Fig. 2From: Change analysis for intermediate disease markers in nutritional epidemiology: a causal inference perspectiveSimulation results for different change analysis models under various extents of unobserved heterogeneity. Parameter \(\theta\) indicates the relation of the unobserved time-invariant individual characteristic and exposure, thus representing the extent of the unobserved confounding. The black dashed line represents the true causal effect of exposure on outcomeBack to article page