The Kepler quarter 4 flux-product difference for KIC 12557548 remains after removal of the 10 percent most changed events

We tested whether a few events explain a known raw-versus-corrected flux difference for KIC 12557548 in Kepler release 25 quarter 4. We paired core and late-egress measurements from 104 events. Removal of the 11 events with the largest product changes leaves a difference of 286537 square ppm in the small noise-subtracted covariance eigenvalue. Its conditional 95 percent paired block interval is 125306 to 466034 square ppm. All four tested baseline choices retain a positive interval at ten-percent removal. Of 104 events, 101 have larger product-change norms than the pooled quarter 5 to 7 95th percentile. The second half of quarter 4 is inconclusive. At twenty-percent removal, quadratic-baseline intervals include zero. Thus, a small event rejection does not remove the tested discrepancy. This result does not identify which product is more accurate, a pipeline fault, or a dust cause. No new object or independent telescope confirmation is reported. The package includes raw inputs, source hashes, code, paired event tables, figures, and offline instructions.

Kepler · KIC 12557548 · dusty transits · paired covariance · influence test · flux correction