cross-sectional studies on this topic, but an ideal study would follow individuals over a much longer time period and thoroughly track individual-level development across large parts of adult life. Finally, although the effects were small (R² < .005), differences were found between participants providing more data (e.g. 6+ waves) compared to participants who provided less data (e.g. 5 or less waves). As a consequence, selective attrition of people in worse health needs to be taken into account when interpreting our findings.

To conclude, our study represents an effort to better understand the vastly different long-term change trajectories of self-rated physical and mental health across adulthood and old age. Data from a national sample in a highly developed nation indicate that typical declines in perceived physical health are apparent throughout adulthood, whereas typical mental health remains relatively stable across adulthood. By more precisely understanding the health changes occurring in adulthood, we can better know when to intervene, and better inform public policy and programs as to when and how resources would be most needed to improve the health and well-being of individuals. We have also shown that personality factors may serve as protective factors against or risk factors for health declines. Of particular note is that some of these associations differ with age and this helps us to know which characteristics of individuals to target for health interventions (e.g. the ones that have greater health implications with age). Our study thus provides further impetus for research targeting the underlying pathways and processes involved in the emergence of health disparities during adulthood.

Footnotes

Funding

The first author was supported by a fellowship from the ISSBD (International Society for the Study of Behavioural Development-Jacobs Foundation Mentored Fellowship for Early Career Scholars). The authors are grateful for the support provided by the National Institute on Aging (Grants RC1 AG035645, R21-AG032379 and R21-AG033109). The content is solely the responsibility of the authors and does not necessarily represent the official views of the funding agencies.

References

Adler

N. E.

Boyce

Chesney

M. A.

Cohen

Folkman

Kahn

R. L.

Syme

S. L.

(1994). Socioeconomic status and health: The challenge of the gradient. American Psychologist, 49, 15–24.

Aldwin

C. M.

Spiro

A. III

Levenson

M. R.

Bosse

(1989). Longitudinal findings from the normative aging study I: Does mental health change with age? Psychology and Aging, 4, 295–306.

Aldwin

C. M.

Spiro

A. III

Levenson

M. R.

Cupertino

A. P.

(2001). Longitudinal findings from the normative aging study III: Personality, individual health trajectories, and mortality. Psychology and Aging, 16, 450–465.

Aldwin

C. M.

Spiro

A. III

Park

C. L.

(2006). Health, behavior, and optimal aging: A life span developmental perspective. In Birren

Schaie

K. W.

(Eds.), Handbook of the psychology of aging, 6th edn (pp. 85–104). San Diego: Academic Press.

Baltes

P. B.

Baltes

M. M.

(1990). Psychological perspectives on successful aging: The model of selective optimization with compensation. In Baltes

P. B.

Baltes

M. M.

(Eds.), Successful aging: Perspectives from the behavioral sciences (pp. 1–34). New York: Cambridge University Press.

Baltes

P. B.

Smith

(2003). New frontiers in the future of aging: From successful aging of the young old to the dilemmas of the fourth age. Gerontology, 49, 123–135.

Berkman

L. F.

Glass

Brisette

Seeman

T. E.

(2000). From social integration to health: Durkheim in the new millennium. Social Science and Medicine, 51, 843–857.

Boggs

T. W.

Roberts

B. W.

(2004). Conscientiousness and health-related behaviors: A meta-analysis of the leading behavioral contributors to mortality. Psychological Bulletin, 130, 887–919.

Bolger

Schilling

E. A.

(1991). Personality and problems of everyday life: The role of neuroticism in exposure and reactivity to daily stressors. Journal of Personality, 59, 356–386.

10.

Booth-Kewley

Vickers

R. R.

(1994). Associations between major domains of personality and health behavior. Journal of Personality, 62, 282–298.

11.

Butterworth

Crosier

(2004). The validity of the SF-36 in an Australian national household survey: Demonstrating the applicability of the Household, Income, & Labour Dynamics in Australia (HILDA) Survey to examination of health inequalities. BMC Public Health, 4, 44–55.

12.

Carstensen

L. L.

Pasupathi

Mayr

Nesselroade

J. R.

(2000). Emotional experiences in everyday life across the adult life span. Journal of Personality and Social Psychology, 79, 644–655.

13.

Case

Paxson

(2005). Sex differences in morbidity and mortality. Demography, 42, 189–214.

14.

Chandola

Ferrie

Sacker

Marmot

(2007). Social inequalities in self reported health in early old age: Follow-up of prospective cohort study. British Medical Journal, 334, 990–993.

15.

