Extended Abstract: Probed Language-Based Assessments Using Spoken Language

Probed language-based assessments, where individuals describe their states using open-ended responses analyzed with Artificial Intelligence, have shown great promise for written language. But emotions are also conveyed through spoken language which might provide further assessment power to this method. Participants (N=645) were randomly assigned to watch a video clip of a church, shopping mall, or park, then describe their emotions using written and spoken language and closed-ended rating scales. Results: Combining written and spoken language increases the convergence with the rating scales. Spoken language demonstrates higher accuracy in classifying which video clip participants saw (AUC=0.80), as compared with written language (AUC=0.69). Thus, spoken language can complement written language responses for assessing affect in communication research.