Page 171 - Emotions through the eyes of our closest living relatives- Exploring attentional and behavioral mechanisms
P. 171

                                Validation of the Pictorial Implicit Association Test
hypothesis, we performed an additional correlation analysis using Bayesian statistics (using the R-packages BayesFactor (Rouder et al., 2009) and bayestestR (Makowski et al., 2019b). We found a BF01 = 5.1 for the correlation between PIAT and SRS scores, and a BF01 = 4,95 for the correlation between WIAT and SRS scores, meaning that in both cases the data are around 5 times more likely under the null hypothesis that there is no correlation between the variables. The results therefore indicate moderate evidence for the null hypothesis (i.e., there is no correlation between the measures (Lee & Wagenmakers, 2013). As such, the implicit biases reflected in the D-scores in both versions of the IAT do not seem to correlate with explicit inter-ethnic biases in our participant pool.
Finally, we also assessed whether performing the SRS in-between tasks affected D-scores on the final IAT. We found that the difference score did not significantly differ from zero (t(137) = -1.14, p = .255), meaning that D-scores in the second task were not significantly higher or lower than the D-scores in the first task. Furthermore, we found no significant effect of First Task (t(137) = -1.27, p = .206) on the difference score, showing that the difference score was not affected by whether a participant first started with the PIAT or the WIAT. In short, performing the SRS in the middle of two IATs did not significantly impact D-scores on the last IAT, and this was regardless of whether participants started with a PIAT or WIAT.
Table 3. Model results for the online PIAT
 Task
PIAT
WIAT
Predictors Estimates SE
(Intercept) .17 .05
Congruency (Congruent first) -.14 .05
Location (“Moroccan descent” Left/ .07 .05 “Dutch descent” right)
(Intercept) .22 .04
Congruency (Congruent first) -.05 .04
Location (“Moroccan descent” Left/ -.04 .04 “Dutch descent” right)
D-score
95% CI
.07 – .27 -.24 – -.03 -.03 – .17
.13 – .31 -.14 – .04 -.13 – .05
t p 7 3.36 .001
-2.63 .009 1.31 .194
4.88 <.001 -1.11 .269 -0.83 .409
    Conclusion
Participants performed similarly on the PIAT and WIAT: in both IATs, participants appeared to have a an implicit bias in the predicted direction. Although the order in which participants completed the critical blocks significantly impacted the D-scores
169











































































   169   170   171   172   173