The measurement of anxiety through the computer analysis of verbal samples

A method for measuring both labile emotional and more permanent psychological events through content analysis of verbal behavior has been developed by Gottschalk and Gleser, and provides an interesting challenge for “artificial intelligence,” the study of thought and action in forms that can be expressed in computer programs.

Abstract

This study aimed to cross-validate the capacity of a computer software program to detect and measure, using a measurement method applied to the content and form analysis of 5-minute speech samples, cognitive impairment and associated comorbid neuropsychiatric psychobiological dimensions in drug-abusing patients. At the University of California-Irvine (UCI) Neuropsychiatric Center, 28 drug-abusing inpatients using illegal drugs were clinically evaluated. Their scores for cognitive impairment derived by the computerized content analysis method were compared with scores derived from selected tests from the Halstead-Reitan Neuropsychological Test Battery, the computerized Automated Neuropsychological Assessment Metric Battery (ANAM), the Alzheimer's Disease Assessment Scale-Cognitive Portion, the Stroop Color and Word Test, the Symbol Digit Modalities Test, and the Controlled Oral Word Association Test. The statistical significance (P value) of the correlations of scores from these different measures with scores obtained from the computerized content analysis measures was less than .05 to .001. The comparative “hit rate,” detecting cognitive impairment above the norms for each measure administered to these drug-abusing patients, for the computerized content analysis measures and some of the ANAM neuropsychological measures was 75% to 89%, and for the other neuropsychological measures, 25% to 64%. In conclusion, the computerized content analysis methodology applied to 5-minute verbal samples is a valid, rapid, easily administered measurement instrument for assessing the magnitude of cognitive impairment and comorbid neuropsychiatric dimensions.