Conference
Predicting Student Performance in a Collaborative Learning Environment
| Τίτλος: | Predicting Student Performance in a Collaborative Learning Environment |
|---|---|
| Γλώσσα: | English |
| Συγγραφείς: | Olsen, Jennifer K., Aleven, Vincent, Rummel, Nikol, International Educational Data Mining Society |
| Πηγή: | Grantee Submission. 2015. |
| Peer Reviewed: | Y |
| Page Count: | 7 |
| Ημερομηνία έκδοσης: | 2015 |
| Sponsoring Agency: | Institute of Education Sciences (ED) |
| Contract Number: | R305A120734 R305B090023 |
| Τύπος εγγράφου: | Speeches/Meeting Papers Reports - Research |
| Education Level: | Grade 4 Intermediate Grades Elementary Education Grade 5 Middle Schools |
| Descriptors: | Educational Environment, Predictive Measurement, Predictor Variables, Cooperative Learning, Regression (Statistics), Models, Problem Solving, Data Collection, Academic Records, Data Processing, Progress Monitoring, Goodness of Fit, Fractions, Grade 4, Grade 5, Prediction, Predictive Validity, Achievement Gains, Learning Strategies, Learning Processes |
| Περίληψη: | Student models for adaptive systems may not model collaborative learning optimally. Past research has either focused on modeling individual learning or for collaboration, has focused on group dynamics or group processes without predicting learning. In the current paper, we adjust the Additive Factors Model (AFM), a standard logistic regression model for modeling individual learning, often used in conjunction with knowledge component models and tutor log data. The extended model predicts performance of students solving problems collaboratively with an ITS. Specifically, we address the open questions: Does adding collaborative features to a standard AFM provide a better fit than the standard AFM? Also, does the impact of these features change based on the nature of the knowledge (conceptual v. procedural) that is being acquired? In our extended AFM models, we include a variable indicating if students are working individually or in pairs. Also, for students working collaboratively, we model both the influence on learning of being helped by a partner and helping a partner. For each model, we analyzed conceptual and procedural datasets separately. We found that both collaborative features (being helped and helping) improve the model fit. In addition, the impact of these features differs between the collaborative and procedural datasets, suggesting collaboration may affect procedural and collaborative learning differently. By adding collaborative learning features into an existing regression model for individual learning over a series of skill opportunities, we gain a better understanding of the impact that working with a partner has on student learning, when working with a step-based collaborative ITS. This work also provides an improved model to better predict when students have reached mastery while collaborating. [For the paper published in "Proceedings of the International Conference on Educational Data Mining (EDM) (8th, Madrid, Spain, June 26-29, 2015)," pages 211-217, see ED560503. For the individual paper published by the International Educational Data Mining Society, see ED560570.] |
| Abstractor: | As Provided |
| Number of References: | 25 |
| IES Funded: | Yes |
| Entry Date: | 2016 |
| Αριθμός Καταχώρησης: | ED564345 |
| Βάση Δεδομένων: | ERIC |
καταχωρήστε σχόλιο πρώτοι!