**Hal R. Varian**: *Causal Inference in Economics and Marketing*: "This is an elementary introduction to causal inference in economics written for readers familiar with machine learning methods. The critical step in any causal analysis is estimating the counterfactual—a prediction of what would have happened in the absence of the treatment...

...The powerful techniques used in machine learning may be useful for developing better estimates of the counterfactual, potentially improving causal inference.... Experiments... it is important to have an estimate of the counterfactual—what would have happened in the absence of the experiment. This is essentially a problem of predictive modeling, an area where machine learning offers several powerful techniques. Regression Discontinuity.... one wants to build a predictive model for behavior near the threshold. We can then use the predicted outcome for the treated group estimated using the training data from the untreated units. Instrumental Variables... variables that are thought to be independent of potential confounders.... There are good reasons why the instruments should enter the predictive model linearly, but the other covariates could easily be nonlinear.... Difference in Differences.... The goal is to estimate a predictive model of what the outcome would be for the treated group if it were not treated. To accomplish this goal, one can use a model, possibly nonlinear, of the observed outcomes of the untreated group in the posttreatment period. In each of these cases, building a predictive model is a key step in identifying the causal impact. Machine-learning tools offer powerful methods for predictive modeling that may prove useful in this context...

```
#shouldread
#statistics
#cognitive
```