The DRIVE Framework, A Practical Pedagogy for Live Decision Simulations

Higher education has no shortage of pedagogy frameworks. What it has historically lacked is a framework that translates cleanly into a runnable simulation. DRIVE is our attempt to close that gap. The five pillars, Decision-led, Reflective, Iterative, Verified, and Experiential, are not rhetorical pillars. Each one drives a specific decision in how we design scenarios, set scoring rubrics, generate AI advisor responses, and gate AI use in student rationales. This post walks through each pillar, the literature it draws on, and what it changes in the runtime.

Why Yet Another Framework

The honest answer is that we did not set out to publish one. DRIVE began as an internal scoring spec; a way to keep the team aligned on what makes a sim 'good' when we have five engineers, three subject matter contributors, and an AI generator producing variants at scale. After a year of using it as a build standard, the same framing kept being useful in conversations with faculty, with procurement teams, and with students explaining what they had just been graded on. Naming it made it shareable; publishing it makes it auditable.

DRIVE is therefore a working framework rather than a research contribution. It synthesises canonical pedagogy (Dewey, Schön, Kolb, Mezirow, Brookfield, Vygotsky, Hattie, Dweck) with current international guidance (UNESCO Ethics of AI, OECD AI Principles, AACSB AI Employability Framework, the UK Professional Standards Framework, PRME). Every pillar links back to literature that is older than us and to guidance that is more authoritative than us.

D, Decision-led

A DRIVE sim earns its name on the first turn. The student is not reading a case study, or answering recall questions, or watching a video; they are making a decision under incomplete information with downstream consequences. The pedagogical foundation here is Dewey's argument that learning is rooted in the structured experience of choice and consequence, and Schön's analysis of how professionals build judgement through reflection on action.

In practice this means three runtime rules. First, every turn must require a decision; no purely informational turns survive editorial review. Second, the decision set must include at least one plausible but suboptimal option, so success requires discrimination rather than recognition. Third, downstream state must visibly update; if the decision does not change anything the student can see, the loop breaks. These three rules are what separates a DRIVE sim from a quiz embedded in a colourful interface.

“We do not learn from experience, we learn from reflecting on experience.”

— John Dewey, Experience and Education, 1938

R, Reflective

Decisions without reflection produce intuition, not judgement. The Reflective pillar draws on Schön's distinction between reflection-in-action (the rapid sense-making that experts do mid-task) and reflection-on-action (the slower analysis that happens afterwards). DRIVE sims build both. Each decision turn includes a rationale field; students explain why they chose what they chose, not as a formality, but because the rationale is graded. After the run, a debrief surfaces the team's decision pattern alongside the canonical reasoning and invites comparison.

The reflective pillar also tracks Mezirow on transformative learning. Many simulation sessions surface a moment where a student realises their default heuristic, the one that has been getting them through coursework, is exactly what the simulation is punishing. That moment is the unit of value; DRIVE sims are designed to produce it on purpose, not by accident.

I, Iterative

Kolb's experiential learning cycle frames learning as a loop: concrete experience, reflective observation, abstract conceptualisation, active experimentation. DRIVE sims are explicitly structured around that loop. A single run is one full traversal; a multi-week course gives students three to six traversals on increasingly difficult variants of the same scenario.

Iteration also pulls in Dweck on mindset and Ericsson on deliberate practice. The platform deliberately makes early attempts uncomfortable; students who replay a sim see their first run's decision log alongside their second, and the gap is the lesson. Hattie's synthesis of educational research places feedback as the single highest-leverage factor in learning gain; iterative sims are essentially a feedback delivery mechanism wearing a business problem as a costume.

V, Verified

Verified is the AI-era pillar. Students enter a workforce where generative AI is ambient; pretending otherwise produces graduates who are either AI-illiterate or AI-credulous, both of which are liabilities. The Verified pillar requires every AI-assisted decision to include a statement of how the AI was used and why the student decided to trust or override it. This is graded separately from the decision itself.

The guidance grounding here is the UNESCO Recommendation on the Ethics of AI, the OECD AI Principles, and the AACSB AI Employability Framework, which all treat verification as a literacy rather than a tool. Halpern on critical thinking transfer and Russell on AI alignment provide the cognitive science layer; the empirical post-2023 research on AI trust calibration in students provides the contemporary evidence base. In runtime terms, the Verified pillar is what produces the 'wrong advisor' mechanic in Boardroom; one of the AI advisors is deliberately wrong on every turn, and a student who fails to catch them loses credibility independent of whether their final decision was correct.

E, Experiential

Experiential is the connective tissue. Dewey, Kolb, Lave and Wenger on situated learning, Vygotsky on the zone of proximal development; the pillar collects the literature that says learning happens in context, with peers, with stakes, and with scaffolding calibrated to where the learner currently is. A DRIVE sim is run live, with a tutor present, with team members holding different roles, with KPIs updating in front of the room. The experience is the medium.

This pillar is why DRIVE sims are not asynchronous LMS modules. We do offer self-paced practice runs, but the graded session is timetabled and facilitated. The empirical research is unambiguous on this point; collaborative performance under pressure cannot be replicated in a self-paced format, and it is one of the top deficits employers cite in new graduates.

What This Means in Practice

Every simulation produced on the MyEdMentor platform maps to all five pillars. The mapping is published on each sim profile page; tutors can see exactly which Decision-led mechanic, which Reflective prompt, which Iterative loop, which Verified gate, and which Experiential affordance the sim instantiates. When an external examiner asks how a simulation session produces assessable evidence comparable to a traditional examination, DRIVE is the answer. The decision log is the decision-led evidence; the rationale field is the reflective evidence; the replay deltas are the iterative evidence; the AI-use statement is the verified evidence; the session-level facilitation record is the experiential evidence.

DRIVE is not a sales pitch, it is a build standard. We use it internally to reject simulation concepts that do not satisfy all five pillars before they reach the editorial pipeline. We publish it because procurement teams, faculty development directors, and accreditation officers reasonably want to know what is underneath the simulation layer; and because students reasonably want to know what they are being graded on. The framework has its own page at myedmentor.com/pedagogy with the full literature map and the scoring rubrics for each pillar.