Occupational Social Class and Personality Traits in Relation to Leisure-Time Physical Activity Level: Cross-Sectional Results From the Copenhagen Aging and Midlife Biobank

Abstract

Objective: To investigate separate and combined associations of occupational social class and personality traits with late midlife leisure-time physical activity duration and intensity. Method: Cross-sectional data from the Copenhagen Aging and Midlife Biobank (N = 4,649) were analyzed using linear regression models with leisure-time physical activity (metric equivalence) as outcome. Results: Low versus high occupational social class was associated with 4% (95% confidence interval [CI] = [3%, 5%]) greater leisure-time physical activity duration, but 2% (CI = [1%, 3%]) lower intensity. Each 10-unit increase in extraversion was associated with 5% (CI = [2%, 8%]) greater duration. Intensity increased by each 10-unit increase in conscientiousness (6%, CI = [4%, 7%]), openness (3%, CI = [1%, 4%]), neuroticism (3%, CI = [1%, 4%]), and extraversion (5%, CI = [4%, 7%]). Conscientiousness was positively associated with duration in low, but not in high, occupational social class (interaction p value = .002). Discussion: Higher occupational social class was associated with lower leisure-time physical activity duration, but higher intensity. Extraversion was positively associated with duration and intensity. Conscientiousness, openness, and neuroticism were positively associated with intensity. Overall, interactions were not consistent.

Keywords

interaction occupational social class personality traits leisure-time physical activity cross-sectional

Introduction

Physical activity level has been shown to be associated with personality traits. A recent systematic review and meta-analysis found conscientiousness and extraversion to be positively associated with physical activity level, while neuroticism was found to be negatively associated with physical activity. The results further suggested a weak positive association with openness (Wilson & Dishman, 2015).

It has been suggested that personality traits may interact with the social environment in a supportive or detrimental way in relation to health outcomes (Chapman, Fiscella, Kawachi, & Duberstein, 2010; Dishman, Sallis, & Orenstein, 1984; Millar et al., 2013; Packard et al., 2012; Roberts, Kuncel, Shiner, Caspi, & Goldberg, 2007). Certain personality traits may help people neutralize potential health risks associated with low socioeconomic position (SEP), while others may exacerbate them (Chapman et al., 2010; Chen, 2004). While people of higher SEP may benefit from a supporting environment promoting an active lifestyle, people of lower SEP may miss out on this and even experience a discouraging or interfering environment (Pampel, Krueger, & Denney, 2010; Seefeldt, Malina, & Clark, 2002; Tucker-Seeley, Subramanian, Li, & Sorensen, 2009).

In 1998, a study suggested that people in low-income households had higher life satisfaction, better perceived health, and less depressive symptoms if they possessed a high sense of control. The moderating role was stronger in the low-income households than in the high-income ones, where less variation in health and well-being was observed (Lachman & Weaver, 1998). Similar results were published more recently in 2014 using educational level as the SEP measure and mortality as the outcome (Turiano, Chapman, Agrigoroaei, Infurna, & Lachman, 2014). Sense of control is related to perceived behavioral control as described in the theory of planned behavior, and studies suggest that perceived control mediates the effects of conscientiousness on physical activity (McEachan, Sutton, & Myers, 2010; Rhodes & Smith, 2006). The hypotheses of an interaction is further supported by results from studies of Danish twin pairs which tested potential explanations for the larger variance in health behaviors (Johnson et al., 2011) and poor physical health (Johnson et al., 2010) among people with low compared with high educational level. The findings suggested that a higher educational level may be associated with a culture that discourages unhealthy behavior (Johnson et al., 2011) and with personal resources that help people to protect their physical health (Johnson et al., 2010).

Only few studies have investigated the interaction between the big five personality traits and SEP in relation to health outcomes, and even fewer in relation to lifestyle.

Using a composite SEP measure of education, household income, and occupational prestige, Elliot, Turiano, and Chapman (2016) found higher levels of conscientiousness to be more protective against high levels of inflammatory markers among people of low SEP (Elliot et al., 2016). Two other studies have investigated interactions between conscientiousness and SEP with respect to lifestyle factors. One study suggested stronger protective effects of conscientiousness at high SEP in relation to smoking using education as SEP measure (Chapman, Fiscella, Duberstein, & Kawachi, 2009) and another in relation to drug use using poverty as SEP measure (Sutin, Evans, & Zonderman, 2013), respectively. Thus, conscientiousness appears to have a stronger positive impact on health-related outcomes at high SEP levels.

The same studies found no interaction between agreeableness and SEP (Chapman et al., 2009; Elliot et al., 2016; Sutin et al., 2013).

Neuroticism interacts with SEP according to some studies, but not all. In the study by Elliot and colleagues (2016), the composite SEP measure showed higher levels of neuroticism to be associated with higher inflammatory levels at low SEP but lower levels at high SEP (Elliot et al., 2016). Another study, also using inflammatory markers as the outcome, but deprivation as SEP marker, found somewhat similar results when comparing deprivations groups (Millar et al., 2013), and so did a study using a composite SEP measure and cardiovascular mortality as the outcome (Hagger-Johnson et al., 2012). With respect to lifestyle factors, one study found no interactions in relation to sense of coherence, self-esteem, self-efficacy, fruit/vegetable consumption, smoking cessation, or aerobic exercise (Packard et al., 2012), and neither did two other studies in relation to drug use (using poverty as SEP measure; Sutin et al., 2013) and smoking (using educational level as SEP measure; Chapman et al., 2009). Thus, neuroticism may not show strong interactions with SEP on lifestyle factors, but it is possible that neuroticism is harmful at low SEP and protective at high SEP levels.

