Are Changes in Somatic Health Reflected Differently in Updated Self-Ratings by Big-Five Personality Traits?

Abstract

Objectives: This longitudinal study tests whether the Big-Five personality traits influence the changes individuals make in self-rated health (SRH) as they adjust their initial level to account for information on concurrent changes in disease burden, activities of daily living (ADLs), and pain. Methods: A bi-variate Latent Growth Curve model was fitted to data to estimate longitudinal associations between SRH and each health measure across up-to-five repeated observations, collected from the year 2006 to 2018 from 13,096 participants in the Health and Retirement Study. Results: Negative longitudinal associations between SRH and all three health reports were significantly stronger for those who are more conscientious. No significant moderation was found for the other four personality traits. Discussion: Compared to the less conscientious, highly conscientious people may assign greater importance to specific health reports when rating and revising their assessments of SRH. This moderating effect was previously tested but not supported.

Keywords

longitudinal analysis personality physical health

Introduction

Self-rated health (SRH) is a common and inexpensive method for screening a person’s well-being in clinical and survey settings. Studies have consistently shown that worse ratings in SRH predict present and future health risks, including those for diseases, fatigue, lethargy, diminished activity, all-cause mortality, and physiological markers outside the normal range (Benyamini, 2011; Huisman & Deeg, 2010; Idler & Benyamini, 1997; Jylhä, 2009; Jylhä et al., 2006).

Even so, much of the variation in individual SRH ratings is left unexplained by self-reported health indicators (Jylhä et al., 2006; Quesnel–Vallée, 2007). One suggested reason for this heterogeneity is that people use different evaluative frameworks for rating their own health, which, in part, may depend on their personality – a unique yet understudied mechanism behind the SRH-health link (Dowd & Zajacova, 2010). Applying a conceptual model put forth by Jylhä (2009) and others (Huisman & Deeg, 2010; Kööts–Ausmees et al., 2016), the current study investigates variability in how older adults longitudinally adjust their SRH as their health condition changes. Using respondents’ repeated reports of three distinct health indicators—functional limitations, disease burden, and pain—the Big-Five personality traits of openness, conscientiousness, extraversion, agreeableness, and neuroticism, are assessed as a possible mechanism influencing respondents’ adjustments of SRH.

To achieve this goal, the study draws on data collected from the Health and Retirement Study (HRS) in the United States, which biennially measures the health reports and SRH for participants aged 50 or above, beginning in 2006 when the Big-Five personality traits were first collected. The study employs Latent Growth Curve Models to simultaneously analyze the cross-sectional and longitudinal associations between each health report and SRH from 2006 to 2018.

SRH and Evaluative Process of Health

In US surveys, self-rated health (SRH) often is measured on a five-point scale of “excellent,” “very good,” “good,” “fair,” and “poor,” and typically read to respondents in that order (Hirdes & Forbes, 1993; Stenholm et al., 2014). According to Jylhä’s model (2009), a person’s self-assessment of health involves three evaluative stages: (1) defining what constitutes health, (2) comparing health to that of others with similar characteristics such as age, and (3) selecting a response category thought to best describe an overall sense of wellness by using norms, ideals, and internalized expectations. Prior research has examined these evaluative stages by analyzing the association between an extensive range of health reports and SRH, as well as demonstrating how those correlates may vary by circumstances, age, birth cohort, and cross-cultural/societal contexts (Dowd & Zajacova, 2010; Henning et al., 2021; Spuling et al., 2015).

Individuals with more pre-existing health issues (e.g. greater pain, dealing with more diseases, struggling with functional limitations) tend to report worse ratings of SRH than those with fewer issues, and they also are likely to downgrade reported SRH after new or additional health challenges occur. This negative correlation between changes in SRH and a variety of health reports was recently confirmed longitudinally, but further examination of these complex relationships also has been facilitated by increased availability of longitudinal data sets that allow researchers to trace changes in multiple measures over time (Kööts–Ausmees et al., 2016; Turiano et al., 2012).

The three health reports used in the current study—pain, functional limitations, and disease burden—are the three strongest correlates of SRH and are health indicators that can be known through assessments and experienced by individuals through related symptoms (Lazarevič & Brandt, 2020). Across many European countries, the strength of the relationship between SRH and functional health has been found to increase in panel data over time, while the association between SRH and disease burden weakens over time (Lazarevič & Brandt, 2020). These findings raise the possibility that certain individual-level factors consistently influence how SRH reports incorporate health reports over time. The Big-Five personality traits have been suggested as one such idiosyncrasy (Henning et al., 2021).

Roles of Big-Five Personality Traits in SRH

Various studies consistently report significant associations among the Big-Five Personality traits and disease burden, functional limitations, and pain as well as SRH (Löckenhoff et al., 2012; Turiano et al., 2012). “Health effects” associated with each personality trait typically refer to significant differences in those health outcomes for individuals showing stronger/weaker tendencies toward each trait. For instance, those who exhibit a weaker, compared to stronger, tendency toward neuroticism have better somatic health and report more positive SRH ratings (Löckenhoff et al., 2012; Stephan et al., 2020). Being highly conscientious as opposed to being less has similar positive effects on both physical health and SRH (Löckenhoff et al., 2012). The accumulated findings are less consistent in health effects associated with a stronger tendency toward extraversion, openness, and agreeableness. Nonetheless, they tend to suggest that those traits have an overall positive impact on health (Kööts–Ausmees et al., 2016; Turiano et al., 2012; Williams et al., 2004).

For each personality trait, different psychological and behavioral mechanisms have been suggested to explain associated health effects. Being less neurotic (having better emotional control) as well as being more conscientious, extraverted, open, and agreeable are said to independently contribute to healthy coping mechanisms of stress and the development of behavioral patterns that are beneficial for health in the short and long run (Löckenhoff et al., 2012; Turiano et al., 2012). Open and extravert tendencies may be helpful seeking out new information to better prevent or address health issues while conscientious and agreeable tendencies can facilitate the internalization of new information and the adoption of new approaches and behavioral adjustments such as diet and exercise (Kööts–Ausmees et al., 2016; Roberts et al., 2007). Maintaining emotional control (i.e. lower levels of neuroticism) is also crucial, as negative emotions can significantly disrupt efforts toward wellness (Williams et al., 2004).

