Concept-Mapping as a Quantitative Measure of Reading Comprehension: Construct Validity and Internal Validity (Poster 17)

• In Event: AERA Poster Session Two
  In Poster Session: Measurement Theory and Application 1

Abstract

To better understand the manner in which students extract information from text and construct meaning from that information, concept maps were used as a measure of reading comprehension. 98 concept maps from students in grade 7-9 were analyzed using Rasch modeling to address issues of: 1) construct validity—item fit and item difficulty compared to the construct, and 2) internal validity—local item dependence and differential item functioning. Number of Nodes and Depth were dropped after the initial model due to local dependence. The resulting model was a reliable predictor of student abilities with an EAP of .75. Differential item functioning for grade, gender, and language proficiency showed that MainIdeas and Details were inversely related and developmentally dependent.

Author

• John Galisky, University of California - Santa Barbara
  Presenting Author