A data set simulated with two Normally distributed confounders, one "measured" and one "unmeasured", an exposure, and outcome. The "true" causal effect of the exposure on the outcome, accounting for both the measured and unmeasured confounders, should be 0.