In the 1970s, Hample developed a successful model of intrapersonal argument. Loosely based on the law of total probability, the model used a normatively correct standard to predict people’s adherence to persuasive claims. That original research used single-item measures that could not be assessed for internal consistency. The present study estimates the reliabilities of the appropriate measures so that corrections for attenuation can be made. In addition, the study exploits the base-rate fallacy to encourage formation of bad premises and perhaps bad reasoning. Results show that Hample’s original model is indeed accurate and that his original results understated people’s rationality in persuasive situations. The base-rate effect was confined to improper premises; given people’s premises, they continued to reason rationally. The model was equally accurate for high experientials and high rationals, in Epstein’s terminology.