Define Treatment and Donor Eligibility

Create an eligibility table before running GeoSC. For every geography record:

• treatment assignment;
• inclusion or exclusion;
• reason;
• decision date and owner;
• known campaign, operational, or measurement contamination;
• geographic adjacency or audience leakage concerns.

Eligible donors should be unaffected by treatment, measured consistently, and plausible counterfactual contributors. Pre-period correlation alone is not an eligibility rule. Exclude geographies for design reasons established without post-treatment outcomes.

Power treats every non-treatment unit in its input as a control. Donor evaluation scores every eligible non-treatment unit in its own input. Inference uses the controls in its prepared panel. GeoSC does not automatically apply the donor recommendations to the other stages. Materialise the approved universe as a versioned panel and reuse it across the stages.

For several treated geographies, confirm that the donor pool can support joint in-space placebo assignments: with \(N_0\) controls and \(N_1\) treated units, the exhaustive assignment count is \(\binom{N_0}{N_1}\), provided \(N_0 \geq N_1\).