Page 172 - Emotions through the eyes of our closest living relatives- Exploring attentional and behavioral mechanisms

Chapter 7
in the PIAT (i.e., participants with congruent trials first had a higher D-score average than participants with incongruent trials first), this effect was not present in the WIAT. The PIAT and WIAT results were significantly correlated and showed good (within- participant) test-retest reliability. Finally, implicit biases measured with the IATS did not correlate with explicit measures of inter-ethnic bias measured through the SRS, nor did the SRS impact performance on the IAT that followed it. The results extend our findings from Experiment 1, showing that the PIAT can tap into the same implicit biases as the more commonly used WIAT.
Discussion
The aim of this study was to design and validate a non-verbal, intuitive pictorial IAT (PIAT). Experiment 1 shows that the PIAT can tap into implicit inter-ethnic attitudes in a large group of participants including adults and children, and can do so reliably using different stimuli. Furthermore, the PIAT can do this outside of a lab setting and on a representative subject population involving participants of different ages (as opposed to only university students). In Experiment 2 using a within-subjects design, the performance of the PIAT was comparable to a more typical word IAT that has been rigorously tested in the last two decades (Dunham et al., 2006; Greenwald et al., 1998, 2003; Kurdi et al., 2019; Nosek et al., 2002a, 2013; Oswald et al., 2015). As such, the PIAT could be standardized tool that enables future studies to make direct comparisons across different cultures, age groups, and potentially also between species.
Internal validity
Although we counterbalanced the order of the presentation of critical blocks, participants who received incongruent trials first and congruent trials second in the PIAT in Experiment 1 showed higher D-scores on average than participants who received a reversed order. Interestingly, this effect was reversed in the PIAT of Experiment 2, and absent in the WIAT. IAT order effects are well documented, and may impact the magnitude of the found IAT effects (Greenwald, 2001). An explanation for their existence is a cost of switching tasks in the two critical blocks, namely in the form of increased reaction time latencies and error rates (Mierke, 2001; Mierke & Klauer, 2003). Furthermore, these task switching costs may remain for quite some time after switching, as switching requires the activation of the appropriate action and suppressing the previous, competing one. Nevertheless, we find diametrically
170