THE UNIVERSITY OF ALABAMA®
myBama
AI-enhanced Educational Game Design
Facilitator Guide · 12 Sessions · Instructor-Facing

Teaching this is mostly
about not rescuing.

A per-session guide for instructors: how the three hours should shape up, what Proficient work looks like in the room (not the rubric), the failure modes that keep repeating, and the moments where you have to choose between intervening and letting the discomfort do the teaching.

THE FACILITATION STANCE

You are running a studio, not delivering a lecture.

The course is built on the assumption that design capability transfers through revision, not reception. Your job is to hold the frame — a clear objective for the session, a visible artifact, a deadline — and to let the learner's first attempt be wrong in public. Most growth happens when a learner ships something they're not proud of and sees, immediately, what a better version would look like. Rescuing short-circuits that loop.

Session 01 · Week 1

Framing

Shift: learners stop saying "make learning fun" and start naming a specific transfer problem.

Open session →
Three-hour shape
0:0015 minStance question discussion: "what is a learning game not?"
0:1540 minReading-led discussion: three failure modes of gamification (points, theme, drill).
0:5515 minBreak.
1:1060 minCasebook walkthrough: three games, name what each actually teaches.
2:1030 minStudio: each learner writes the engagement–transfer problem they intend to solve.
2:4020 minExit ticket: stance commitment + capstone framing draft.
What "in-session Proficient" looks like
  • Stance commitment names a specific objective type, not a domain ("discrimination in clinical reasoning," not "medical education").
  • Learner can state what the game is not trying to do — a visible scope edge.
  • At least one casebook game is named with a concrete mechanic-to-objective link.
Common pitfalls
  • "Make X more engaging." Push for the specific cognition that has to change.
  • Choosing a domain before an objective type. Objective type is the design-decisive lens; domain is decoration.
  • Defaulting to quiz games because they feel safe. Note it, do not forbid — the mechanics weeks will do the work.
Decision points
  • When a learner's framing is too broad: ask "which of these three cognitions?" and accept whichever they pick. Narrowing matters more than correctness this week.
  • When a learner wants to change topic mid-week: fine before S03. Past S03, they owe the class a revised crosswalk before the switch.
Session 02 · Week 2 · D1 due

Learner & context

Shift: learners write about a real person with real constraints, not "the student."

Open session →
Three-hour shape
0:0010 minWarm-up: name three things about your learner you cannot Google.
0:1030 minMini-lecture: learner-context-constraint-measurable-shift structure.
0:4060 minStudio: write D1 v0.
1:4015 minBreak.
1:5540 minPeer critique in triads — structured protocol, one voice per round.
2:3525 minRevision pass + log what changed and why.
What "in-session Proficient" looks like
  • Learner description includes prior knowledge, motivation, and binding constraint — all three.
  • The measurable shift is stated as a behavior a stranger could verify, not an internal state.
  • Revision log names which specific feedback drove which change.
Common pitfalls
  • Composite learners. "Students who are either X or Y" — kill it. One population, one context.
  • "Increase engagement" as the measurable shift. Ask what they would do differently.
  • Vague constraints. "Limited time" is not a constraint; "30 minutes, once, on a phone, in waiting rooms" is.
Decision points
  • If peer critique devolves into "I like it": stop the room, re-read the protocol together. Do not let it drift.
  • If a D1 passes four criteria and blocks on one: Proficient floor is non-negotiable. Return it; do not average.
Session 03 · Week 3 · D2 due

Objectives & crosswalk

Shift: learners stop picking a genre and start picking a mechanic family that fits an objective type.