Integration of Personality Traits into the Evaluative Model of SRH

While the accumulated evidence is converging on how each personality tendency may harm or benefit overall health, the specific roles of each trait in connecting somatic health symptoms to SRH remain sharply disputed. The current study specifies those roles as moderators in how closely individuals align their SRH ratings with known health issues. By testing whether this alignment varies depending on people’s stronger/weaker tendencies toward each trait, our findings provide new insight into personality-based tendencies in the evaluative process of SRH. Prior research scarcely speaks to such moderation roles for personality traits, but a series of psychological studies indicate some personality-based tendencies/dispositions that are likely reflected in one’s internal-assessment of well-being.

In self-reports, those with stronger neurotic tendencies are said to exaggerate their perceived pain and confuse it with co-occurring clinical needs (Aiken-Morgan et al., 2014; Denovan et al., 2019; Feldman et al., 1999). This phenomenon, often referred to as “health neuroticism” or “illness anxiety” (Kööts–Ausmees et al., 2016), can operate in two conflicting ways whereby such neurotic biases affect the alignment between SRH and somatic health reports. Some see pain conflation as suggestive of a closer alignment because those with neurotic tendencies appear to rate their SRH more negatively for a given health issue (Henning et al., 2021). Others argue the contrary position—that neurotic tendencies can weaken the alignment between SRH and specific physical health reports by negatively skewing SRH ratings relative to their perceptions, ideals, and beliefs about health rather than their actual health (Aiken-Morgan et al., 2014; Feldman et al., 1999). A greater bias in SRH for those with stronger neurotic tendencies may indeed weaken the alignment, which would be indicated by the larger amount of individual-level variability in SRH among those with stronger neurotic tendencies when self-reports of one’s own health are held constant (Kööts–Ausmees et al., 2016).

Moderation by conscientiousness also has been hypothesized through two conflicting pathways. On the one hand, a stronger tendency toward conscientiousness implies a propensity toward more premeditative, adaptive, and resourceful behaviors that can promote wellness and help protect against pain, functional limitations, diseases, and mortality (Kööts–Ausmees et al., 2016). But these same dispositions, which are associated with better health, may also facilitate greater psychological resilience and perseverance despite evidence of health problems. This resilience associated with high conscientiousness might lead in turn to more positive SRH reports than a less conscientious person would offer given the same health problems. This speculation, however, lacks empirical support from prior research. Although the association between higher conscientiousness and better SRH was supported, conscientiousness does not moderate the relationship between health conditions and either the level or change n SRH. In other words, the alignment between SRH and somatic health symptoms was not weaker at higher levels of conscientiousness (Henning et al., 2021). Alternatively, being more conscientious may counter bias such that sequential assessments of SRH more accurately reflect the physical changes that occur. In this view, a smaller bias in SRH associated with greater conscientiousness would translate into a stronger alignment between SRH and health reports. In fact, prior research suggested that those with conscientious tendencies are able to effectively use heath information to address clinical needs, maintain more robust health, and engage in more health-conscious behaviors, which could be due to their having a more accurate perception of health and well-being (Bogg & Roberts, 2013).

The theoretical view that conscientiousness may strengthen the internal alignment between SRH and physical health can be extended to other personality traits that are positively related to health, such as being more extraverted, open, and agreeable (Turiano et al., 2012). For example, behavioral or psychological characteristics associated with health-positive traits, such as greater social connectedness and adaptability to new information, may help explain why some people’s sense of well-being is more closely aligned with known health indicators (Gomez et al., 2002; Harkins et al., 1989; Rafienia et al., 2008; Sočan & Bucik, 1998; Wade et al., 1992).

Current Study

We identified a single study that empirically assesses how the Big-Five traits influence the relationship between self-rated health (SRH) and physical health. This study used Swedish data to estimate correlations between changes in SRH and somatic health and compared these correlations across different levels of the Big-Five personality traits (Henning et al., 2021). The current study also uses this moderation framework to investigate the roles of the Big-Five personality traits in either strengthening or weakening the association between SRH and physical health reports. High levels of conscientiousness may be associated with a more significant decline in SRH than that reported by the less conscientious when facing similar health issues. This pattern of moderation suggests a strengthening of the link for the more conscientious. Similarly, traits like extraversion, agreeableness, and openness, which also are positively related to health, are expected to adjust their reports of SRH to mirror diagnoses of diseases and conditions more closely, thereby strengthening the link. In contrast, neurotic tendencies may introduce a different type of bias in SRH ratings that could exaggerate diagnosed conditions, thereby weakening the association.

To test the moderation roles of the Big-Five personality traits, the current study uses US-based data collected from older adults aged between 50 and 85+ across an 8-year period. The HRS data provide greater within-(longer period) and between-(wider age range) variability in both SRH and somatic health changes, thus allowing more statistical power for the analyses of changes over time. Furthermore, this study extends its analysis to all Big-Five traits rather than focusing only on conscientiousness and neuroticism, as all five traits are likely intercorrelated with one another and thus could co-influence SRH and the health reports. In parallel to the previous analysis, this study separately examines the longitudinal associations between SRH and each health report in two ways: (a) baseline health report predicting subsequent change in SRH, (b) subsequent change in health report correlated with subsequent change in SRH. For the personality traits found significantly moderating the associations (a) and/or (b), an extended model was constructed to simultaneously estimate moderation in the associations (a) and (b), with the aim of adjusting for possible regression toward the mean and ceiling/bottoming-out longitudinal effects that both health reports and the personality traits may have on SRH (Henning et al., 2021).

We propose the following hypotheses based on the theoretical perspective that positive health traits such as conscientiousness contribute to a closer alignment between health reports and SRH, while negative health traits such as neuroticism can interfere with this alignment:

With higher levels in conscientiousness, extraversion, openness, and agreeableness, negative longitudinal associations between SRH and each health report strengthen.

With higher levels neuroticism, negative longitudinal associations between SRH and each health report weaken.

