Chu's only real-world-network test (Facebook-100, n = 769) runs the quantum model under an approximation mathematically identical to Friedkin–Johnsen

The sole real-world numerical experiment in Chu's 2026 quantum opinion-dynamics preprint (arXiv:2607.01452) uses the largest connected component of the Caltech network from the Facebook-100 dataset — n = 769 agents, friendship edges from fall 2005. This run is performed "under the product state approximation (33)" — the same approximation the paper shows (§III.D) reduces the quantum governing equations to exactly the classical continuous-time Friedkin–Johnsen model: "We apply Tr(· Ô) to (33) and obtain the governing equations for opinions... which is exactly the continuous-time FJ model."

The Facebook-100 experiment therefore reports opinion quantities the reduction guarantees will match an FJ baseline by construction, plus quantum-only side quantities (coherence decay, order-effect asymmetry) that have no FJ counterpart to compare against, since FJ does not define them at all. No comparison against an independently-run exact quantum simulation on this network, and no comparison against real polarization or survey data, is reported. This is the structural reason the falsifiability question (question-chu-quantum-opinion-model-falsifiable-beyond-friedkin-johnsen) stays open: the one place the model touches real data is exactly the place it cannot, by its own math, diverge from the classical baseline — the divergence in claim-chu-quantum-fj-divergence-shown-only-on-toy-networks lives entirely on synthetic graphs.