No Return from Heavy Distrust: A Pre-Specified Summary-Level Test of Mindsponge Theory on COVID-19 Vaccine Intent Transitions
Description
Mindsponge theory (MT) holds that information met with very heavy distrust is discharged almost immediately rather than held in the buffer zone for later evaluation. We tested a summary-level consequence of this claim, prediction P3c of the MT Testing Program. If the claim is right, people at the extreme of distrust should show near-zero later uptake once uptake becomes useful, while people who were only holding the decision should show clearly positive uptake. The decision rule, threshold, and admission criterion were written and fingerprinted before any outcome table was read. Two candidate sources were not admitted because neither reported the extreme category on its own. The admitted source was a two-wave transition table from the Understanding America Study covering 1,967 adults who were unvaccinated in April/May 2021. Of those who were very unlikely to vaccinate, 4.4% (95% CI 3.2% to 6.0%) were vaccinated by June/July. Among those who were somewhat unlikely or unsure, 12.1% (9.6% to 15.1%) were vaccinated, a difference of 7.7 percentage points (4.7 to 10.9). The prediction was supported under the pre-specified rule. Exact binomial calculations show that, at this sample size, the rule would have returned “supported” with probability at most 0.025 had the true extreme rate been 10% or more. The extreme group also kept its stated intent far more often than the held groups (83% against 42% to 56%), which bears on the evaluation process itself and not only on behavior. The result is, however, weakly diagnostic. Uptake declined steadily across all five intent levels, a pattern that a graded account without a discharge boundary can also produce, and the 10% threshold for “near zero” is our operational choice rather than the book’s. We describe a multi-wave test that could separate the two accounts.