Inductive Reasoning and its Underlying Structure: Support for Difficulty and Item Position Effects

This paper reports an investigation of the influence of method effects on the measurement of reasoning and of the relationships of these effects to basic cognitive processes. For this purpose, the variation due to the item-position and difficulty effects was separated from the variation due to the measured latent source of inductive reasoning. Data were collected by means of inductive reasoning items and cognitive tasks measuring working memory (WM) updating, rule learning, and automatization. Confirmatory factor analysis models served the decomposition of the variation of inductive reasoning data into a purified version of inductive reasoning, item-position, and difficulty components. The investigation of the relationships of corresponding latent variables and basic cognitive processes revealed two major associations: (a) the purified version of reasoning correlated with WM updating and (b) the item-position effect correlated with variants of learning. These results could be interpreted as signifying a two-dimensional structure of reasoning associated with executive functioning and learning processes.