Tag Archives: Open-SOEP

Personality and Health- Satisfaction in the SOEP

Research on personality and health has a long tradition. Some research provides support for the concept of a hypochondriac personality. That is, some people react to physical symptoms with extreme distress and they tend to exaggerate the severity of symptoms or their consequences (see Seinfeld episode as an example, clip). Watson and Pennebaker (1989) found that neuroticism is consistently related to subjective health perceptions. Not surprisingly, neuroticism is also a predictor of health-satisfaction ratings (Brief et al., 1993).

One open question in wellbeing science is how personality is related to domain satisfaction (Diener, Lucas, & Oishi, 2018). One possibility is that personality traits like neuroticism influence global life-satisfaction judgments and that global life-satisfaction influences satisfaction with specific life domains. According to this model, life-satisfaction would mediate the relationship between neuroticism and health satisfaction (Heller, Watson, & Illies, 2004). The alternative model assumes that personality influences domain satisfaction and that satisfaction with important life domains leads to higher overall life-satisfaction (Brief et al., 1993; Schimmack, Oishi, & Diener, 2002). So far, empirical studies have been unable to settle these opposing views of the relationship between life-satisfaction and domain satisfaction. The SOEP data provide a unique opportunity in making progress on this front because personality, life-satisfaction and domain satisfaction have been assessed in three waves over an eight-year period.

I already posted analyses of life-satisfaction and job satisfaction (Schimmack, 2019a, 2019b). Here, I present the results for health satisfaction. These results will be used to build a larger model with multiple domains in a single model. The model is identical to the model that was used to analyze life-satisfaction (see OSF for code and detailed results; https://osf.io/vpcfd/ ). Model fit was acceptable, CFI = .97, RMSEA = .022, SRMR = .030.

Results

Observed stability was r = .54 from 2005 to 2009, r = .54 from 2009 to 2013, and r = .48 from 2005 to 2009. It is remarkable that the retest correlation spanning 8 years is just slightly lower than the 4-year retest correlations. Using Heise’s formula, this implies low reliability and high stability; REL = .54*.54/.48 = .61, 8-year stability = .48/.61 = .79. The reliability estimate is consistent with estimates based on annual assessments (Schimmack, Schupp, & Wagner, 2008). Thus, health satisfaction is rather stable and it is worthwhile to examine the predictors of stability in health satisfaction.

Personality measured at Time 1 was used as a predictor of health satisfaction at times 1 to 3. If personality contributes to stability in health satisfaction, personality traits should predict health satisfaction concurrently and prospectively. The results in Table 1 show that this was the case for neuroticism. The remaining personality traits were weak predictors of health satisfaction. Halo bias also predicted stability in health satisfaction but the effect was small and decreased over time. Overall, these results are consistent with the hypochondriac hypothesis.

JS-T1JS-T2JS-T3
Neuroticism-039-0.32-0.31
Extraversion0.080.080.08
Openness
Agreeableness0.090.090.07
Conscientiousness0.030..040..03
Halo0..200.190.14
Acquiescence

Table 2 examines whether changes in personality predict changes in health satisfaction. To do so, health satisfaction was regressed on the residual variances in personality at times 2 and 3.

JS-T2JS-T3
Neuroticism-0.14-0.22
Extraversion 0.010.06
Openness
Agreeableness-0.06-0.02
Conscientiousness0.020.01
Halo0.190.24
Acquiescence

As before, neuroticism and halo bias were the only notable predictors of change in health satisfaction. The results for halo bias show that health satisfaction ratings change as respondents tendencies to respond positively change. The results for neuroticism are more dfficult to interpret. Maybe changes in health status produce changes in neuroticism or changes in neuroticism produce changes in health perceptions. More complex models are needed to disentangle these complex relationships.

The final result was the stability of the residual variance in job satisfaction that is not explained by personality – as measured in the SOEP. Stability estimates were r = .86 and r = .86 over the 4-year intervals with an implied stability of r = .75 over the 8-year interval. Thus, personality is just one predictor of stability in health satisfaction and it contributes a relatively small amount to stability in job satisfaction. Other factors like objective health status may also contribute to stability in health satisfaction.

Conclusion

The results are largely consistent with previous evidence that neuroticism is the main predictor of health satisfaction (Brief et al., 1993). The results show that this relationship holds concurrently and prospectively over an eight-year period and that it holds while controlling for shared method variance in personality and health ratings. These results will be used for a more complex model that can distinguish between top-down and bottom-up effects of health satisfaction and life-satisfaction (Diener et al., 2018).

Personality and Job-Satisfaction in the SOEP

Research on job satisfaction has a long history in applied or industrial/organizational psychology. One line of research examines how environmental factors (job characteristics) influence job satisfaction. Another line of research examines the influence of personality on job satisfaction. Finally, a third line of research examines person x job interaction effects.

Timothy Judge is a leading researcher on personality influences on job satisfaction (see Judge & Kammeyer-Mueller, 2012, for a review). Evidence for personality influences on job satisfaction comes from two lines of research. First, longitudinal studies of job satisfaction show moderate stability over time even when employers change jobs. However, similarity between jobs may contribute to this stability. The second line of research relates measures of personality to job satisfaction. A meta-analysis suggested that the Big Five predict job satisfaction with neuroticism (r = -.29), extraversion (r = .26), and conscientiousness (r = .26) being the strongest predictors (Judge, Heller, & Mount, 2002).

