How DreamSpace Teaches

The method, the standard, and what real classes changed

One of three essays on how our education work and the rest of DreamSpace Academy fit together.

In short. At DreamSpace Academy, our written method for the STEAM programmes fixes a few things and leaves the rest flexible. It sets when a piece of work counts as finished and how a learner moves forward, and it has changed in response to what happened in real classes.

Scope. The STEAM programmes are one part of our education. We also run business and entrepreneurship training, digital security training, and workplace and teacher training. The method below is for the STEAM programmes. We have not yet written up the rest of our teaching in the same detail.

Evidence marks. The method comes from our written plans. The evidence about teaching comes from our own reports on what was actually taught. Marks: 🟢 backed by original documents or by more than one source; 🟡 supported only in part, or by one source; 🔴 stated, and still being tested.

---

1. How the work is organised: pillars and verbs

Our work rests on three pillars, Education, Innovation and Entrepreneurship, on a foundation of research. We aim for everything we run to rest on evidence, though that evidence is uneven. Peacebuilding runs through all three pillars rather than standing as a fourth.

Wherever they run, the STEAM programmes offer four kinds of activity, called verbs because each is something the learner does:

Verb	What it means	How well supported
Learn	Digital education, AI literacy, electronics and fabrication, taught through challenge	🟢 well evidenced
Make	Prototype, 3D printing, CNC and laser, robotics, electronics workbench	🟢 well evidenced
Solve	Turn real local challenges into projects	🟡 supported in outline
Earn	Income and work, in three tiers, see below	🟡 / 🔴

Earn is the weakest of the four. It has three tiers:

• Tier 1: a skilled maker sells things they have made for a small income. This is near-term and low-risk. Making produces things people can buy, so nothing needs to be proved about cause and effect 🟡.
• Tier 2: a lasting livelihood, either by starting a business or through a job or freelance work. This is a hypothesis to measure, not a result to claim 🔴.
• Tier 3: the venture employs, contracts or buys locally. This tier is the least proven, not the strongest: we have never counted jobs created, so we publish no figure for them.

The four verbs run through every stage. They are not steps in order, and nobody graduates from Learn into Make.

---

2. Stage and Level: two separate measures of progress

The Changemaker Journey is the path learners follow in the STEAM programmes. It keeps apart two things that almost every other framework joins into one.

• Stage is purpose: what the making is for. Stages are named, never numbered and never graded. We have not decided whether our other teaching will use the four stages.
• Level is how deep a skill goes in one track (one subject, such as robotics or music). Levels are numbered, and they belong to a course, never to a person.

The four stages of the Changemaker Journey:

Stage	Identity	Core question	Enter when
Explorer	I find out what’s possible	“What can I do?”	Walk in, always open
Maker	I turn ideas into things	“Can I make something?”	Chooses a track
Innovator	I take on real problems	“Can I make it better?”	Skilled in at least one track
Changemaker	I change my community	“Can I make a difference?”	Has produced a grassroots innovation

A Changemaker’s record of their work is not better than a Maker’s. It is a record made for a different purpose. So stages are never compared against a skills grid. Levels within a track are.

Levels belong to courses. Robotics & Intelligent Systems Level 3 is a course a learner enrols in. How far a person has gone in each track is shown in their Maker Profile, for example Robotics 4, AI 2, Music 1. The Maker Profile is the record that follows a learner over time.

The number of levels in a track is worked out, not chosen. Start from the entry point (no making experience) and the exit point (a Skilled Maker, who completes a full piece of assessed work on a project they chose, without help). Then divide the distance by the size of one block of teaching. For robotics this comes out at 4 units of 48 hours, about 192 hours of study, or about 4 credits on the Sri Lanka Qualifications Framework (SLQF). That gives four levels, because robotics has four real steps up in skill.

Learners follow a default order, and can skip ahead. Every track publishes its usual order, so a beginner always knows where they are. Any learner can skip ahead by passing the placement check for the course they want to join. Showing the skill counts for more than waiting, and earlier learning counts however it was gained.

Age does not decide placement. It is not an entry rule or an age band, and it is not printed in publicity either: a typical age printed on a brochure is read as a rule by exactly the people the programme is trying not to turn away. Age is used only where safeguarding and duty of care require it. The STEAM programmes are for everyone, and placement is by evidence.

The earlier scale. Until July 2026 our method used a five-level scale tied to age. It packed five separate things into one word: age, depth of thinking, depth of the course, the certificate and the group a learner was in. That led to levels whose difficulty ran in the wrong order, three middle levels that could not be sorted out, and constant argument about the age bands. Keeping Stage and Level apart removed all three problems.

