Different Evaluative Well-being Measures, Different Conclusions? Evidence across Regions

Life satisfaction (LS) and the Cantril ladder (CL) are widely used as measures of evaluative well-being. Because the two measures are strongly correlated, they are frequently treated as interchangeable when analysing the determinants of subjective well-being, comparing countries and informing public policy. However, recent evidence suggests that respondents may not interpret these questions similarly across regions. We examine whether LS and the CL measure the same underlying concept using pooled data from the first two waves of the Global Flourishing Study, covering 23 countries (including one territory). We first compare respondents' answers to the two measures across all regions by analysing country averages and individual-level correlations. We then examine the within-person LS-CL difference and estimate separate XGBoost models on pooled observations from Waves 1 and 2, alongside pooled OLS models with country and wave fixed effects, focusing on the two regions that display the largest contrast: Western countries and Sub-Saharan Africa (SSA). Finally, we compare country rankings for all regions across our 23 countries. Our findings show that LS and the CL should not be treated as universally interchangeable measures of evaluative well-being. Correlation analyses between the two measures at the country and individual levels, both within and between regions, vary substantially. Focusing on the Western and SSA regions, respondents differ much more in their evaluations of the two measures in SSA than in the West. The features associated with each measure also differ. In the Western sample, mental health is the most important predictor of both measures, but is substantially more important for LS, while perceived income is relatively more important for the CL. Loneliness is considerably more important for LS in SSA, while country is among the most important features for both SSA outcomes. Perceived income is also important in both regions. Therefore, the distinction between LS and the CL is not simply that different variables matter in each case. Many of the same features are important for both measures, but the relative weight placed on them differs between the two measures and, more substantially, between the two regions. Furthermore, the standard set of variables used to explain LS and the CL accounts for substantially more variation in the West than in SSA, suggesting that important features of evaluative well-being in SSA remain unobserved or suffer from significant measurement error. Therefore, treating the two measures as interchangeable, particularly in country comparisons that extend beyond Western countries, risks drawing conclusions that are not representative of all countries and regions. We conclude that the choice of evaluative well-being measure has substantive consequences for conclusions and ranking of countries and should be guided by the research question and regional context.