On the limits of judgmental correction: Differences of fluency effects in bias awareness and naïve theories about bias direction.

点击翻译按钮获取中文摘要

In times of fake news and massive media manipulation, it is of highest societal relevance to explore how individuals can correct their judgments from unwanted influences. When people need to make a judgment about a stimulus in their environment, they are influenced not only by judgment-relevant characteristics of that stimulus and the content of their thoughts but also by the metacognitive feelings that accompany stimulus perception and thinking. One potent metacognitive feeling is the subjective experience of processing fluency, that is, the feeling of ease or difficulty with which information is processed. While fluency can be useful because it enables fast and effortless judgments, it is also often a source of bias with harmful results for judgment accuracy. This led to a vast amount of research on the processes that allow the correction of unwanted fluency biases, and classical research outlined two requirements of judgmental correction: a) awareness of the source of bias; and b) valid knowledge of the direction of the biasing effect. However, recent evidence on the correction of fluency effects shows that some fluency effects can be corrected for but others cannot, even if the above-mentioned requirements were seemingly fulfilled. The aim of this proposal is therefore to explore a new moderating variable of fluency correction success. I propose that different fluency effects vary in a) the degree of bias salience and thus in bias awareness; and b) the degree to which the relevant naïve theories individuals use for correction entail valid knowledge about the direction of the biasing effect. This will be tested in four clusters involving fluency effects that have generated mixed evidence regarding correction success in previous studies: word pronounceability and word length effects on trustworthiness judgments, which could not be corrected for in previous studies; and repetition and color contrast effects on truth judgments, which could be partially corrected or even overcorrected in previous studies. After replicating the correction performance for all these effects in Cluster A, Cluster B will test the bias salience and Cluster C will test participants’ lay theories about the direction of a possible influence of these fluency variations on their judgments. Finally, Cluster D will demonstrate that a successful correction of these fluency effects is only possible when both the bias is made salient and participants have a valid knowledge about the direction of influence.