Improving eLearning through Statistical Feedback Bachelors Thesis - PowerPoint PPT Presentation

Improving eLearning through Statistical Feedback Bachelor’s Thesis – Final Talk Student: Beatris Burdeva Supervisor: Prof. Dr.-Ing. Georg Carle Advisers: Marc-Oliver Pahl, Stefan Liebald Garching, 10. July 2017

Overview I. Motivation II. Reasoning behind our Implementation Choices - Analysis, Initial Ideas, Related Work III. Implemented Artifacts - Input - Processing - Visualizations IV. Future Work 2 Beatris Burdeva (TUM) | Bachelor‘s thesis | Final talk | Improving eLearning through statistical feedback

I. Motivation • labsystem • eLearning platform for content creation and course management • ilab1 and ilab2 • How to improve the learning process of ilab participants by making use of the statistical data that is being gathered? • Additional features (feedback elements)? • What data? • What visualizations? 3 Beatris Burdeva (TUM) | Bachelor‘s thesis | Final talk | Improving eLearning through statistical feedback

II. Reasoning behind our Implementation Choices -Analysis Best practice: Explanation: Simplicity [1] easy to understand and use focus on the content, not organization Interactivity [1, 2] collaboration Best practices different perspectives for creating a good learning social and cooperative skills process Diversity [3] diverse strategies to communicate new contents and to assess learners' progress Focus - on the Authenicity [1] relevant problems learners learning-by-doing make mistakes and learn from them Adaptivity [2, 4] good practices from traditional education Privacy [5] respective and protective of personal data 4 Beatris Burdeva (TUM) | Bachelor‘s thesis | Final talk | Improving eLearning through statistical feedback

II. Reasoning behind our Implementation Choices -Analysis (Related Work) What is a good learning process? o Simplicity o Interactivity o Diversity o Implementation features: o Authenticity o graphical representation of o Adaptivity o the answers (e.g. with Privacy o boxes) coloring of the answers o Basic set of feedback elements: o ... o email o forum o chat o multiple choice questions/quizzes o labsystem, Coursera, edX, Moodle free input fields/peer assessments o (analytical) visualizations o 5 Beatris Burdeva (TUM) | Bachelor‘s thesis | Final talk | Improving eLearning through statistical feedback

II. Reasoning behind our Implementation Choices -Analysis -Related Work Best practice: Features, contributing to fulfilling the practice: Coursera edX Moodle labsystem Simplicity [1] Option to sort Separation of posts Restrict chat to a Link each element (e.g. posts by e.g. votes in categories certain group of question) to an email body users Interactivity [1, 2] Comments section, visible only Option to follow Option to upload files Time left until posts finish-deadline to tutors shown Diversity [3] Red/green coloring of answers Option for learners Check/X marks for Option to choose to specify if they the right/wrong certainty level want to receive answers emails from organizations that represent the course Authenicity [1] More than one attempt possible Explanation to the More than one Graphical given answers attempt possible to representation of to answer a question (why correct/ answer a question number of incorrect) registrations per country Adaptivity [2, 4] Blended learning Blended learning Blended learning Blended learning Privacy [5] Visibility rights Privacy policy Privacy policy Privacy policy 6 Beatris Burdeva (TUM) | Bachelor‘s thesis | Final talk | Improving eLearning through statistical feedback

II. Reasoning behind our Implementation Choices - Related Work Main Findings: Missing basic feedback elements in the labsystem: o Chat and forum o  However: ticket system, email system Visualizations by the other three platforms: o Coursera and edX o OS, browser, cookies, etc. o No information of how exactly they process and visualize this data. o Moodle – users per site, registrations per country, etc. o Conclusion. o More suitable for large scale courses. o No information found about visualizations relevant for us. o Stay with our initial ideas. o 7 Beatris Burdeva (TUM) | Bachelor‘s thesis | Int ermediate talk | Improving eLearning through statistical feedback

