Language intensity: Testing a content‐based metric

This investigation assessed two versions (original, elaborated) of a content‐based metric for coding and calculating composite verbal intensity values for naturalistic discourse. Intensity scores were calculated for 50 sample talking turns using two versions of the scheme and compared with subjects’ intensity ratings of the 50 turns in two conditions. In the content condition subjects (n = 65) read each message and rated its intensity along a five‐point scale, while in the audio condition subjects (n = 65) listened to the recorded messages and indicated their perceived intensity using the same five‐point scale. Results indicate that composite intensity scores calculated via a revised (elaborated) version of the metric were not significantly different from subjects’ intensity ratings in the two conditions and were characterized by a significant positive correlation with those ratings. However, the composite intensity scores produced via the original version of the metric were significantly different from the subjects’ ratings and were marked by a low positive, yet nonsignificant correlation with those ratings.