For openness, a study found no interaction with SEP in relation to inflammatory markers (Elliot et al., 2016), but the study by Chapman et al. (2009) indicated that low openness was protective against smoking in the high education group, but not in the lower (Chapman et al., 2009).

Finally, the findings for extraversion are mixed. Higher extraversion was found to be associated with less hopelessness, greater sense of coherence, and greater self-esteem, especially among the most deprived group, but no interactions between extraversion and SEP were observed in relation to self-efficacy, fruit/vegetable consumption, smoking cessation, or aerobic exercise (Packard et al., 2012). Other studies relating extraversion to inflammatory markers, smoking behavior, and mortality, respectively, did not find any interaction with SEP (Chapman et al., 2009; Elliot et al., 2016; Hagger-Johnson et al., 2012; Millar et al., 2013). Overall, there appears little evidence of interactions between extraversion and SEP with respect to lifestyle factors.

To our knowledge, no previous studies have investigated if personality traits and SEP interact in relation to leisure-time physical activity level, and previous studies using other outcomes of health or health behavior may not be generalizable to physical activity. However, based on the available evidence, we hypothesize that interactions may be demonstrated for conscientiousness and possibly also neuroticism and openness. Analyses of interactions between personality traits and SEP with respect to physical activity are of particular importance, to people of late-middle age for at least three reasons. First, the age-related physical functional decline increases the risk of functional disability. According to the World Health Organization (WHO), a third of the adult population in high-income countries is insufficiently physically active (<150 min/week of moderate intensity) with the highest prevalence in older age (WHO, 2014). However, the importance of remaining physically active in midlife and later to prevent early onset functional limitations has been demonstrated (Patel et al., 2006; Peterson et al., 2009), and a prerequisite for prevention of physical inactivity is to know potential predictors. Second, physical activity reduces the risk of a number of common non-communicable diseases, such as diabetes and cardiovascular disease (WHO, 2014). Third, interaction analyses could help identifying vulnerable groups with high risk of physical inactivity and related morbidity.

The aim of this study was to investigate the separate and combined associations of occupational social class and personality traits with duration and intensity of leisure-time physical activity among late midlife men and women active at the labor market.

Method

This study is based on cross-sectional data from the Copenhagen Aging and Midlife Biobank (CAMB; Avlund et al., 2014; Lund et al., 2016). CAMB was established as a late midlife follow-up (age 49-62 years) of three cohorts: the Metropolit Cohort (MC; comprising men born in Copenhagen in 1953; Osler, Lund, Kriegbaum, Christensen, & Andersen, 2006), the Copenhagen Perinatal Cohort (CPC; comprising men and women born at the National University Hospital in Copenhagen in 1959-1961; Zachau-Christiansen, 1972), and the Danish Longitudinal Study on Work, Unemployment, and Health (DALWUH; comprising a random sample of men and women from the Danish population born in 1949 or 1959; Christensen et al., 2004). From April 2009 to March 2011, a total of 5,521 participants answered two comprehensive questionnaires, one at home and one at the study clinic (Avlund et al., 2014). From this cohort, 872 participants (18%) were excluded due to the following reasons: outside the labor market (n = 590), on benefits (n = 119), economically active without social class information (n = 6), student or housewife/househusband (n = 11) or missing information about occupational status (n = 21), personality traits (n = 53), leisure-time physical activity level (n = 131), age (n = 2), and/or severe depressive symptoms (n = 10). The final sample size had 4,649 participants.

Occupational Social Class

Occupational social class was categorized according to the Danish occupational social class measurement (Christensen et al., 2014). Each participant in CAMB was manually assigned to one of eight groups based on self-reported job title according to educational requirements, required management skills, and/or subordinates or economical assets. Due to the selection of labor market attached participants, only, occupational social class groups I-V were included in this study (I: jobs requiring top-level educational attainment with ≥4 years of university or a similar training/white-collar jobs implying management control of big organizations/owners of large-scale companies with >50 subordinates/self-employed with >20 subordinates; II: white-collar jobs requiring approximately 3 years of theoretical training or which imply management responsibilities for 11-50 subordinates/owners of medium-scale companies with 6+ employees; III: non-manual white-collar jobs requiring approximately 1½ years of theoretical training and practical skills/white-collar jobs requiring expertise at basic level but with management responsibilities for 1-10 subordinates/self-employed with business with 0-5 employees; IV: manual white-collar jobs requiring up to 1 year theoretical training and practical training/blue-collar workers with up to 1 year of theoretical training and practical training; V: manual jobs requiring little theoretical and practical training/semi- or unskilled workers (Christensen et al., 2014). To avoid small groups in the interaction analyses, a dichotomized variable was created as high (I-II) versus low (III-V) occupational social class.

Personality Traits

Personality traits were assessed with the short version of the Danish 240 item Revised NEO Personality Inventory, the NEO Five-Factor Inventory described in Mortensen et al. (2014). The instrument assesses the big five personality traits, each defined in terms of six specific facets: conscientiousness (competence, order, dutifulness, achievement striving, self-discipline, deliberation), agreeableness (trust, straightforwardness, altruism, compliance, modesty, tender-mindedness), neuroticism (anxiety, angry hostility, depression, self-consciousness, impulsiveness, vulnerability), openness (fantasy, aesthetics, feelings, actions, ideas, values), and extraversion (warmth, gregariousness, assertiveness, activity, excitement-seeking, positive emotions) (Costa & McCrae, 1989).