The existing evidence has several limitations. First, meta-analysis combine ad-hoc convenience samples, which makes it unclear how much these results generalize to the general population. Second, meta-analytic studies can be biased by publication bias. Third, simple correlations tend to overestimate effect sizes because they fail to control for shared method variance in personality and job satisfaction measures. Finally, most studies are cross-sectional and do not examine the contribution of personality to stability in job satisfaction.

To address these concerns, I examined the relationship between personality and job satisfaction in the SOEP; a longitudinal panel study with annual assessments of job satisfaction. Personality was assessed three times, four years apart (2005, 2009, 2013). For the present analysis, I limited the data analysis to individuals who reported job satisfaction on all three waves (N = 4,064). This ensures that the sample focuses on employed individuals. For these initial analyses, I did not distinguish between individuals who changed employers and those who stayed in the same job.

The data were analyzed with a three-wave latent variable model that models retest correlations as a function of reliability and stability (Heise, 1969). This model makes it possible to estimate the reliability of single-item measures. Thus, the substantial correlations between personality and job satisfaction are corrected for unreliability and occasion-specific influences on job satisfaction. The same model was used to analyze personality and life-satisfaction (Schimmack, 2019; see OSF for syntax and detailed results https://osf.io/vpcfd/ ).

Results

Observed stability was r = .38 from 2005 to 2009, r = .39 from 2009 to 2013, and r = .36 from 2005 to 2009. It is remarkable that the retest correlation spanning 8 years is just slightly lower than the 4-year retest correlations. Using Heise’s formula, this implies low reliability and high stability; REL = .38*.39/.36 = .41, 8-year stability = .36/.41 = .88. The reliability estimate is lower than the reliability estimate based on annual assessments (Schimmack, Schupp, & Wagner, 2008). Thus, job satisfaction seems to fluctuate reliably form year to year. However, there is also a stable component that is highly stable over an eight year period. The main question is how much personality contributes to this stable component.

Personality measured at Time 1 was used as a predictor of job satisfaction at times 1 to 3. If personality contributes to stability in job satisfaction, personality traits should predict job satisfaction concurrently and prospectively. The results in Table 1 show that this was the case for neuroticism and for halo. Effect sizes for the other traits were very small. The effect size for neuroticism was comparable to Judge et al.’s meta-analysis (r = -.29), but these results do not replicate the estimates for extraversion or conscientiousness. One reason for this discrepancy could be that the current model controlled for evaluative biases in personality ratings, while effect size estimates in the meta-analysis were inflated by halo bias. The finding that halo bias was a stable predictor of job satisfaction ratings is consistent with this interpretation.

JS-T1JS-T2JS-T3
Neuroticism-0.24-0.18-0.23
Extraversion0.060.070.06
Openness
Agreeableness0.080.110.09
Conscientiousness0.100.060.04
Halo0.400.220.28
Acquiescence

Table 2 examines whether changes in personality predict changes in job satisfaction. To do so, job satisfaction was regressed on the residual variances in personality at times 2 and 3.

Table 3 shows the relationship between residual variances in personality and life-satisfaction at times 2 and 3. These results show whether changes in personality predict changes in well-being. The coefficients in Table 3 cannot be directly compared to those in Table 2 because they are standardized coefficients and the residual variance in personality is much smaller than the stable variances. However, the results do provide seminal information whether changes in personality can predict changes in life-satisfaction.

JS-T2JS-T3
Neuroticism-0.43-0.31
Extraversion-0.010.11
Openness
Agreeableness-0.07-0.12
Conscientiousness0.080.02
Halo0.160.24
Acquiescence

As before, neuroticism and halo bias were the only notable predictors of change in job satisfaction. The results for halo bias show that job satisfaction ratings change as respondents tendency to respond positively changes. The results for neuroticism are more interesting. One possible explanation for this finding is that neuroticism measures are not pure trait measures and that changes in anxiety are related to worries in the job domain. Another explanation is that neuroticism changes for other reasons and influences job satisfaction. To disentangle these explanations it is important to examine other domain satisfactions. If personality changes cause changes in job satisfaction, one would expect similar changes for other domains like health satisfaction.

The final result was the stability of the residual variance in job satisfaction that is not explained by personality – as measured in the SOEP. Stability estimates were r = .87 and r = .86 over the 4-year intervals with an implied stability of r = .75 over the 8-year interval. Thus, personality is just one predictor of stability in job satisfaction and it contributes a relatively small amount to stability in job satisfaction.

Conclusion

The results of this investigation only partially replicate previous results. Neuroticism is a moderate predictor of job-satisfaction that contributes to stability in job satisfaction. However, extraversion and conscientiousness play a much smaller role than previous meta-analyses suggested. A plausible reason is that meta-analyses relied on observed correlations that do not control for shared method variance. In contrast, the present study modeled method variance and found that evaluative bias (halo) in personality ratings also influences job-satisfaction ratings.

The present results also showed that job satisfaction is highly stable over an 8-year period even after removing the influence of personality. This suggests that job characteristics can also have lasting effects on job satisfaction. A practical implication of this finding is to identify these factors. The results suggest that helping people to find jobs that fit their personality or to improve job characteristics that influence job satisfaction can have lasting effects on job satisfaction and wellbeing. The SOEP data provide ample opportunity to explore additional predictors of stable variance in job satisfaction.

The results also have implications for theories of well-being. McCrae and Costa (1991) suggested that conscientiousness increases well-being because conscientiousness is instrumental for good job performance. However, job performance and job satisfaction are different constructs. While conscientiousness is a robust predictor of job performance, the relationship to job satisfaction is rather weak. Presumably, conscientious individuals work harder because they have higher performance goals which makes it harder to be satisfied. Whatever the reasons, the current results suggest that theories of personality and wellbeing need to be revised.