Week 13 — E-Learning Systems & Educational Technology¶

Course Objective: CO8 — Design and evaluate technology-mediated learning environments, applying instructional design principles and analyzing learning outcomes through data-driven frameworks.

Learning Objectives¶

By the end of this week, you should be able to:

• [x] Trace the historical evolution of e-learning from CBT through AI-adaptive systems
• [x] Compare major LMS platforms (Canvas, Blackboard, Moodle, D2L) on architecture, features, and market position
• [x] Explain SCORM and xAPI standards and articulate why content interoperability matters
• [x] Apply the ADDIE instructional design model to plan a digital learning experience
• [x] Evaluate MOOC business models and the factors affecting completion rates
• [x] Design gamification elements appropriate for corporate or academic e-learning contexts
• [x] Audit an e-learning course against Section 508 / WCAG accessibility requirements
• [x] Interpret learning analytics dashboards and identify at-risk students using early alert indicators

1. E-Learning Evolution: CBT to AI-Adaptive Systems¶

1.1 The Historical Arc¶

E-learning did not emerge fully formed with the internet — it evolved over five decades, each era building on the technological and pedagogical innovations of the previous one.

1.2 The Pendulum of Educational Philosophy¶

E-learning technology choices are never value-neutral — they embed particular educational philosophies:

2. Learning Management Systems¶

2.1 LMS Architecture¶

A Learning Management System is a software platform that manages, delivers, and tracks educational content and learner progress. The core components of any LMS include:

┌───────────────────────────────────────────────────────────┐
│                    PRESENTATION LAYER                      │
│  Student Browser  │  Instructor Browser  │  Mobile App    │
└──────────────────────────┬────────────────────────────────┘
                           │
┌──────────────────────────▼────────────────────────────────┐
│                   APPLICATION LAYER                        │
│  ┌──────────────┐  ┌─────────────┐  ┌──────────────────┐  │
│  │Course & Content│ │Assessment  │  │Communication     │  │
│  │Management    │  │Engine      │  │(Discussion,      │  │
│  │              │  │            │  │ Announcements)   │  │
│  └──────────────┘  └─────────────┘  └──────────────────┘  │
│  ┌──────────────┐  ┌─────────────┐  ┌──────────────────┐  │
│  │Gradebook &   │  │Analytics & │  │Integration APIs  │  │
│  │Reporting     │  │Dashboards  │  │(LTI, REST)       │  │
│  └──────────────┘  └─────────────┘  └──────────────────┘  │
└──────────────────────────┬────────────────────────────────┘
                           │
┌──────────────────────────▼────────────────────────────────┐
│                   DATA LAYER                               │
│  Relational DB (PostgreSQL/MySQL)  │  File Store (S3)     │
│  Video Platform (Kaltura/Panopto)  │  Search Index        │
└───────────────────────────────────────────────────────────┘

2.2 Canvas LMS¶

Canvas (Instructure), launched in 2011, has become the most widely adopted LMS in higher education in the United States. As of 2024, Canvas is used by over 4,000 educational institutions and approximately 30 million learners.

Canvas differentiating features: - SpeedGrader: Side-by-side submission review and rubric-based grading; supports video feedback recording - Outcomes and Mastery: Built-in competency tracking aligned to learning objectives - Canvas Studio (formerly Arc): Integrated video creation, hosting, and in-video quiz insertion - Commons: Shared repository of open educational resources that instructors can import and remix - Blueprint Courses: Master course template syncing to multiple child course sections - API-first architecture: Rich REST API enabling deep integration with other institutional systems

Canvas API example — fetching student submissions:

import requests

canvas_url = "https://your-institution.instructure.com"
api_token = "your_api_token"

headers = {"Authorization": f"Bearer {api_token}"}

# Get all submissions for a specific assignment
response = requests.get(
    f"{canvas_url}/api/v1/courses/12345/assignments/67890/submissions",
    headers=headers,
    params={"include[]": ["user", "rubric_assessment"], "per_page": 50}
)

submissions = response.json()
for submission in submissions:
    print(f"Student: {submission['user']['name']}, Score: {submission['score']}")

2.3 Blackboard Learn¶

Blackboard (now owned by Anthology following a 2021 merger with Anthology Inc.) was the dominant LMS from the early 2000s through the 2010s. Despite losing market share to Canvas and Moodle, it remains widely deployed, particularly in larger universities that made substantial investments in the platform.

