Page 31 - Demo

29How attractiveness affects implicit cognition 2Figure 1. Trial outline of the dot-probe task. Stimuli from the ChicagoFace Database (D. S. Ma et al., 2015). Copyright 2015 by University ofChicago, Center for Decision Research. Adapted with permission.a dependent variable in all further analyses.After the experiment, participants validated all 40 stimuli (presentedin a random order) by rating their attractiveness on a 7-point ordinal scale(very unattractive, fairly unattractive, somewhat unattractive, neutral, somewhat attractive, fairly unattractive, very unattractive). We used these scoresto determine whether the ratings of the participants aligned well with thepre-determined attractiveness categories (attractive, intermediate, unattractive).Statistical AnalysesWe first filtered out extremely fast or slow responses. For fast trials, weexcluded all trials with RTs < 250ms. The upper exclusion level was determined per subject. Specifically, we computed the median RT and the medianabsolute deviation (MAD; Leys, Ley, Klein, Bernard, & Licata, 2013) persubject. The following conservative filter was applied per subject (upperlimit RT = Median + 2 ⇥ MAD). The lower and upper filter resulted inexclusion of 4.7% overall. Hereafter, we mean-centered the reaction timesby subject (i.e., how fast did the participant react relative to their own meanRT).All analyses were done in R statistics version 4.2 (R Core Team, 2018).We fitted Bayesian mixed models using the brms package (Bürkner, 2017,2018). Bayesian analyses have gained popularity over the last few years,because they have a number of benefits compared to frequentists analyses(Kruschke, Aguinis, & Joo, 2012; Makowski, Ben-Shachar, Chen, & Lüdecke,2019). While frequentist methods (e.g., p-value null-hypothesis testing; seeIliana Samara 17x24.indd 29 08-04-2024 16:34