Abstract

Background

Associations between personality traits, mental wellbeing and good health behaviours were examined to understand further the social and psychological context of the health divide.

Methods

In a cross-sectional study, 666 subjects recruited from areas of high and low socioeconomic deprivation had personality traits and mental wellbeing assessed, and lifestyle behaviours quantified. Regression models (using deprivation as a moderating variable) assessed the extent to which personality traits and mental wellbeing predicted health behaviour.

Results

Deprived (vs. affluent) subjects exhibited similar levels of extraversion but higher levels of neuroticism and psychoticism, more hopelessness, less sense of coherence, lower self-esteem and lower self-efficacy (all P< 0.001). They ate less fruit and vegetables, smoked more and took less aerobic exercise (all P< 0.001). In the deprived group, personality traits were significantly more important predictors of mental wellbeing than in the least deprived group (P< 0.01 for interaction), and mental wellbeing and extraversion appeared more strongly related to good health behaviours.

Conclusions

Persistence of a social divide in health may be related to interactions between personality, mental wellbeing and the adoption of good health behaviours in deprived areas. Effectiveness of health messages may be enhanced by accommodating the variation in the levels of extraversion, neuroticism, hopelessness and sense of coherence.

Introduction

Associations between personality, mental wellbeing, socioeconomic status (SES) and health have been well documented.1,2 Personality is assessed commonly using either a five-factor3 (comprising neuroticism, extraversion, agreeableness, openness to experience and conscientiousness) or a three-factor (neuroticism, extraversion and psychoticism) model4 and links can be seen with health related decision-making.5,6 Personality traits show consistent associations with smoking, diet and exercise. Smoking has been associated with higher levels of neuroticism,7 with extraversion8,9 and psychoticism.9 Lower levels of agreeableness and conscientiousness also predict smoking behaviour.7 A greater propensity to exercise has been associated with lower levels of neuroticism10,11 and with higher extraversion.10 Similarly, higher conscientiousness is associated with a greater likelihood of exercising and consuming fruit and vegetables.8 The influence of neuroticism is more complex: it has been associated with higher scores on dietary restraint12 but also with a tendency to increase emotional eating.13 These relationships between personality and health behaviours may explain, in part, why personality predicts mortality and morbidity,14–18 possibly independent of SES and level of social support.19,20

Personality factors associated with health-related behaviours may also help explain why certain sub-groups within the population experience significantly better, or worse, health than others. It may be significant that low SES is associated with high levels of neuroticism,21,22 low levels of conscientiousness,21 higher hostility23 and depression,24 the latter dispositions reflecting lower mental wellbeing. The concept of sense of coherence (SoC)25 is of particular interest in this context because of its significant association both with mental wellbeing26 and SES.27 SoC does not define a personality type but is rather a disposition which characterizes the individual's confidence that their internal and external environments are comprehensible, manageable and meaningful. High SoC is associated with lower levels of psychological morbidity, lower trait anxiety and better self-reported health status26,27 The fact that income, education and extended social networks all contribute positively to SoC28,29 would confirm the relevance of SES to the factor: broadly, low SES is associated with lower SoC.27 Whilst SoC shows relative stability in adulthood,30 it can change if significant and enduring events occur in personal circumstances.31 It has been proposed that the potential for SoC to change may be of relevance in health promotion in that suitably tailored interventions may then be devised to provide positive experiences, and support, that allow strengthening of an individual's SoC with positive consequences for health.28,29

The rationale for the present study was to investigate to what extent personality traits and mental wellbeing influence uptake of positive health behaviours, and to test the hypothesis that these characteristics have a differential impact on affluent and deprived communities. Subjects participating in the ‘Psychological, Social and Biological Determinants of Ill-health’ (pSoBid) study were recruited from the ends of the socioeconomic gradient in a large Scottish city,32 and we have reported previously that those from deprived areas fare significantly less well than their more affluent counterparts on a range of biological measures that influence health.33,34 The current analysis investigates the association between personality traits, mental wellbeing and the uptake of healthy living advice (eating fruit and vegetables, smoking cessation and aerobic exercise) in the two social groups.

Methods

Ethical approval

The study was approved by the Glasgow Royal Infirmary Research Ethics Committee and all participants gave written informed consent.

Study population and protocol

The design of the pSoBid study, including the recruitment strategy, response rates and study protocol has been described elsewhere.32–34 Briefly, selection of subjects was based on the Scottish Index of Multiple Deprivation (SIMD) 2004,35 which ranks small areas on the basis of multiple deprivation indicators. Subjects were recruited from five general practices that served the bottom 5% of SIMD (i.e. relatively deprived) and a further five practices in areas classified as being in the top 20% of the SIMD (i.e. relatively affluent). Between December 2005 and May 2007 we recruited approximately equal numbers from both areas, equal numbers of males and females and equal numbers from each age group (35–44, 45–54 and 55–64 years old).

At Visit 1, participants completed lifestyle and psychology questionnaires. The lifestyle questionnaire included questions on physical activity, alcohol intake, dietary habits and smoking behaviour. Psychological questionnaires completed at this visit examined the affective state and coping/control. Assessment included completion of the General Health Questionnaire-28 (GHQ-28),36 the Generalized Self-Efficacy Scale (GSS),37 the SoC Scale25 and the Beck Hopelessness Scale (BHS).38

Two weeks later at Visit 2, participants completed the Rosenberg Self-esteem Scale (RSES)39 and Eysenck Personality Questionnaire (EPQ-R)4 which comprises measures of extraversion, neuroticism and psychoticism. The scale includes a so-called ‘Lie’ scale to detect those who may attempt to answer so as to portray themselves in a socially acceptable way.

