A Change Laboratory That Cannot Fail Is Not a Change Laboratory

I have sat in a lot of rooms that were called something other than what they were. The invitation says "collaborative design session." There is a facilitator, a wall of sticky notes, and real data about a real problem. Everyone participates. Something gets decided. It is, by any ordinary standard, a good workshop.

It is also almost always the case that the outcome was known before anyone walked in. Leadership had a direction. The session was where people came to arrive at it.

The Change Laboratory — a specific methodology from cultural-historical activity theory, developed at the University of Helsinki and first described by Engeström, Virkkunen, Helle, Pihlaja, and Poikela (1996) — looks superficially like that workshop. Same room, same wall, same data. It is a fundamentally different thing, and the difference sits in one commitment that institutions find very hard to make.

The mechanics, briefly

A Change Laboratory runs as a series of sessions with the practitioners of an activity — teachers, clinicians, advisors — rather than about them.

It begins with mirror data: video, documents, incident records, interview excerpts, aggregate figures. Raw material from the participants' own practice, including the disturbances and breakdowns, held up so the group can look at what it actually does rather than what it believes it does (Virkkunen & Newnham, 2013).

Then come conceptual tools — the activity-system triangle, models of contradiction, a germ-cell model — offered as instruments the group can pick up to analyze what the mirror is showing.

This two-part structure is not incidental. It instantiates Vygotsky's principle of double stimulation, in which a subject facing a genuinely difficult situation (the first stimulus) takes up an external artifact (the second stimulus) to gain control over their own action (Vygotsky, 1978). Worth a precise footnote here, because it is often garbled: the "waiting experiment" usually invoked in this connection was originally designed in Kurt Lewin's research group and discussed by Vygotsky as one instance among a range of double-stimulation experiments (Sannino, 2016). Vygotsky discussed it; he did not invent it.

Over the sessions, the group analyzes contradictions in its own activity, models an alternative, and — in the strong cases — reconceptualizes what the activity is for.

The commitment

Engeström (2011) drew the line that matters, contrasting formative interventions with design experiments. In a design experiment, the researcher specifies the intended outcome in advance and studies how well it is achieved. In a formative intervention, the participants take over the second stimulus and reformulate it — Engeström calls the result "thirdness" — and what emerges is genuinely not what the researcher designed.

Sannino, Engeström, and Lemos (2016) put it in terms of who holds the pen: in a formative intervention the learners drive the design, and the aim is a generative solution rather than a scalable one.

Read those two sentences again with an institutional ear and you can hear the problem.

"A Change Laboratory that cannot fail is not a Change Laboratory."

Unpredictability of outcome is not a tolerated side effect of the method. It is constitutive. If the object of the activity cannot change during the intervention — if the deliverable was pre-approved, if the strategic plan already names the destination, if there is a version of the finding that would get the sponsor in trouble — then whatever is happening in that room, it is not this.

This is the honest reason so few things called change labs are one. Not incompetence, and not bad faith. Institutions are built to produce predictable outputs from resourced activities, and a genuine Change Laboratory asks a sponsor to fund a process that might conclude that the sponsor's premise was wrong. That is a real thing to ask, and it should be asked plainly rather than smuggled in under the word "collaborative."

Four questions before you call it one

• Is there mirror data the group has not seen, including the embarrassing parts? If the data has been curated to protect anyone, the mirror is a poster.
• Can the object change? Not the method, not the workflow — the answer to "what is this activity actually for." If that is fixed, you are optimizing, which is fine, but name it.
• Who owns the outcome? If the facilitator can predict the conclusion, the participants are not driving.
• What would it look like for this to fail? If you cannot answer, there was no risk, and without risk there was no intervention.

It does work — and the evidence is in schools

None of this is a counsel of despair. When the commitment is real, the method produces documented change, and the recent education literature is worth knowing.

Hopwood, Palmer, and Castelli (2025) ran a Change Laboratory in an Australian secondary school around the contradiction between the exchange value and use value of senior certification — the collision between credential-as-currency and learning-as-purpose that anyone who has worked near a high-stakes exam will recognize. Aggregate and individualized student data served as the first stimulus; diagrammatic models as the second; concrete curricular and assessment changes followed.

