Reasoning Ability but not the Item-Position Effect is Related to Implicit Learning

Previous research on reasoning tests has revealed relatively consistent evidence for an item-position effect (IPE). To explain this effect, the learning hypothesis has been put forward, positing that learning the rules underlying the items leads to increasing individual differences from the first to the last item of a reasoning test. For a better understanding of the type of learning contributing to the IPE, the current study examined whether implicit learning was associated with the IPE in a reasoning test. For this purpose, 193 participants completed a Figural Analogies Test with 21 items, which were not ordered according to their difficulty, and a Serial Reaction Time (SRT) task comprising 2x8 blocks of ten 10-trial sequences each. Using bifactor modeling, a latent reasoning and a latent IPE variable could be identified in the reasoning test, with the latter characterized by monotonically increasing factor loadings. Implicit learning was evidenced by a significant reduction in response times across the blocks of the SRT task, where the same 10-trial series was always repeated (experimental condition), but not in the eight blocks of the control condition. From the response times in the experimental condition, two latent variables were extracted. Factor loadings of the first latent variable were fixed at the same value; factor loadings of the second latent variable increased strictly monotonically across the learning blocks, representing individual differences in the continuous (implicit) learning process. Latent reasoning ability was significantly but weakly related to implicit learning, whereas the correlation between the IPE and implicit learning just failed to reach statistical significance. Therefore, the present results supported the assumption of an association between reasoning ability and implicit learning. The IPE in the reasoning test, however, was unrelated to implicit learning, suggesting that it does not rely on implicit learning.