Quantification of health behaviours

A score for the consumption of fruit and vegetables was calculated from participants' self-reported intake of a range of 21 food categories.40 Responses for each question ranged from the number of portions consumed per day, weekly or monthly. Participants selected one response per food category. Responses to four questions from the food frequency questionnaire relating to fruit and vegetable intake were aggregated to give an overall monthly diet score, i.e. frequency of intake of fresh fruit, cooked green vegetables, cooked root vegetables and raw vegetables or salad.

Questions on habitual physical activity at work and in recreation allowed participants to be classified as inactive, moderately inactive, moderately active or active.41 For the present analysis, the number of hours per month when each participant undertook vigorous physical activity (defined as undertaking activities vigorous enough to cause sweating or a faster heartbeat) was also calculated. Participants' smoking behaviours were assessed according to whether they had ever smoked regularly (at least one cigarette a day for 12 months or more), the material smoked and the age at which they had started and stopped smoking if applicable.

Statistical analysis

Descriptive statistics are presented as mean (SD) for continuous variables and count (%) for categorical variables. Continuous variables were compared between the most deprived and the least deprived group using t-tests or Wilcoxon tests as appropriate. Categorical variables were compared between the most deprived and the least deprived groups using Fisher's exact test. The BHS score was log-transformed. Associations between measures of mental wellbeing (outcomes) and personality traits (predictors), and then between health behaviours (outcomes) and mental wellbeing scores and personality traits (predictors) were assessed using linear or logistic regression models, including interactions between predictor variables and deprivation. All models were adjusted for age, sex and years of education and included deprivation as a moderator variable. Results are presented as effect estimates and 95% confidence intervals within each deprivation group, and P-values for tests of interaction. Analyses were conducted in R for Windows v. 2.9.42

Results

Of the 2712 subjects invited, 666 participated and attended both study visits, giving an overall response rate of 24.6%. By design, there were approximately equal numbers of men and women in each of the three age groups: 342 were drawn from the least deprived areas and 324 from the most deprived. For the least deprived group as a whole the response rate was 33.9% and for the most deprived group of 19.0%.

Significant differences (Table 1) were found between the most and least deprived groups in health behaviours (smoking, exercise and diet indices) and in indicators of mental wellbeing (SoC, self-esteem, hopelessness and self-efficacy, GHQ scores). In terms of personality factors deduced from the responses to the EPQ,4 the most deprived group showed significantly higher levels of neuroticism and psychoticism than those from least deprived areas. The groups did not differ in mean extraversion or in ‘lie’ scores.

Table 1

Description of basic demographics, socioeconomic status and mental wellbeing by area deprivation category

 SIMD least deprived (n= 342) SIMD most deprived (n= 324) Pa 
Study population demographics 
 Age (years) 51.77 (8.03)b 51.46 (8.48) 0.63 
 Gender (male/female) 171/171 156/168 0.64 
Markers of individual SES 
 Average household income (£) £41 699 £16 461 <0.0001 
 Total education (years) 16.1 (3.64) 11.8 (2.49) <0.0001 
 Current home status (owner/tenant) (%) 97.7/2.3 29.9/70.1 <0.0001 
 Participant occupation categoryc (I and II/III/IV and V/unemployed) (%) 74/22/3.6/0.3 20/43/33/0.6 <0.0001 
Indices of health behaviour 
 Cigarette smoker (never/former/current) (%) 66.3/27.4/6.3 28.8/26.8/44.6 <0.0001 
 Physical activity (inactive/mod inactive/mod active/active) (%)d 23.9/24.6/25.4/26 49.4/11.4/21.9/17.3 <0.0001 
 Fruit and vegetable consumption (portions per month) 95.7 (51.5) 59.9 (50.4) <0.0001 
EPQ-R scorese 
 Neuroticism 4.06 (3.19) 5.96 (3.79) <0.0001 
 Extraversion 7.49 (3.41) 7.34 (3.61) 0.58 
 Psychoticism 1.26 (1.30) 2.58 (2.02) <0.0001 
 Lie 5.35 (2.68) 5.34 (2.78) 0.95 
Mental wellbeing scores 
  BHS (missing data n= 38) 2.82 (3.24) 5.12 (4.81) <0.0001 
 SoC (missing data n= 12) 70.31 (11.34) 59.63 (15.33) <0.0001 
 RESf (missing data n= 17) 17.49 (4.48) 20.78 (5.32) <0.0001 
 GSS (missing data n= 7) 32.74 (4.42) 30.08 (6.14) <0.0001 
 GHQ totalg (missing data, n= 27) 2.53 (4.06) 5.19 (6.87) <0.0001 
 SIMD least deprived (n= 342) SIMD most deprived (n= 324) Pa 
Study population demographics 
 Age (years) 51.77 (8.03)b 51.46 (8.48) 0.63 
 Gender (male/female) 171/171 156/168 0.64 
Markers of individual SES 
 Average household income (£) £41 699 £16 461 <0.0001 
 Total education (years) 16.1 (3.64) 11.8 (2.49) <0.0001 
 Current home status (owner/tenant) (%) 97.7/2.3 29.9/70.1 <0.0001 
 Participant occupation categoryc (I and II/III/IV and V/unemployed) (%) 74/22/3.6/0.3 20/43/33/0.6 <0.0001 
Indices of health behaviour 
 Cigarette smoker (never/former/current) (%) 66.3/27.4/6.3 28.8/26.8/44.6 <0.0001 
 Physical activity (inactive/mod inactive/mod active/active) (%)d 23.9/24.6/25.4/26 49.4/11.4/21.9/17.3 <0.0001 
 Fruit and vegetable consumption (portions per month) 95.7 (51.5) 59.9 (50.4) <0.0001 
EPQ-R scorese 
 Neuroticism 4.06 (3.19) 5.96 (3.79) <0.0001 
 Extraversion 7.49 (3.41) 7.34 (3.61) 0.58 
 Psychoticism 1.26 (1.30) 2.58 (2.02) <0.0001 
 Lie 5.35 (2.68) 5.34 (2.78) 0.95 
Mental wellbeing scores 
  BHS (missing data n= 38) 2.82 (3.24) 5.12 (4.81) <0.0001 
 SoC (missing data n= 12) 70.31 (11.34) 59.63 (15.33) <0.0001 
 RESf (missing data n= 17) 17.49 (4.48) 20.78 (5.32) <0.0001 
 GSS (missing data n= 7) 32.74 (4.42) 30.08 (6.14) <0.0001 
 GHQ totalg (missing data, n= 27) 2.53 (4.06) 5.19 (6.87) <0.0001 

