Personality Development Across the Life Span:
Longitudinal Analyses with a National Sample from Germany
Supplemental Gender Analyses
Richard E. Lucas
M. Brent Donnellan
Michigan State University
Previous investigations, including our original cross-sectional analysis of the GSOEP data (Donnellan & Lucas, 2008) have examined how overall trends in personality traits across the lifespan vary by gender. Because of length considerations, we did not include similar analyses of gender differences in the current report. However, we present these here for researchers who have a particular interest in these results and for future meta-analytic reviews.
To evaluate gender differences, we extended the multi-group modeling strategy used in the main text. Specifically, we took the final model from Table 2 of the main text and expanded the multi-group modeling approach to also include gender. This created twice as many groups (34 instead of 17, with 17 age groups for men and 17 age groups for women). Factor loadings and intercepts that were constrained to be equal across age groups in the original analyses were also constrained across genders. Loadings and intercepts that were not constrained to be equal across groups were only constrained to be equal across genders. For four of the five traits (all except Openness), model fit for this initial model was acceptable (see Table S1). For Openness, an additional model modification needed to be made to achieve a model that met the criteria we set. Specifically, in the original Openness model, the intercept for Item 2 (“I see myself as someone who values artistic, aesthetic experiences”) could not be constrained across groups. When we initially constrained this intercept to be equal across genders, it became clear that the fit of the model for the middle-aged male cohorts was negative affected. Therefore, we removed the cross-gender constraint for this item intercept (which was already freed across cohorts), which led to an acceptable model fit (see Table S1). It is important to note, however, that we were still able to achieve partial measurement invariance, which allows for comparisons at the latent level.
Results for the mean-level analyses are presented in Figure S1. In addition, to allow for future meta-analytic reviews, the underlying data for this figure are presented in Table S2. The figure shows that the primary differences between men and women are in mean levels that seem to stay rather consistent across the life span. Consistent with past analyses of these data, women score higher on Extraversion, Agreeableness, and Neuroticism; there are only small and inconsistent differences for Conscientiousness and Openness. More importantly for the purposes of this paper, lifespan trajectories for men and women are mostly similar. Men appear to decline in Extraversion more quickly early in the lifespan, whereas the trajectory is flatter later in life. For women, this pattern is opposite, with larger changes after age 60 than before. For the other four traits, however, the patterns are mostly similar across genders.
Results for the stability analyses are presented in Figure S2. Again, to allow for future meta-analytic work, the underlying data for this figure are presented in Table S3. Simple quadratic models are presented rather than the locally smoothed models from the original paper because with the smaller sample sizes, the lines became somewhat less stable. These plots show that peak stability is similar for men and for women, and both genders show the pattern of increasing stability early in the lifespan, followed by decreasing stability in later life. The curves are more rounded for men, but this is primarily due to some very low stability coefficients in late life for the oldest male cohorts. However, these should be interpreted with caution as the sample sizes get somewhat small in these groups (see Table 1 in the original paper) and men who live to this age are somewhat atypical given the life expectancies for men born in those particular years.
Figure S1.
Figure S2. Stabilities across the lifespan, separated by gender.