The experiment is aimed to test various hypotheses related to the scale distortion theory of anchoring (Frederick & Mochon, 2012; Mochon & Frederick, 2013). First, we will try to replicate the effect from Frederick and Mochon (2012) by comparing target numerical judgments of participants from condition (1) and (2). Mochon and Frederick (2013) found that participants given judgments with low and a medium answers before the target judgment estimated the target attribute to be lower than those who made only one of the judgments. We will follow up on this result and test whether the first or the second judgment are more influential in this effect by comparing conditions (4) and (5). We will also test whether the anchoring effect can be debiased by imagining high and low points on the scale on which the judgment is made by comparing the condition (6) with the condition (1) and the condition (7) with the condition (2). We will also test whether the instruction to imagine high and low points on the scale helps to reduce the anchoring effect more than just giving the participants both low and high anchors (i.e., a comparison with (4) and (5)). Next, we will test whether making the comparison judgment used in the standard anchoring paradigm increases the anchoring effect by comparing conditions (1) and (8) and conditions (2) and (9). Mochon and Frederick (2013) suggest that the comparison question itself is not sufficient to produce the anchoring effect – which will be tested by comparing (10) and (11) – while Harris and Speekenbrink (2016) show that the comparison question can influence the use of the absolute judgment as an anchor. The hypothesis is therefore aimed to show whether the comparison question may increase the anchoring effect even in the case of the sequential anchoring paradigm. The analysis will be computed using mixed-effect models. The target judgment will be transformed to z-scores using the distribution of the control conditions for each scale. That is, we will compute what percentile would a given value of the target judgment be if it was an answer in the control condition. We will use this percentile to compute the z-score afterward. Using these z-scores, we will use a mixed-effect model with all the conditions, apart from the control condition, as predictors without including an intercept. The estimated coefficient for each condition will therefore test the difference of the given condition from the control condition. We will include random intercepts for a participant and an item (i.e., scale). The next analysis will be similar, but will not include all the conditions, and it will test the particular comparisons described above. <br><br> Frederick, S. W., & Mochon, D. (2012). A scale distortion theory of anchoring. Journal of Experimental Psychology: General, 141(1), 124-133. Harris, A. J. L., & Speekenbrink, M. (2016). Semantic cross-scale numerical anchoring. Judgment and Decision Making, 11(6), 572-581. Mochon, D., & Frederick, S. (2013). Anchoring in sequential judgments. Organizational Behavior and Human Decision Processes, 122(1), 69-79.
OSF does not support the use of Internet Explorer. For optimal performance, please switch to another browser.