Open session →
Three-hour shape
0:0010 minWarm-up: classify five objectives from the casebook against the five-type taxonomy.
0:1040 minMini-lecture + guided practice: objective type → signature mechanic family.
0:5060 minStudio: write D2 crosswalk, 2–4 objectives × mechanic × rationale × risk.
1:5015 minBreak.
2:0535 minStructured peer review: defend one mechanic choice against a plausible alternative.
2:4020 minMark at least one objective out of scope and say why. Log in D2.
What "in-session Proficient" looks like
  • Every objective is classified by type before a mechanic is chosen.
  • Rationale is a two-sentence argument, not a label.
  • Declined alternative is named and its reason for decline is specific to this learner and context.
Common pitfalls
  • Mechanic before objective. "I'll use a card game" before deciding what cognition it should exercise. Return it.
  • Five objectives. Scope creep. Keep to 2–4; the rubric explicitly caps this.
  • Rationale that's really a description. "Cards are fun and engaging" is not rationale.
Decision points
  • If an objective resists classification: it's almost certainly two objectives fused. Split it.
  • If a learner wants to use the Orbit Sum Lab pattern: good — add it as a seventh casebook case with their own mechanic/risk/rationale.
Session 04 · Week 4

Mechanics I — challenge, feedback, failure

Shift: learners see that difficulty is a zone you tune, not a number you pick.

Open session →
Three-hour shape
0:0015 minWarm-up: name the four kinds of feedback by timing and grain.
0:1545 minMini-lecture: target difficulty zone, adjustment knobs, feedback taxonomy.
1:0015 minBreak.
1:1555 minStudio: design a failure loop that teaches or argue why this objective should not have one.
2:1050 minRevision of D2: annotate every row with challenge / feedback / failure risks.
What "in-session Proficient" looks like
  • Difficulty is named as a zone between two specific boundaries, with the knob to move within it.
  • Feedback is differentiated by timing and grain, not just "immediate."
  • D2 annotations identify at least one plausible failure mode per row.
Common pitfalls
  • Difficulty as single number. "Medium difficulty" — not useful. What's the zone, what moves it?
  • Failure loops everywhere. Some objectives (safety-critical procedures) should not have failure baked in. Name it.
  • Feedback = score. Score is summative. Most learning feedback is process-grain, not outcome-grain.
Decision points
  • If a learner's mechanic has no failure state: decide together whether that's principled or avoidance. Either answer is fine — but it must be named.
Session 05 · Week 5

Mechanics II — narrative, role, social

Shift: narrative stops being wrapper and starts being a mechanic that lowers threshold-to-try.

Open session →
Three-hour shape
0:0015 minWarm-up: for a casebook game, is the narrative wrapper or mechanic? Justify.
0:1540 minMini-lecture: narrative-as-mechanic, roles, collaboration structures.
0:5560 minStudio: role brief — a teammate could implement it.
1:5515 minBreak.
2:1050 minIntegrate narrative/role/social choices into D2 — revise rationale column.
What "in-session Proficient" looks like
  • Role brief specifies what the role does, decides, and cannot do.
  • Collaboration structure fits both objective type and D1 constraints (not just "looks fun").
  • Learner can state one failure mode of this role and how they'd scaffold around it.
Common pitfalls
  • Narrative-as-theme. Space pirates doing arithmetic. Nice, but not a mechanic.
  • Four-person cooperative for a solo objective. Social scale mismatched to learning target.
  • Roles with no teeth. If every role can do everything, there is no role.
Decision points
  • If narrative and mechanic conflict: mechanic wins. Theme can flex; cognition cannot.
Session 06 · Week 6

Facilitator design

Shift: the game includes the person running it, and that person is not always you.

Open session →
Three-hour shape
0:0015 minWarm-up: three non-negotiable sections of a facilitator guide.
0:1545 minMini-lecture: opening, mid-play, debrief — word-level scripts.
1:0015 minBreak.
1:1565 minStudio: write v0.1 of facilitator guide — specific words, not summaries.
2:2040 minTable-read: peer runs your guide without asking questions. Revise where they stumbled.
What "in-session Proficient" looks like
  • Guide includes opening script, mid-play decision tree, debrief questions — all three present.
  • A peer runs it on the first try without asking clarifying questions.
  • At least three failure modes have recovery lines.
Common pitfalls
  • "Explain the rules." Not a guide. What words exactly?
  • Debrief = "what did you think?" The debrief is where transfer happens. Write three specific questions that point at the objective.
  • No mid-play decisions. What does the facilitator do when it's boring? When it's too hard? Name two intervention points.
Decision points
  • If the table-read fails silently (peer shrugs through it): that's a worse signal than explicit confusion. Return for revision even if nothing was "wrong."
Session 07 · Week 7 · D3 due

Low-fi prototyping

Shift: learners stop designing on paper and start playing on paper.

