interventions

While SE CE, it is not generally true thatSE = CE. Consider now an intervention onX0. In the example, this only happens when we set the weights, means and variances to very particular values. Here we present a slightly adapted version of Invariant Causal Prediction [27]. Under this approach, the complexity oftesting asingle setofpredictors ofsize k is the cost of performing a least-squares regression and computing the residuals (O(k2N)) and the cost of performing the t-test and F-test over each split of thee environments (O(eN)).