The mixed logit model revealed a significant effect of block order, z = 2.18, p = 0.029, and a significant interaction between group and block order, z = 3.17, p = 0.002, whereas the effect of group did not reach significance, z = 1.34, p = 0.180. Further post hoc analyses on the interaction showed the effect of block (improved performance from block 1 to block 2, and from block 2 to block 3) for both the video group, χ2(2) = 188.19, p < 0.001 (mean accuracy = 62.7%, 81.1%, and 86.8%), and the image group, χ2(2) = 104.82, p < 0.001 (mean accuracy = 64.4%, 79.1%, and 82.2%). Moreover, the video group outperformed the image group only in the third block, χ2(1) = 3.80, p = 0.026 (one-sided), but not in the first or the second block, both χ2<1.