Blackboard's evolution challenges: - Legacy architecture (originally built on early-2000s J2EE stack) created performance and usability problems - Blackboard Ultra: Complete UX rewrite launched 2016; still being adopted by institutions migrating from Classic - High licensing costs relative to open-source alternatives - Acquisition history (WebCT, ANGEL, Moodlerooms) created a fragmented product portfolio

2.4 Moodle¶

Moodle (Modular Object-Oriented Dynamic Learning Environment), created by Martin Dougiamas and first released in 2002, is the world's most widely deployed LMS by installation count — over 400 million users in 242 countries — primarily because it is free and open source.

Moodle's strengths: - Zero licensing cost (hosting and customization costs remain) - Highly extensible plugin architecture (1,900+ plugins in the directory) - Strong pedagogical philosophy (constructivism-influenced design) - Active global community; localized in 150+ languages - MoodleCloud: Hosted SaaS option for institutions without IT capacity

Moodle's challenges: - Default UI has historically been criticized for being visually dated - Customization requires PHP development expertise - Total cost of ownership (self-hosting, IT staff, customization) can exceed proprietary alternatives - Update cycle management complexity for highly customized installations

2.5 D2L Brightspace¶

Desire2Learn (D2L) Brightspace occupies a premium position in the LMS market, particularly strong in K-12 and corporate learning. D2L's differentiator is its Intelligent Agents feature and its early investment in learning analytics.

D2L distinctive capabilities: - Intelligent Agents: Rule-based automation that triggers actions when learner behavior patterns match conditions (e.g., "if student has not logged in for 5 days AND has <60% grade, send motivational email and notify advisor") - Learning Outcomes: Comprehensive competency framework aligned to curriculum standards - Performance+: Predictive analytics dashboard showing at-risk students with recommended interventions - Brightspace Creator+: Authoring tool for interactive content creation within the LMS

2.6 LMS Market Comparison¶

3. Content Interoperability Standards¶

3.1 SCORM: The Foundation¶

SCORM (Sharable Content Object Reference Model) was developed by ADL (Advanced Distributed Learning) initiative of the U.S. Department of Defense, first released in 1999. SCORM solved a critical problem: e-learning content created for one LMS couldn't run on another.

SCORM defines: 1. Packaging: How to bundle course files into a ZIP archive with a imsmanifest.xml describing the content structure 2. Runtime communication: How a course (SCO — Sharable Content Object) communicates with the LMS via a JavaScript API 3. Sequencing and navigation: Rules for how learners progress through course components

SCORM runtime API — key data elements:

// SCORM 2004 API communication (simplified)
var API_1484_11 = window.parent.API_1484_11;  // Get LMS API

// Initialize communication
API_1484_11.Initialize("");

// Record learner progress
API_1484_11.SetValue("cmi.completion_status", "completed");
API_1484_11.SetValue("cmi.success_status", "passed");
API_1484_11.SetValue("cmi.score.raw", "85");
API_1484_11.SetValue("cmi.score.min", "0");
API_1484_11.SetValue("cmi.score.max", "100");
API_1484_11.SetValue("cmi.score.scaled", "0.85");

// Set session time (ISO 8601 duration)
API_1484_11.SetValue("cmi.session_time", "PT1H23M45S");

// Save and terminate
API_1484_11.Commit("");
API_1484_11.Terminate("");

SCORM versions: - SCORM 1.1 (2000): Rarely used; superseded quickly - SCORM 1.2 (2001): Most widely deployed version; used by majority of existing SCORM content - SCORM 2004 (2004–2009): 4 editions; advanced sequencing/navigation capabilities; less widely adopted due to complexity