Materials and Methods

Data and Sample

To test these hypotheses, this study uses HRS data from 2006 to 2018. Begun in 1992, the HRS is one of the longest running longitudinal studies in the U.S. using a nationally representative sample of people aged 50 and older, who were living in non-institutional settings and interviewed biennially. Starting in 2006, a subpanel of HRS participants was randomly selected to participate in a psychosocial and lifestyle questionnaire that includes a battery of psychological items including those for the Big-Five personality traits (Smith et al., 2017). Of those panel participants, our analysis included 16,045 persons who provided responses at a minimum of two data points and did not get dropped from the HRS later due to death (as of 2023), to SRH and the health report measures calibrated to be consistent across 1992–2018 HRS years (as of 2023) by the RAND Corporation and have complete information on covariates listed below.

Between the years 2006 and 2018, seven HRS surveys (2006, 2008, 2010, 2012, 2014, 2016, and 2018) were conducted. However, to maintain consistency with the previous study and prevent potential estimate biases due to high attrition rates (Enders & Bandalos, 2001), this study used five observations instead of seven. In the sixth and seventh observations, the attrition rates exceeded 60%. The retention rates for the second through fifth observations were 98%, 89%, 82%, and 68%, respectively, which were comparable to the Swedish data used in the previous study (Henning et al., 2021). To address the missing data, this study employed the Full-Information Maximum Likelihood method in Mplus version 8.4 (Mazza et al., 2015), which retained the participants with missing information in the analyses without using imputation techniques.

To determine the selectivity of this study sample, a statistical comparison in the descriptive statistics was conducted between the study sample (Table 1) and an alternative sample of 22,116 HRS respondents (Appendix A) that includes those who did not meet the inclusion criteria or were dropped from the HRS due to death at some point. In short, demographic and personality differences between the two samples are statistically insignificant or marginal. In terms of SRH and the health reports, overall health profiles of the alternative sample were more negative than the study sample (e.g. worse SRH, more ADL issues). The gap between the two samples decreased over time and became statistically insignificant by the 5th observation.

Table 1.

The Weighted Mean, Standard Error, and 95% CI for all Variable by Demographic Characteristics and Time for the Study Sample (n = 16,045).

	Demographic characteristics						Health report by t observation
	Mean	SE	95% CI				Mean	SE	95% CI		N
N for demographic variables = 16045					Pain	t1	.620	.010	.600	.641	16,045
Baseline age	61.191	.075	61.044	61.339		t2	.616	.010	.595	.636	15,803
Male	.491	.006	.480	.502		t3	.660	.011	.638	.682	14,237
Female	.509	.006	.498	.520		t4	.669	.012	.646	.692	13,230
Non-HispanicWhite	.755	.004	.746	.764		t5	.721	.013	.695	.746	11,138
Non-HispanicBlack	.108	.003	.102	.113	ADLs	t1	.158	.006	.146	.170	16,045
Hispanic	.094	.003	.088	.100		t2	.169	.006	.156	.182	15,803
Others	.043	.003	.038	.048		t3	.195	.007	.182	.209	14,237
Less than college	.630	.006	.619	.641		t4	.200	.007	.185	.214	13,230
Some college +	.370	.006	.359	.381		t5	.250	.009	.233	.267	11,138
Kept working	.442	.006	.431	.453	Disease Burden	t1	.967	.011	.947	.988	16,045
Never worked	.098	.003	.093	.104		t2	1.099	.011	1.076	1.121	15,803
Stopped work (H)	.229	.005	.220	.238		t3	1.262	.013	1.237	1.287	14,237
Stopped working	.231	.005	.222	.240		t4	1.381	.013	1.355	1.407	13,230
						t5	1.557	.015	1.527	1.587	11,138
Openness	2.998	.006	2.986	3.010	SRH	t1	3.372	.012	3.349	3.396	16,045
Extraversion	3.201	.006	3.189	3.213		t2	3.330	.012	3.307	3.353	15,803
Conscientiousness	3.079	.004	3.071	3.088		t3	3.290	.012	3.266	3.314	14,237
Neuroticism	2.373	.005	2.363	2.384		t4	3.248	.013	3.224	3.273	13,230
Agreeableness	3.507	.006	3.496	3.517		t5	3.199	.014	3.172	3.227	11,138

Note. Stopped working = having stopped working at some point after baseline for reasons other than health. Stopped working (H) = having stopped working at some point after baseline for health-related reasons including disability. This study uses 5 waves of HRS data from 2006 to 2018. Rather than referencing survey years, successive observations are sequenced from t1 through t5 in a roughly 2-year interval. In any given observation, information may or may not be available due to various missing patterns. In other analyses, the original indexes used to measure extraversion, conscientiousness, and neuroticism are converted into z-scores with the mean of 0 and variance of 1.

Rather than referencing survey years, the Tables, Figures, and subsequent text refer to observation t from 1 to 5. Within each tth observation, SRH and the three health reports were measured in the same survey year and roughly every 2 years following the t1 observation. In any given observation t for the study members, information may or may not be available due to missing data or because they were permanently dropped from the HRS (e.g., for refusal or other unobserved reasons).

Measures

SRH

A question on self-rated health is asked in every survey year. Participants are asked to use five categories in response to the question, “would you say your health is excellent, very good, good, fair, or poor?”, coded 1–5 respectively. We preserve this coding, which is consistent with remaining health indicators for which lower numbers describe better health, higher numbers describe worse health.

Pain, functional limitations, disease burden

For pain, the respondents were first asked “are you often troubled with pain?” Upon responding “yes” to the question, they were then asked, ‘‘how bad is the pain most of the time: mild, moderate, or severe?” Combining the responses to both questions, the pain index was created on a scale of 0–4 (0 = not troubled with pain, 1 = mild, 2 = moderate, and 4 = severe). Functional limitations were measured with five items from Activities of Daily Living (ADLs), each of which asks the respondents if they can complete a routine self-care task without help. The index (scaled 0–5) for functional limitations counts the number of tasks that they cannot complete independently. The five tasks are eating, bathing, dressing, getting in and out of bed, and walking across the room. Last, Disease Burden (scaled 0–7) is the number of chronic conditions (high blood pressure, diabetes, cancer, lung disease, heart disease, stroke, and arthritis) a doctor has told respondents they have had at any point prior to the interview.

