The linked paper actually refutes this number based on the experiments described. It suggests the number is closer to 5% of adults who knowingly believe in some sort of emission theory, directly arguing against the Winer et al. finding.