✳flâneur — a map of the web's best reading

Causal Inference The Mixtape - 8 Panel Data

mixtape.scunning.com · 6,722 words · saved by 1 readers

One of the most important tools in the causal inference toolkit is the panel data estimator. The estimators are designed explicitly for longitudinal data—the repeated observing of a unit over time. Under certain situations, repeatedly observing the same unit over time can overcome a particular kind of omitted variable bias, though not all kinds. While it is possible that observing the same unit over time will not resolve the bias, there are still many applications where it can, and that’s why this method is so important. We review first the DAG describing just such a situation, followed by discussion of a paper, and then present a data set exercise in R and Stata.1 Before I dig into the technical assumptions and estimation methodology for panel data techniques, I want to review a simple DAG illustrating those assumptions. This DAG comes from Imai and Kim (2017). Let’s say that we have data on a column of outcomes 𝑌 𝑖 , which appear in three time periods. In other words, 𝑌 𝑖 1 ,

8 Panel Data – Causal Inference The Mixtape Causal Inference: The Mixtape. Buy the print version today: Buy from Amazon Buy from Yale Press \[ % Define terms \newcommand{\Card}{\text{Card }} \DeclareMathOperator*{\cov}{cov} \DeclareMathOperator*{\var}{var} \DeclareMathOperator{\Var}{Var\,} \DeclareMathOperator{\Cov}{Cov\,} \DeclareMathOperator{\Prob}{Prob} \newcommand{\independent}{\perp \!\!\! \perp} \DeclareMathOperator{\Post}{Post} \DeclareMathOperator{\Pre}{Pre} \DeclareMathOperator{\Mid}{\,\vert\,} \DeclareMathOperat

Explore this link on the map →

Causal Inference The Mixtape - 8 Panel Data

related reading