Are hopeful narratives more convincing? A laboratory experiment

Assessing the causal impact of narratives on beliefs and behaviors remains an empirical challenge for social scientists, largely due to endogeneity and cultural factors. To address these limitations, we present the results of a novel, content-neutral laboratory experiment. In this experiment, participants (i) engage in a zero-sum game against a non-strategic robot, where the final outcome is determined with equal probability either by their choices or by randomness, and (ii) are exposed to either hopeful or passive narratives. These narratives differ in how ambiguous evidence is presented, suggesting whether or not participants can actively determine the final outcome of the game through their choices. Our findings reveal that, regardless of the narrative they are exposed to, participants consistently form beliefs and make choices under the illusion that they can influence the final outcomes. When provided with unambiguous evidence disproving this illusion, participants adjust their beliefs accordingly, although their choices take longer to align with these updated beliefs. Furthermore, exposure to the passive narrative reduces the inconsistency between beliefs and choices when participants mistakenly believe their choices determine the final outcome. Finally, presenting unambiguous evidence that contradicts the narrative's content increases the proportion of random and unpredictable choices.