aValues are presented as mean (SD) for all participants or as percentages for categorical variables, adjusted for age and sex.

bP relates the comparison between the two groups. Categorical variables were compared using Fisher's exact test and continuous variables were compared using t-tests or Wilcoxon tests as appropriate.

cParticipants occupational category, data unclassifiable: least deprived 0.3% (n= 1); most deprived 5% (n= 16). Occupation classified using the Registrar General Social Class Classification. Occupational Social Class classified on the basis of current job or, if not currently working, on the basis of participants' last paid job. Only those who had never been in paid employment were classed as ‘unemployed.’ I, professional occupations; II, managerial and technical occupations; III, manual and non-manual skilled occupations; IV, partly skilled occupations; V, unskilled occupations.

dThe physical activity level is a combination of activity at work and recreational exercise. Activity at work classified as sedentary, moderately active or active. Recreational exercise classified as none, moderate (>0, <0.25 h/day average), active (>0.25, <0.5 h/day average) and very active (>0.5 h/day average). Overall activity is classified as inactive (sedentary occupation and no recreational exercise), moderately inactive (sedentary work and moderate exercise or moderately active work and no exercise), moderately active (either sedentary work and recreational exercise active or moderately active work and recreational exercise moderate or active or active work and no recreational exercise) or active (everyone else).

ePersonality trait scores were self-reported, each on a scale of 1–12.

fOn the RSES, high scores relate to lower self-esteem.

gHigher GHQ scores reflect poorer mental health.

The distribution of extraversion scores (Supplementary data, Fig. S1) was similar in the groups, while a greater proportion of the most deprived subjects scored higher on neuroticism. In the case of psychoticism, both groups were characterized by low scores but there was a skew towards higher scores in the deprived group. The distribution of mental wellbeing scores also differed between the two groups with the least deprived group having lower levels of hopelessness and self-esteem (note that on the RSES high scores denote low self-esteem), a greater SoC and increased self-efficacy.

Personality, deprivation and mental wellbeing

Figure 1 and Table 2 examine the extent to which personality traits predicted mental wellbeing in affluent and deprived subjects. The impact of personality, extraversion in particular, appeared different in the two groups and formal tests of interaction with deprivation category were significant for hopelessness versus neuroticism and extraversion (P= 0.0017 and P< 0.001, respectively), for SoC (P< 0.001 for extraversion) and self-esteem (P= 0.006 for extraversion). High levels of neuroticism and low levels of extraversion were associated strongly in both groups with increased hopelessness, reduced SoC, reduced self-esteem and reduced self-efficacy. However, it was also seen that in the most deprived, individuals who were high in neuroticism or low in extraversion reported lower mental wellbeing (a greater degree of hopelessness, lower SoC and less self-esteem) than their counterparts in the least deprived group. It was noteworthy that those in the deprived group with low levels of neuroticism (score of < 4) and those with high levels of extraversion (scores > 6) exhibited a degree of mental wellbeing similar to that in the least deprived group.

Table 2

Relationship of personality traits and deprivation to mental wellbeing

Mental wellbeing scale EPQ trait Effect (95% CI)a
 
Interaction P-value 
Least deprived Most deprived 
BHS Neuroticism 0.066 (0.043, 0.089) 0.103 (0.082, 0.124) 0.017 
Extraversion −0.025 (−0.047, −0.003) −0.092 (−0.114, −0.069) <0.001 
Psychoticism 0.019 (−0.041, 0.080) 0.039 (−0.004, 0.081) 0.605 
SoC Neuroticism −2.158 (−2.528, −1.787) −2.418 (−2.753, −2.084) 0.297 
Extraversion 0.535 (0.135, 0.935) 1.567 (1.166, 1.968) <0.001 
Psychoticism −0.603 (−1.705, 0.499) −0.869 (−1.619, −0.118) 0.693 
RSES Neuroticism 0.749 (0.612, 0.887) 0.910 (0.787, 1.033) 0.082 
Extraversion −0.368 (−0.511, −0.225) −0.652 (−0.795, −0.510) 0.006 
Psychoticism −0.093 (−0.498, 0.312) 0.169 (−0.116, 0.454) 0.293 
GSS Neuroticism −0.524 (−0.685, −0.364) −0.724 (−0.866, −0.581) 0.064 
Extraversion 0.430 (0.274, 0.586) 0.524 (0.370, 0.678) 0.400 
Psychoticism 0.344 (−0.089, 0.777) −0.110 (−0.413, 0.193) 0.089 
Mental wellbeing scale EPQ trait Effect (95% CI)a
 
