Factorial Difference-in-Differences: A Framework for Causal Moderation Analysis

Yiqing Xu published a methodological framework for factorial difference-in-differences (FDID) research designs in 2026. The framework extends canonical difference-in-differences to settings where an event affects all units, using a factorial design with baseline and exposure factors. Supplementary materials for reproducing the work are available online.

Use Cases

• Estimating effect modification in panel data based on the baseline factor G and exposure Z framework.
• Testing the factorial parallel trends assumption for identifying causal moderation.
• Applying the FDID estimator to empirical applications, such as the role of social capital in famine relief in China mentioned in the description.
• Extending regression-based implementations to repeated cross-sectional data or continuous baseline factors.

Strengths

• The framework is formally defined with clear assumptions like no anticipation and parallel trends.
• The methodology is demonstrated with an empirical application on famine relief in China.
• Supplementary materials for reproducing the work are available online.

Limitations

• The dataset's row count and column-level documentation are unknown, requiring manual inspection after download.
• The primary files are PDF and ZIP formats (1.4 MB total), which may not contain raw, analysis-ready tabular data.
• The description is methodological; the availability and structure of the underlying empirical data are unclear.

Provenance

Source

figshare, authored by Yiqing Xu.

Collection Method

Likely contains methodological exposition and supplementary materials for an academic paper.

Freshness

Last updated 2026-06-04 16:13:27.

Geography

The empirical application references famine relief in China.