Charles

S. T.

Gatz

Kato

Pedersen

N. L.

(2008). Physical health 25 years later: The predictive ability of neuroticism. Health Psychology, 27, 369–378.

16.

Charles

S. T.

Reynolds

C. A.

Gatz

(2001). Age-related differences and change in positive and negative affect over 23 years. Journal of Personality and Social Psychology, 80, 136–151.

17.

Cohen

(2004). Social relationships and health. American Psychologist, 59, 676–684.

18.

Cohen

Tyrrell

D. A. J.

Smith

A. P.

(1993). Negative life events, perceived stress, negative affect and susceptibility to the common cold. Journal of Personality and Social Psychology, 64, 131–140.

19.

Costa

P. T.

McCrae

R. R.

(1987). Neuroticism, somatic complaints, and disease: Is the bark worse than the bite? Journal of Personality, 55, 299–316.

20.

Crimmins

E. M.

Kim

J. K.

Sole-Auro

(2010). Gender differences in health: Results from SHARE, ELSA, and HRS. The European Journal of Public Health, 20, 1–11.

21.

Diener

Suh

(1998). Age and subjective well-being: An international analysis. Annual Review of Gerontology and Geriatrics, 17, 304–324.

22.

Duberstein

P. R.

Sörensen

Lyness

J. M.

King

D. A.

Conwell

Seidlitz

Caine

E. D.

(2003). Personality is associated with perceived health and functional status in older primary care patients. Psychology and Aging, 18, 25–37.

23.

Dyrenforth

P. S.

Kashy

D. A.

Donnellan

M. B.

Lucas

R. E.

(2010). Predicting relationship and life satisfaction from personality in nationally representative samples from three countries: The relative importance of actor, partner, and similarity effects. Journal of Personality and Social Psychology, 99, 690–702.

24.

Friedman

H. S.

Kern

M. L.

Hampson

S. E.

Duckworth

A. L.

(2012). A new life-span approach to conscientiousness and health: Combining the pieces of the causal puzzle. Developmental Psychology. Advance online publication

25.

Gatz

Zarit

S. H.

(1999). A good old age: Paradox or possibility. In Bengtson

V. L.

Ruth

J. E.

Schaie

K. W.

(Eds.). Theories of gerontology (pp. 396–416). New York: Springer.

26.

Gerstorf

Ram

Estabrook

Schupp

Wagner

G. G.

Lindenberger

(2008). Life satisfaction shows terminal decline in old age: Longitudinal evidence from the German Socio-Economic Panel Study (SOEP). Developmental Psychology, 44, 1148–1159.

27.

Gerstorf

Ram

Mayraz

Hidajat

Lindenberger

Wagner

G. G.

Schupp

(2010). Late-life decline in well-being across adulthood in Germany, the United Kingdom, and the United States: Something is seriously wrong at the end of life. Psychology and Aging, 25, 477–485.

28.

Hampson

S. E.

Friedman

H. S.

(2008). Personality and health: A life span perspective. In John

O. P.

Robins

Pervin

(Eds.), The handbook of personality (pp. 770–794). New York: Guilford Press.

29.

Hemingway

Stafford

Stansfeld

S. A.

Shipley

Marmot

(1997). Is the SF-36 a valid measure of change in population health? Results from the Whitehall II Study. British Medical Journal, 315, 1273–1279.

30.

Holmes

T. H.

Rahe

R. H.

(1967). The social readjustment rating scale. Journal of Psychosomatic Research, 11, 213–218.

31.

Hopman

W. M.

Berger

Joseph

Towheed

VandenKerkhof

Anastassiades

Papadimitropoulos

E. A.

(2006). The natural progression of health-related quality of life: Results of a five-year prospective study of SF-36 scores in a normative population. Quality of Life Research, 15, 527–536.

32.

Idler

E. L.

Benyamini

(1997). Self-rated health and mortality: A review of twenty-seven community studies. Journal of Health and Social Behavior, 38, 21–37.

33.

Infurna

F. J.

Ram

Gerstorf

(in press). Level and change in perceived control predict 19-year mortality: Findings from the Americans’ Changing Lives Study. Developmental Psychology.

34.

Jerram

K. L.

Coleman

P. G.

(1999). The big five personality traits and reporting of health problems and health behaviour in old age. British Journal of Health Psychology, 4, 181–192.

35.

Johnson

P. O.

Neyman

(1936). Tests of certain linear hypotheses and their applications to some educational problems. Statistical Research Memoirs, 1, 57–93.

36.