Outcome Measures

Physical activity level was quantified in metric equivalence (MET) utilizing the Danish version of the International Physical Activity Long Form Questionnaire (Craig et al., 2003; Hansen, Blangsted, Hansen, Søgaard, & Sjøgaard, 2010). MET takes into account both duration and intensity of physical activity. Activity levels correspond to multiples of resting metabolic rate and one MET corresponds to resting metabolic rate during quiet sitting which is considered to be no activity (for more details, see www.ipaq.ki.se; Craig et al., 2003; Hansen et al., 2010).

Duration of leisure-time MET-minutes (MET-min) per week were estimated as the weekly sum in hours of all reported 24-hr vigorous (e.g., running, soccer, tennis, and aerobics corresponding to six MET-min), moderate (e.g., gardening, heavy housekeeping, gymnastics, dancing, and weight training corresponding to five MET-min), and low intensity (e.g., strolls, light cleaning, yoga, or bowling corresponding to three MET-min) leisure-time physical activities including active transportation defined as bicycling when commuting to work (corresponding to four MET-min).

Intensity of mean leisure-time MET-min/hour was derived from the weekly score by dividing with the sum of reported hours of leisure-time activities during 1 week.

Covariates

Confounders were identified by directed acyclic graph methods (Greenland, Pearl, & Robins, 1999). The adjusted models included sex, age (years), cohort of origin (MC/CPC/DALWUH), and severe depressive symptoms (yes/no as identified by the Major Depression Inventory; Olsen, Jensen, Noerholm, Martiny, & Bech, 2003) as these could potentially affect both exposures and outcome.

Statistical Analyses

Descriptive results are presented as frequencies with column percentages (%) for categorical variables and medians with interquartile ranges (IQRs) for continuous variables by occupational social class and for the full study sample.

Associations between occupational social class, personality traits, and physical activity were assessed using crude and adjusted linear regression models. Adjusted analyses included sex, age, cohort, and severe depressive symptoms. In addition, adjusted analyses of associations between personality traits and leisure-time physical activity level were adjusted for the remaining four personality traits (i.e., analysis of conscientiousness was adjusted for agreeableness, neuroticism, openness, and extraversion, and similarly for the other traits). Tests for trend were performed by including occupational social class as a continuous variable.

The interaction of each personality trait with dichotomized occupational social class (high vs. low) in relation to leisure-time physical activity level were analyzed in crude models as well as adjusted for sex, age, cohort, severe depressive symptoms, and the remaining four personality traits as previously described.

Model assumptions were assessed by fit diagnostics and probability plots. Due to skewed residuals, the two outcome variables MET-min/week (duration) and MET-min/hour (intensity) were transformed using the natural logarithm. Results were back-transformed for presentation and are thus interpreted as the relative increase (labeled “factor”) in the outcome (“FAQ How Do I Interpret a Regression Model When Some Variables Are Log Transformed?” 2016). For analyses of personality traits, measures were rescaled so that a one-unit increase in the MET-score corresponded to a 10% increase in the considered personality trait score. Results from the interaction analyses are presented as crude plots of simple slopes of personality trait score by physical activity level with 95% confidence interval (CI) bands for high and low occupational social class group. Adjusted interaction analyses are presented in the online appendix.

Sensitivity Analyses

In additional analyses, occupational social class was categorized according to non-manual (I-III) versus manual work (IV-V) to investigate whether physical strain explained differences in leisure-time physical activity level differently.

Furthermore, an alternative outcome measure of general intensity of leisure-time physical activity was used to test if participants’ self-perceived leisure-time physical activity were similarly associated with exposure measures as the calculated measures of MET-min. The question was phrased as follows: “What would you say best describes your leisure-time physical activities?” and response categories ranged from “competitive sport regularly and several times a week”; “physical training or heavy house or garden work at least 4 hr per week”; “walks, biking, or other kinds of light exercise at least 4 hr a week (including Sunday excursions, lighter garden work, and biking/walking to and from work)” to “read, watch television, or have other sedentary activities”. General intensity of leisure-time physical activity was included in linear regression models in its categorical form and test for trend were performed by including it in its continuous form.

Finally, the analyses were rerun under adjustment for number of previous or current diseases. Each participant answered the question, “Below is a list of health problems or diseases. Do you have or have you had any of the following diseases?” with “Yes, I have now,” “Yes, I have had” or “No, I have never had” for each of the following diseases: asthma, allergy (not asthma), diabetes, cataract, hypertension, also called high blood pressure, myocardial infarction or angina pectoris, stroke, chronic bronchitis, emphysema (large lungs, smoker’s lungs), osteoarthritis, rheumatoid arthritis, osteoporosis, peptic ulcer, cancer including leukemia, migraine or frequent headaches, chronic anxiety/depression, other mental disorder/bad nerves, back disease, urinary incontinence, difficulty urinating, tinnitus, kidney stones, and gallstones.

Data management and descriptive analyses were performed using SAS software, version 9.4 for Windows, while regression models and graphical presentations were conducted in the R version 3.2.4.

Ethical Approval

The local ethical committee and Danish Data Protection Agency have approved CAMB as a database combining three cohorts (Approval no. H-A-2008-126 and no. 2013-41-1814).

Results

The majority of the participants was men of age 48 to 58 years and belonged to occupational social classes II and III. Participants of lower compared with higher occupational social class had lower scores on conscientiousness, openness, and extraversion, and higher scores on neuroticism (Table 1). Lower occupational social class participants more frequently reported sedentary or light leisure-time activities compared with high occupational social class participants (results not shown in tables).

Table 1.

Characteristics of for participants in the Copenhagen Aging and Midlife Biobank with complete information on all parameters according to occupational social class in late midlife.