Big-five personality traits

As with the control measures, the HRS conducted an extensive item analysis for each set of items designed to measure each Big-Five trait (see Smith et al., 2017). For this study, we quantified each of the five dimensions by averaging responses within each set. Subsequent item analysis used here shows that agreeableness (α = 0.79) and openness (α = 0.79) are most reliable, while conscientiousness (α = 0.70) and neuroticism (α = 0.70) have the lowest level of internal consistency (Smith et al., 2017). On a scale of one to four, the personality index created for each trait quantifies how strongly respondents are aligned with the trait. One the original scale of 1–4, higher (positive) values indicate a stronger tendency (e.g. more agreeable, more neurotic); lower (negative) scores indicate a weaker tendency.

Covariates

In parallel to the earlier work, the following covariates were included: age at baseline (t1), sex, race/ethnicity (Non-Hispanic Whites, Non-Hispanic Black, Hispanic, and Others), tertiary education (whether they had at least some college education), and retirement status. Based on available information in 1st-5th observations, individual work status was measured with the following categories: (1) “never worked,” (2) “kept working throughout,” (3) “ever stopped working for health,” and (4) “ever stopped working for other reasons.” Prior research shows SRH and health reports at baseline and their changes are significantly differentiated by these categories (Bloemen et al., 2017; Gustman & Steinmeier, 2000; Oi, 2020). Using the cross-Wave variable RwLBRF constructed by RAND (Bugliari et al., 2019, p. 1568), the respondents are flagged as “kept working” or “never worked” when they reported that they worked for pay or did not in each available observation. Those in the category, “stopped working due to health,” are identified when they reported as disabled or not working and due to any health reasons in any of the available observations. If they did not, they were identified as “stopped for other reasons” instead.

Analytic Methods

This study uses a bivariate Latent Growth Curve Model which simultaneously estimates both baseline associations among the measures (SRH, pain, ADLs, and disease burden) and changes in each measure. The model includes structural equations that regress changes in SRH on both the baseline and changes in each health measure. To achieve this, the baseline and changes of each measure are estimated as latent variables in the metric of their respective measure (e.g. SRH on the range of 1 as poor and 5 as excellent) with two subcomponents: latent baseline (hence force referred to as SRH*₁ and all three health reports interchangeably as Health*₁) and latent slope (referred to as ∆SRH* and ∆Health*). The individual slopes of each measure represent net change over the entire span between t1 and t5. The t-specific change is parameterized in the “free” form described in Appendix B, as well as the estimated covariance between the baseline and slope of each measure. This latent treatment of repeated measures, a key adjustment in longitudinal analysis of changes in those measures, accounts for the time-constant measurement error associated with each measure (McArdle, 2009).

To examine the relationships among constructs, the structural equation component of this model simultaneously estimates a separate series of associations between SRH and each health report taken in turn. For the baseline association, the initial health reports (Health*₁) predict the initial SRH*_1. In the first set of longitudinal associations, the health reports at baseline are included as predictors of ∆SRH* while in the second set, changes in the health reports are used as predictors. All covariates including the Big-Five Traits are regressed on all constructs at once.

To evaluate the hypotheses, these structural equations are extended to estimate the moderation of the relationships between Health* and SRH* by the Big-Five traits with three different specifications. The first model includes interaction terms between the Big-Five Traits and the baseline of each health measure (Health*₁) as a series of moderators to the baseline association between Health*₁ and SRH*₁, as well as to the first set of longitudinal associations between Health*₁ and ∆SRH*. The second model includes interaction terms between the Big-Five traits and the slope of each health measure (∆Health*). The third model includes both sets of the interaction terms to predict SRH*₁ and ∆SRH*, but only for the Big-Five traits found to be significantly interacting with at least one of the health measures in the first or/and second models. The hypotheses were then assessed based on moderation effects predicting ∆SRH* from the third model. The first model emulates the one used in the previous study (Henning et al., 2021, p. 738), while the second model follows another (Henning et al., 2021, tbls. S11-13). The third combines both features of the first two, extending the previous analysis.

Hypothesis H1

specifies that greater tendencies toward conscientiousness, extraversion, openness, and agreeableness strengthen the negative coefficients of Health*₁ and ∆Health* predicting ∆SRH*. The negative coefficients represent the degree of downward adjustment to SRH that a person makes after baseline, for every one unit increase in the baseline health report (Health*₁) and/or one unit increase in the health report over time since baseline (∆Health*). Negative coefficients for the interaction quantify the degree to which those average effects are more negative per one standard deviation increase in tendencies toward those traits.

Hypothesis H2

specifies that greater tendencies toward neuroticisms weakens the expected negative coefficients of Health*₁ and ∆Health* predicting ∆SRH*. In this hypothesis, positive coefficients for the interaction with neuroticism are expected, meaning that the average main effects of Health*₁ and ∆Health* are less negative per one standard deviation increase in neuroticism.

Results

Table 1 shows the weighted average and its 95% confidence interval for all variables. For repeatedly measured variables, descriptive statistics refer to observations (t1-5). The study sample members are about 62 years old on average at baseline, predominately White with 10% non-Hispanic Black, 9% Hispanic, 4% identified as none of these racial/ethnic groups (“others”). About 37% of the sample have at least some college education. Nearly 44% of the sample reported that they worked in all observations they appeared in, while 10% never worked. About 23% stopped working at some point for health or other reasons. Turning to the repeated health measures, a steady increase in the average of each measure (pain, Disease Burden, ADLs) indicates a gradual health decline. On the other hand, the average rating of SRH hovers around 3.2.

