Teresa Hurley. 2. 31/03/2012. Attrition in on-line learning. On-line learning dynamic and potentially enriching but attrition remains a serious problem.Reported attrition from on-line learning as high as 70 - 80% (Flood 2002, Forrester 2000, in Dagger and Wade, 2004).. Teresa Hurley. 3. 31/03/2012
Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author.While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server.
1. “MotSaRT”- Motivation Strategies: A Recommender Tool for On-line Learning Facilitators
National College of Ireland
2. Teresa Hurley 2 31/03/2012 Attrition in on-line learning On-line learning dynamic and potentially enriching but attrition remains a serious problem.
Reported attrition from on-line learning as high as 70 - 80% (Flood 2002, Forrester 2000, in Dagger and Wade, 2004).
3. Teresa Hurley 3 31/03/2012 Monitoring Motivation Levels Classrooms teachers infer learners’ motivation from:
speech; behavior; attendance; body language and feedback.
Teachers intervene appropriately depending on learner behavior.
Intelligent Tutoring Systems (ITS) need to:
be able to recognize when the learner is becoming demotivated.
intervene with effective motivational strategies.
4. Teresa Hurley 4 31/03/2012 Intelligent Tutoring Systems (ITS) ITS would comprise two main components:
an assessment mechanism that infers the learner’s level of motivation from observing the learner’s behavior.
an adaptation component that selects the most appropriate intervention strategy to increase motivation.
5. Teresa Hurley 5 31/03/2012 Research Question Can intervention strategies increase motivation in adaptive online learning?
The focus of this research is intervention strategies to increase motivation and reduce attrition.
6. Teresa Hurley 6 31/03/2012 Eliciting Intervention Strategies from On-Line Facilitators Learner model created based on Bandura’s Social Cognitive Theory constructs: self-efficacy, goal orientation, locus of control and perceived task difficulty.
21 profiles were selected from possible 48 as the most likely to experience demotivation (Table 1).
7. Teresa Hurley 7 31/03/2012 Selected Persona Profiles
8. Teresa Hurley 8 31/03/2012 Example Persona Persona 1: “Chris is an intelligent student who enjoys learning for its own sake. She is motivated to learn new things and enjoys being challenged (GO:Mastery). She believes she can do very well in her studies as she has a very good understanding of her subject (SE:High). Chris believes hard work will conquer almost any problem and lead to success (LOC:Internal). However, she finds that she becomes bored when she has to work on a concept which she already understands well (PTD:Low)”.
9. Teresa Hurley 9 31/03/2012 On-line Survey To identify rules to determine which intervention strategy is the most appropriate for each learner’s persona, on-line facilitators were surveyed.
required to have at least two years experience teaching on-line, average was five years.
The facilitators were asked to
select the strategies they would Highly Recommend, Recommend or considered Not Applicable for each persona.
suggest any further strategies that they find useful in the case of each persona type.
10. Teresa Hurley 10 31/03/2012 Intervention Strategies Table 2. Intervention strategies
11. Teresa Hurley 11 31/03/2012 Who participated? Sixty participants completed the surveys
each persona got a minimum of six and a maximum of fourteen responses.
Emails, listservs – EDTECH and DOES, personal contacts, colleges offering on-line courses.
Ireland, America, Sweden, Australia, Austria, UK, New Zealand.
12. Teresa Hurley 12 31/03/2012 Survey Analysis Using Weka data mining tool set, five algorithms applied to predict whether a strategy was marked as recommended by the facilitators.
All experiments were run with a 10-fold cross validation. J48 decision trees turned out to provide the best predictions (Table 4).
13. Teresa Hurley 13 31/03/2012 Correct predictions (%) of the J48 decision tree algorithm Table 4
14. Teresa Hurley 14 31/03/2012 MotSaRT – Motivational Strategies: A Recommender Tool Using the recommendations derived from study, developed a recommender tool, MotSaRT (Fig 1) to support online facilitators when motivating learners.
Its functionality enables the facilitator to specify the learner’s motivation profile.
MotSaRT then recommends the most likely intervention strategies to increase motivation for any particular profile.
15. Teresa Hurley 15 31/03/2012 MotSaRT – Motivational Strategies: A Recommender Tool Technically, MotSaRT is a Java Applet and can be integrated into most L[C]MS fairly easily (Fig 2).
Observing the activities of learners and interacting with them synchronously or asynchronously through instant massaging, email or fora, facilitators assess learners’ self-efficacy, goal-orientation, locus of control and perceived task difficulty.
MotSaRT classifies the profile and sorts the strategies in terms of their applicability.
Facilitators can then plan their interventions according to these recommendations.
16. Teresa Hurley 16 31/03/2012 Screenshots of MotSaRT Fig 1
Demonstration on Moodle
As you can see the interface now separates the strategy (middle field) and an explanation of a strategy (lower field) which appears when you click on a strategy. The question marks would give an explanation of the concepts. So you could think of additional explanations or examples in the lower field for each strategy if you like. For demonstration purposes at EdTech the whole thing would be embedded in a mock-up Moodle page. As you can see the interface now separates the strategy (middle field) and an explanation of a strategy (lower field) which appears when you click on a strategy.
17. Teresa Hurley 17 31/03/2012 High Level Architecture Fig 2. MotSaRT will implement the path on left
hand side of the figure outlined in blue.
18. Teresa Hurley 18 31/03/2012 Future Perspectives Vision is to develop an automated tool which can be used in a fully automatic system, a semi-automatic system or in a manual system.
Diagnosis may be made either by a facilitator or by automatic assessment.
The diagnosis will be fed into the learner model.
MotSaRT can then be used to either make recommendations to the facilitator or to make an automatic intervention.
19. Teresa Hurley 19 31/03/2012 Can intervention strategies increase motivation in on-line learning? Problem – High attrition rates from on-line courses
Motivation improves attrition and increases learning
Which intervention strategies will motivate learners
Learner Model based on Social Cognitive Theory constructs
Phase 1 - Extraction of Intervention Strategies
Online facilitators’ survey
Validate Intervention Strategies