Occupational social class
	I	II	III	IV	V	Total
Total	835	1,348	1,200	838	428	4,649
Sex, n (%)
Women	160 (19.2)	443 (32.9)	378 (31.5)	306 (36.5)	123 (28.7)	1,410 (30.3)
Men	675 (80.8)	905 (67.1)	822 (68.5)	532 (63.5)	305 (71.3)	3,239 (69.7)
Originating cohort, n (%)
MC	445 (53.3)	602 (44.7)	519 (43.3)	355 (42.4)	191 (44.6)	2,112 (45.4)
DALWUH	170 (20.4)	324 (24.0)	287 (23.9)	196 (23.4)	95 (22.2)	1,072 (23.1)
CPC	220 (26.4)	422 (31.3)	394 (32.8)	287 (34.3)	142 (33.2)	1,465 (31.5)
Age, n (%)
48-52	307 (36.8)	601 (44.6)	546 (45.5)	383 (45.7)	201 (47.0)	2,038 (43.8)
55-58	445 (53.3)	602 (44.7)	520 (43.3)	355 (42.4)	191 (44.6)	2,113 (45.5)
59-62	83 (9.9)	145 (10.8)	134 (11.2)	100 (11.9)	36 (8.4)	498 (10.7)
Severe depressive symptoms, n (%)
No	826 (98.9)	1,324 (98.2)	1,167 (97.3)	811 (96.8)	411 (96.0)	4,539 (97.6)
Yes	9 (1.1)	24 (1.8)	33 (2.8)	27 (3.2)	17 (4.0)	110 (2.4)
Midlife personality traits, median (IQR)
Conscientiousness	35.0 (33.0-39.0)	34.0 (31.0-37.0)	34.0 (31.0-37.0)	34.0 (30.0-37.0)	33.0 (29.0-36.0)	34.0 (31.0-37.0)
Agreeableness	32.0 (29.0-36.0)	34.0 (30.0-37.0)	33.0 (30.0-37.0)	34.0 (31.0-38.0)	34.0 (31.0-37.0)	33.0 (30.0-37.0)
Neuroticism	15.0 (11.0-19.0)	16.0 (12.0-21.0)	16.0 (12.0-21.0)	17.0 (14.0-22.0)	18.0 (14.0-23.0)	16.0 (12.0-21.0)
Openness	31.0 (27.0-34.0)	30.0 (25.0-34.0)	27.0 (23.0-32.0)	26.5 (23.0-31.0)	26.0 (22.0-30.0)	28.0 (24.0-33.0)
Extraversion	33.0 (28.0-36.0)	32.0 (28.0-36.0)	31.0 (27.0-35.0)	30.0 (26.0-33.0)	30.0 (25.0-33.0)	31.0 (27.0-35.0)

Note. MC = Metropolit Cohort; DALWUH = Danish Longitudinal Study on Work, Unemployment, and Health; CPC = Copenhagen Perinatal Cohort; IQR = interquartile range.

A comparison of the study sample and the 872 excluded participants revealed that a larger proportion among the non-participants were female (37.7% vs. 30.3% in study sample), of older age (56.6 years vs. 56.1 years), of low occupational social class (64.9% vs. 53.0%), had severe depressive symptoms (14.2% vs. 2.4%), and reported an overall lower self-perceived leisure-time physical activity level. In addition, non-participants had lower scores on conscientiousness, openness, and extraversion, while scores on agreeableness and especially neuroticism were higher (results not shown in tables).

Occupational Social Class and Physical Activity

Compared with the highest occupational social class, being in a lower occupational social class was associated with 10% to 19% longer duration of leisure-time physical activity, with a 4% increase by each step down in occupational social class (Table 2). The opposite association was observed in relation to intensity. Lower occupational social class (III-V) was associated with 4% to 8% lower intensity of leisure-time physical activity with a 2% reduction by each step. No substantial changes were observed from crude to adjusted results.

Table 2.

Crude and adjusted results from linear regression analyses of late midlife occupational social class and personality traits (by 10-unit increase) in relation to duration and intensity of leisure-time physical activity level in metabolic equivalents (MET-min/week and MET-min/hour, respectively). Results are for participants in the Copenhagen Aging and Midlife Biobank and are based on complete cases with information on all parameters (N=4,649).

	Crude		Adjusted^a
	Factor	95% CI	Factor	95% CI
Duration (MET^b-min/week)
Occupational social class
I	1	Reference	1	Reference
II	1.10	[1.06, 1.15]	1.10	[1.06, 1.15]
III	1.09	[1.05, 1.14]	1.10	[1.05, 1.14]
IV	1.18	[1.13, 1.23]	1.19	[1.14, 1.24]
V	1.18	[1.12, 1.25]	1.19	[1.13, 1.25]
Trend	1.04	[1.03, 1.05]	1.04	[1.03, 1.05]
Conscientiousness^c	1.02	[1.00, 1.05]	1.01	[0.99, 1.05]
Agreeableness^c	1.01	[0.99, 1.04]	1.02	[0.99, 1.04]
Neuroticism^c	0.99	[0.97, 1.01]	1.02	[0.99, 1.04]
Openness^c	0.99	[0.97, 1.01]	0.98	[0.95, 1.00]
Extraversion^c	1.04	[1.01, 1.06]	1.05	[1.02, 1.08]
Intensity (MET^b-min/hr)
Occupational social class
I	1	Reference	1	Reference
II	0.96	[0.94, 0.98]	0.96	[0.94, 0.98]
III	0.94	[0.92, 0.96]	0.94	[0.92, 0.96]
IV	0.94	[0.91, 0.96]	0.94	[0.91, 0.96]
V	0.92	[0.89, 0.95]	0.92	[0.89, 0.95]
trend	0.98	[0.97, 0.99]	0.98	[0.97, 0.99]
Conscientiousness^c	1.06	[1.05, 1.08]	1.06	[1.04, 1.07]
Agreeableness^c	0.99	[0.97, 1.00]	0.99	[0.97, 1.00]
Neuroticism^c	0.98	[0.97, 0.99]	1.03	[1.01, 1.04]
Openness^c	1.05	[1.03, 1.06]	1.03	[1.01, 1.04]
Extraversion^c	1.07	[1.06, 1.08]	1.05	[1.04, 1.07]

Note. CI = confidence interval; MET = metabolic equivalents.

