Individuals who vary their preferred hand when performing different types of manual activities, so-called mixed handers (MH), have been frequently reported to outperform individuals with a consistent (right) hand preference (cRH) on tasks assessing declarative-memory functions. For example, in one influential study, this MH advantage extended to incidental learning from presumed "deep" semantic processing of verbal stimuli but not from "shallow" phonemic or structural processing. In the present study, we aimed to replicate this research finding in two separate participant samples. First, in a pre-registered and sample-size planned experiment we confronted 49 participants (23 MH; 26 cRH) with "phonemic" and "semantic" word evaluation tasks (using a within design), followed by a surprise delayed recognition test. In a second experiment, we repeated the same procedure with 65 other participants (31 MH, 34 cRH). A mixed-effect analyses of variance found a significant main effect of Encoding Condition (phonemic vs. semantic tasks) in both experiments (effect size: ηp2 = .81 to .85), indicating the classical level-of processing effect with higher recognition hits and sensitivity (d') for words that followed semantic versus phonemic encoding. However, the predicted interaction effect of Encoding Condition with Handedness Group was not statistically significant for either sample (all ηp2 < .03), nor was the main effect of Handedness Group. Thus, our findings conflicted with those of the original study in two independent samples. As we had sufficient statistical power to be confident in our failure to detect a genuine group difference, we cannot confirm the previously reported MH over cRH advantage in incidental learning of verbal material. We discuss possible reasons for these contradictory results and the theoretical implications of this discovery.
Read full abstract