**Overview** The purpose of this study is to determine whether the relationship between MOSS and reduced bias in judgment depends on the identifiability of the bias. In studies 1-7 of Hawkins' dissertation, a common theme emerged from the DVs that MOSS predicted and the DVs that MOSS did not predict - identifiability of the bias. The MOSS predicted DVs with relatively identifiable objective outcomes (resisting self enhancement in the above average effect, resisting favoring one's own group in the minimal group paradigm) but did not predict DVs with relatively unidentifiable objective outcomes (resisting the self-serving bias and false consensus effect). This led to the hypothesis that high-MOSS individuals may only show reduced bias in judgment and behavior when the objective response is relatively identifiable. The purpose of this study is to manipulate the identifiability of the objective response in the same paradigm. In Study 5, MOSS predicted reduced above-average effect at r = -.24. Pronin, Lin, & Ross (2002) manipulated (for other purposes) the identifiability of the AAE with the following instructions: Studies have shown that on the whole, people show a “better than average” effect when assessing themselves relative to other members within their group. That is, 70-80% of individuals consistently rate themselves “better than average” on qualities that they perceive as positive, and conversely, evaluate themselves as having “less than average” amounts of characteristics they believe are negative. We will use these instructions to make the objective response more identifiable and predict that high-MOSS participants will show increase reduction in the AAE when the objective response is relatively more identifiable. Pronin et al. (2002) demonstrated that participants still showed the bias blind spot in regard to the AAE even after they were told what the AAE was with these instructions. However, we would be measuring whether the relationship between MOSS and bias decreased after these instructions. An alternative strategy would be to manipulate the identifiability of the objective outcome for a bias that MOSS did not predict (e.g., false consensus effect), and hypothesize that MOSS would predict reduced bias in judgment when the bias was identifiable, but not when it wasn't. The risk is that there may be any number of reasons why MOSS did not predict reduced bias in those paradigms in addition to identifiability, so a null result would not answer the research question. **Procedure** Project Implicit volunteers will consent, read a brief instructions page, and then are told they will complete some ratings of their personality. Half of the participants will then be given instructions about what the AAE effect is. Then, participants will rate themselves on 10 positive personality traits (randomized), complete the MOSS, an IAT assessing associations between Good and Bad with Self and Truth, and finally, be debriefed. Study link here: https://dw2.psyc.virginia.edu/implicit/showfiles.jsp?user=chawkins&study=accuracy8 **Hypotheses and Analysis Plan** In a past study, the trait ratings were reliable (alpha = .77) and provided they are again, AAE will be calculated by averaging the 10 trait ratings. Significant AAE will be reflected by an average self trait rating that is significantly larger than 4 (the midpoint of the scale) and will be tested with a one-sample t-test (H0 = 4). The 21 MOSS items will be averaged to create a composite score. AAE scores will be regressed on the main effects of Condition (identifiable AAE vs. control; between participants factor) and MOSS (measured variable; centered on its mean) and their interaction. The main hypothesis is a significant interaction effect -- MOSS should negatively predict AAE scores, and more strongly in the identifiable condition than control. Two hundred fifty participants will provide 95% power to detect this interaction assuming a moderate effect size (F = .25 for the interaction effect; Faul et al., 2009), whereas 1550 people would be necessary for 95% power for detecting a small effect (F = .10) and a small-to-medium effect size (F = .17) would be detectable with 95% power with 530 participants. Given these numbers, we'll end data collection with 550 participants, which provides 55% for a small effect (F = .10) and >99% power for a medium effect (F = .25).