Advancing the Age-Happiness Debate: How Can We Reconcile Graphical and Numerical Approaches?

While in the past, especially numerical approaches such as regression models and p-values were utilized to investigate whether the functional form between age and happiness is linear or U-shaped in a given country, recent advances have shown that such approaches can be misleading, and graphical analyses should amend the analyses. However, these have the downside that they rely to some extent on subjective interpretations and are hardly quantifiable. If applied carelessly, they can be misleading as well. We suggest two easily computed statistics to combine graphical and numerical approaches. We demonstrate their usage with ESS data (N > 440,000) and show how they enable a more nuanced investigation of functional forms. Furthermore, we discuss how statistical uncertainty can be handled in the age-happiness debate and emphasize practical significance, which needs to be kept in mind.