Page 81 - Demo

Attractiveness modulates attention794SD = 1 for the fixed effects, default Student t priors with 3 degrees of freedom for the thresholds, and default half Student t priors with 3 degrees of freedom for the random effects and residual standard deviation.To test our main hypothesis, we created a model that used by-subject meancentered RT as the dependent variable and the interaction between condition (attractive vs. intermediate or unattractive vs. intermediate) and probe location (behind intermediate or behind (un)attractive stimulus). Furthermore, to explore the effect of sex and age, we created two more complex models that included the three-way interaction between condition, probe location, and sex and age, respectively. All categorical fixed effects were sum-to-zero coded, and age was z-transformed. In all models, we added random intercepts per subject and trial number (to control for order effects) and allowed slopes of the interaction between condition and probe location to vary by subject. We used regularizing Gaussian priors with M = 0 and SD = 5 for all fixed effects, a Gaussian prior with M= 0 and SD = 10 for the intercept, and default half Student t priors with 3 degrees of freedom for the random effects and residual standard deviation, which were weakly informative.We used multiple measures to summarize the posterior distributions for each variable: (a) the median estimate and the median absolute deviation of this estimate, (b) the 89% credible interval (CI; McElreath, 2018), and (c) the probability of direction (pd). The 89% CI indicates the range within which the effect falls with 89% probability, while the pd indicates the proportion of the posterior distribution that is of the median’s sign (Makowski et al., 2019). We have chosen an 89% CI instead of the conventional 95% to reduce the likelihood that the CIs are interpreted as strict hypothesis tests (McElreath, 2018). Instead, the main goal of the credible intervals is to communicate the shape of the posterior distribution.Furthermore, we used leave-one-out cross-validation (PSIS-LOO-CV; Vehtari et al., 2017) to compare the predictive accuracy of the more complex models that include sex and age, respectively, to that of the simpler model. Using PSISLOO-CV, we calculated the expected log predictive density (elpdLOO), which quantifies predictive accuracy, for each model. Then, we calculated the difference in elpdLOO (ΔelpdLOO) between the models and the standard error of the difference. If ΔelpdLOO is small (< 4) and the SE is large relative to the difference, this suggests that models have similar predictive performance.Tom Roth.indd 79 08-01-2024 10:41