The (In)flexibility of Evolved Frequency Representations for Statistical Reasoning: Cognitive Styles and Brief Prompts Do Not Influence Bayesian Inference

What happens when format manipulations improve Bayesian reasoning? One view is that naturally sampled frequencies help induce a privileged representational system that is relatively specific in its operation. A contrasting view is that naturally sampled frequencies are but one way to induce a more general process of appreciating nested set relationships. This later view implies that fairly brief and immediate interventions (e.g., simple directives) should produce improvement, whereas the former view implies that more extensive interventions and/or more insightful understanding are necessary for improvement. The present research indicates that neither brief and immediate interventions nor pre-existing representational biases or representational flexibility facilitate performance. Some evidence emerged, on the other hand, that frequentist problem interpretation can improve statistical reasoning performance and increase confidence in responses at times. These results support the privileged representational system view.