Keyes

C. L. M.

(2005). Mental illness and/or mental health? Investigating axioms of the complete state model of health. Journal of Consulting and Clinical Psychology, 73, 539–548.

37.

Lahey

B. B.

(2009). Public health significance of neuroticism. American Psychologist, 64, 241–256.

38.

Lawton

M. P.

Lawrence

R. H.

(1994). Assessing health. In Lawton

M. P.

Teresi

J. A.

(Eds.), Annual review of gerontology and geriatrics (pp. 23–56). New York: Springer.

39.

Littell

R. C.

Milliken

G. A.

Stroup

W. W.

Wolfinger

R. D.

Schabenberger

(2006). SAS for mixed models (2nd edn). Cary, NC: SAS Institute.

40.

Little

R. J. A.

Rubin

D. B.

(1987). Statistical analysis with missing data. New York: John Wiley.

41.

Lockenhoff

C. E.

Sutin

A. R.

Ferrucci

Costa

P. T.

(2008). Personality traits and subjective health in the later years: The association between NEO-PI-R and the SF-36 in advanced age is influenced by health status. Journal of Research in Personality, 42, 1334–1346.

42.

Lucas

R. E.

(2007). Long-term disability is associated with lasting changes in subjective well being: Evidence from two nationally representative longitudinal studies. Journal of Personality and Social Psychology, 92, 717–730.

43.

Lucas

R. E.

Donnellan

M. B.

(2009). Age differences in personality: Evidence from a nationally representative Australian sample. Developmental Psychology, 45, 1353–1363.

44.

McArdle

J. J.

Nesselroade

J. R.

(2003). Growth curve analysis in contemporary psychological research. In Schinka

Velicer

(Eds.), Comprehensive handbook of psychology, Volume II: Research methods in psychology. New York: Pergamon Press.

45.

McCullough

M. E.

Laurenceau

J. P.

(2004). Gender and the natural history of self-rated health: A 59-year longitudinal study. Health Psychology, 23, 651–655.

46.

Mishra

G. D.

Ball

Dobson

A. J.

Byles

J. E.

(2004). Do socioeconomic gradients in women’s health widen over time and with age? Social Science & Medicine, 58, 1585–1595.

47.

Moen

(1996). Gender, age, and the life course. In Binstock

R. H.

George

L. K.

(Eds.), Handbook of aging and the social sciences, 4th edn (pp. 230–238). New York: Academic Press.

48.

Mroczek

D. K.

Almeida

D. M.

(2004). The effects of daily stress, age, and personality on daily negative affect. Journal of Personality, 72, 354–378.

49.

Mroczek

D. K.

Kolarz

C. M.

(1998). The effect of age on positive and negative affect: A developmental perspective on happiness. Journal of Personality and Social Psychology, 75, 1333–1349.

50.

Mroczek

D. K.

Spiro

(2007). Personality changes influences mortality in older men. Psychological Science, 18, 371–376.

51.

Neeleman

Bijl

Ormel

(2004). Neuroticism, a central link between somatic and psychiatric morbidity: A path analysis of prospective data. Psychological Medicine, 34, 521–531.

52.

Nolen-Hoeksema

(1999). Gender differences in depression. Current Directions in Psychological Science, 10, 173–176.

53.

Preacher

K. J.

Curran

P. J.

Bauer

D. J.

(2006). Computational tools for probing interactions in multiple linear regression, multilevel modeling, and latent curve analysis. Journal of Educational and Behavioral Statistics, 31, 437–448.

54.

Quinn

M. E.

Johnson

M. A.

Poon

L. W.

Martin

(1999). Psychosocial correlates of subjective health in sexagenarians, octogenarians, and centenarians. Issues in Mental Health Nursing, 20, 151–171.

55.

Ram

Grimm

(2007). Using simple and complex growth models to articulate developmental change: Matching theory to method. International Journal of Behavioral Development, 31, 303–316.

56.

Rook

K. S.

Charles

S. T.

Heckhausen

(2007). Aging and health. In Friedman

H. S.

Silver

R. C.

(Eds.), Foundations of health psychology (pp. 234–262). Oxford: Oxford University Press.

57.

Rowe

J. W.

Kahn

R. L.

(1997). Successful aging. The Gerontologist, 37, 433–440.

58.

Ryff

Singer

(1998). The contours of positive human health. Psychological Inquiry, 9, 1–28.

59.

Sacker

Head

Gimeno

Bartley

(2009). Social inequality in physical and mental health comorbidity dynamics. Psychosomatic Medicine, 71, 763–770.

60.

