The deactivation design's bundling limitation is empirically bounded because the same design detected polarization effects in 2018 despite identical substitution patterns, confirming the 2024 null reflects genuine attenuation rather than substitution masking.

[DEFENSE-REBUT → 3d5efa] The earlier Allcott et al. (2020) deactivation experiment detected a statistically significant 0.16 SD reduction in political polarization despite users substituting toward watching TV alone, demonstrating the design can detect effects through substitution. Allcott et al. (2025) extended the 2024 deactivation results to the average user including lighter users and found that the null effects on polarization are largely similar, ruling out heavy-user selection as an explanation. If substitution to TV were fully masking a platform polarization effect in 2024, it should have similarly masked the effect detected in 2018, since substitution patterns were comparable. Therefore, The deactivation design's bundling limitation is empirically bounded because the same design detected polarization effects in 2018 despite identical substitution patterns, confirming the 2024 null reflects genuine attenuation rather than substitution masking. (Warrant: When the same deactivation design with similar substitution patterns detects effects in one period but not another, the null reflects genuine attenuation of the platform effect rather than substitution masking.)

• The earlier Allcott et al. (2020) deactivation experiment detected a statistically significant 0.16 SD reduction in political polarization despite users substituting toward watching TV alone, demonstrating the design can detect effects through substitution.
• Allcott et al. (2025) extended the 2024 deactivation results to the average user including lighter users and found that the null effects on polarization are largely similar, ruling out heavy-user selection as an explanation.
• If substitution to TV were fully masking a platform polarization effect in 2024, it should have similarly masked the effect detected in 2018, since substitution patterns were comparable.

These are challenges this argument’s reasoning pattern must still withstand. Answering them on Isonomia strengthens the argument.

• Is there a plausible alternative hypothesis that has not been considered or that would explain the facts at least as well?Open
• Could the facts be jointly explained by a conjunction of weaker causes rather than a single dominant H?Open
• Could the body of facts F itself be an artifact of selection, measurement, or reporting bias rather than a real phenomenon needing causal explanation?Open
• Does H actually explain the full body of facts F, or only a salient subset?Open
• Are the criteria used to judge H 'best' (scope, simplicity, mechanism, prior probability) appropriate for this domain, and are they applied consistently across the alternatives?Open

No one has cited this argument yet.

No arguments cite this one yet — no one has built on or contested it. That is an absence of engagement, not a finding of soundness. Build on or contest it on Isonomia to change that.