Analyses of occupational social class are adjusted for sex, age, cohort, and severe depressive symptoms. Analyses of personality traits are additionally adjusted for other personality traits (i.e., analysis of conscientiousness is adjusted for agreeableness, neuroticism, openness, and extraversion).

Metabolic equivalents (one MET corresponds to resting metabolic rate during quiet sitting).

The regression coefficients are interpreted as the relative increase in physical activity (% change in MET-min/week or hour, respectively) by each 10-unit increase in the personality trait score.

Categorization of occupational social class as manual versus non-manual yielded similar results to using the low versus high dichotomization. Moreover, utilizing the alternative outcome of self-perceived activity showed a lower overall leisure-time physical activity level with lower occupational social class (−3% by each step; 95% CI = [−5%, −2%]; results not shown in tables).

Personality Traits and Physical Activity

Overall, extraversion was most consistently associated with leisure-time physical activity level. For each 10-unit increase in extraversion score (ranging from 2 to 48), both duration and intensity increased by 5% (Table 2). The associations between the remaining personality traits and duration of physical activity appeared weaker and less consistent as not all results pointed in the same direction.

In relation to intensity, positive associations were observed for conscientiousness and openness (adjusted 6% and 3%, respectively, by each 10-unit increase). Neuroticism was negatively associated with intensity in crude analyses, but after adjustment, each 10-unit increase was associated with a 3% increase (Table 2).

Results from utilizing the alternative outcome measure of general intensity of leisure-time physical activity showed that conscientiousness and extraversion were positively associated with activity level, adjusted 13% (95% CI = [9%, 17%]) and 16% (95% CI = [12%, 19%]), respectively; Online Appendix Table 4.

Combined Effect of Occupational Social Class and Personality Traits

Figures 1 and 2 show the crude results from linear regression analyses of occupational social class and personality traits in relation to leisure-time physical activity level in duration and intensity, respectively. Conscientiousness was positively associated with duration of activity in the low occupational social class group, but not in the high (p value for interaction = .002 (Online Appendix Table 3)). None of the remaining results for either duration or intensity were statistically significant. Adjusting for confounding changed results for neuroticism in relation to intensity as the associations became positive for both occupational social class groups (without any interaction). Besides this, adjustment for confounding only led to minor changes in results (Online Appendix Tables 3 and 5).

Figure 1.

Crude results from linear regression analyses of midlife occupational social class and midlife personality traits in relation to duration of leisure-time physical activity level in metabolic equivalents (MET-min/week). Results are for participants in the Copenhagen Aging and Midlife Biobank and are based on complete cases with information on all parameters (N=4,649).

Figure 2.

Crude results from linear regression analyses of midlife occupational social class and midlife personality traits in relation to intensity of leisure-time physical activity level in metabolic equivalents (MET-min/hour). Results are for participants in the Copenhagen Aging and Midlife Biobank and are based on complete cases with information on all parameters (N=4,649).

Discussion

We found participants of higher compared with lower occupational social class to have a lower duration, but a higher intensity, of leisure-time physical activity. Extraversion was the personality traits most consistently and positively associated with higher duration of physical activity, while conscientiousness, extraversion, and surprisingly also neuroticism and openness were positively associated with intensity. Personality traits and occupational social class showed no clear interactions in relation to leisure-time physical activity level, the only exception being conscientiousness which was found to be associated with a greater duration in the low, but not in the high occupational social class.

The higher intensity but lower duration of leisure-time physical activity among the high occupational social class could be explained by this group engaging more in strenuous exercise or sports activities which usually are characterized with a greater intensity but shorter duration than everyday activities (Bauman et al., 2009; Craig et al., 2003). The high occupational social class includes managers of large- to medium-scale companies, owners of companies, and so on, and they might spend more time on work-related activities leaving less time to leisure-time physical activity. To compensate, they might prioritize strenuous physical activity for health reasons. Considering our sensitivity analyses using the participants’ self-reported overall leisure-time physical activity, the findings support such hypothesis and are in line with an older review showing that people with a high educational level engage more in leisure-time strenuous exercise (e.g., supervised) compared with blue-collar workers (Dishman et al., 1984).

Our findings for the main associations between personality traits and physical activity levels are partly in line with previous findings. Conscientiousness and extraversion were both positively associated with physical activity levels in our study, which is consistent with the review by Wilson and Dishman (2015). They further found a weak positive association with openness and no association with agreeableness (Wilson & Dishman, 2015). We found a positive relationship between openness and intensity, but no association with duration, and no associations for agreeableness. These minor differences between our results and the review by Wilson and Dishman (2015) could be a consequence of different outcome measures, but their finding of a negative relationship between neuroticism and physical activity only fits with our crude results. Interestingly, after adjustment for other traits, sex, age, and cohort, we found an overall positive association, suggesting that the negative association could be due to confounding by other traits or characteristics. A health-beneficial effect of high neuroticism is previously suggested by others (Elliot et al., 2016; Hagger-Johnson et al., 2012; Jokela & Keltikangas-Järvinen, 2011; Korten et al., 1999; Weiss & Costa, 2005), but due to different outcomes and covariates, it is not possible to compare the details of the findings to ours. Future studies could focus on facet-level analyses to enhance understanding of the relationship between personality and physical activity, as some facets may show a stronger association than others.

