CHICAGO — A Measurement Quantization evidence audit distinguishes theoretical derivations, internal consistency checks, retrospective comparisons and prospective predictions. The classification identifies what each result can establish and where multiple results share calibration or observational information. It is the research program’s own evidence classification, not a new experiment or an independent certification of the framework.

Several MQ results use common physical constants, cosmological normalization or overlapping galaxy data. Repeated agreement across outputs that share those foundations does not automatically supply independent evidence. Figure 1 summarizes the distinction between consistency, external comparison and the boundary on combining results.

A target can be external to a derivation without every comparison being statistically independent. The audit classifies a CMB temperature calculation using a CMB-informed age as cross-scale closure. It treats the two frame-resolved Hubble comparisons as sharing the same scalar, age normalization and mapping structure. Neither example supplies two independent confirmations merely because it has multiple outputs.

Galaxy tests make the distinction tangible. A stellar-mass substitution can expose source sensitivity while reusing the same SPARC velocities. A separate LITTLE THINGS population provides a different target population, yet still uses the same MQ law and upstream normalization. The 99 high-quality SPARC galaxies are a subset of the 175-galaxy catalog, not a second independent survey.

Statistics also answer different questions. A speed residual, a goodness-of-fit probability, a likelihood ratio and an information criterion cannot be used as interchangeable scores. In the gravitational-constant comparison, testing a fixed MQ mode gap against exact zero is different from fitting an arbitrary nonzero gap. Agreement with the fixed value does not uniquely identify its physical cause.

Timing remains part of the evidence. MQ’s numerical gravitational realizations predate the 2026 NIST result, while assigning the particular free-deflection and electrostatic-servo modes to those realizations was retrospective. An earlier broad statement and a later apparatus-specific interpretation must retain their separate dates and scope.

The audit does not assign a universal probability or discovery significance to the combined MQ portfolio. A defensible synthesis would require explicit hypotheses, a joint treatment of dependencies and uncertainty, and a record of which choices were made before target inspection. Posterior odds would additionally require stated priors and alternatives.

The companion Brief provides the classification table, a shared-input diagram, numerical galaxy examples and a worked fixed-target likelihood comparison. It also identifies prospective tests that must not be counted as completed confirmations.

# # #