In the search for answers to causal questions, assuming parametric models can be dangerous. With a seemingly small amount of confounding and model misspecificaton, they can give biased answers. One ...