Table 2 summarizes the coefficients of the health reports, the personality traits, and the covariates predicting SRH from the bi-variate Latent Growth Curve Model without moderation effects estimated. These coefficients are parameterized on the five-point SRH scale (1 poor–5 excellent). Appendix C lists the coefficients of the personality traits predicting each health report and discusses converging findings that higher levels of conscientiousness and extraversion are associated with better somatic health (especially in terms of ADLs and Disease Burden) both cross-sectionally and longitudinally, while greater neuroticism is detrimental. Unexpectedly, openness and agreeableness are found to be negatively related to somatic health reported at baseline. Across all the health reports, the model fit statistics are above acceptable levels, .95 or larger for CFI/TFL, and smaller than .06 for RMSEA (Xia & Yang, 2019).

Table 2.

Coefficients From the Bi-variate Latent Growth Curve Model Predicting the Latent Baseline (intercept) and Change (Slope) in SRH (1 Poor−5 Excellent).

	SRH intercept			SRH slope
	B	SE	P	B	SE	P
Pain intercept	−.582	.015	<.000	.063	.021	.003
Pain slope				−.790	.059	<.000
Disease burden intercept	−.360	.009	<.001	.063	.011	<.001
Disease burden slope				−.298	.015	<.001
Activities of daily living intercept	−.597	.028	<.001	.188	.036	<.001
Activities of daily living slope				−.551	.033	<.001
Extraversion	.105	.012	<.001	−.025	.015
Conscientiousness	.095	.010	<.001	−.025	.014
Neuroticism	−.090	.009	<.001	.016	.012
Openness	.046	.011	<.001	.016	.015
Agreeableness	−.026	.012	.031	.022	.015
Baseline age	−.010	.001	<.001	−.007	.002	<.001
Female	.097	.018	<.001	.029	.024
Non-hispanic black¹	−.313	.024	<.001	.068	.032	.035
Hispanic¹	−.404	.031	<.001	.125	.041	.002
Others¹	−.190	.047	<.001	−.034	.071
Some college or more	.207	.020	<.001	−.042	.026
Never worked²	−.103	.028	<.001	−.033	.039
Stopped working (H)²	−.430	.027	<.001	.018	.036
Stopped working	.000	.022		−.015	.029

Note. p-values above 0.05 are not shown. One unit change in each personality trait refers to one z-score below/above the sample mean. The coefficient of latent health baseline predicting both the baseline/change counterpart (slope) of SRH. ¹ Non-Hispanic White as ref, ²kept working as ref, (H) = stopped working for health reason.

Extraversion, conscientiousness, and openness are positively associated, and neuroticism is negatively associated with SRH. However, none of those traits significantly affected longitudinal change in SRH. In contrast, negative correlations between the health reports and SRH were confirmed both cross-sectionally and longitudinally. The latent baseline of each health report is positively associated with the slope of SRH, likely indicating bottoming-out effects and truncated measurement scales. If respondents were experiencing multiple health problems at baseline, they also were likely to report worse SRH at baseline, which leaves less room for SRH to decline further. Alternatively, this pattern could be interpreted as a known tendency to weigh earlier health problems as more important than subsequent additional problems in calibrating repeated ratings of SRH (i.e. regression toward the mean or adaptive responses).

Table 3 presents the results of three different models examining whether personality traits moderate the cross-sectional and longitudinal associations between the health reports and SRH (under the columns titled “intercept” and “slope,” respectively). Model 1 shows some significantly negative interaction terms, indicating that for each unit increase in baseline levels of pain and disease burden, the highly conscientious downgrade SRH more steeply over time, or tend to shift their SRH toward the “poor” category to a greater degree, given the same level of pain or disease burden than the less conscientious. Moderation of the baseline association between ADL and SRH is significantly negative for conscientiousness and positive for neuroticism.

Table 3.

Coefficients From the Bi-variate Latent Growth Curve Model Predicting the Latent Baseline (intercept) and Change (Slope) in SRH (1 Poor−5 Excellent) With interaction Terms Between Big-Five Personality Factors and each Heath Measure (Pain, Disease Burden, and ADLs).

	SRH intercept			SRH slope
	B	SE	P	B	SE	P
Model 1
Pain intercept	−.580	.016	<0.001	.192	.020	<0.001
✖ Agreeable	.024	.014		−.001	.017
✖ Extravert	−.001	.015		.001	.018
✖ Open	−.028	.015		.021	.017
✖ Conscience	.005	.013		−.031	.016	0.048
✖ Neurotic	.010	.011		.012	.014
Disease burden intercept	−.355	.009	<0.001	.066	.011	<0.001
✖ Agreeable	.002	.010		−.018	.012
✖ Extravert	.000	.011		.010	.012
✖ Open	−.013	.010		.015	.012
✖ Conscience	−.015	.009		−.022	.011	0.037
✖ Neurotic	.011	.007		.015	.009
ADL intercept	−.619	.031	<0.001	.252	.031	<0.001
✖ Agreeable	.002	.024		−.001	.028
✖ Extravert	−.040	.026		.016	.029
✖ Open	−.048	.026		.021	.027
✖ Conscience	−.052	.022	<0.001	−.019	.022
✖ Neurotic	.042	.016	0.017	.012	.020
Model 2
Pain intercept				.064	.019	0.001
Pain slope				−.779	.050	<0.001
✖ Agreeable				−.053	.043
✖ Extravert				.005	.044
✖ Open				−.051	.042
✖ Conscience				.000	.045
✖ Neurotic				−.024	.037
Disease burden intercept				.063	.011	<0.001
Disease burden slope				−.299	.015	<0.001
✖ Agreeable				−.013	.018
✖ Extravert				−.006	.019
✖ Open				.003	.018
✖ Conscience				.002	.019
✖ Neurotic				−.002	.015
ADL intercept				.176	.032	<0.001
ADL slope				−.574	.030	<0.001
✖ Agreeable				.008	.026
✖ Extravert				−.033	.028
✖ Open				−.022	.026
✖ Conscience				−.057	.023	0.012
✖ Neurotic				.025	.019
Model 3
Pain
Pain intercept	−.593	.015	<.001	.102	.031	.001
Pain slope				−.752	.040	<.001
Conscience	.162	.012	<.001	.004	.016
✖ Conscience (I)	−.016	.011		−.039	.019	.049
✖ Conscience (S)				−.053	.038
Neurotic	−.108	.012	<.001	.014	.015
✖ Neurotic(I)	.007	.009		.016	.021
✖ Neurotic(S)				−.021	.035
Disease burden
DB intercept	−.367	.009	<.001	.063	.011	<.001
DB slope				−.296	.015	<.001
Conscience	.179	.014	<.001	−.011	.016
✖ Conscience (I)	−.019	.007	.008	−.018	.009	.037
✖ Conscience (S)				.004	.014
Neurotic	−.152	.013	<.001	.015	.016
✖ Neurotic(I)	.013	.007		.011	.009
✖ Neurotic(S)				.004	.014
Activities of daily living
ADLs intercept	−.662	.032	<.001	.180	.034	<.001
ADLs slope				−.567	.030	<.001
Conscience	.178	.010	<.001	−.025	.012	.027
✖ Conscience (I)	−.091	.016	<.001	.006	.018
✖ Conscience (S)				−.072	.017	<.001
Neurotic	−.151	.010	<.001	.028	.011	.012
✖ Neurotic(I)	.051	.016	.001	.007	.020
✖ Neurotic(S)				.027	.019