The interaction we found, suggesting conscientiousness to be positively related to duration of physical activity only in low occupational social class adds support to previous findings from studies using different indicators of personality, SEP, and health outcomes (Elliot et al., 2016; Lachman & Weaver, 1998; Turiano et al., 2014). However, conflicting results are also published suggesting a protective relation only in high SEP (Chapman et al., 2009; Sutin et al., 2013) or no interaction (Jokela & Keltikangas-Järvinen, 2011; all using different measures of SEP and other outcomes). We did not find any interaction in relation to intensity of physical activity which could be due to study limitations (described later). However, previous research actually support our findings by reporting that both physically active and sedentary individuals perceive exercise as healthy, but it is mainly those who believe they are in control of their health who engage in physical activity and that this would most often be light to moderate activities and not strenuous exercise (Dishman et al., 1984). Future studies could further investigate if control beliefs interact with personality traits, namely conscientiousness and extraversion, in relation to leisure-time physical activity level.

Our finding of no interaction for agreeableness is in line with previous findings using other outcomes (Chapman et al., 2009; Elliot et al., 2016; Jokela & Keltikangas-Järvinen, 2011; Sutin et al., 2013).

Overall, we found no interactions for either neuroticism, or extraversion, which is also in line with previous findings using other outcomes (Elliot et al., 2016; Hagger-Johnson et al., 2012; Millar et al., 2013). Our findings may not be convincing null-findings because tendencies can be observed for both traits in relation to duration of activity, suggesting neuroticism to be negatively related to duration of activity in low occupational social class, only, and extraversion to be more positively associated with duration in the low occupational social class than in the high. In addition, other studies actually suggest similar relations for both neuroticism (Elliot et al., 2016; Hagger-Johnson et al., 2012; Jokela & Keltikangas-Järvinen, 2011; Millar et al., 2013; Packard et al., 2012) and extraversion (Jokela & Keltikangas-Järvinen, 2011; Packard et al., 2012), with some evidence even suggesting that neuroticism could be health protective at high SEP (Elliot et al., 2016; Hagger-Johnson et al., 2012; Millar et al., 2013). This should be further elucidated in future studies able to overcome some of the shortcomings of this study (see later discussion of limitations).

Previous findings regarding the potential interaction between openness and SEP in relation to health outcomes are sparse. With one study showing a harmful association in high SEP (Chapman et al., 2009) and two studies not finding any interaction (Elliot et al., 2016; Jokela & Keltikangas-Järvinen, 2011), more research is necessary to draw conclusions, in particular, because the available studies use different SEP and outcome measures compared with our study.

This study includes detailed measures of occupational social class, personality, and physical activity, and the study population is larger than most previous studies. However, findings are also compromised by some limitations which must be taken into account. The relatively low participation rate in CAMB suggests that the findings of this study may be affected by selection bias (Avlund et al., 2014; Christensen et al., 2014; Lund et al., 2016; Mortensen et al., 2014). Highly conscientious people and those of higher occupational social class are more likely to participate in surveys, especially if they comply with a socially acceptable lifestyle and are in good health. Non-participation will reduce statistical power and ability to detect interactions, and the selection can also have led to an underestimation of the association between the exposures and the outcomes as loss-to-follow-up was more pronounced in the low SEP group, as indicted by our attrition analysis and a previous analyses of the full CAMB cohort (Lund et al., 2016).

Information bias can apply if specific personality profiles such as those high on conscientiousness are more eager to report socially accepted health behavior (Wilson & Dishman, 2015) including a high leisure-time physical activity level. If this applies mostly to the high occupational social class, for example, due to greater knowledge about official recommendations, this will result in a strengthening of the detected interaction in relation to duration of leisure-time physical activity and bias toward no difference between groups in relation to intensity. This bias should be investigated using objective measures of leisure-time physical activity.

Furthermore, it is not possible to assess causal directions from observational cross-sectional data. However, a large number of longitudinal studies have found SEP to be related to later life physical activity level (Ball, Carver, Downing, Jackson, & O’Rourke, 2015; Fransson et al., 2012), while the support for the opposite hypothesis appears weaker (Fransson et al., 2012). Investigating the latter relationship, one must take account of the well-established selection into leisure-time physical activity groups based on early life SEP (Lindström, Hanson, & Ostergren, 2001; Mäkinen et al., 2010). Furthermore, we did not find prior evidence to support that physical activity level should differentially influence development of specific personality traits.

We cannot completely reject that our finding of one statistically significant interaction between occupational social class and level of conscientiousness in relation to duration of leisure-time physical activity level could be due to chance. As most of the sensitivity analyses (besides the one adjusted for childhood SEP) could not support this results, this finding should be interpreted with caution, and the hypotheses should be tested in other studies that are able to overcome the discussed shortcomings. However, the lack of statistical significance could also be due to low statistical power. This is of particular relevance in the intensity analyses where the variance and differences between occupational social class groups were smaller.

Despite the limitations, our findings add nuances to the consistently observed social gradient in leisure-time physical activity level by emphasizing the importance of differentiation between duration and intensity and considering the potential impact of personality traits. Conscientiousness might be particularly relevant for duration of leisure-time physical activity in the lower occupational social class.

