My work samples come from one project carried through all five ADDIE phases. The performance problem was real: a newly formed team of curriculum alignment specialists started work in the wrong course record nearly 30% of the time. That happened because they chose the wrong Course Identification Number (CID) from a catalog of more than 150,000. I used the Morrison, Ross, Kalman and Kemp model from Designing Effective Instruction as my working method inside the ADDIE framework. That model treats design as a flexible cycle of interdependent decisions rather than a strict linear sequence.
At a glance
| ADDIE phase | What I did | Where it shows |
|---|---|---|
| Analyze | Observed the team’s error rate, surveyed and interviewed learners, refined goals, analyzed learners and context, and broke the task into procedural and topical steps | Design document, sections 1–3 |
| Design | Wrote five performance objectives, classified each, aligned two assessment items to each, and planned sequencing, pre-instructional, and practice strategies | Design document, sections 4–5 |
| Develop | Built a self-paced, menu-driven eLearning module with a checklist job aid, narrated screencasts, branching practice, and a posttest | eLearning course |
| Implement | Planned release timing around sprint planning and remote, self-paced delivery | Design document, section 2 |
| Evaluate | Designed a four-stage formative evaluation: SME review, one-to-one, small group, and field trial, with five instruments | Design document, section 6 |
Analysis: confirming training was the right fix
I started by checking whether this was a training problem at all. Observation showed the error rate. An eight-item survey and follow-up interviews captured the team’s felt, expressed, and anticipated needs. Five of six specialists had been locating CIDs for only about three months, and two-thirds rated their confidence 3 or lower on a 0–5 scale. Five of six said training paired with documentation would help most. That pointed to a skill and knowledge gap, not a motivation or tool problem, which is the condition where training is the right intervention.
The data shaped the design directly:
- Length and format. Every respondent was willing to take a 15-minute self-paced module, so I capped instruction at roughly 15–20 minutes.
- Scope. Five of six wanted to learn both search methods, so the module covers both tools.
- A job aid. The preference for documentation led to a reusable CID Checklist that anchors the whole module.
- Practice. One learner asked for “examples that we need to go and find,” so every topic ends with a find-the-CID practice item.
Goal analysis narrowed eleven brainstormed goals to five. Each pass asked one question: is this a direct step in locating the correct CID? Useful reference material, such as course lists, was cut. Learner analysis described experienced, certified educators who already understood standards alignment but were new to the course system’s data model. Contextual analysis covered the orienting context (why they’d care), the instructional context (remote, mid-sprint), and the transfer context (applying it at the next sprint planning). Finally, a combined procedural and topical task analysis captured both the clicks an expert makes and the concepts behind them, such as how a base course differs from a linked course.
Design: aligning objectives, assessment, and strategy
Objectives and alignment. Each of the five goals became one measurable performance objective with a condition, behavior, and criterion. One example: “Given a set of course characteristics and the CID Checklist, the learner will identify the correct CID in a Course Manager results screen with 100% accuracy.” I classified each objective using Merrill’s content–performance matrix (concept or procedure, recall or application). I then wrote two posttest items per objective at the same level. Application objectives are tested by having learners find the CID in a realistic results screen, not by asking them to define terms.
Sequencing. I used task-expertise sequencing (English and Reigeluth), which moves from simplest to most complex. Learners first get the checklist, the decision framework used in every later step. Next they apply it in the simpler, form-based Course Manager tool, then repeat the same locate-then-verify pattern in the more complex Course Metadata Report spreadsheet.
Pre-instructional strategies. An overview states what the module covers, why it matters, and how long it takes. The objectives are shown up front so learners can see how each topic connects to the posttest.
Instructional strategies. Each topic follows the same rhythm:
- Initial presentation. For the checklist this is integration, presenting each item with its meaning. For the tool skills it is demonstration, a narrated screencast plus annotated screens.
- Generative strategy. Elaboration: learners paraphrase the checklist or try the search in their own system.
- Practice with feedback. A realistic find-the-CID question that branches to “Correct” or “Try again.”
Development and implementation: building the prototype
The contextual analysis set the delivery constraints: a fully remote team, working mid-sprint, using tools the company already had. I built the original prototype as a self-paced, non-linear module in PowerPoint, using hyperlinks and slide branching in place of an authoring tool. It includes:
- A navigation tutorial and main menu, so learners can jump to any topic or come back later as a refresher. The survey asked for refreshers every six months to a year.
- A course progression map showing all five topics, which reduces uncertainty about how long the module is.
- Consistent topic structure: a title slide, key content, a narrated screencast, then a practice question. Learners always know what comes next.
- Branching feedback after each practice item, with a path back to retry.
- A summary and a posttest in Google Forms, mapped item by item to the objectives.
For implementation, I planned to release the module just before sprint and user-story planning so learners could use the checklist on real tickets right away. The web version of the course on this site rebuilds the same design in a scrolling, block-based format.
Portfolio note: the screencasts were recorded inside a proprietary system, so in this version they are replaced by labeled placeholders, and every system screen is a mock-up with fictional data.
