A further analysis of the four lure types in a 2 × 2 × 4 (relatedness × consistency × region) repeated measurements ANOVA did not reveal any main effect (consistency: F (1; 23) < 1; MSe = 2.048; relatedness: F (1; 23) = 1.476; MSe = 1.595; p =.236) nor any interaction (consistency × relatedness: F (1; 23) < 1; MSe = 2.742; consistency × relatedness × region: F (3; 69) = 1.813; MSe =.366; p =.16).
Most importantly, we also obtained an interaction between relatedness and consistency (F (1; 23) = 9.69; MSe = 995.733; p <.01): While there was a consistency effect of 34 ms in the related conditions (t (23) = 3.23; p <.01), we found no such consistency effect (-6 ms) for unrelated trials (t (23) = 0.86; p =.40).
Only the three-way interaction SOA × relatedness × consistency failed to reach significance for the error data.
A 2 × 2 × 4 (relatedness × consistency × region) repeated measurements ANOVA revealed main effects for the factor relatedness (F (1; 23) = 4.608; MSe = 3.718; p <.044) and for the factor consistency (F (1; 23) = 8.465; MSe = 1.58; p <.009).
Furthermore, an outlier trimming procedure was applied which excluded all responses for which RT deviated by more than three standard deviations from the mean of the particular condition (relatedness × consistency × SOA).
Finally, we observed a three-way interaction between all three factors (SOA, relatedness, and consistency).
In the analysis of arcsine-transformed error rates, the same main effects for SOA, relatedness, and consistency reached significance.
A more detailed analysis of electrophysiological responses to different lure types revealed a distinctive influence of the factors relatedness and consistency.
For both factors (relatedness and consistency) the distance of the lures from the correct result (split) was matched (mean split for related consistent: 5.4; related inconsistent: 5.4; unrelated consistent: 5.1; unrelated inconsistent: 5.1).
Finally, there was a three-way interaction between SOA, relatedness, and consistency (F (1; 23) = 15.07; MSe = 500.198; p =.001), which was further explored by separate ANOVAs for the different SOAs (see Table 1).
Overall, the findings obtained for lures, which varied according to the factors relatedness and consistency, indicate that both factors influenced the mean voltage changes in the verification of arithmetic problems.