Open session →
Three-hour shape (studio-heavy)
0:0010 minDemo: instructor builds a paper loop in 5 minutes, live.
0:1025 minCycle 1: build paper prototype. Timer hard-stop.
0:3515 minCycle 1 playtest: peer-as-proxy, instructor observes silently.
0:5010 minCycle 1 revision based on one specific observation.
1:0015 minBreak.
1:1575 minCycles 2 + 3: same loop, faster. Iteration log required.
2:3030 minSubmit D3: surviving version + facilitator guide v0.3 + what-I-cut note.
What "in-session Proficient" looks like
  • By end of hour 1, something is being played, not being planned.
  • Iteration log names, for each cycle: what changed, why, and the observation that drove it.
  • What-I-cut note identifies at least one idea that didn't survive contact with play.
Common pitfalls
  • "I'll build it nicely first." No. Ugly and playable beats beautiful and static.
  • Instructor-as-observer narrates during playtest. Silence. The designer must see the silence too.
  • Cycles 2 and 3 re-make the whole thing. They should tune, not rebuild.
Decision points
  • If a learner insists "it's not ready to playtest": enforce the 25-min cycle. Ready-feeling is the problem this session solves.
  • If the loop genuinely breaks: fine — document it and pick a simpler loop for Cycle 2. Failing the cycle is not failing the session.
Session 08 · Week 8

Interaction spec

Shift: the paper loop becomes something a developer could build without asking.

Open session →
Three-hour shape
0:0015 minWarm-up: draw your paper loop as three states.
0:1540 minMini-lecture: state machines, event→feedback maps, Three.js bridge.
0:5560 minStudio part 1: formal state machine (entry, exit, transitions).
1:5515 minBreak.
2:1040 minStudio part 2: event→feedback map. Launch Electric Circuit Lab as worked example.
2:5010 minExit ticket: say no to 3D in one scene. Explicitly.
What "in-session Proficient" looks like
  • State machine is readable without the author present. Every arrow has a label.
  • Event→feedback map is a table, not prose. Event column, feedback column, timing column.
  • Learner can identify one scene where 3D adds nothing and cuts it in the spec.
Common pitfalls
  • State machines that are flowcharts. States vs. steps. Push.
  • "3D everywhere." 3D is a bridge tool, not a theme. Force the no-3D scene.
  • Events without grain. "Player wrong" is not an event. "Player submits answer X when correct was Y, within 3s" is.
Decision points
  • If a learner has no dev background: the spec is still the goal. Pair them in-session with a peer who does — handoff is the artifact.
Session 09 · Week 9 · D4 due

Playtest design

Shift: learners stop asking "did they like it?" and start asking "what did they do?"

Open session →
Three-hour shape
0:0015 minWarm-up: separate observation from interpretation in three transcripts.
0:1540 minMini-lecture: protocol design, recruitment (target-learner, not peers), finding taxonomy.
0:5560 minStudio: write protocol a stranger could run.
1:5515 minBreak.
2:1035 minPeer simulation: run each other's protocol. Observe silently.
2:4515 minCommit to three target-learner sessions before next week.
What "in-session Proficient" looks like
  • Protocol is step-by-step executable by someone with no context.
  • Findings vocabulary separates observation, interpretation, recommendation — three columns, not one.
  • Recruitment plan names actual people or a realistic channel, not "I'll find three students."
Common pitfalls
  • Peer playtests as "data." Peers are rehearsal, not evidence.
  • Likert surveys. Feeling is not behavior. Can they do the thing?
  • Leading questions. "Was it fun?" → "What did you try first?"
Decision points
  • If recruitment of three target learners is infeasible: acceptable substitute is two target learners + one practitioner review — but it must be named as a limit in D4.
Session 10 · Week 10

Audit — accessibility, ethics, UDL

Shift: the design gets stress-tested against learners the designer didn't imagine.

