ASEE Presentation

ExplaNet: A Framework to Manage and Analyze Student-AuthoredCourse Content: 

ExplaNet: A Framework to Manage and Analyze Student-Authored Course Content University of California at Santa Cruz June 2003 Jessica Masters Tara Madhyastha Ali Shakouri

Introduction and Motivation: 

Introduction and Motivation Students think critically and explain difficult concepts have access to a variety of unique explanations are presented with an individualized choice of explanations Instructors utilize the framework for content development assess student comprehension of difficult concepts trace the flow of information throughout the class

ExplaNet: A Screen Shot: 

ExplaNet: A Screen Shot

The Experiment: 

The Experiment Computer Science 12B: Introduction to Data Structures Winter 2003 150 students This gateway course traditionally has at least a 25% failure rate. Students had one deadline to submit first answer, could only then view other answers and then submit their final answer.

Results: Student Improvement: 

Results: Student Improvement The average improvement was 0.75 points.

Results: Student Opinion Poll: 

Results: Student Opinion Poll What resources did you use in between your first and second submissions? Avg Grade: 2.40 Avg Improvement: 0.78 Avg Grade: 2.25 Avg Improvement 0.76 Avg Grade: 2.26 Avg Improvement: 0.22 Avg Grade: 2.37 13% 19% 53% 15%

Results: Students Can Identify Quality Answers: 

Results: Students Can Identify Quality Answers The correlation between the students’ rank of an answer and the grade of that answer is 0.54. Of all the answers that passed (grade >= 3), 79% of those answers had a rank of at least 3. Of the failing answers (grade < 3), 57% of those answers had a rank of less than 3. The best answers are viewed almost twice as much as all answers on average, and more than twice as much as failing answers. The most popular answers (those viewed twice as much) are of a better quality than all answers on average.

Conclusions and Future Work: 

Conclusions and Future Work Conclusions: improvement linked to viewing other answers students enjoy and learn from the framework students can be responsible for identifying the best answers The Future: a new interface input answers directly to the screen multiple choice question capability to automatically filter out incorrect answers intelligent collaborative filtering system for personalizing each student’s list of explanations http://www.collage.soe.ucsc.edu

Results: Content Development: 

Results: Content Development

Results: Improvement Between Submissions: 

Results: Improvement Between Submissions 91% of students who resubmitted answers viewed other students’ answers in between submissions. On average, students viewed 11 answers in between their first and second submission. The average improvement was 0.75 points (15%).

Results: Student Opinion Poll: 

Results: Student Opinion Poll Was the process of looking at other answers helpful to you? 58% 16% 20% 6%

Results: Student Opinion Poll: 

Results: Student Opinion Poll Did you learn more than from a traditional quiz? 62% 22% 16% “Yes. I don’t think quizzes teach anything so much as they test your level of understanding. Diagnostics allow you to challenge yourself.” “Yes I did learn more by completing the diagnostics because they push you to do more research on the question that you are trying to answer. Therefore, you better understand the concept. With a quiz I think I would have been more worried about the grade than learning.” 12%: Quizzes are better because they are just about the right and wrong answer.