Heatmap from Scientific Research

Comparison of the nine social class indicators for each relevant hypothesis in the pooled sample. This figure presents the standardized coefficient estimates from fixed-effects regressions (with country fixed effects) testing the link between each social class indicator and outcome (excluding the original control variables for comparability). The sign of the hypothesized effect of social class is indicated in parentheses after the outcome in the first column. Five hypotheses were excluded from the analysis because social class was not the predictor (H14 and H16), they pertained to an indirect effect (H20) or they predicted a null moderation (H32–H33). Darker green cases indicate more positive estimates, darker red cases indicate more negative estimates and grey cases indicate non-significant estimates (after correction for multiple comparisons). Education is the planned contrast (-0.5, lowest group; 0, middle; 0.5, highest). Income was log-transformed and rescaled within countries to account for differences in monetary units (that is, 1 unit = 1 s.d.(log(income)) within ). Financial scarcity was reverse-coded, with a higher score indicating higher social class. Objective indicators are education, income, occupation and scarcity. \(\overline{\beta }\) indicates the average standardized coefficient across social class indicators. *** P < 0.001; ** P < 0.01; * P < 0.05 (significance levels were adjusted for multiple testing for each hypothesis using a sequential Bonferroni procedure; the exact P values can be found in Supplementary Table 11 ).

DOI