Open session →
Three-hour shape
0:0015 minWarm-up: name a UDL principle that your design currently violates.
0:1545 minMini-lecture: UDL 3.0 crosswalk, accessibility audit checklist, ethics prompts.
1:0015 minBreak.
1:1555 minStudio: self-audit against UDL + accessibility. Write three concrete changes.
2:1050 minPeer audit: score each other's design against the checklist. Revise D3/D4 entries accordingly.
What "in-session Proficient" looks like
  • Self-audit names at least one violation honestly, not just "we meet all of them."
  • Three concrete changes are implementable this week, not aspirational.
  • Peer audit produces at least one blind spot the designer missed.
Common pitfalls
  • "It's accessible because it's a card game." Not a finding. Which specific affordance for which specific learner?
  • Treating UDL as a checklist. It's a design lens; most violations are in the debrief, not the mechanic.
  • Peer audit as rubber-stamp. Require at least one concrete critique per pair.
Decision points
  • If a change is required but infeasible this semester: acceptable to log as "known limit" in D5, but only if a mitigation is proposed.
Session 11 · Week 11

Revision studio

Shift: learners prioritize revisions by evidence strength, not by what was easiest to fix.

Open session →
Three-hour shape (studio-only)
0:0010 minCheck-in: what's the one thing you're afraid your final is going to be judged on?
0:1080 minStudio block 1: revise based on evidence ranking from D4 + audit findings.
1:3015 minBreak.
1:4560 minStudio block 2: D5 spec completion. Instructor floats, 5-min slots per learner.
2:4515 minDry-run outline for final presentation — 6 slides, no more.
What "in-session Proficient" looks like
  • Every revision is tied to a specific piece of evidence from D4 or audit — not "we thought it'd be better."
  • At least one revision is deferred and justified (not enough evidence to act).
  • D5 is 80%+ complete by end of session.
Common pitfalls
  • Revising everything. Weak evidence + big change = wasted week. Push for triage.
  • "We'll revise based on final demo feedback." No — the final is a presentation, not a playtest.
  • Instructor-as-consultant. 5-min slots per learner, not 30. Studio, not office hours.
Decision points
  • If a learner's D5 is blocking: ask whether the block is conceptual (help) or mechanical (they can finish this week). Do not do the mechanical work for them.
Session 12 · Week 12 · D5 due

Final presentations

Shift: learner defends the handoff as something a team could pick up on Monday.

Open session →
Three-hour shape
0:0010 minGround rules: critique the spec, not the presenter.
0:10150 min12 presentations × 10 min each (6 talk, 4 Q&A). Rotating role: dev / learner / skeptic.
2:4015 minBreak between blocks as needed — keep presentations on timer.
2:5515 minCredential sign-off: instructor confirms rubric pass on all D1–D5 or names the open item.
What "in-session Proficient" looks like
  • Presenter can walk a dev through D5 without reading slides.
  • Known limits section is present and honest.
  • Q&A reveals the presenter can update the spec live when a real gap is surfaced.
Common pitfalls
  • Demo-as-finale. This is not a demo. Playtests happened in S09.
  • "We're still iterating." Fine, but credential asks what you are handing over.
  • Audience in default-positive mode. Assign the skeptic role explicitly, rotate it.
Decision points
  • If a learner misses Proficient on a single criterion: one-revision-per-D appeal is available. Sign off conditionally; close the open item within two weeks.
  • If multiple learners cluster on the same miss: the curriculum has a gap, not the cohort. Log it for program review.
Program-level note. If you see more than 30% of a cohort miss the same criterion, do not tighten the rubric — tighten the session that owns that criterion. The rubric is calibrated; the session is the lever. Log the signal and route it to the evaluation plan at program review.