Interaction P-value 
Least deprived Most deprived 
BHS Neuroticism 0.066 (0.043, 0.089) 0.103 (0.082, 0.124) 0.017 
Extraversion −0.025 (−0.047, −0.003) −0.092 (−0.114, −0.069) <0.001 
Psychoticism 0.019 (−0.041, 0.080) 0.039 (−0.004, 0.081) 0.605 
SoC Neuroticism −2.158 (−2.528, −1.787) −2.418 (−2.753, −2.084) 0.297 
Extraversion 0.535 (0.135, 0.935) 1.567 (1.166, 1.968) <0.001 
Psychoticism −0.603 (−1.705, 0.499) −0.869 (−1.619, −0.118) 0.693 
RSES Neuroticism 0.749 (0.612, 0.887) 0.910 (0.787, 1.033) 0.082 
Extraversion −0.368 (−0.511, −0.225) −0.652 (−0.795, −0.510) 0.006 
Psychoticism −0.093 (−0.498, 0.312) 0.169 (−0.116, 0.454) 0.293 
GSS Neuroticism −0.524 (−0.685, −0.364) −0.724 (−0.866, −0.581) 0.064 
Extraversion 0.430 (0.274, 0.586) 0.524 (0.370, 0.678) 0.400 
Psychoticism 0.344 (−0.089, 0.777) −0.110 (−0.413, 0.193) 0.089 

aRegression analyses are presented as effect estimates and 95% confidence intervals. P-values for test of interaction of personality trait and deprivation. Each model included deprivation and personality trait and their interaction and additionally adjusts for age, sex and years of education.

Fig. 1

Interaction of personality traits with deprivation in determining mental wellbeing. Regression models were constructed, which related the outcomes of predicted value for hopelessness (BHS), SoC scale, self-esteem (RSES) and self-efficacy (GSS) to scores for neuroticism, extraversion and psychoticism. The models were adjusted for age, sex and years of education. Deprivation was included as a moderating variable. The BHS score was log-transformed. P value quoted is for a test of interaction between the two subject groups.

Fig. 1

Interaction of personality traits with deprivation in determining mental wellbeing. Regression models were constructed, which related the outcomes of predicted value for hopelessness (BHS), SoC scale, self-esteem (RSES) and self-efficacy (GSS) to scores for neuroticism, extraversion and psychoticism. The models were adjusted for age, sex and years of education. Deprivation was included as a moderating variable. The BHS score was log-transformed. P value quoted is for a test of interaction between the two subject groups.

Mental wellbeing, deprivation and positive health behaviours

Consumption of fruit and vegetables

The most deprived group ate on average one-third fewer portions of fruit and vegetables per month than the least deprived (Table 1). While none of the tests of interaction with deprivation was significant in the regression models in Table 3, consumption of fruit and vegetables was lower in those expressing increased hopelessness and greater in those with higher SoC, self-esteem and self-efficacy. In general, within the least deprived group, mental wellbeing scores were not linked significantly to fruit and vegetable consumption, whereas in the most deprived group these associations were statistically significant.

Table 3

Relationship between mental wellbeing and personality traits to health behaviours in affluent and deprived groups

 Effect (95% CI)b
 
Interaction P-value 
Least deprived Most deprived 
Fruit and vegetable consumptiona 
 Mental wellbeing scale       
  BHS −0.878 (−2.525, 0.769) −2.041 (−3.211, −0.870) 0.258 
  SoC 0.067 (−0.404, 0.539) 0.414 (0.050, 0.778) 0.251 
  RSES −0.372 (−1.579, 0.835) −1.298 (−2.354, −0.242) 0.255 
  GSS 0.900 (−0.308, 2.109) 1.655 (0.751, 2.559) 0.322 
 Eysenck personality trait 
  Neuroticism −0.206 (−1.954, 1.542) −1.460 (−3.005, 0.085) 0.283 
  Extraversion 1.078 (−0.526, 2.683) 2.431 (0.861, 4.000) 0.237 
  Psychoticism −1.323 (−5.522, 2.877) −0.987 (−3.817, 1.843) 0.895 
Smoking cessationc 
 Mental wellbeing scale 
  BHS 0.006 (−0.155, 0.66) −0.040 (−0.104, 0.025) 0.604 
  SoC −0.032 (−0.080, 0.015) 0.024 (0.002, 0.045) 0.034 
  RSES 0.001 (−0.110, 0.111) −0.031 (−0.088, 0.026) 0.619 
  GSS −0.018 (−0.126, 0.090) 0.053 (0.002, 0.105) 0.244 
 Eysenck personality trait 
  Neuroticism 0.045 (−0.101, 0.192) −0.001 (−0.084, 0.082) 0.584 
  Extraversion 0.047 (−0.092, 0.186)  0.043 (−0.040, 0.126) 0.963 
  Psychoticism −0.11. (−0.448, 0.223) −0.092 (−0.250, 0.066) 0.910 
Vigorous aerobic exercised 
 Mental wellbeing scale 
  BHS −0.001 (−0.069, 0.066) −0.093 (−0.148, −0.037) 0.040 
  SoC 0.002 (−0.017, 0.021) 0.019 (0.003, 0.034) 0.180 
  RSES −0.043 (−0.093, 0.006) −0.071 (−0.116, −0.025) 0.422 
  GSS 0.035 (−0.015, 0.085) 0.048 (0.008, 0.088) 0.687 
 Eysenck personality trait 
  Neuroticism −0.008 (−0.078, 0.062) −0.051 (−0.115, 0.012) 0.363 
  Extraversion 0.007 (−0.058, 0.071) 0.087 (0.020, 0.154) 0.090 
  Psychoticism −0.007 (−0.176, 0.163) −0.161 (−0.290, −0.032) 0.151 
       
