Counteracting Narratives: Evidence from An Online Experiment
通过在线实验发现，即使人们知道叙事有偏向且随机分配，仍会受其影响；后续提供平衡信息也无法抵消初始叙事的效果，只有直接呈现对立叙事才能完全对抗。
Abstract Can people counteract biased narratives with subsequent information? Using an online experiment where counteracting may have the best odds by design, we investigate this question by first randomly assigning subjects to read different narratives that contain the same facts, and then offering them the opportunity to acquire and process balanced arguments. We document three main findings. First, subjects shift their attitudes towards the standpoint of the randomly assigned narrative, knowing that the narrative is slanted and randomly assigned. Second, the opportunity to read additional arguments does not prompt subjects to counteract the persuasion effects of the initial narratives. Third, when evaluating subsequent arguments, participants find arguments aligned with the randomly assigned narrative more convincing. These findings remain qualitatively similar in additional treatments where the balanced arguments are provided two weeks after initial exposure to narratives. Only when we replace these arguments with the exact opposing narratives that subjects do not initially see are they able to fully counteract the effects. Taken together, our results highlight the importance of balanced and complete exposure at the outset in counteracting the influence of biased narratives.