Difference-in-Differences and Changes-in-Changes with Sample Selection
Sample selection arises endogenously in causal research when the treatment affects whether certain units are observed. It is a common pitfall in longitudinal studies, particularly in settings where treatment assignment is confounded. In this paper, I highlight the drawbacks of one of the most popular identification strategies in such settings: Difference-in-Differences (DiD). Specifically, I employ principal stratification analysis to show that the conventional ATT estimand may not be well defined, and the DiD estimand cannot be interpreted causally without additional assumptions. To address these issues, I develop an identification strategy to partially identify causal effects on the subset of units with well-defined and observed outcomes under both treatment regimes. I adapt Lee bounds to the Changes-in-Changes (CiC) setting (Athey & Imbens, 2006), leveraging the time dimension of the data to relax the unconfoundedness assumption in the original trimming strategy of Lee (2009). This setting has the DiD identification strategy as a particular case, which I also implement in the paper. Additionally, I explore how to leverage multiple sources of sample selection to relax the monotonicity assumption in Lee (2009), which may be of independent interest. Alongside the identification strategy, I present estimators and inference results. I illustrate the relevance of the proposed methodology by analyzing a job training program in Colombia.
Code (0)
등록된 구현이 없습니다.
Similar Papers 제목 키워드 기반
Decomposition of Differences in Distribution under Sample Selection and the Gender Wage Gap
I address the decomposition of the differences between the distribution of outcomes of two groups when individuals self-select themselves into participation. I differentiate between the decomposition for participants and…
quantile regressionregressionvalidDifference-in-Differences with Compositional Changes
This paper studies difference-in-differences (DiD) setups with repeated cross-sectional data and potential compositional changes across time periods. We begin our analysis by deriving the efficient influence function and…
Optimal Transport Features for Morphometric Population Analysis
Brain pathologies often manifest as partial or complete loss of tissue. The goal of many neuroimaging studies is to capture the location and amount of tissue changes with respect to a clinical variable of interest, such …
What's the Difference? Supporting Users in Identifying the Effects of Prompt and Model Changes Through Token Patterns
Prompt engineering for large language models is challenging, as even small prompt perturbations or model changes can significantly impact the generated output texts. Existing evaluation methods of LLM outputs, either aut…
Language ModelingLanguage ModellingPrompt EngineeringWordWars: A Dataset to Examine the Natural Selection of Words
There is a growing body of work on how word meaning changes over time: mutation. In contrast, there is very little work on how different words compete to represent the same meaning, and how the degree of success of words…