Note. p-values above 0.05 are not shown. Interaction coefficients refer to change in SRH per one z-score increase in each personality score for one unit change in the intercept (I)/slope (S) of each health measure. Each personality trait interacts with the intercept of each health measure in Model 1 (e.g. ×Conscience(I)), with the slope in Model 2 (e.g. ×Conscience (S)), and with both intercept and slope in Model 3.

Model 2 examines how change in each health reports after baseline is associated with change in SRH confirms expectations for conscientiousness. The negative association between changes in both SRH and ADLs is stronger for highly conscientious individuals, meaning that they adjust SRH downward to a greater degree for each additional ADL limitation that occurs. No expected patterns were observed for the other traits, and no significant moderation by neuroticism was found.

In Model 3, the moderation of the longitudinal associations (from Models (1) and (2) is re-estimated simultaneously for neuroticism and conscientiousness only, with the other nonsignificant personality traits excluded. This extended model confirms significant moderation by conscientiousness as seen in the previous models, indicating a tendency among highly conscientious individuals to rate their SRH worse for a given increase in the baseline of pain and Disease Burden and an ADL limitation over time. Model 3 also indicated that neuroticism and conscientiousness are associated with both the baseline and slope of SRH but in opposite directions, findings which were previously interpreted as regression toward the mean. For example, the highly neurotic tend to rate SRH worse at baseline but adjust their SRH upward subsequently, rather than downward. Note that these patterns were observed only in Model 3 with ADLs, which conditions the slope of ALDs to be 0.

In summary, the strengthening of the negative associations between SRH and the health reports is longitudinally confirmed for conscientious individuals. The significant moderation effect on the association between the slopes of SRH and ADLs is considered the most robust in supporting Hypothesis 1, as it is influenced less by individual-level time-constant idiosyncrasies and best conceptually captures personality-based differences in evaluative SRH adjustments in response to contemporaneous changes in somatic health. Figure 1 illustrates the longitudinal moderation effects observed in Model 3, contrasting the average change in SRH for zero change, one unit increase, and two unit increase in ADLs, with a variation of +/− 1 standard deviation from the average level of conscientiousness. The steeper line with higher levels of conscientiousness (+1 SD) means that the correlation between the slopes of SRH and ADLs is strengthened for the highly conscientious.

Figure 1.

Predicted change in SRH for zero change, one unit increase, and two unit increase in ADLs, with a variation in levels of conscientiousness. Note: Based on the estimates from Model 3 shown in Table 3. +1SD = one standard deviation above the average (considered highly conscientious). −1SD = one standard deviation below the average (considered as less conscientious). The Steeper line means a closer alignment between concurrent changes in SRH and ADLs.

Discussion

Prior research has established that personality traits such as conscientiousness and neuroticism have significant impacts on health and SRH. However, what remains unsettled is how individuals showing strong tendencies toward those traits may report SRH differently based on known health reports (Henning et al., 2021; Kööts–Ausmees et al., 2016). This study tests the theoretical plausibility drawn from the evaluative framework of SRH (Jylhä et al., 2006). According to this framework, those with positive health traits such as conscientiousness tend to evaluate their sense of wellness in a way that aligns more closely with the health reports, while those with negative health traits such as neuroticism may experience more misalignment. This view receives partial supported from some studies, but conflicting evidence also exists (Gomez et al., 2002; Harkins et al., 1989; Henning et al., 2021; Rafienia et al., 2008; Sočan & Bucik, 1998; Wade et al., 1992).

With no significant moderation effects found for extraversion, agreeableness, and openness, the focus point of the findings revolves around conscientiousness. Firstly, it confirms that conscientiousness is a positive health trait—those with highly conscientious tendencies not only rate their SRH higher, but also report lower levels of functional limitations and Disease Burden both at baseline and subsequent observations (see Appendix C). In the evaluative process translating those reports to SRH, those who are highly conscientious make greater downward adjustments to SRH based on previously reported levels of pain and Disease Burden. Further supporting the prior speculations regarding conscientiousness (Gomez et al., 2002; Harkins et al., 1989; Rafienia et al., 2008; Sočan & Bucik, 1998; Wade et al., 1992), the highly conscientious also make greater downward changes in SRH in response to an increase in ADLs over time. Taken together, robust support for H1 is found for conscientiousness across the health measures and the model specifications.

Moreover, neuroticism is confirmed as a negative health trait, associated not only with lower SRH ratings, but also with higher levels of functional limitations and Disease Burden both cross-sectionally and longitudinally (See Appendix C). Contrary to Hypothesis 2, however, this study produced limited evidence to support weaker negative associations between the health reports and SRH as neuroticism increased. Only in the cross-sectional association between SRH and ADLs were higher levels of neuroticism associated with weaker alignment. We note that the absence of significant findings for neuroticism in the US fails to support the earlier study in Sweden, which reported the strengthening of the negative longitudinal association between SRH and somatic health (Henning et al., 2021). However, far from settling the question, these divergent sets of findings suggest several testable hypotheses as to why findings might differ.

