Analisi log-lineare di un limite attentivo

A common procedure that is pervasively adopted in the analysis of accuracy data collected through attentional experiments is to apply the analysis of variance (ANOVA) to an estimate of the frequency of correct responses vs. incorrect responses produced by the participants. An in-depth discussion of the implications, both theoretical and methodological, subtended in the use of this procedure is reported in the present article. Criticisms to this common procedure are followed by the proposal of an analysis based on log-linear modeling as a better methodological expedient for drawing correct inferences from accuracy results obtained in attentional tasks.