3.2 xAPI (Tin Can API / Experience API)¶

xAPI (also called Tin Can API or Experience API) was released in 2013 as a fundamentally different approach to learning data capture. Instead of course-to-LMS communication via a JavaScript API, xAPI uses simple HTTP POST statements in JSON format to record any learning experience to a Learning Record Store (LRS).

xAPI statement structure:

{
  "actor": {
    "mbox": "mailto:jsmith@example.com",
    "name": "John Smith",
    "objectType": "Agent"
  },
  "verb": {
    "id": "http://adlnet.gov/expapi/verbs/completed",
    "display": {"en-US": "completed"}
  },
  "object": {
    "id": "https://example.com/courses/introduction-to-sql",
    "definition": {
      "name": {"en-US": "Introduction to SQL"},
      "type": "http://adlnet.gov/expapi/activities/course"
    },
    "objectType": "Activity"
  },
  "result": {
    "score": {"scaled": 0.92, "raw": 92, "min": 0, "max": 100},
    "completion": true,
    "success": true,
    "duration": "PT2H15M"
  },
  "timestamp": "2024-03-15T14:30:00.000Z"
}

What xAPI enables that SCORM cannot: - Track mobile offline learning (statements submitted when connection restores) - Track simulations, games, physical training, on-the-job experiences - Track reading (time spent on article pages) - Track video watching (play, pause, seek, complete events) - Aggregate learning data from multiple systems into one LRS

4. Instructional Design Models¶

4.1 ADDIE¶

ADDIE (Analysis, Design, Development, Implementation, Evaluation) is the foundational instructional design process model. Developed in the 1970s at Florida State University for the U.S. Army, it remains the most widely taught and practiced framework.

┌─────────────┐    ┌─────────────┐    ┌─────────────┐
│   ANALYSIS  │───▶│   DESIGN    │───▶│ DEVELOPMENT │
│             │    │             │    │             │
│ • Who are   │    │ • Learning  │    │ • Build     │
│   learners? │    │   objectives│    │   content   │
│ • What are  │    │ • Sequence  │    │ • Create    │
│   the gaps? │    │ • Assessment│    │   media     │
│ • Context?  │    │   strategy  │    │ • Alpha test│
└─────────────┘    └─────────────┘    └─────────────┘
                                              │
┌─────────────┐    ┌─────────────┐           │
│ EVALUATION  │◀───│IMPLEMENTATION│◀──────────┘
│             │    │             │
│ • Kirkpatrick│   │ • Pilot     │
│   levels    │   │ • Deploy    │
│ • Analytics │    │ • Support   │
│ • Iteration │    │ • Maintain  │
└─────────────┘    └─────────────┘

ADDIE phases in detail:

4.2 Bloom's Taxonomy for Learning Objectives¶

Well-designed e-learning must specify measurable learning objectives at the appropriate cognitive level. Bloom's Revised Taxonomy (Anderson & Krathwohl, 2001) provides a six-level hierarchy:

4.3 Kirkpatrick Model for Evaluation¶

Donald Kirkpatrick's four-level training evaluation model is the standard framework for measuring e-learning effectiveness:

5. MOOC Platforms and Business Models¶

5.1 The MOOC Landscape¶

Massive Open Online Courses emerged from a 2008 experiment by George Siemens and Stephen Downes (Connectivism and Connective Knowledge, 1,900+ external students enrolled alongside 25 for-credit students at University of Manitoba). The modern MOOC era launched in fall 2011 when Stanford opened three CS courses to the public simultaneously.

