Kepler DR25 recovery near rotation-period ratios differs by +3.0 points in 268 matched pairs, with a 95% interval of −2.2 to +8.2

We tested pipeline recovery near ratios 0.5, 1, and 2 of injected orbital period to measured stellar rotation period. We joined the Kepler DR25 on-target pixel-injection table to the McQuillan et al. rotation catalog and the DR25 stellar table. There were 2,882 eligible injections with orbital periods of 3–100 days, expected MES of 6–20, durations of 1.5 to less than 15 hours, and dwarf-star cuts. We saved the plan and selected pairs before reading joined recovery results. The match used nine controls and the same nearest ratio, with no star reused. It retained 268 of 354 near signals. Recovery was 181/268 (67.5%) near the ratios and 173/268 (64.6%) in comparison windows farther from them. The difference was +3.0 percentage points. The 95% paired bootstrap interval was −2.2 to +8.2 points; the exact McNemar probability was 0.332. A pair and sky-group variance check gave −1.8 to +7.8 points. Under the pair sampling assumptions, the pooled interval excludes losses of 5 and 10 points. It includes zero and smaller losses. Separate ratio tests were less precise, and most pairs had nearest ratio 2. Tighter matching, a narrower near window, and alternative recovery-flag rules gave the same general result. The result concerns a selected sample with measured rotation, before candidate vetting. It does not prove equal recovery, a causal effect, or a planet occurrence rate. We preserve source provenance, hashes, code, pair identifiers, all planned results, three figures, and reproduction checks. The external rotation catalog has a fetch script with a required content hash. We report no new planet.

Kepler DR25 · transit injection · detection efficiency · stellar rotation · matched pairs · planet search bias · reproducible research