This previous study suggests that highly neurotic individuals can be characterized with “health neuroticism” or “illness anxiety” (Kööts–Ausmees et al., 2016) and tend to perceive a given challenge in somatic health as more threatening to their overall sense of well-being. The US study does not support this claim, but instead suggests that conscientiousness may play a similar role. In taking a more proactive stance toward wellness, highly conscientious individuals may be more alert to and knowledgeable about known and foreseeable health challenges (Kööts–Ausmees et al., 2016). When deciding which SRH category best represents their sense of well-being, those individuals may prioritize their knowledge and information of their own health rather than subjective reactions and projections.

This study also finds a more consistent moderation pattern concerning conscientiousness when compared to the previous study (Henning et al., 2021). Possible explanations for the disagreements are many and may include cultural as well as design differences. For instance, the items and scales used to measure ADLs, Disease Burden, and pain differ between the two studies, with all three measures scaled in a wider range for the Swedish study (e.g. 0–34 for Disease Burden vs. 0–7). Moreover, although both study samples are selective with regard to health, this study’s sample may be more robust in terms of health, as it includes those who survived until an older age, possibly with fewer cases of highly neurotic and frail individuals in subsequent observations. Nonetheless, the reader is reminded that no differences in personality were detected between the study sample and those who were dropped from the HRS after baseline.

If design differences are ruled out, the observed discrepancies between the studies may be interpreted to reflect cultural and ecological variations in the roles of conscientiousness and neuroticism in SRH. Notably in many countries, highly conscientious individuals tend to be less neurotic, and vice versa (Declerck et al., 2006; Klimstra et al., 2013). It could be that neuroticism plays a dominant role in the evaluative process of well-being among Swedish people whereas conscientiousness is more dominant in the US population. However, future research is necessary to confirm if conscientiousness and neuroticism play similar or opposite roles in the internal alignment between SRH and somatic health across different societal and cultural contexts.

Limitations and Future Directions

The estimation of change itself as well as the covariates were necessarily simplified despite the depth, complexity, and intra-variability afforded by the data. Future studies may employ cross-lagged designs instead, to better examine the internal link between SRH and somatic health while considering individual histories and various events over time, as well as correlated changes in the personality traits. Due to this limited detailing of time-varying changes, the reader is urged to view these findings as suggestive. We also emphasize the need for future research that considers distinct personality profiles that may emerge from concurrent and inter-related changes across the Big-Five personality traits in later life (Allen & DeYoung, 2017).

Supplemental Material

Supplemental Material - Are Changes in Somatic Health Reflected Differently in Updated Self-Ratings by Big-Five Personality Traits?

Supplemental Material for Are Changes in Somatic Health Reflected Differently in Updated Self-Ratings by Big-Five Personality Traits? by Katsuya Oi, and Melissa Hardy in Journal of Aging and Health

Footnotes

Declaration of Conflicting Interests

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

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Katsuya Oi

Supplemental Material

Supplemental Material for this article is available online.

References

Aiken-Morgan

A. T.

Bichsel

Savla

Edwards

C. L.

Whitfield

K. E.

(2014). Associations between self-rated health and personality. Ethnicity and Disease, 24(4), 418–422.

Allen

T. A.

DeYoung

C. G.

(2017). Personality neuroscience and the five factor model. In Oxford handbook of the five factor model (p. 52). Oxford Library of Psychology.

Benyamini

(2011). Why does self-rated health predict mortality? An update on current knowledge and a research agenda for psychologists.

Bloemen

Hochguertel

Zweerink

(2017). The causal effect of retirement on mortality: Evidence from targeted incentives to retire early. Health Economics, 26(12), Article e204–e218. https://doi.org/10.1002/hec.3493

Bogg

Roberts

B. W.

(2013). The case for conscientiousness: Evidence and implications for a personality trait marker of health and longevity. Annals of Behavioral Medicine: A Publication of the Society of Behavioral Medicine, 45(3), 278–288. https://doi.org/10.1007/s12160-012-9454-6

Bugliari

Campbell

Chan

Hayden

Hayes

Hurd

Karabatakis

Main

Mallett

McCullough

(2019). RAND HRS detailed imputations file 2016 (V1) documentation. RAND Center for the Study of Aging. https://Www.Rand.Org/Content/Dam/Rand/Www/External/Labor/Aging/Dataprod/Randhrsimp1992_2016v1.Pdf

Declerck

C. H.

Boone

De Brabander

(2006). On feeling in control: A biological theory for individual differences in control perception. Brain and Cognition, 62(2), 143–176. https://doi.org/10.1016/j.bandc.2006.04.004

Denovan

Dagnall

Lofthouse

(2019). Neuroticism and somatic complaints: Concomitant effects of rumination and worry. Behavioural and Cognitive Psychotherapy, 47(4), 431–445. https://doi.org/10.1017/S1352465818000619

Dowd

J. B.

Zajacova

(2010). Does self-rated health mean the same thing across socioeconomic groups? Evidence from biomarker data. Annals of Epidemiology, 20(10), 743–749. https://doi.org/10.1016/j.annepidem.2010.06.007

10.

Feldman

P. J.

Cohen

Doyle

W. J.

Skoner

D. P.

Gwaltney

J. M.

Jr. (1999). The impact of personality on the reporting of unfounded symptoms and illness. Journal of Personality and Social Psychology, 77(2), 370–378. https://doi.org/10.1037//0022-3514.77.2.370

11.

Gomez

Cooper

(2002). Neuroticism and extraversion as predictors of negative and positive emotional information processing: Comparing Eysenck’s, Gray’s, and Newman’s theories. European Journal of Personality, 16(5), 333–350. https://doi.org/10.1002/per.459

12.

Gustman

A. L.

Steinmeier

T. L.

(2000). Retirement outcomes in the health and retirement study. National Bureau of Economic Research.

13.

Harkins

S. W.

Price

D. D.

Braith

(1989). Effects of extraversion and neuroticism on experimental pain, clinical pain, and illness behavior. Pain, 36(2), 209–218. https://doi.org/10.1016/0304-3959(89)90025-0

14.