---

3. Definition of Done: the one fixed standard

There is no fixed set of steps for making. Some learners decide on an aim first. Some play with materials until an idea appears. Trainers order their sessions in whatever way suits the group. Other maker organisations work the same way. None of the established ones we looked at publishes a fixed set of steps: Maker Ed’s Approaches, Fab Academy, NuVu and the Dutch Platform Maker Education.

What is fixed is the standard for when work is finished. An assessed piece of work is done when six kinds of evidence exist. They can be produced in any order:

Evidence	Means
Aim	The intent is stated (a problem, a brief or an idea) with criteria to judge by
Options	More than one route was considered before committing
Made	The thing exists
Evaluated	It was judged against the aim’s own criteria
Shared	There is a record and an audience saw it
Refined	At least one improvement was made from what evaluation or sharing surfaced

These are items on a checklist, not steps. No order is set or marked. Explorer work is exempt: it is never assessed, and runs as hands-on exploration with goals the learners set themselves. The standard applies wherever certificates are given: the Maker stage and above.

The six items are what three traditions have in common: engineering (the US Next Generation Science Standards, NGSS), the arts (the US National Core Arts Standards, NCAS) and making (Papert and the Tinkering Studio). Each item has a one-line meaning for each track, so the same standard fits every track:

	Robotics & Intelligent Systems	Music	Visual Arts
Aim	What problem, for whom	What piece, what feeling, what occasion	What idea, what response wanted
Options	Sketch 2 or 3 mechanisms	Try several melodic / rhythmic ideas	Thumbnails, references, compositions
Made	Built and programmed	Composed, arranged, recorded	The work produced
Evaluated	Tested against spec, does it work?	Rehearsed and listened back	The crit, does it read as intended?
Shared	Demo + build log	Performance + session notes	Exhibit + process journal
Refined	Redesigned from test data	Arrangement revised, re-recorded	Reworked from the crit

What the standard replaced. Until July 2026 our method described a fixed six-step loop and called it the best-evidenced part of the method. That was wrong. The loop began with “Identify a problem”, which suits engineering but not music or art. It was also meant to run at every stage, and what happened in Explorer sessions showed that it did not.

---

4. Certificates: three pieces of evidence

A certificate at any level needs all three:

1. The artifact: the finished work, meeting the Definition of Done.
2. The talk-through: a short, structured conversation in which the learner explains how and why the work works. It is based on Piaget’s method of questioning a child about their reasoning, as adapted for judging learning by making 🟢. It costs only the trainer’s time and works the same for a robot, a song or a painting. It also evens out the difference in how easily work can be judged in digital tracks and in craft tracks.
3. The reflection: what did I learn, and what would I do differently? This is the direct evidence for the soft results (personal qualities rather than technical skills) that the skills grid certifies.

We have not proven this yet. Soft results are hard to check, and the Dutch platform’s own teachers say so 🟡. The talk-through and reflection together are how we try to deal with this, not a proof. Certificates for soft results stay marked as unverified until our own teaching shows otherwise.

---

5. Teaching: taught classes, open making and learning by inquiry

Two ways of running sessions, and both are protected:

• Taught Instruction: structured, following a syllabus, and fitted to learning by making. This is what a university or company partner pays for.
• Open Making: unstructured time where learners choose what to work on and learn from each other. We treat it as what keeps learners coming back 🟢, though we have not yet measured how many stay. Open making time is protected on the timetable.

Learning by inquiry, matched to each stage. In inquiry-based teaching, learners work things out through questions and projects rather than being told the answers. The approach changes as they move through the stages:

Explorer     Hands-on exploration                       no problem given yet; the materials invite it
Maker        Guided Project Work                        the facilitator sets the brief
Innovator    PBL (Project-Based), inside a CBL frame     the learner chooses a problem within a domain
Changemaker  CBL + action                               the community supplies the problem; Engage, Investigate, Act

Our method does not use Problem-Based Learning at all. For us, PBL always means Project-Based Learning.

Challenge-Based Learning (CBL) runs through every stage rather than being one of them. It is how we teach at every point on the pathway. Learners rotate through labs, join workshops that mix subjects and take on challenge projects before discovery, during innovation and inside venture building.

