Measuring Cognitive Load: A Comparison of Self-report and Physiological Methods

i ABSTRACT This study explored three methods to measure cognitive load in a learning environment using four logic puzzles that systematically varied in level of intrinsic cognitive load. Participants' perceived intrinsic load was simultaneously measured with a self-report measure—a traditional subjective measure—and two objective, physiological measures based on eye-tracking and EEG technology. In addition to gathering self-report, eye-tracking data, and EEG data, this study also captured data on individual difference variables and puzzle performance. Specifically, this study addressed the following research questions: 1. Are self-report ratings of cognitive load sensitive to tasks that increase in level of intrinsic load? 2. Are physiological measures sensitive to tasks that increase in level of intrinsic load? 3. To what extent do objective physiological measures and individual difference variables predict self-report ratings of intrinsic cognitive load? 4. Do the number of errors and the amount of time spent on each puzzle increase as the puzzle difficulty increases? Participants were 56 undergraduate students. Results from analyses with inferential statistics and data-mining techniques indicated features from the physiological data were sensitive to the puzzle tasks that varied in level of intrinsic load. The self-report measures performed similarly when the difference in intrinsic load of the puzzles was the most varied. Implications for these results and future directions for this line of research are discussed. ii ACKNOWLEDGMENTS I would like to thank my advisor Dr. Robert Atkinson for the inspiration to pursue my research interests in the LSRL. The experience was invaluable as he gave me the independence to explore and the encouragement and feedback necessary to grow as a researcher. I had access to a wealth of lab resources, a team of intelligent colleagues, and the opportunity for rich and meaningful discussions about measurement and data analysis. I will forever be grateful for these opportunities. Lin for his friendship and discussions about statistics. I also appreciate Robert Christopherson for supporting the technology in the lab.