Footnotes

Acknowledgements

The authors thank the staff at Department of Public Health and National Research Centre for the Working Environment, who undertook the data collection. Further thanks to Kirsten Avlund†, Helle Brunsgaard, Nils-Erik Fiehn, Poul Holm-Pedersen, and Merete Osler, who initiated and established the Copenhagen Aging and Midlife Biobank from 2009-2011. The authors acknowledge the crucial role of the initiators and steering groups of The Metropolit Cohort, The Copenhagen Perinatal Cohort; and The Danish Longitudinal Study on Work, Unemployment, and Health. The authors would further like to thank the Social Inequalities in Ageing (SIA) project, funded by NordForsk, Project 74637, for valuable collaboration relating to this project.

Declaration of Conflicting Interests

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

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by a grant from Center for Healthy Aging provided by the Nordea Foundation. The Copenhagen Aging and Midlife Biobank has been supported by a generous grant from the VELUX FOUNDATION (Grants VELUX26145 and 31539).

Supplemental Material

Supplemental material is available for this article online.

References

Avlund

Osler

Mortensen

E. L.

Christensen

Bruunsgaard

Holm-Pedersen

. . . Lund

(2014). Copenhagen Aging and Midlife Biobank (CAMB): An introduction. Journal of Aging and Health, 26, 5-20.

Ball

Carver

Downing

Jackson

O’Rourke

(2015). Addressing the social determinants of inequities in physical activity and sedentary behaviours. Health Promotion International, 30(Suppl. 2), ii8-ii19.

Bauman

Bull

Chey

Craig

C. L.

Ainsworth

B. E.

Sallis

J. F.

. . . Pratt

(2009). The international prevalence study on physical activity: Results from 20 countries. International Journal of Behavioral Nutrition and Physical Activity, 6, Article 21.

Chapman

B. P.

Fiscella

Duberstein

Kawachi

(2009). Education and smoking: Confounding or effect modification by phenotypic personality traits? Annals of Behavioral Medicine, 38, 237-248.

Chapman

B. P.

Fiscella

Kawachi

Duberstein

P. R.

(2010). Personality, socioeconomic status, and all-cause mortality in the United States. American Journal of Epidemiology, 171, 83-92.

Chen

(2004). Why socioeconomic status affects the health of children. Current Directions in Psychological Science, 13, 112-115.

Christensen

Krølner

Nilsson

C. J.

Lyngbye

P. W.

Hougaard

C. Ø.

Nygaard

. . . Lund

(2014). Addressing social inequality in aging by the Danish occupational social class measurement. Journal of Aging and Health, 26, 106-127.

Christensen

Lund

Damsgaard

M. T.

Holstein

B. E.

Ditlevsen

Diderichsen

. . . Lynch

(2004). Cynical hostility, socioeconomic position, health behaviors, and symptom load: A cross-sectional analysis in a Danish population-based study. Psychosomatic Medicine, 66, 572-577.

Costa

P. T.

McCrae

R. R.

(1989). NEO PI-R/NEO-FFI manual [Supplement]. Odessa, FL: Psychological Assessment Resources.

10.

Craig

C. L.

Marshall

A. L.

Sjöström

Bauman

A. E.

Booth

M. L.

Ainsworth

B. E.

. . . Oja

(2003). International Physical Activity Questionnaire: 12-country reliability and validity. Medicine and Science in Sports and Exercise, 35, 1381-1395.

11.

Dishman

R. K.

Sallis

J. F.

Orenstein

D. R.

(1985). The determinants of physical activity and exercise. Public Health Reports, 100, 158-171.

12.

Elliot

A. J.

Turiano

N. A.

Chapman

B. P.

(2016). Socioeconomic status interacts with conscientiousness and neuroticism to predict circulating concentrations of inflammatory markers. Annals of Behavioral Medicine, 51, 240-250.

13.

FAQ How do I interpret a regression model when some variables are log transformed? (2016). Retrieved from https://stats.idre.ucla.edu/other/mult-pkg/faq/general/faqhow-do-i-interpret-a-regression-model-when-some-variables-are-log-transformed/

14.

Fransson

E. I.

Heikkilä

Nyberg

S. T.

Zins

Westerlund

Westerholm

. . . Kivimäki

(2012). Job strain as a risk factor for leisure-time physical inactivity: An individual-participant meta-analysis of up to 170,000 men and women. American Journal of Epidemiology, 176, 1078-1089.

15.

Greenland

Pearl

Robins

J. M.

(1999). Causal diagrams for epidemiologic research. Epidemiology, 10, 37-48.

16.

Hagger-Johnson

Roberts

Boniface

Sabia

Batty

G. D.

Elbaz

. . . Deary

I. J.

(2012). Neuroticism and cardiovascular disease mortality. Psychosomatic Medicine, 74, 596-603.

17.

Hansen

A. M.

Blangsted

A. K.

Hansen

E. A.

Søgaard

Sjøgaard

(2010). Physical activity, job demand-control, perceived stress-energy, and salivary cortisol in white-collar workers. International Archives of Occupational and Environmental Health, 83, 143-153.

18.

Johnson

Kyvik

K. O.

Mortensen

E. L.

Skytthe

Batty

G. D.

Deary

I. J.

(2010). Education reduces the effects of genetic susceptibilities to poor physical health. International Journal of Epidemiology, 39, 406-414.

19.

Johnson

Kyvik

K. O.

Mortensen

E. L.

Skytthe

Batty

G. D.

Deary

I. J.

(2011). Does education confer a culture of healthy behavior? Smoking and drinking patterns in Danish twins. American Journal of Epidemiology, 173, 55-63.

