Conceptual Replication of the Color- Erasure Random- Assignment Paradigm with Food Images

The present study constitutes a conceptual replication of the strict causality test conducted by Maier et al. (2025) with color stimuli – but now applied to the domain of food perception. Analogous to the random-assignment color study, the present study removes the deliberate confound present in the food-image preregistration (Refaeli & Maier, preregistered) and instead implements a completely random assignment of food images to conditions on the participant level. This means that for each participant, images are randomly drawn from the full stimulus set and randomly assigned to the stored vs. erased condition, ensuring an unconfounded, strict causal test of the erasure manipulation. Given that this study represents a strict causal test of a proposedly acausal non-classical correlation, the NT-axiom predicts that the erasure effect should not be robustly documentable. Specifically, we predict an “effect and decline” data pattern: initially strong Bayesian evidence for the erasure effect should appear, followed by a decline of the effect with continued data collection.