Evaluation: testing before scaling
The formative evaluation plan follows the classic staged approach described by Dick, Carey and Carey and adopted in Morrison et al. Each stage tests a revised version with more learners than the last:
| Stage | Testers | Purpose | Instruments |
|---|---|---|---|
| SME review | 1 SME | Content accuracy and alignment to objectives | SME notes form |
| One-to-one trials | 1–2 learners | Catch obvious errors and confusing screens | Observation notes, interview protocol, attitude survey |
| Small-group trials | 2–3 learners | Find strengths and weaknesses | Observation notes, interview protocol, attitude survey |
| Field trial | All 6 learners | Test the revised module in its real context | Attitude survey, posttest |
The plan answers four stakeholder questions: whether the objectives were met, whether the content is sufficient, whether it needs to expand, and whether other teams should take it. Posttest results are analyzed item by item, so a missed question points to a specific gap in the content rather than a vague “scores were low.” In Kirkpatrick’s terms, the attitude survey measures reaction (Level 1) and the posttest measures learning (Level 2). The business measure that started the project, the 30% wrong-CID rate, is the natural behavior and results (Levels 3–4) check after launch.
Adult learning theory in the design
My learners were experienced, certified educators, so the design leans on andragogy, Knowles’ model of how adults learn. Each of its six assumptions shows up in a concrete decision:
| Andragogy assumption | How it shaped the module |
|---|---|
| Need to know: adults want to know why before they learn | The overview opens with the cost of a wrong CID (rework, failed reviews, edits to live courses), and objectives are shown up front |
| Self-concept: adults are self-directed | Self-paced, menu-driven navigation lets learners skip what they know and return as needed |
| Experience: adults bring prior knowledge | Assumed what they already knew (standards alignment, system navigation) and taught only the new data model, keeping the module short |
| Readiness: adults learn what their job requires now | Released right before sprint planning, when the skill is immediately needed |
| Problem-centered orientation: adults learn best through real tasks | Organized around one job task, finding the right CID, with authentic find-the-CID practice instead of definitions |
| Intrinsic motivation | Framed around self-sufficiency: the skill removes the team’s dependency on another team |
Other theories guided specific choices:
- Gagné’s Nine Events of Instruction. The module moves through gaining attention and stating objectives (overview, objectives), presenting content (checklist, annotated screens), providing guidance (screencasts), eliciting performance with feedback (branching practice), assessing (posttest), and enhancing retention and transfer (a reusable checklist, timed release).
- Cognitive load theory and Mayer’s multimedia principles. I split the content into five short topics (segmenting). I used red callouts on screens to show where to look (signaling), paired narration with on-screen demonstrations (modality), and kept a fixed topic rhythm so learners spend effort on content, not navigation.
- Social learning theory (Bandura). Narrated screencasts let learners watch an expert perform the task before trying it, which is modeling.
- Situated learning and transfer. Practice uses realistic course data and real search scenarios, and the checklist goes back to the job with the learner.
- Keller’s ARCS model. Attention comes from a real problem and interactive branching. Relevance comes from authentic work examples. Confidence comes from short topics, practice before the test, and “Try again” feedback. Satisfaction comes from immediate correct-answer feedback and a skill learners use the same week.
- Bloom’s taxonomy and Merrill’s content–performance matrix. I used them to keep each objective, its practice, and its posttest items at the same level (remember vs. apply).
Looking back: what I’d do differently today
I completed this project in 2022, and my practice has grown since. If I rebuilt it now, I would:
- Build it in a dedicated authoring tool such as Articulate Storyline or Rise, with SCORM or xAPI tracking, instead of PowerPoint hyperlinks.
- Design for accessibility from the start: captions and transcripts for every screencast, meaningful alt text, keyboard-friendly navigation, and color contrast checked against WCAG.
- Make the checklist a standalone, printable job aid learners can keep open beside the system.
- Add scenario-based assessment that asks learners to explain why a CID is correct, not just pick it, to capture their reasoning.
- Measure transfer by tracking the team’s wrong-CID rate for a sprint or two after launch. That Level 3–4 evidence is what tells a business the training worked.
References
- Bandura, A. (1977). Social learning theory. Prentice Hall.
- Dick, W., Carey, L., & Carey, J. O. (2015). The systematic design of instruction (8th ed.). Pearson.
- Gagné, R. M., Wager, W. W., Golas, K. C., & Keller, J. M. (2005). Principles of instructional design (5th ed.). Wadsworth.
- Keller, J. M. (2010). Motivational design for learning and performance: The ARCS model approach. Springer.
- Kirkpatrick, J. D., & Kirkpatrick, W. K. (2016). Kirkpatrick’s four levels of training evaluation. ATD Press.
- Knowles, M. S., Holton, E. F., III, & Swanson, R. A. (2015). The adult learner (8th ed.). Routledge.
- Lave, J., & Wenger, E. (1991). Situated learning: Legitimate peripheral participation. Cambridge University Press.
- Mayer, R. E. (2021). Multimedia learning (3rd ed.). Cambridge University Press.
- Merrill, M. D. (1983). Component display theory. In C. M. Reigeluth (Ed.), Instructional-design theories and models (pp. 279–333). Erlbaum.
- Morrison, G. R., Ross, S. M., Morrison, J. R., & Kalman, H. K. (2019). Designing effective instruction (8th ed.). Wiley.
- Sweller, J. (1988). Cognitive load during problem solving: Effects on learning. Cognitive Science, 12(2), 257–285.