---

6. Innovation & Enterprise: the strand that runs through everything

This strand runs through the curriculum and the venture pathway in three areas, each building on the one before:

1. Research & Development, make it work. Research → hypothesis → prototype → test → document → iterate. (The Solve verb.)
2. Entrepreneurship, make it matter. Opportunity → customer discovery → value proposition → MVP → pitch → network. (The Earn verb.)
3. Business & Financial Literacy, make it last. Costing, pricing, bookkeeping, marketing, operations, quality.

In one line: R&D → Entrepreneurship → Business, make it work, make it matter, make it last.

Business topics open up when a learner has the knowledge they build on, not at a set age. The research on child development shows why:

• abstract business ideas (ownership, profit versus revenue, how a business is structured) do not reliably make sense to children before about age 10 to 13 🟢;
• saving and selling can be learned from about age six, in a hands-on, playful setting 🟢;
• children understand profit only once they have been taught to compare prices 🟢.

The best randomised trial in this area (BizWorld, about 2,751 children aged 11 to 12) shows that learning entrepreneurship by doing it changes mindset (persistence, creativity, planning ahead) but not business knowledge 🟢.

What those studies point to is knowledge a learner needs first. Cognitive Load Theory explains why: material pitched above what a learner already knows does not stick. The studies used age as a stand-in for that knowledge, because they studied school classes. We measure the knowledge itself, through the placement check.

So early stages teach the entrepreneurial mindset through making, plus simple activities in selling and saving. Profit versus revenue, pricing and building a venture come in later stages, never at a set birthday.

---

7. Tracks taught within the STEAM programmes

A track is one subject area, taught at increasing levels. The tracks currently written up for the STEAM programmes are below. Some of them, such as software and AI, overlap with our digital skills training, and we have not decided where they belong.

#	Track	Footing
1	Maker & Fabrication, electronics and soldering, 3D printing, CNC/laser basics, robotics foundations	🟢
2	Digital Education, digital literacy, hands-on skills, work readiness, responsible digital use	🟢
3	AI Programmes, intro to AI and generative AI, responsible/ethical use, hands-on AI, AI-assisted creativity	🟢
4	AI + Electronics, sensors and automation, hardware plus smart-system projects	🟢
5	Software & Tech-Entrepreneurship, coding → web/app → shipping a product, freelancing, tech venture	🟢 curriculum; taught unevenly
6	Creative Arts, visual art and craft, music, media	🟡 really taught, not yet a structured multi-level curriculum
7	Fitness & Sports, a fitness studio, sport and athletics including chess	🔴 a design direction, not yet taught regularly

The AI track has grown over five years: No-Code AI (2021), then AI Camp (2023), then AI & Data Science Fundamentals (2024), then GenAI/LLM JumpStart (2025) 🟢.

Our written curriculum covers less than we actually teach. The teaching has usually come first and the handbooks after:

• the core electromechanics course was already being taught in May 2019, more than three years before it was first written down;
• cardboard prototyping is still used in workshops, although it is not on the written list of lab courses.

---

8. Sessions in practice

This is what our method looks like in real sessions:

• A standard session in the robotics batches lasts 180 minutes. Robotics runs over four levels, each ending in a themed build such as an Innovation Farm, a Mini Factory or a Space Station 🟢:
  1. Basics of Electronics (3 months): atoms, current, battery, LED and copper-tape circuits, breadboards, series and parallel, buzzers, motors, LDR.
  2. Electronics in the Factory / Microcontroller: sensors, Scratch/mBlock, Arduino, and a first look at 3D printing.
  3. IoT: ESP8266/ESP32 boards and cloud platforms.
  4. AI: machine learning, vision, sound, face and colour recognition.
• Learners write code in Tamil as a teaching step. In a “Human Arduino” role-play they act out the program with their bodies before typing it 🟢.
• An EIS session has a published shape: a 10-minute icebreaker, then theory 30%, hands-on 60% and reflection 10%. ICT runs for 20 hours over four months, two sessions a week 🟢.
• Student projects start from local problems, not textbook exercises: a robot that detects landmines on its own, lake flood monitoring and alerts, rain-activated protection for drying clothes, energy-saving library lights, an LED early flood warning, and a home-made animal repellent to protect crops 🟢. The list reflects Batticaloa after the war, and a place prone to floods.
• Learning is meant to be measured at the start and the end: EIS is designed to test pupils before teaching begins and compare the final results with that starting point, with monthly, mid-term and final impact reports 🟢.

---

9. What makes someone a Changemaker?

Progress inside the top two stages is marked by named milestones, which are recorded and never graded.

• Innovator: problem chosen (a real one, outside the classroom), then first user feedback, then solution validated with the person it was for.
• Changemaker: grassroots innovation produced (the milestone that makes someone a Changemaker), then route chosen, then measured impact beyond the programme.