 Effect (95% CI)b
 
Interaction P-value 
Least deprived Most deprived 
Fruit and vegetable consumptiona 
 Mental wellbeing scale       
  BHS −0.878 (−2.525, 0.769) −2.041 (−3.211, −0.870) 0.258 
  SoC 0.067 (−0.404, 0.539) 0.414 (0.050, 0.778) 0.251 
  RSES −0.372 (−1.579, 0.835) −1.298 (−2.354, −0.242) 0.255 
  GSS 0.900 (−0.308, 2.109) 1.655 (0.751, 2.559) 0.322 
 Eysenck personality trait 
  Neuroticism −0.206 (−1.954, 1.542) −1.460 (−3.005, 0.085) 0.283 
  Extraversion 1.078 (−0.526, 2.683) 2.431 (0.861, 4.000) 0.237 
  Psychoticism −1.323 (−5.522, 2.877) −0.987 (−3.817, 1.843) 0.895 
Smoking cessationc 
 Mental wellbeing scale 
  BHS 0.006 (−0.155, 0.66) −0.040 (−0.104, 0.025) 0.604 
  SoC −0.032 (−0.080, 0.015) 0.024 (0.002, 0.045) 0.034 
  RSES 0.001 (−0.110, 0.111) −0.031 (−0.088, 0.026) 0.619 
  GSS −0.018 (−0.126, 0.090) 0.053 (0.002, 0.105) 0.244 
 Eysenck personality trait 
  Neuroticism 0.045 (−0.101, 0.192) −0.001 (−0.084, 0.082) 0.584 
  Extraversion 0.047 (−0.092, 0.186)  0.043 (−0.040, 0.126) 0.963 
  Psychoticism −0.11. (−0.448, 0.223) −0.092 (−0.250, 0.066) 0.910 
Vigorous aerobic exercised 
 Mental wellbeing scale 
  BHS −0.001 (−0.069, 0.066) −0.093 (−0.148, −0.037) 0.040 
  SoC 0.002 (−0.017, 0.021) 0.019 (0.003, 0.034) 0.180 
  RSES −0.043 (−0.093, 0.006) −0.071 (−0.116, −0.025) 0.422 
  GSS 0.035 (−0.015, 0.085) 0.048 (0.008, 0.088) 0.687 
 Eysenck personality trait 
  Neuroticism −0.008 (−0.078, 0.062) −0.051 (−0.115, 0.012) 0.363 
  Extraversion 0.007 (−0.058, 0.071) 0.087 (0.020, 0.154) 0.090 
  Psychoticism −0.007 (−0.176, 0.163) −0.161 (−0.290, −0.032) 0.151 
       

aMonthly fruit and vegetable consumption.

bRegression analyses are presented as effect estimates and 95% confidence intervals. P-values for test of interaction of personality trait/mental wellbeing scale and deprivation. Each model included deprivation and personality trait/mental wellbeing scale and their interaction and additionally adjusts for age, sex and years of education.

cPredicted possibility in those who ever smoked of becoming a former smoker.

dNumber of hours per month of vigorous aerobic exercise.

Extraversion was related significantly to fruit and vegetable consumption only in the most deprived group (P= 0.002), but again a test of interaction was not significant (P= 0.237) (Table 3). Neither neuroticism nor psychoticism was linked with fruit and vegetable consumption in either social group. In a multivariate model of fruit and vegetable consumption, only higher self-efficacy in both groups (P= 0.003) and greater extraversion in people living in deprivation (P= 0.031) emerged as statistically significant predictors (Supplementary data, Table S1).

Smoking

Table 1 shows that 66.3% of the least deprived group had never smoked, while 71.4% of the most deprived had smoked at some time and 44.6% were current smokers. Among those who had smoked at some time in their lives, 81.4 and 37.6% had stopped smoking in the least and most deprived groups, respectively (P< 0.001).

In the deprived subject group those with a higher SoC had greater success in giving up (P= 0.034 for interaction with deprivation category, Table 3). A similar association seemed to be present for those with greater self-efficacy; however, interaction was not statistically significant for this variable. In general, individuals living in less deprived areas were more likely to stop smoking but this tendency showed no association with any of the tested predictors.

Aerobic exercise

Only the association of aerobic exercise with hopelessness showed a statistically significant interaction between deprivation groups (P= 0.040, Table 3). In the least deprived group, participation in regular exercise was high and unrelated to mental wellbeing or personality. Levels of aerobic exercise in the most deprived group were lower (Table 1), particularly so for those with higher levels of hopelessness, lower SoC, self-esteem or self-efficacy, lower extraversion or higher psychoticism.

Discussion

Main findings of this study

In general the pSoBid cohort exhibited the same relationships between personality traits,4 health behaviours and SES that have been reported previously.7–11 Participants from areas of multiple deprivation had significantly higher levels of neuroticism and psychoticism than their more affluent counterparts. Similarly, a significantly lower level of mental wellbeing among the more deprived group, and its association with high neuroticism and low extraversion, is consistent with existing evidence.21,22 The higher prevalence of harmful health behaviours (smoking, poor diet and lack of exercise) in the most deprived group, again, replicates previous research.43–47

A novel finding is that personality traits appeared to have a significantly greater impact on mental wellbeing among participants from more deprived circumstances. Further, personality and wellbeing impacted more on the pattern of health behaviours in this group compared with their more affluent counterparts. In the more deprived group, mental wellbeing—low hopelessness and a high degree of self-esteem, SoC and self-efficacy—and high extraversion were significant predictors of consumption of fruit and vegetables. In contrast no personality trait or aspect of mental wellbeing appeared to predict this health behaviour in the more affluent group. Success in smoking cessation was associated in the most deprived (but not the least deprived) with SoC and self-efficacy. We did not see an influence of personality trait on smoking cessation in either group. Engagement in aerobic exercise was lower in the more deprived group and was associated positively in that group with high extraversion, SoC, self-esteem and self-efficacy and inversely to high psychoticism and hopelessness.