II. Reasoning behind our Implementation Choices -Initial Ideas Artifact Contributes to: Audience 1. Time Analysis Simplicity, -recognition of time Tutors and Interactivity, related issues advisers Diversity -overview of the (learners - course workload indirect) 2. Time Left Simplicity, Interactivity, -increase self-refection Learners Diversity -assess performance Future work -better time planning 3. Worst Answered Simplicity, Diversity -recognition of tasks Tutors and Multiple Choice (MC) with worst achieved advisers and Input Questions results (learners - -adaptation of indirect) Implemented contents 4. Additional MC Simplicity, Diversity Tutors and -emphasis on Statistics (“clicks” per advisers misunderstood anwer) (learners - contents indirect) 5. Emotional Feedback - Simplicity, Diversity, -improve the emotional Tutors, advisers, like/dislike buttons Authenticity engagement of the learners D. Dimova learners -point out preferred contents 8 Beatris Burdeva (TUM) | Bachelor‘s thesis | Final talk | Improving eLearning through statistical feedback

II. Reasoning behind our Implementation Choices 1. Time Analysis 3. Worst Answered MC and Input Questions 4. Additional MC Statistics • existing overview page (Figure 1) - results in average, not just per team • comparison between terms Figure 1: “My statistics” overview page, demo 9 Beatris Burdeva (TUM) | Bachelor‘s thesis | Final talk | Improving eLearning through statistical feedback

Learning elements in the labsystem - Analysis • page (p), • multiple choice (m), • input (i), • collection (c/C), • lab (l), • schedule (s) Logic: p,m,i -> c -> C -> prelab/lab -> virtual page 10 Beatris Burdeva (TUM) | Bachelor‘s thesis | Final talk | Improving eLearning through statistical feedback

III. Implemented Artifacts 3. Worst Answered Multiple Choice (MC) Questions -Visualization 11 Beatris Burdeva (TUM) | Bachelor‘s thesis | Final talk | Improving eLearning through statistical feedback

III. Implemented Artifacts 3. Worst Answered Multiple Choice (MC) Questions -Input -Processing 12 Beatris Burdeva (TUM) | Bachelor‘s thesis | Final talk | Improving eLearning through statistical feedback

III. Implemented Artifacts 3. Worst Answered Input Questions -Visualization 13 Beatris Burdeva (TUM) | Bachelor‘s thesis | Final talk | Improving eLearning through statistical feedback

III. Implemented Artifacts 3. Worst Answered Input Questions -Input -Processing 14 Beatris Burdeva (TUM) | Bachelor‘s thesis | Final talk | Improving eLearning through statistical feedback

III. Implemented Artifacts 4. Additional MC statistics - clicks per answer -Input -Processing 15 Beatris Burdeva (TUM) | Bachelor‘s thesis | Final talk | Improving eLearning through statistical feedback

III. Implemented Artifacts 4. Additional MC statistics - clicks per answer -Visualization • gChartPHP [8] - PHP wrapper for the Google Chart API [10], published under the Apache License [11] 16 Beatris Burdeva (TUM) | Bachelor‘s thesis | Final talk | Improving eLearning through statistical feedback

III. Implemented Artifacts -Visualization 1. Time Analysis • gChartToolPHP [9] - PHP wrapper for the Google Chart API [10], • published under GNU General Public License [12] 17 Beatris Burdeva (TUM) | Bachelor‘s thesis | Final talk | Improving eLearning through statistical feedback

Already existing: Time spent per credit Figure 2: Time spent per credit [7] 18 Beatris Burdeva (TUM) | Bachelor‘s thesis | Final talk | Improving eLearning through statistical feedback

III. Implemented Artifacts 1. Time Analysis -Input -Processing 19 Beatris Burdeva (TUM) | Bachelor‘s thesis | Final talk | Improving eLearning through statistical feedback

IV. Future Work • Implementation of the features offered by Coursera, edX and Moodle (e.g. forum) • “Emotional feedback” - like/dislike buttons • Time Left • Adaptation/ complete new implementation of the existing time tracking algorithm - advantageous • Implementation of features, which make the time analysis easier and more precise. For example, buttons "pause", "revise" and "end" 20 Beatris Burdeva (TUM) | Bachelor‘s thesis | Final talk | Improving eLearning through statistical feedback