Applying Network Analysis to Think-Aloud and Eye-Tracking Data

• In Event: Sunday Roundtable 9:45 am
  In Roundtable Session: Innovative Approaches to Applying Technology in Science Education Research (Table 31)

Abstract

We explored application of network analysis (NA) to science text think-aloud (TA) and eye tracking (ET) data. We constructed knowledge networks for 30 undergraduate students based on subject-verb-object phrases during TAs, and analyzed scanpaths of ET sequences. For TA metrics learners with larger and more interconnected networks comprehended better. The ET data revealed one metric (diameter) negatively correlated with comprehension. Stepwise regression showed TA number of nodes and network eigencentrality, and mean distance from ET data, were significant. Students who picked up more nouns while reading, connected them in straighter causal chains, and had more interconnected ET patterns did better. Overall, this bimode NA approach demonstrated its potential in explaining substantial variance in reading comprehension for multiple text sets.

Authors

• Jennifer G. Cromley, University of Illinois at Urbana-Champaign

• Andrea J. Kunze, Delta State University