Kaup and Dau (2024) worked with six Danish educators integrating computational thinking into mathematics teaching, and found the collaboration between sessions — not only within them — did substantial developmental work. Postholm (2025) tracked eight sessions across two years linking a teacher-education program with a school, and found the interventionist researcher's own role expanded, oscillating between process leader and collaboration partner. Hyrkkö and Kajamaa examined leadership-related expansive learning actions in a Finnish teacher-training school's Change Laboratory. [Hyrkkö & Kajamaa, Mind, Culture, and Activity, 31(3–4), 238–255, https://doi.org/10.1080/10749039.2025.2491543 — online-first 2025 vs. print-issue 2024 discrepancy; confirm the year before citing.]

Hopwood (2025) has published a critical and formative review of twenty-five years of school-based Change Laboratories, which is the obvious place to start a serious literature review. [I have not read the full text — paywalled at the time of writing — so I am not characterizing its conclusions. Read before citing its findings.]

One caution I want to record accurately. I looked specifically for a source in which Engeström himself warns that Change Laboratories get diluted or misapplied in institutional practice, because that would be the natural citation for the argument I am making here. I could not find one. Engeström and Sannino (2010) include self-reflective caveats — that expansive learning necessarily carries the possibility of disintegration and regression, that outcomes are not linear — but that is not the same claim, and third-party critiques of Engeström should not be attributed to him. The argument in this post is mine, and it is not yet, as far as I can establish, his.

The claim

The Change Laboratory is not a facilitation technique with a theory attached. It is a wager that the people inside an activity can grasp the contradictions in it and redesign it themselves, and that the redesign will be something nobody in the room could have specified in advance.

Strip out the unpredictability and you keep the furniture and lose the method. So before booking the room, ask the sponsor one question: are we willing to find out that the thing we are trying to improve is the wrong thing?

A yes is rarer than the literature's enthusiasm suggests. But a no, said honestly, is worth more than a change laboratory that was never going to change anything.

The opening scene is a composite; no individual or institution is described.

References

Engeström, Y. (2011). From design experiments to formative interventions. Theory & Psychology, 21(5), 598–628. https://doi.org/10.1177/0959354311419252

Engeström, Y., & Sannino, A. (2010). Studies of expansive learning: Foundations, findings and future challenges. Educational Research Review, 5(1), 1–24. https://doi.org/10.1016/j.edurev.2009.12.002

Engeström, Y., Virkkunen, J., Helle, M., Pihlaja, J., & Poikela, R. (1996). The change laboratory as a tool for transforming work. Lifelong Learning in Europe, 1(2), 10–17.

Hopwood, N. (2025). Twenty-five years of change laboratories in schools: A critical and formative review. Educational Action Research, 33(4), 571–598. https://doi.org/10.1080/09650792.2024.2379343

Hopwood, N., Palmer, T.-A., & Castelli, B. (2025). Driving educational improvement through transformative agency by double stimulation in a high school change laboratory. Journal of Educational Change, 26(3), 475–497. https://doi.org/10.1007/s10833-025-09531-6

Kaup, C. F., & Dau, S. (2024). New perspectives on collective collaboration within and between Change Laboratory sessions. Education Sciences, 14(11), Article 1159. https://doi.org/10.3390/educsci14111159

Postholm, M. B. (2025). Expanding the change laboratory and the interventionist researcher's role in teacher education–school collaboration. Frontiers in Education, 10, Article 1588752. https://doi.org/10.3389/feduc.2025.1588752

Sannino, A. (2016). Double stimulation in the waiting experiment with collectives: Testing a Vygotskian model of the emergence of volitional action. Integrative Psychological and Behavioral Science, 50(1), 142–173. https://doi.org/10.1007/s12124-015-9324-4

Sannino, A., Engeström, Y., & Lemos, M. (2016). Formative interventions for expansive learning and transformative agency. Journal of the Learning Sciences, 25(4), 599–633. https://doi.org/10.1080/10508406.2016.1204547

Virkkunen, J., & Newnham, D. S. (2013). The change laboratory: A tool for collaborative development of work and education. SensePublishers.

Vygotsky, L. S. (1978). Mind in society: The development of higher psychological processes (M. Cole, V. John-Steiner, S. Scribner, & E. Souberman, Eds.). Harvard University Press.