Participatory horizons and the cosmic dipole anomaly II: a cross-survey analysis

The dipole anisotropy in extragalactic source number counts exceeds the kinematic prediction from the CMB rest frame by a factor of two to three across multiple wavebands, with no settled explanation from local structure or survey systematics. The participatory-horizon programme proposes that the observer's causal diamond acts as an information aperture whose dipolar ($L{=}1$) component couples the source-count monopole into the observed dipole along the velocity axis, producing a universal additive excess $\Delta D_\parallel = \mathcal{A}\beta$. In an earlier companion paper, we anchored this framework to the CatWISE2020 quasar catalogue ($\mathcal{A} = 5.3 \pm 1.3$, $4.0\sigma$) and advanced two cross-survey predictions: a parameter-free $G$-ratio test and a transverse-null diagnostic. Here we test both by applying the same $\hat{\beta}$-projected decomposition to three independent surveys spanning five orders of magnitude in observing frequency: CatWISE (3.4\,$\mu$m, mid-infrared quasars), RACS-low (887\,MHz, radio galaxies), and NVSS (1.4\,GHz, radio galaxies). For RACS and NVSS we project the Bayesian Poisson dipole posteriors (48\,000+ MCMC samples each) onto the CMB velocity axis. The parameter-free $G$-ratio between CatWISE and RACS is consistent with the prediction at $0.5\sigma$. As a supporting diagnostic, $D_\parallel$ is stable across the three surveys (coefficient of variation 13\%) while the transverse component $D_\perp$ is more than twice as variable (CV 30\%), consistent with a $\hat{\beta}$-aligned excess and survey-dependent off-axis contamination. Two additional catalogues were tested as stress tests: Quaia is excluded on systematic grounds ($\mathcal{A} \approx 38$, direction unstable under Galactic latitude cuts), and WISE$\times$SuperCOSMOS is excluded because local large-scale structure dominates at $z_{\mathrm{med}} \approx 0.2$. These results provisionally support a universal, population-independent dipole excess of $\mathcal{A} \approx 5$--$7$ across mid-infrared and radio source populations, and motivate further empirical tests of the hypothesis.