Email-запись: Consistent causal effect estimation under dual misspecification and implications for confounder selection procedures