The univariate tests of between-subjects effects revealed a marginally significant effect of role ( F (1, 383) = 3.772, p = .053, n 2 p = .010) and significant interaction between phase and role ( F (2, 383) = 3.257, p = .040, n 2 p = .017) for the number of blinks (see Fig. 2 a).
Positive trials were rated higher than both the neutral and the negative; however, Conover post-hoc tests revealed only a trend toward significance ( p = .063) after Bonferroni and Holm adjustments, despite medium effect sizes for positive-neutral and positive-negative comparisons (r rb = -0.549 and r rb = -0.562, respectively).
Interestingly, while blink synchronization decreased significantly from baseline to audio, the synchrony of pupil dilation velocity peaks showed a descriptive increase across the same transition, although it did not reach statistical significance.