Time weights: 14 additional papers
Complete-case or earlier-cohort rescues, with author-default time weights and an optimized reference at the same penalty. No penalty retuning or cross-validation. Original results are preserved.
Inspect a cohort and outcome
The two charts share a y-axis scale within each comparison. Gray bars are at or below the support cutoff; the underlying vectors are never thresholded or renormalized by this control. Hover bars for source-period labels and weights.
Inspect the selected weight vectors
One preselected example per paper
Examples were selected before fitting: first available specification, median eligible cohort, preferring at least five pre-periods. Click a paper to inspect it. Support uses weight > 10⁻⁶. The available papers, outcomes and cohorts are not an independent or representative sample of studies.
How to read these comparisons
- Defaults: author sparsification and 10,000 final iterations; improvement tolerance 10⁻⁵ × noise scale.
- Optimized reference: the identical full-simplex convex objective, not a different SDID penalty. The independent objective-gap target is 10⁻⁸ after normalizing by noise variance. Twenty low-donor or initializer-sensitive blocks receive 80-digit active-set solutions; other blocks retain the double-precision QP. All original solver variants remain available.
- Near-flat directions: the normalized ridge coefficient is only 10⁻¹². Small objective gaps alone do not prove the weights themselves are accurate or stable. The high-precision check resolved six initializer-sensitive cases. Passing objective and initializer checks alone is not a universal certificate of weight accuracy.
- Large donor matrices: an orthogonal QR reduction preserves the time objective without subsampling donors. Direct author-solver checks are in the numerical report.
- Numerical zeros: QP solutions can have tiny roundoff entries. Raw weights and the small feasibility repair are exported; strictly positive support is not directly comparable across solvers.
- Scope: time weights only. No unit-weight fitting, new ATTs or bootstrap. This tranche attempts 418 blocks; 58 zero-noise blocks remain without substitute weights. Sample/window changes are explicit. Fiscal Capacity retains its mixed-design label and E-governance its pending label. Original results are unchanged.
Findings and comparison profiles · Base numerical report · High-precision refinement · Profile PDF · Default and QP vectors · Diagnostics · Paired comparisons · Optimized reference vectors · Initializer checks · Support thresholds · Sample manifest · 14-paper summary · Unavailable fits · Quarto source