Cloze Procedure and Type Token Ratio: an aid to diagnosis

Abstract

SummaryDisturbances in the verbal communication patterns of manic and schizophrenic patients have been identified as two tools borrowed from quantitative linguistics: the Cloze Procedure and the Type Token Ratio (TTR). The Cloze Procedure, which measures the communicability of a message, consists of suppressing systematically every “nth” word of a text and in asking raters to try and guess the missing words. The Cloze Score of each text is then evaluated. The TTR measures the index of repetitiveness of a text or, in other words, its verbal richness. The degree of communicability, or overall comprehension, and the index of verbal richness are obtained through the analysis of a corpus of oral texts which have been recorded and transcribed. The patients are all paired with suitable controls. It was shown that low Cloze Scores (CS) and low TTR identify schizophrenic patients, whereas low CS and high TTR indicate a manic state. Control subjects show both high CS and high TTR. These results suggest that the Cloze Procedure linked with TTR can provide substantial aid to the differential diagnosis of mania and schizophrenia.