Sargent-Cox

K. A.

Anstey

K. J.

Luszcz

M. A.

(2010). Patterns of longitudinal change in older adults’ self-rated health: The effect of the point of reference. Health Psychology, 29, 143–152.

61.

Saucier

(1994). Mini-markers: A brief version of Goldberg’s unipolar Big-Five markers. Journal of Personality Assessment, 63, 506–516.

62.

Singer

J. D.

Willett

J. B.

(2003). Applied longitudinal data analysis: Methods for studying change and event occurrence. New York: Oxford University Press.

63.

Sliwinski

Hoffman

Hofer

S. M.

(2010). Evaluating convergence of within-person change and between-person age differences in age-heterogeneous longitudinal studies. Research in Human Development, 7, 45–60.

64.

Smith

Baltes

M. M.

(1998). The role of gender in very old age: Profiles of functioning and everyday life patterns. Psychology and aging, 13, 676–695.

65.

Smith

Borchelt

Maier

Jopp

(2002). Health and well-being in old age. Journal of Social Issues, 58, 715–732.

66.

Spiro

III . (2007). The relevance of a lifespan developmental approach to health. In Aldwin

C.M.

Park

C. L.

Spiro

III (Eds.), Handbook of health psychology and aging (pp. 75–93). New York: Guilford Press.

67.

Staudinger

U. M.

Kunzmann

(2005). Positive adult personality development: Adjustment and/or growth? European Psychologist, 10, 320–329.

68.

Steel

Schmidt

Shultz

(2008). Refining the relationship between personality and subjective well-being. Psychological Bulletin, 134, 138–161.

69.

Steinhagen-Thiessen

Borchelt

(1999). Morbidity, medication, and functional limitations in very old age. In Baltes

P. B.

Mayer

K. U.

(Eds.), The Berlin Aging Study: Aging from 70 to 100 (pp. 131–166). New York: Cambridge University Press.

70.

Stone

A. A.

Schwartz

J. E.

Broderick

J. E.

Deaton

(2011). A snapshot of the age distribution of psychological well-being in the United States. Proceedings of the National Academy of Sciences, 107, 9885–9990.

71.

Suzman

R. M.

Willis

D. P.

Manton

K. G.

(1992). The oldest old. Oxford: Oxford University Press.

72.

Turiano

N. A.

Pitzer

Armour

Karlamangla

Ryff

C. D.

Mroczek

D. K.

(2012). Personality trait level and change as predictors of health outcomes: Findings from a national study of Americans (MIDUS). Journals of Gerontology Series B: Psychological & Social Sciences, 67, 4–12.

73.

Ware

J. E.

Kosinski

Bayliss

M. S.

McHorney

C. A.

Rogers

W. H.

Raczek

(1995). Comparison of methods for the scoring and statistical analysis of SF-36 health profile and summary measures: Summary results from the Medical Outcomes Study. Medical Care, 33, AS264–AS279.

74.

Ware

J. E.

Kosinski

Keller

S. D.

(1994). SF-36 physical and mental health summary scales: A user’s manual. Boston, MA: The Health Institute, New England Medical Center.

75.

Ware

J. E.

Snow

K. K.

Kosinski

Gandek

(1993). SF-36 health survey manual and interpretation guide. Boston, MA: New England Medical Center.

76.

Watson

Hubbard

(1996). Adaptational style and dispositional structure: Coping in the context of the five-factor model. Journal of Personality, 64, 737–774.

77.

Watson

Pennebaker

J. W.

(1989). Health complaints, stress, and distress: Exploring the central role of negative affectivity. Psychological Review, 96, 234–254.

78.

Watson

(2010). HILDA user manual – Release 8. Melbourne: University of Melbourne.

79.

Wilson

R. S.

Krueger

K. R.

Bienias

J. L.

Mendes de Leon

C. F.

Evans

D. A.

(2005). Neuroticism, extraversion, and mortality in a defined population of older persons. Psychosomatic Medicine, 67, 841–845.

80.

Wilson

R. S.

Mendes de Leon

C. F.

Bienas

J. L.

Evans

D. A.

Bennett

D. A.

(2004). Personality and mortality in old age. Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 59, 110–116.

81.

Zarit

S. H.

(2010). A good old age: Theories of mental health and aging. In Bengston

V. L.

Gans

Putney

Silverstein

(Eds.), Handbook of theories of aging, 2nd edn (pp. 675–691). New York: Springer.

82.

Zarit

S. H.

Johansson

Berg

(1993). Functional impairment and co-disability in the oldest old: A multidimensional approach. Journal of Aging and Health, 5, 291– 305.