Utilizing the RADSM Process: Developing an Unobtrusive Measure of Cohesion

Abstract

In this study, verbal indicators produced following the RADSM process were used to code transcript data collected from 62 three-person teams playing a cooperative bridge crew simulator. A training dataset, consisting of 88 mission transcripts, and a testing dataset, consisting of 39 mission transcripts, were created from the data. Multilevel modeling was employed to develop a linear-regression algorithm to predict/measure team task and social cohesion. The findings revealed that task cohesion perceptions are linked to the proportion of speech dedicated to information requests and instructions, whereas social cohesion perceptions are associated with the proportion of speech dedicated to amusing and humoring others. Although the regression equations did not demonstrate convergent validity, these findings provide a foundation for future iterations.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Dec 01, 2021
Accession Number
AD1230006

Entities

People

  • Zachary L. Rahner

Tags

Readers

  • Geospatial Intelligence and Artificial Intelligence Analytics
  • Regression Analysis.
  • Team-Based Human-Centered Cognitive Task Decision Making and Information Performance.