The analysis of interactions in experiments on the specificity of schizophrenic thought disorder: a reply to Phillips

Phillips (1977) has criticized the investigation of statistical interactions in a series of related papers concerned with the specificity hypothesis of schizophrenic thought disorder. In a discussion of these criticisms, some are accepted and some rejected. However, when the relevant data are reanalysed by a correct method for the investigation of nominated interactions, the specificity hypothesis is broadly supported. A new finding to emerge from the reanalysis is that, in two experiments where level of 'difficulty' of constructs was controlled, the specificity effect appears much reduced or non-existent when intensity scores are employed while remaining intact for consistency scores.