Post-hoc pairwise comparisons (with Tukey correction for multiple comparisons) revealed a highly significant difference in accuracy between monolinguals and bilinguals on the illusion condition (p < 0.001), and no significant difference between the groups on the filler condition (p = 0.96; Table 2b in the S1 File ).
Post-hoc pairwise comparisons revealed that bilinguals were numerically slower than the monolinguals in the illusion condition when they judged the sentences as ‘wrong’ (i.e., the target answer), but the difference did not reach statistical significance (p = 0.17).
Post-hoc pairwise comparisons (Table 4b in the S1 File ) showed that the bilinguals were significantly slower than the monolinguals when they responded ‘correct’ (p = 0.002), marginally but not significantly slower when they responded ‘wrong’ (p = 0.07), and with a numerical trend in the predicted direction, but no significant difference when they responded ‘neither’ (p = 0.68).