What is already known on this topic

The nearest parallel report on diet is the finding that a high level of conscientiousness predicts fruit and vegetable consumption in a sample of college students.8 The three-factor model of personality applied in the present study does not permit extraversion to be decomposed into ‘sub-facets’ or traits as in the case of the five-factor model3,7 so it is more difficult to determine the basis of the association. However, some insight is provided by the measures of mental wellbeing which variously reflect the traits of optimism, assertiveness and goal-directed activity associated with high extraversion and which may plausibly underlie the effect.

Our findings on smoking contrast with those of Terracciano and Costa7 who found that current smokers scored higher on neuroticism than those who had never smoked. Similarly, Arai et al.9 found ex-smokers to give higher scores for extraversion and psychoticism than those who had never smoked, and our results showed that personality was not a predictor of ‘never smoking’. Neither of these other studies considered the factor of SES.

The beneficial influence of extraversion on the propensity to exercise has been reported before.10 The lower level of exercise associated with greater hopelessness in the deprived group is consistent with the apathy and inactivity associated with negative affect.10,11 Personality traits and measures of mental wellbeing appeared not to be related to involvement in exercise in the affluent group.

What this study adds

Our findings provide support for the proposal11,14,18 that more attention should be paid to personality traits and parameters of mental wellbeing when designing health promotion activities. In the case of personality, the practicality of such a proposal might be questioned in the light of the commonly held assumption that adult personality is stable and largely unchangeable, and hence impervious to interventions that might modify those traits associated with negative health behaviours. However, this objection has been well addressed by Roberts et al.14 who cite evidence that personality traits do change and may be modifiable.48,49 They suggest that a focus upon social factors that may modify personality traits might have broad-reaching effects across the individual's activities. Although their proposal is cogently made, the practical, and indeed ethical, complexities of developing and implementing such an intervention programme might seem formidable.

There might, however, be greater scope for less controversial intervention through the manipulation of resources and experiences that influence the cognitions that underlie some of the measures of mental wellbeing employed in this study. For example, it was evident, within the most deprived group, that an individual's SoC was a significant determinant of their success in smoking cessation, engaging in aerobic exercise and uptake of a healthy diet. Moreover, it was noted above that while SoC shows some stability in adulthood, it is amenable to change in the light of significant and influential life experiences;30,31 Nilsson et al.31 have provided data to support the conclusions of a 2006 systematic review28 that factors which are known to promote health, such as educational provision and social support, may act to enhance SoC with positive consequences for self-reported health.

It is recognized that the adoption of more health-enhancing behaviours may be more difficult to achieve in poorer communities among individuals expressing low extraversion, high hopelessness and a low SoC. However, the present results may add further support to the importance of accounting for individual differences. Interventions may be more effective when they are adapted to certain personality characteristics and have a focus upon supporting and enhancing those aspects of mental wellbeing such as SoC which have a demonstrated positive association with health.

Limitations of this study

Selected from ends of the socioeconomic gradient,28 subjects in our sample may not represent the population as a whole. Further, there is possible response bias, particularly due to the difficulties in recruiting younger men from the most deprived areas. To explore this, we examined the characteristics of non-respondents and found that within each age, sex and socioeconomic stratum participants were comparable to non-participants.25 The use of the three-factor measure of personality was also a limitation in that we were unable to assess the impact of conscientiousness which has been reported to be an important determinant of many health behaviours. Finally, as a cross-sectional study it is not possible to infer causality from the observed statistical relationships, nor the differences in associations between deprivation groups, not all of which constituted statistically significant interactions, and which were not adjusted for multiple statistical testing. Nevertheless, the consistent pattern of relatively stronger associations within the more deprived group, compared with no or weaker associations in the less deprived group, adds weight to the hypothesis that personality traits and mental wellbeing are more important determinants of health behaviours within areas of high socioeconomic deprivation. We propose this should be the focus of future research.

Supplementary data

Supplementary data are available at the Journal of Public Health online.

Authors' contributions

C.J.P., J.C., J.S.M., A.M., C.M.M., G.D.B., H.B., K.A.D., N.S., P.G.S., Y.N.V., C.T. and K.M. contributed equally to conception, design and final approval of the version of the manuscript. C.J.P., K.M. and J.S.M. have been involved in drafting the manuscript and revising it critically for important intellectual content. A.M. and C.M.M. performed the statistical analysis. Y.N.V. supervised the recruitment of subjects and data collection.

Funding

This work was supported by the Glasgow Centre for Population Health which is a partnership between NHS Greater Glasgow and Clyde, Glasgow City Council and the University of Glasgow, supported by the Scottish Government.

Acknowledgements

Thanks are due to administrative staff in the Glasgow Centre for Population Health; Robertson Centre for Biostatistics, University of Glasgow for data management, statistical support and analysis; Health Information & Technology section of NHS Greater Glasgow & Clyde for sample selection and analysis; Alasdair Buchanan for GPASS data and analysis; all members of GP Practices who participated in the study and the participants themselves.

