Free survey constants retain at least 67% of additive magnitude dipole information in Pantheon+ at zHD > 0.0233

We test how constant brightness shifts between surveys affect an additive sky dipole in Pantheon+. The primary sample has 1,365 non-calibrator observations at zHD > 0.0233. We use the full released covariance and fit a global magnitude constant and a local matter-density term. Free constants for all 20 survey IDs retain 67.49% to 93.38% of the baseline dipole Fisher information in every parameter direction. If each survey shift has absolute value at most 0.010 magnitude, the largest fitted dipole change is 0.0101276 magnitude. Exhaustive cube-vertex enumeration and an independent convex-hull calculation give the same bound. The primary residual dipole is consistent with zero under the tested Gaussian linear model, with nominal p = 0.515. A nonlinear density fit changes each dipole component by at most 0.0164 millimagnitude. Redshift and covariance controls are reported. The chosen shift bound is a sensitivity scale, not a measured calibration uncertainty. The result does not establish cosmic anisotropy and does not test velocity-model errors. We release response matrices, numerical witnesses, figures, code, input hashes, and reproduction instructions.

Pantheon+ · Type Ia supernovae · dipole · survey calibration · covariance · sensitivity analysis · cosmic isotropy · reproducible research