Didactics

Learning objective before medium

Technology follows the learning objective, not the other way round: first the objective, then the learning context, then the method (Goertz, Hagenhofer & Fehling, 2021). A meta-analysis of 184 studies shows that neither the headset nor the input device nor game elements explain learning success – what matters is how well task and medium fit (Howard, Gutworth & Jacobs, 2021). Experienced technicians did not learn more with VR than with conventional training (Gavish et al., 2015).

When XR pays off: D.I.C.E.S.S.

XR pays off where learning on the real system reaches its limits:

Criterion	In refrigeration and air-conditioning
Dangerous	Fault-finding on high-pressure and propane systems is dangerous at first contact.
Impossible	The thermodynamic inner workings of closed components remain invisible in the workshop.
Counter-productive	Long waiting times, for example during evacuation, eat into practice time.
Expensive	Industrial systems cannot be provided for every apprentice.
Social	Large groups lack the space for joint problem-solving at the real equipment.
Spatial	Invisible 3D processes and complex plant layouts are hard to grasp spatially on fixed equipment.

The D.I.C.E. scheme – dangerous, impossible, counter-productive, expensive – comes from Jeremy Bailenson, director of Stanford University’s Virtual Human Interaction Lab (Bailenson, 2018). He restated it in the OECD Digital Economy Outlook 2024 (Bailenson & Lesher, 2024). Others have added one S for sensory or social experiences (Virtofy, 2024; see also Reframe XR, 2026). XR2ACH distinguishes two S – Social and Spatial – because each plays its own role in refrigeration training.

Types of virtual learning environments

Virtual learning environments can be distinguished by their didactic approach (Buder & Schwan, 2002; Weise & Zender, 2017). Beckmann, Menke and Weber (2018) linked these types to the revised Bloom’s taxonomy (Anderson & Krathwohl, 2001) for training in the HVAC trades: which type of knowledge is taught, and which cognitive process is required?

Knowledge Dimension Cognitive Process	Remember	Understand	Apply	Analyse	Evaluate	Create
Factual Knowledge	Expository · Exploratory	Expository · Exploratory	Expository · Exploratory	Expository · Exploratory	Expository · Exploratory	Expository · Exploratory
Conceptual Knowledge	Expository · Exploratory	Expository · Exploratory	Expository · Exploratory	Experimental · Constructive	Experimental · Constructive	Experimental · Constructive
Procedural Knowledge	Training · Expository · Exploratory	Training · Expository · Exploratory	Training	Training	Training	Training
Metacognitive Knowledge	–	–	–	–	–	–

Which learning environment suits which learning objective. Our own representation of the mapping by Beckmann, Menke and Weber (2018), who combine the revised taxonomy (Anderson & Krathwohl, 2001) with the learning environments of Buder and Schwan (2002) and Weise and Zender (2017). None of the environments covers metacognitive knowledge well.

Learning
Environments

Expositional
Guided knowledge transfer in a fixed order.
Example in XR2ACH: A trainer explains a compressor model to the group in the shared virtual room.
Explorational
Self-directed exploration in one’s own order and pace; understanding, factual knowledge.
Example in XR2ACH: Learners explore a CAD model or a system freely.
Training
Practising procedures with guidance and immediate feedback; procedural knowledge.
Example in XR2ACH: Leak detection and commissioning, step by step.
Experimental
Changing parameters, exploring cause and effect.
Example in XR2ACH: Testing systems in different, including critical, operating states; seeing how the log p,h diagram responds to a misadjusted expansion valve.
Constructional
Learning by design; checking knowledge while creating.
Example in XR2ACH: Assembling a compressor from its parts; planning a refrigeration system from a component library.

From nugget to course

A course consists of modules, lessons and small learning units. A VR nugget fills one of these units: short, self-contained and built on a proven pattern – show, compare, practise a procedure, assemble (Horst & Dörner, 2019; Horst et al., 2022). The rest remains lecture, workshop and conversation.

A presentation with two embedded VR nuggets

Learning units at macro, meso and micro level: course, modules, lessons, nuggets

Top: a presentation with two embedded VR nuggets. Bottom: learning units at macro, meso and micro level. Figures from Horst et al. (2022), Figs. 1 and 2, CC BY 4.0.

This produces a learning journey: for each unit, the right format is chosen – together or alone, with VR, with AR or without XR. The InnoVET PLUS project XR-Upskill plans its courses this way in “swimlanes”, following an approach based on the ADDIE model (Sturman et al., 2026). At the train-the-trainer workshop in Maintal, trainers sketched such a learning journey for one of their own courses.

References