References

1
George
LK
The impact of personality and social status factors upon levels of activity and psychological well-being
J Gerontol
 , 
1978
, vol. 
33
 (pg. 
840
-
7
)
2
Chapman
BP
Fiscella
K
Kawachi
I
, et al.  . 
Personality, socioeconomic status and all-cause mortality in the United States
Am J Epidemiol
 , 
2010
, vol. 
171
 (pg. 
83
-
92
)
3
Costa
PT
McCrae
RR
Revised NEO Personality Inventory (NEO PI-R) and NEO-Five-Factor Inventory (NEO-FFI)
 , 
1992
Odessa
Psychological Assessment Resources Inc
4
Eysenck
HJ
Eysenck
SBG
Manual of the Eysenck Personality Scales
 , 
1991
London
Hodder and Stoughton
5
Flynn
KE
Smith
MA
Personality and health care decision-making style
J Gerontol B
 , 
2007
, vol. 
41
 (pg. 
772
-
86
)
6
Chapman
BP
Shah
M
Friedman
B
, et al.  . 
Personality traits predict emergency department utilisation over 3 years in older patients
Am J Ger Psych
 , 
2009
, vol. 
17
 (pg. 
526
-
35
)
7
Terracciano
A
Costa
PT
Jr
Smoking and the five-factor model of personality
Addiction
 , 
2004
, vol. 
99
 (pg. 
472
-
81
)
8
Raynor
DA
Levine
H
Associations between the five-factor model of personality and health behaviors among college students
J Am Coll Health
 , 
2009
, vol. 
58
 (pg. 
73
-
81
)
9
Arai
Y
Hosokawa
T
Fukao
A
, et al.  . 
Smoking behaviour and personality: a population-based study in Japan
Addiction
 , 
1997
, vol. 
92
 (pg. 
1023
-
33
)
10
De Moor
MHM
Beem
AL
Stubbe
JH
, et al.  . 
Regular exercise, anxiety, depression and personality: a population-based study
Prev Med
 , 
2006
, vol. 
42
 (pg. 
273
-
9
)
11
Yeung
RR
Hemsley
DR
Personality, exercise and psychological well-being: static relationships in the community
Person Ind Diff
 , 
1997
, vol. 
22
 (pg. 
47
-
53
)
12
Provencher
V
Bégin
C
Gagnon-Girouard
MP
, et al.  . 
Personality traits in overweight and obese women: associations with BMI and eating behaviours
Eat Behav
 , 
2008
, vol. 
9
 (pg. 
294
-
302
)
13
Elfhag
K
Morey
LC
Personality traits and eating behaviour in the obese: poor self-control in emotional and external eating but personality assets in restrained eating
Eat Behav
 , 
2008
, vol. 
9
 (pg. 
285
-
93
)
14
Roberts
BW
Kuneel
NR
Shiner
B
, et al.  . 
The power of personality. The comparative validity of personality traits, socioeconomic status, and cognitive ability for predicting important life outcomes
Perspect Psychol Sci
 , 
2007
, vol. 
2
 (pg. 
313
-
45
)
15
Terracciano
A
Löckenhoff
CE
Zonderman
AB
, et al.  . 
Personality predictors of longevity: activity, emotional stability and conscientiousness
Psychosom Med
 , 
2008
, vol. 
70
 (pg. 
621
-
7
)
16
Nabi
H
Kivimäki
M
Zins
M
, et al.  . 
Does personality predict mortality? Results from the GAZEL French prospective cohort study
Int J Epidemiol
 , 
2008
, vol. 
37
 (pg. 
386
-
96
)
17
Hudek-Knezević
J
Kardum
I
Five-factor personality dimensions and three health-related personality constructs as predictors of health
Croat Med J
 , 
2009
, vol. 
50
 (pg. 
394
-
402
)
18
Shipley
BA
Weiss
A
Der
G
, et al.  . 
Neuroticism, extraversion and mortality in the UK Health and Lifestyle Survey: a 21 year prospective cohort study
Psychosom Med
 , 
2007
, vol. 
69
 (pg. 
923
-
31
)
19
Korten
AE
Jorm
AF
Letenneur
L
, et al.  . 
Health, cognitive and psychosocial factors as predictors of mortality in an elderly community sample
J Epidemiol Com Health
 , 
1999
, vol. 
53
 (pg. 
83
-
8
)
20
Weiss
A
Costa
PT
Domain and facet personality predictors of all-cause mortality among Medicare patients aged 65 to 100
Psychosom Med
 , 
2005
, vol. 
67
 (pg. 
724
-
33
)
21
Jonassaint
CR
Siegler
IC
Barefoot
JC
, et al.  . 
Low life course socioeconomic status (SES) is associated with negative NEO PI-R personality patters
Int J Behav Med
 , 
2011
, vol. 
18
 (pg. 
13
-
21
)
22
Bosma
H
van de Mheen
HD
Mackenbach
JP
Social class in childhood and general health in adulthood: questionnaire study of contribution of psychological attributes
BMJ
 , 
1999
, vol. 
318
 (pg. 
18
-
22
)
23
Kubzansky
L
Kawachi
I
Sparrow
D
Socioeconomic status, hostility, and risk factor clustering in the normative aging study: any help from the concept of allostatic load?
Ann Behav Med
 , 
1999
, vol. 
21
 (pg. 
330
-
8
)
24
Harper
S
Lynch
J
Hsu
WL
, et al.  . 
Life course socioeconomic conditions and adult psychosocial functioning
Int J Epidemiol
 , 
2002
, vol. 
31
 (pg. 
395
-
403
)
25
Antonovsky
A
The structure and properties of the sense of coherence scale
Soc Sci Med
 , 
1993
, vol. 
36
 (pg. 
725
-
33
)
26
Pallant
JF
Lae
L
Sense of coherence, coping and personality factors: further evaluation of the sense of coherence scale
Person Ind Diff
 , 
2002
, vol. 
33
 (pg. 
39
-
48
)
27
Larsson
G
Kallenberg
KO
Sense of coherence, socio-economic conditions and health
Eur J Pub Health
 , 
1996
, vol. 
6
 (pg. 
175
-
80
)
28
Eriksson
M
Lindström
B
Antonovsky's sense of coherence scale and the relation with health: a systematic review
J Epidemiol Com Health
 , 
2006
, vol. 
60
 (pg. 
376
-
81
)
29
Richardson
CG
Ratner
PA
Sense of coherence as a moderator of the effects of stressful life events on health
J Epidemiol Com Health
 , 
2005
, vol. 
59
 (pg. 
979
-
84
)
30
Eriksson
M
Lindström
B
Validity of Antonovsky's sense of coherence scale: a systematic review
J Epidemiol Com Health
 , 
2005
, vol. 
59
 (pg. 
460
-
6
)
31
Nilsson
KW
Leppert
J
Simonsson
B
, et al.  . 
Sense of coherence and psychological well-being: improvement with age
J Epidemiol Com Health
 , 
2010
, vol. 
64
 (pg. 
347
-
52
)
32
Velupillai
YN
Packard
CJ
Batty
GD
, et al.  . 
Psychological, social and biological determinants of ill health (pSoBid): study protocol of a population based study
BMC Public Health
 , 
2008
, vol. 
8
 pg. 
126
 
