밤에 쓰는 편지 밤에 쓰는 편지

https://mixtape.scunning.com/09-difference_in_differences#inference1. The Importance of Placesbos in DDThere are several tests of the validity of a DD strategy. I have already discussed one—comparability between treatment and control groups on observable pre-treatment dynamics. Next, I will discuss other credible ways to evaluate whether estimated causal effects are credible by emphasizing the u..

https://mixtape.scunning.com/09-difference_in_differences1. John Snow's Cholera HypothesisWhen thinking about situations in which a difference-in-differences design can be used, one usually tries to find an instance where a consequential treatment was given to some people or units but denied to others “haphazardly.” This is sometimes called a “natural experiment” because it is based on naturally..

https://mixtape.scunning.com/05-matching_and_subclassification#example-the-nsw-job-training-programThere are several ways researchers can estimate average treatment effects using an estimated propensity score.  As there are different ways in which the weights are incorporated into a weighting design, I discuss a few canonical versions of the method of inverse probability weighting and associated..

https://mixtape.scunning.com/05-matching_and_subclassification#exact-matchingPropensity score methodspropensity score matching is an application used when a conditioning strategy can satisfy the backdoor criterion. But how exactly is it implemented? Propensity score matching takes those necessary covariates, estimates a maximum likelihood model of the conditional probability of treatment (usuall..

ttps://mixtape.scunning.com/05-matching_and_subclassification#exact-matchingSubclassification uses the difference between treatment and control group units and achieves covariate balance by using the K probability weights to weight the averages. The subclassification method is using the raw data, but weighting it so as to achieve balance. We are weighting the differences, and then summing over t..

https://mixtape.scunning.com/05-matching_and_subclassificationInsofar as there exists a conditioning strategy that will satisfy the backdoor criterion, then you can use that strategy to identify some causal effect. We now discuss three different kinds of conditioning strategies. They are subclassification, exact matching, and approximate matching. Subclassification is a method of satisfying the ..

https://mixtape.scunning.com/04-potential_outcomes#methodology-of-fishers-sharp-nullFisher's sharp null & Steps to a p-valueYou can never let the fundamental problem of causal inference get away from you: we never know a causal effect. We only estimate it. And then we rely on other procedures to give us reasons to believe the number we calculated is probably a causal effect. Randomization infere..

https://mixtape.scunning.com/04-potential_outcomes#physical-randomization Simple difference in means decompositionIt’s biased because the individuals units were optimally sorting into their best treatment option, creating fundamental differences between treatment and control group that are a direct function of the potential outcomes themselves. Independence assumptionIf treatment is randomly ass..