Fig. 2From: Within- and between-group regression for improving the robustness of causal claims in cross-sectional analysisConceptual model of the C8 Health Project. Footnote: Dotted rectangles indicate larger not directly observable causal constructsBack to article page