Major MOOC platforms:

5.2 MOOC Completion Rate Problem¶

MOOC completion rates are notoriously low — typically 3-15% of enrolled students complete a course. This number requires careful interpretation:

Factors improving MOOC completion rates: 1. Paid enrollment (financial commitment increases follow-through) 2. Cohort-based delivery with fixed deadlines and peer interaction 3. Short overall course length (4 weeks vs. 12 weeks) 4. Strong community features (discussion forums, study groups) 5. Practical, job-relevant content with clear career outcomes 6. Peer accountability mechanisms

6. Assessment Design in LMS¶

6.1 Assessment Types in Online Learning¶

6.2 Academic Integrity: Turnitin Integration¶

Turnitin is the leading plagiarism detection service, analyzing submitted text against: - 70+ billion web pages - 69+ million published academic works - Turnitin's own database of previously submitted student work

How Turnitin works:

Student submits paper
       │
       ▼
Turnitin fingerprints text (rolling hash comparison)
       │
       ▼
Similarity report generated (0-100% similarity score)
       │
       ▼
Instructor reviews matches highlighted in original document
       │
       ├── High similarity to web source → potential web plagiarism
       ├── High similarity to database paper → potential academic plagiarism
       └── High similarity to own prior work → potential self-plagiarism

6.3 Quiz Banks and Randomization¶

Effective online assessment uses question banks with randomization to reduce the effectiveness of sharing answers among students:

Question Pool: 50 questions on Topic A
Question Pool: 40 questions on Topic B
Question Pool: 30 questions on Topic C

Student quiz drawn from:
  - 10 random questions from Topic A pool
  - 8 random questions from Topic B pool
  - 7 random questions from Topic C pool
  = 25 question quiz; no two students get identical questions

7. Gamification in E-Learning¶

7.1 Gamification Theory¶

Gamification is the application of game design elements to non-game contexts to increase engagement and motivation. In e-learning, gamification taps into intrinsic and extrinsic motivation mechanisms identified by Self-Determination Theory (Deci & Ryan): autonomy, competence, and relatedness.

7.2 Gamification Elements in E-Learning¶

8. Accessibility in E-Learning¶

8.1 Section 508 for Educational Technology¶

Educational institutions receiving federal funding (virtually all U.S. colleges and universities) must ensure their technology is accessible under Section 508 of the Rehabilitation Act. For e-learning, this means:

• LMS accessibility: The platform itself must be accessible (VPAT — Voluntary Product Accessibility Template — documents compliance)
• Course content accessibility: Instructor-created content must also be accessible
• Third-party tool accessibility: Every integrated tool (Turnitin, VoiceThread, Kahoot, etc.) must have documented accessibility compliance

8.2 E-Learning Specific Accessibility Requirements¶

9. Learning Analytics and AI-Adaptive Learning¶

9.1 Learning Analytics Framework¶

Learning analytics is the measurement, collection, analysis, and reporting of data about learners and their contexts to optimize learning and the environments in which it occurs (Society for Learning Analytics Research, 2011).

Levels of learning analytics:

┌─────────────────────────────────────────────────────────────┐
│  INSTITUTIONAL ANALYTICS: Enrollment patterns, completion   │
│  rates, program-level outcomes, accreditation reporting     │
├─────────────────────────────────────────────────────────────┤
│  COURSE-LEVEL ANALYTICS: Engagement by module, quiz item    │
│  analysis, discussion participation patterns               │
├─────────────────────────────────────────────────────────────┤
│  LEARNER-LEVEL ANALYTICS: Individual progress, time-on-task,│
│  performance trajectories, at-risk identification          │
└─────────────────────────────────────────────────────────────┘

9.2 Early Alert Systems¶

Early alert systems (also called early warning systems) use predictive analytics to identify students showing behavioral indicators of academic risk before they fail — when intervention is still effective.

Common early alert indicators: - Login frequency below course average - Days since last login (e.g., no login in 7+ days) - Assignment submission rate (missing or late submissions) - Quiz/assessment scores declining over time - Discussion forum non-participation - Grade currently below course passing threshold

# Simplified early alert algorithm
def calculate_risk_score(student_data):
    risk_score = 0

    # Days since last login (max 30 points)
    days_inactive = student_data['days_since_login']
    risk_score += min(days_inactive * 2, 30)

    # Missing assignments (max 40 points)
    missing_pct = student_data['missing_assignments'] / student_data['total_assignments']
    risk_score += missing_pct * 40