20.

Jokela

Keltikangas-Järvinen

(2011). The association between low socioeconomic status and depressive symptoms depends on temperament and personality traits. Personality and Individual Differences, 51, 302-308.

21.

Korten

A. E.

Jorm

A. F.

Jiao

Letenneur

Jacomb

P. A.

Henderson

A. S.

. . . Rodgers

(1999). Health, cognitive, and psychosocial factors as predictors of mortality in an elderly community sample. Journal of Epidemiology and Community Health, 53, 83-88.

22.

Lachman

M. E.

Weaver

S. L.

(1998). The sense of control as a moderator of social class differences in health and well-being. Journal of Personality and Social Psychology, 74, 763-773.

23.

Lindström

Hanson

B. S.

Ostergren

P. O.

(2001). Socioeconomic differences in leisure-time physical activity: The role of social participation and social capital in shaping health related behaviour. Social Science & Medicine, 52, 441-451.

24.

Lund

Mortensen

E. L.

Christensen

Bruunsgaard

Holm-Pedersen

Fiehn

N.-E.

. . . Osler

(2016). Cohort profile: The Copenhagen Aging and Midlife Biobank (CAMB). International Journal of Epidemiology, 45, 1044-1053.

25.

Mäkinen

Kestilä

Borodulin

Martelin

Rahkonen

Prättälä

(2010). Effects of childhood socio-economic conditions on educational differences in leisure-time physical activity. European Journal of Public Health, 20, 346-353.

26.

McEachan

R. R. C.

Sutton

Myers

(2010). Mediation of personality influences on physical activity within the theory of planned behaviour. Journal of Health Psychology, 15, 1170-1180.

27.

Millar

Lloyd

S. M.

McLean

J. S.

Batty

G. D.

Burns

Cavanagh

. . . Tannahill

(2013). Personality, socio-economic status and inflammation: Cross-sectional, population-based study. PLoS ONE, 8(3), e58256.

28.

Mortensen

E. L.

Flensborg-Madsen

Molbo

Christensen

Osler

Avlund

Lund

(2014). Personality in late midlife: Associations with demographic factors and cognitive ability. Journal of Aging and Health, 26, 21-36.

29.

Olsen

L. R.

Jensen

D. V.

Noerholm

Martiny

Bech

(2003). The internal and external validity of the Major Depression Inventory in measuring severity of depressive states. Psychological Medicine, 33, 351-356.

30.

Osler

Lund

Kriegbaum

Christensen

Andersen

A. M. N.

(2006). Cohort profile: The Metropolit 1953 Danish male birth cohort. International Journal of Epidemiology, 35, 541-545.

31.

Packard

C. J.

Cavanagh

McLean

J. S.

McConnachie

Messow

C. M.

Batty

G. D.

. . . Millar

(2012). Interaction of personality traits with social deprivation in determining mental wellbeing and health behaviours. Journal of Public Health, 34, 615-624.

32.

Pampel

F. C.

Krueger

Denney

(2010). Socioeconomic disparities in health behaviors. Annual Review of Sociology, 36, 349-370.

33.

Patel

K. V.

Coppin

A. K.

Manini

T. M.

Lauretanim

Bandinelli

Ferrucci

Guralnik

J. M.

(2006). Midlife physical activity and mobility in older age: The InCHIANTI Study. American Journal of Preventive Medicine, 31, 217-224.

34.

Peterson

M. J.

Giuliani

Morey

M. C.

Pieper

C. F.

Evenson

K. R.

Mercer

. . . Simonsick

E. M.

(2009). Physical activity as a preventative factor for frailty: The health, aging, and body composition study. The Journals of Gerontology, Series A: Biological Sciences & Medical Sciences, 64A, 61-68.

35.

Rhodes

R. E.

Smith

N. E. I.

(2006). Personality correlates of physical activity: A review and meta-analysis. British Journal of Sports Medicine, 40, 958-965.

36.

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, 2, 313-345.

37.

Seefeldt

Malina

R. M.

Clark

(2002). Factors affecting levels of physical activity in adults. Sports Medicine, 32, 143-168.

38.

Sutin

A. R.

Evans

M. K.

Zonderman

A. B.

(2013). Personality traits and illicit substances: The moderating role of poverty. Drug and Alcohol Dependence, 131, 247-251.

39.

Tucker-Seeley

R. D.

Subramanian

S. V.

Sorensen

(2009). Neighborhood safety, socioeconomic status, and physical activity in older adults. American Journal of Preventive Medicine, 37, 207-213.

40.

Turiano

N. A.

Chapman

B. P.

Agrigoroaei

Infurna

F. J.

Lachman

(2014). Perceived control reduces mortality risk at low, not high, education levels. Health Psychology: Official Journal of the Division of Health Psychology, American Psychological Association, 33, 883-890.

41.

Weiss

Costa

P. T.

(2005). Domain and facet personality predictors of all-cause mortality among Medicare patients aged 65 to 100. Psychosomatic Medicine, 67, 724-733.

42.

Wilson

K. E.

Dishman

R. K.

(2015). Personality and physical activity: A systematic review and meta-analysis. Personality and Individual Differences, 72, 230-242.

43.

World Health Organization. (2014). Global status report on non-communicable diseases 2014. Geneva, Switzerland: WHO Library Cataloguing-in-Publication Data.

44.

Zachau-Christiansen

(1972). Development during the first year of life. Helsingør, Denmark: Poul Andersens Forlag.

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

0.29 MB