Two reductions do not resolve added even-shape variance in 78 selected K2-22 Campaign 1 transits

We test whether two K2 Campaign 1 reductions resolve a second even-shape coefficient in 78 selected K2-22b orbits. We fit exposure-integrated Gaussian, width-derivative and timing-derivative templates with an orbit baseline. After measurement covariance subtraction, the main residual width-coefficient variance is -0.1585 parts per thousand squared for K2SFF and 0.0125 for EVEREST. Their conditional 95 percent block intervals are -0.3295 to -0.0006 and -0.1324 to 0.1972. Negative variance flags limits of the noise correction. No tested baseline, width or noise law gives an interval wholly above zero. Narrow-template resamples are often undefined. A fixed 46-minute box transit with variable amplitude gives a negative coefficient slope close to the measured slopes, without physical width changes. Thus this test does not establish a width law or a constant dust shape. The reductions share photons, and independent confirmation of this statistic is absent. Code, aggregate results, a figure, hashes and external-data retrieval steps are included. No new planet is reported.

astronomy · K2-22 · dust transit · exposure integration · shape variability · noise covariance · null result