Earthquake Impact on Educational Attainment with Multi-Valued Treatment Matching

Magdalena Bennett of Columbia University presents a methodological paper and accompanying simulation data for a new matching algorithm for observational studies. The approach is designed to handle multi-valued treatments, create representative samples, and scale to hundreds of thousands of observations. The method is illustrated with an observational study on the impact of an earthquake on school attendance and college admission test scores.

Use Cases

• Evaluating causal effects of multi-level exposures based on the described matching methodology
• Studying the impact of natural disasters on educational outcomes based on the earthquake case study
• Benchmarking new matching algorithms against the linear-sized mixed integer formulation described
• Creating transparent graphical displays for matched results as demonstrated in the article

Strengths

• Method is demonstrated to handle datasets with hundreds of thousands of observations
• Algorithm is characterized as effective for matching with distributional covariate balance under certain assumptions
• Includes a concrete application case studying earthquake exposure effects

Limitations

• Row count, column definitions, and sample data are unknown; data quality requires manual inspection after download
• Last update date is unknown; freshness unverified
• The dataset's primary purpose is methodological illustration; its scale and features for direct analysis are unspecified

Provenance

Source

Magdalena Bennett, Columbia University

Collection Method

Likely generated for a simulation study and derived from an observational study on earthquakes and education.