A grassroots innovation has to meet three conditions, all at once:

The problem came from the community, a real user who is not on the project team has tested it, someone depends on the outcome.

The output can be a product, a service or an initiative; the form does not decide it.

Not every Innovator becomes a Changemaker. A build that is technically excellent, but did not come from the community and has no real user outside the project team, leaves its maker an Innovator. That is a complete place to finish.

We apply the same test to our own past work. When we checked the past projects we had described as innovations against this stricter definition, few passed. For a larger group we could not tell, because we know nothing about them after the build. Several were innovations but not grassroots innovations.

The Changemaker stage opens onto an open field of outcomes, not a funnel. The routes seen so far are examples, not a boundary, and all are valid: an impact venture, research, activism or community leadership, employment or freelancing, and Grow Your Own. Not everyone is steered toward founding a company. The research warns that pushing students into starting businesses, when they would not otherwise have chosen to, can lower their outcomes 🟡.

---

10. What real classes changed in the method

We treat our own history as design-based research: try something, watch what happens, improve it, and learn general lessons from it. Our approach has gone through four versions, one after another, rather than four separate programmes:

1. Bootcamps that introduce the subject: awareness and first skills. Limitation: nothing to continue with after the event.
2. Groups that develop ideas together: ideas were produced. Limitation: few ideas were carried through to a build.
3. A growing network of makers learning from each other: mentoring begins. Insight: in the later stages of building and growing, learning from peers worked better than structured teaching.
4. A mixed model: structured teaching plus informal learning. Some learners now stay with it longer.

The core finding: the path from maker to entrepreneur cannot be run as a single programme from start to finish. It grows out of structured teaching combined with learners helping each other, and it is improved one round at a time. The key change is from running programmes to supporting a network of people who learn and build together.

A trainer’s own written notes across many sessions changed how sessions are designed 🟢:

“The biggest issue is not the lack of facilities, laptops, internet, or tools… The biggest issue is the mismatch between the expected level of the session and the actual readiness of the participants.”

What we decided to change as a result:

Finding	Change
The biggest problem is learners not being ready for the planned level, not a lack of resources	Plan each session around measured readiness, and teach the basics first
Asking people whether they already know a topic does not show whether they do	Replace the “do you know it? / yes” phone check with a short task to do before joining and a clear statement of commitment
Attendance drops at the point where the work gets harder	Short tasks, demonstrations, group work and tools learners can use straight away; make projects easier rather than adding more content
Sessions of 2 to 3 hours outlast learners’ attention and the trainer’s energy	Sessions of 90 minutes to 2 hours
Hands-on work does not automatically lead to learning	Making without readiness or drive did not produce learning, so the claim that people learn by making carries that limit
Feedback only comes from people who stayed	Build a way to collect feedback from people who drop out; feedback from those who finished is vaguely positive and does not show what went wrong
Adult sessions fail when the session and the participants want different things	Agree what the session should produce with the participants or their managers before designing it

These are findings with a list of changes, not changes already made. We have watched and analysed. We have not yet made the changes and tested them again.

---

11. What we count

• For each course: learners enrolled, how many earn a certificate at the talk-through, and how many move on to the next level or stage.
• For each stage: how many reach each milestone, and the routes Changemakers take.
• For each learner, each term: how many pieces of work meet the Definition of Done. This is the unit used to count finished, assessed work.
• For each learner over time: the Maker Profile.

Two things we do not yet measure, and we are working on both. We do not yet know how many people who join the pathway become Changemakers, because we have not counted the number who start. And we do not yet follow up learners who drop out, at any stage. Until we do, there is no way to know how many stay, and every outcome we have published so far comes from the people who stayed.

---

12. Open questions

• Earlier certificates: we have not yet settled how certificates issued under the earlier scheme match the current stages.
• Number of levels in each track: worked out as each track is written; robotics first.
• Milestone names: a working set, not yet tested in real teaching.
• What learners hand in at each level: the curriculum table places content that was really taught onto the levels as a first guide. It is not yet a set of finished course outlines.
• Certificates for soft results: these remain unverified.

---

A note on sources

The method comes from our written plans: the Changemaker Journey and Definition of Done, the curriculum, the DreamSpace Lifecycle, the innovation and enterprise strand, and the agreed list of terms. The evidence about teaching comes from our own reports: our lists of training and of course materials, the account of how our approach changed over time, and case studies on learner readiness.

Short definitions of every term we use in a specific sense are in the glossary.