Henning

Berg

A. I.

Cengia

Hansson

Spuling

S. M.

Wettstein

(2021). Do neuroticism and conscientiousness interact with health conditions in predicting 4-year changes in self-rated health among Swedish older adults? Psychology and Aging, 36(6), 730–743. https://doi.org/10.1037/pag0000626

15.

Hirdes

J. P.

Forbes

W. F.

(1993). Factors associated with the maintenance of good self-rated health. Journal of Aging and Health, 5(1), 101–122. https://doi.org/10.1177/089826439300500105

16.

Huisman

Deeg

D. J.

(2010). A commentary on Marja Jylhä’s “what is self-rated health and why does it predict mortality? Towards a unified conceptual model”(69: 3, 2009, 307–316). Social Science and Medicine, 70(5), 652–654. https://doi.org/10.1016/j.socscimed.2009.11.003

17.

Idler

E. L.

Benyamini

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

18.

Jylhä

(2009). What is self-rated health and why does it predict mortality? Towards a unified conceptual model. Social Science and Medicine, 69(3), 307–316. https://doi.org/10.1016/j.socscimed.2009.05.013

19.

Jylhä

Volpato

Guralnik

J. M.

(2006). Self-rated health showed a graded association with frequently used biomarkers in a large population sample. Journal of Clinical Epidemiology, 59(5), 465–471. https://doi.org/10.1016/j.jclinepi.2005.12.004

20.

Klimstra

T. A.

Bleidorn

Asendorpf

J. B.

Van Aken

M. A.

Denissen

J. J.

(2013). Correlated change of Big five personality traits across the lifespan: A search for determinants. Journal of Research in Personality, 47(6), 768–777. https://doi.org/10.1016/j.jrp.2013.08.004

21.

Kööts–Ausmees

Schmidt

Esko

Metspalu

Allik

Realo

(2016). The role of the five–factor personality traits in general self–rated health. European Journal of Personality, 30(5), 492–504. https://doi.org/10.1002/per.2058

22.

Lazarevič

Brandt

(2020). Diverging ideas of health? Comparing the basis of health ratings across gender, age, and country. Social Science and Medicine, 267, 112913. https://doi.org/10.1016/j.socscimed.2020.112913

23.

Löckenhoff

C. E.

Terracciano

Ferrucci

Costa

P. T.

Jr. (2012). Five-factor personality traits and age trajectories of self-rated health: The role of question framing. Journal of Personality, 80(2), 375–401. https://doi.org/10.1111/j.1467-6494.2011.00724.x

24.

McArdle

J. J.

(2009). Latent variable modeling of differences and changes with longitudinal data. Annual Review of Psychology, 60, 577–605. https://doi.org/10.1146/annurev.psych.60.110707.163612

25.

(2020). Disuse as time away from a cognitively demanding job; how does it temporally or developmentally impact late-life cognition? Intelligence, 82, 101484. https://doi.org/10.1016/j.intell.2020.101484

26.

Quesnel–Vallée

(2007). Self-rated health: Caught in the crossfire of the quest for ‘true’health? International Journal of Epidemiology, 36(6), 1161–1164. https://doi.org/10.1093/ije/dym236

27.

Rafienia

Azadfallah

Fathi-Ashtiani

Rasoulzadeh-Tabatabaiei

(2008). The role of extraversion, neuroticism and positive and negative mood in emotional information processing. Personality and Individual Differences, 44(2), 392–402. https://doi.org/10.1016/j.paid.2007.08.018

28.

Roberts

B. W.

Kuncel

N. R.

Shiner

Caspi

Goldberg

L. R.

(2007). The power of personality: The comparative validity of personality traits, socioeconomic status, and cognitive ability for predicting important life outcomes. Perspectives on Psychological Science: A Journal of the Association for Psychological Science, 2(4), 313–345. https://doi.org/10.1111/j.1745-6916.2007.00047.x

29.

Smith

Ryan

Sonnega

Weir

(2017). Psychosocial and lifestyle questionnaire 2006–2016. Survey Research Center, Institute for Social Research.

30.

Sočan

Bucik

(1998). Relationship between speed of information-processing and two major personality dimensions—extraversion and neuroticism. Personality and Individual Differences, 25(1), 35–48. https://doi.org/10.1016/s0191-8869(98)00031-2

31.

Spuling

S. M.

Wurm

Tesch-Römer

Huxhold

(2015). Changing predictors of self-rated health: Disentangling age and cohort effects. Psychology and Aging, 30(2), 462–474. https://doi.org/10.1037/a0039111

32.

Stenholm

Pentti

Kawachi

Westerlund

Kivimäki

Vahtera

(2014). Self-rated health in the last 12 years of life compared to matched surviving controls: The Health and Retirement Study. PloS One, 9(9), Article e107879. https://doi.org/10.1371/journal.pone.0107879

33.

Stephan

Sutin

A. R.

Luchetti

Hognon

Canada

Terracciano

(2020). Personality and self-rated health across eight cohort studies. Social Science and Medicine, 263, 113245. https://doi.org/10.1016/j.socscimed.2020.113245

34.

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). The Journals of Gerontology. Series B, Psychological Sciences and Social Sciences, 67(1), 4–12. https://doi.org/10.1093/geronb/gbr072

35.

Wade

J. B.

Dougherty

L. M.

Hart

R. P.

Rafii

Price

D. D.

(1992). A canonical correlation analysis of the influence of neuroticism and extraversion on chronic pain, suffering, and pain behavior. Pain, 51(1), 67–73. https://doi.org/10.1016/0304-3959(92)90010-9

36.

Williams

P. G.

O'Brien

C. D.

Colder

C. R.

(2004). The effects of neuroticism and extraversion on self-assessed health and health-relevant cognition. Personality and Individual Differences, 37(1), 83–94. https://doi.org/10.1016/j.paid.2003.08.001

37.

Xia

Yang

(2019). RMSEA, CFI, and TLI in structural equation modeling with ordered categorical data: The story they tell depends on the estimation methods. Behavior Research Methods, 51(1), 409–428. https://doi.org/10.3758/s13428-018-1055-2

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.39 MB