33
Deans
KA
Bezylak
V
Ford
I
, et al.  . 
Area-based socioeconomic differentials in atherosclerosis are not explained by traditional or emerging risk factors in the psychosocial and biological determinants of Ill-health (pSoBid) study—a cross-sectional population-based study
BMJ
 , 
2009
, vol. 
339
 pg. 
b4170
 
34
Packard
CJ
Bezylak
V
McLean
JS
, et al.  . 
Early life socioeconomic adversity is associated in adult life with chronic inflammation, carotid atherosclerosis, poorer lung function and decreased cognitive performance: a cross-sectional, population-based study
BMC Public Health
 , 
2011
, vol. 
11
 pg. 
42
 
35
Scottish Index for Multiple Deprivation
2004
 
Scottish Government http://www.scotland.gov.uk/Topic/Statistics/SIMD/Overview (April 2005, date last accessed)
36
Goldberg
DP
The Detection of Psychiatric Illness by Questionnaire
 , 
1972
London
Oxford University Press
37
Johnston
M
Wright
SJ
Weinman
J
Generalised Self-efficacy Scale. Measures in Health Psychology: A User's Portfolio
 , 
1995
Windsor
: NFER-Nelson
38
Beck
TA
Steer
RA
Beck Hopelessness Scale
 , 
2007
London
Harcourt Assessment
39
Rosenberg
M
Society and the Adolescent Self-image
 , 
1965
Princeton, NJ
Princeton University Press
40
Shiels
PG
McGlynn
LM
McIntyre
A
, et al.  . 
Accelerated telomere attrition is associated with relative household income, diet and inflammation in the pSoBid cohort
PLoS ONE
 , 
2011
, vol. 
6
 pg. 
e22521
 
41
Khaw
KT
Jakes
R
Bingham
S
, et al.  . 
Work and leisure time physical activity assessed using a simple, pragmatic, validated questionnaire and all-cause mortality in men and women: the European prospective investigation into cancer in Norfolk prospective population study
Int J Epidemiol
 , 
2006
, vol. 
35
 (pg. 
1034
-
43
)
42
R Development Core Team.
R: A Language and Environment for Statistical Computing
 , 
2009
Vienna, Austria
R Foundation for Statistical Computing
 
ISBN 3-900051-07-0 http://www.R-project.org (August 2011, date last accessed).
43
Lantz
PM
House
JS
Lepkowski
JM
, et al.  . 
Chen J
Socioeconomic Factors, Health Behaviours and Mortality. JAMA
 , 
1998
, vol. 
279
 (pg. 
1703
-
8
)
44
Adler
NE
Newman
K
Socioeconomic disparities in health: pathways and policies
Health Aff
 , 
2002
, vol. 
21
 (pg. 
60
-
76
)
45
Wilkinson
MG
Wilkinson
RG
Social Determinants of Health
 , 
2005
London
Oxford University Press
46
Sabia
S
Nabi
H
Kivimaki
M
, et al.  . 
Health behaviours from early to late midlife as predictors of cognitive function: the Whitehall II study
Am J Epidemiol
 , 
2009
, vol. 
170
 (pg. 
428
-
37
)
47
Stringhini
S
Dugravot
A
Shipley
M
, et al.  . 
Health behaviours, socioeconomic status and mortality: further analyses of the British Whitehall II and the French GAZEL prospective cohorts
PLoS Med
 , 
2011
, vol. 
8
 pg. 
e10000419
 
48
Roberts
BW
Walton
KE
Viechtbauer
W
Patterns of mean-level change in personality traits across the life course: a meta-analysis of longitudinal studies
Psychol Bull
 , 
2006
, vol. 
132
 
1
(pg. 
1
-
25
)
49
De Fruyt
F
Van Leeuwen
K
Bagby
RM
, et al.  . 
Assessing and interpreting personality change and continuity in patients treated for major depression
Psychol Assess
 , 
2006
, vol. 
18
 
1
(pg. 
71
-
80
)