All-Subset Regression as a Means for Selection of Self-Regulated Learning Processes Measured Using Think Aloud Protocol Data

During the 1990s computers were placed into most educational classrooms; however, they sat underused or not used at all. One reason for this is students lacked the skills to use computers effectively. One set of skills that can help students make use of computers is self-regulated learning. By using think aloud protocol analysis while students complete a task on the computer, a trace of their cognition, metacognition, and behavior can be created. Analyzing these traces, however, has proven difficult due to the high number of variables compared to the typical number of participants. A solution to dealing with this problem is to analyze all possible combinations of variables. In this thesis, I compared the results of two pre-existing variable reduction methods and Best All Subset Regression. It was found that Best All Subset Regression outperformed the existing methods, by fitting better models without diagnostic problems or extensive time demands. Best All Subset Regression also retained more information than the prior methods, so I suggest using it moving forward instead of the aggregation-based methods used previously. [The dissertation citations contained here are published with the permission of ProQuest LLC. Further reproduction is prohibited without permission. Copies of dissertations may be obtained by Telephone (800) 1-800-521-0600. Web page: http://www.proquest.com/en-US/products/dissertations/individuals.shtml.]