    # Current grade below threshold (max 30 points)
    if student_data['current_grade'] < 70:
        grade_deficit = 70 - student_data['current_grade']
        risk_score += min(grade_deficit, 30)

    return risk_score  # 0-100; >60 triggers alert

# Alert levels
def get_alert_level(risk_score):
    if risk_score >= 75:
        return "RED - Immediate intervention required"
    elif risk_score >= 50:
        return "YELLOW - Proactive outreach recommended"
    else:
        return "GREEN - No action required"

9.3 AI-Adaptive Learning Systems¶

AI-adaptive learning uses machine learning algorithms to continuously adjust the learning experience based on individual learner performance, pace, and behavior patterns.

Khan Academy's adaptive engine: Khan Academy's mastery-based system uses a knowledge graph of learning dependencies and learner performance data to determine: - Which skill to present next based on prerequisite mastery - When a student has demonstrated sufficient mastery to progress - What type of hint to provide when a student is struggling

Duolingo's algorithms: Duolingo uses Bayesian knowledge tracing to model the probability that a learner has truly learned a concept (vs. lucky guessing). The spaced repetition algorithm determines when to review previously learned content based on forgetting curve models.

P(learned | correct answer) = P(learned) × P(correct | learned) /
                               [P(learned) × P(correct | learned) +
                                P(not learned) × P(correct | not learned)]

10. Corporate Training and eLearning ROI¶

10.1 Corporate E-Learning Market¶

Corporate training represents a massive application of e-learning technology. U.S. companies spent $101.8 billion on corporate training in 2023 (Statista), with approximately 42% delivered through e-learning platforms.

Corporate LMS leaders (different from academic LMS): - Cornerstone OnDemand: Full talent management suite with integrated LMS - SAP SuccessFactors Learning: Deep HCM integration for global enterprises - Workday Learning: Native integration with Workday HCM/Payroll - LinkedIn Learning: Content library of 20,000+ courses plus LMS capabilities - Docebo: AI-driven enterprise LMS with skills inference

10.2 Measuring E-Learning ROI¶

The Kirkpatrick model (discussed in Section 4.3) provides the framework, but corporate learning ROI requires converting Level 3 and Level 4 data to financial metrics:

E-Learning ROI Formula:

ROI (%) = [(Benefits - Costs) / Costs] × 100

Where:
  Benefits = (improvement in performance metric × business value per unit)
  Costs = development + delivery + learner time + technology + administration

Example:
  Sales training increased average deal size from $45,000 to $52,000
  100 salespeople trained × $7,000 additional revenue = $700,000 benefit
  Training cost: $150,000 development + $50,000 delivery = $200,000

  ROI = ($700,000 - $200,000) / $200,000 × 100 = 250%

Key Vocabulary¶

Review Questions¶

Further Reading¶

• Clark, R. C., & Mayer, R. E. (2016). E-Learning and the Science of Instruction (4th ed.). Wiley.
• Means, B., Bakia, M., & Murphy, R. (2014). Learning Online: What Research Tells Us About Whether, When and How. Routledge.
• Siemens, G. (2013). "Learning Analytics: The Emergence of a Discipline." American Behavioral Scientist, 57(10), 1380-1400.
• Nonaka, I., & Takeuchi, H. (1995). The Knowledge-Creating Company. Oxford University Press. (Chapter 3 on knowledge spiral)
• ADL Initiative. (2023). xAPI Specification. https://github.com/adlnet/xAPI-Spec
• IMS Global. (2023). LTI Advantage Specification. https://www.imsglobal.org/lti-advantage-overview
• Anderson, L. W., & Krathwohl, D. R. (2001). A Taxonomy for Learning, Teaching, and Assessing: A Revision of Bloom's Taxonomy. Longman.
• Kirkpatrick, J. D., & Kirkpatrick, W. K. (2016). Kirkpatrick's Four Levels of Training Evaluation. ATD Press.
• Kapp, K. M. (2012). The Gamification of Learning and Instruction. Pfeiffer/Wiley.

← Week 12 | Course Index | Week 14 →