The within-person variation of personality therefore explains nearly double the explanation of the demographic factors included in the regression from column 1. Our 3rd column extends our model by including both personality traits and demographic factors. As with the cross-sectional regressions there again appears to be an additive effect. The coefficients remain similar in magnitude and the total explanation of the within-person variation in life satisfaction rises to around 3%. This suggests (a) that personality, traditionally considered as fixed and non-changing, does change and (b) that the individual changes in personality are more predictive of life satisfaction changes than economic characteristics that would typically be seen as variable, such as income or employment status.