Phase 313 weeksSix intervention schools, six controls
The intervention
Weekly gameplay in six Addis Ababa high schools, run entirely inside the students' lunch break. Designed for fourteen weeks and delivered in thirteen, against a school calendar that fought back the whole way.
13
Weeks delivered
6
Intervention schools
50
Target students per session
15–30
Minutes of play per session
73.2%
Best school retention
40.6%
Lowest school retention
Operations
One school a day, five days a week
Scientific consistency required a fixed rotation. Every school got the same games in the same order in the same week, which meant the schedule had no slack in it at all.
Monday School 1
Tuesday School 2
Wednesday School 3
Thursday School 4
Friday School 5 and School 6
Game Masters assembled at the Creative Hub from 11:00. Supervisors checked materials and staffing before transit. Teams arrived 20 to 35 minutes ahead to set up a classroom, which bought an uninterrupted 15 to 30 minutes of actual play.
First-day sessions ran Beyaynetu Level 1 at every school — the fastest game in the set, chosen deliberately to open on energy.
Operational findingThe focal person decided everything
Observed
Session quality correlated directly with how involved each school's designated focal person was.
Evidence
Where focal persons mobilised students, secured venues like the library and communicated with administration, attendance was stable and delays were minimal. Where communication was weak, students forgot the schedule and Game Masters lost setup time hunting for participants.
Design decision
Future interventions must build rigid accountability and incentive structures for school-side focal persons. This is not a soft variable — it produced the widest spread in the whole dataset.
Thirteen weeks
The behavioural arc
How students related to the games, the facilitators and the content evolved sharply over the weeks. We divide it into four phases.
Phase A · Weeks 1–2
Initiation, and a visible knowledge gap
Week 1 introduced Beyaynetu Level 1 — rapid categorisation of basic food items. Initial confusion was high; most students adapted after two rounds.
Week 2 moved to Gursha and portion control, and immediately exposed a competency gap. Students who had attended week 1 integrated smoothly. Students who had missed it could not balance portions, because they could not yet name the food groups.
Why this mattered
The gap was the first proof the curriculum was actually teaching something. Miss week 1's food groups, and week 2's portioning is impossible to execute. Learning was cumulative — which made absence genuinely costly.
Phase B · Weeks 3–6
Emotional investment, and students teaching students
As Digis and Beyaynetu Level 2 arrived, emotional investment deepened sharply. During Digis, students showed real frustration at losing points to danger and event cards — volatility the team read as high intrinsic motivation, not a management problem.
The milestone of this phase had nothing to do with scores. Experienced students began teaching the rules and food groups to teammates who were absent or new. Knowledge transfer had shifted from facilitator-to-student to student-to-student — the guidelines were entering ordinary conversation.
Mid-intervention sessions. By this point rule explanations were getting shorter every week.
Phase C · Weeks 7–10
Mastery, strategy, and a bit of cheating
By the midpoint, rule explanations were no longer necessary. Students had achieved mechanical mastery and moved entirely to strategic play — forming alliances to help friends win, and at School 1, competing intensely enough to produce illicit card swapping.
At School 2, players started deliberately avoiding street-food locations in Digis to dodge penalty cards. The clearest evidence in the whole study: choosing to avoid a food source to prevent a consequence means the logic of the guidelines was understood, not memorised.
They transitioned from simply memorizing food names to understanding the simulated consequences of dietary choices.
Project insight, weeks 7–10
This phase also took the heaviest hit from the academic calendar. National exams, holidays and election-related absences broke up team cohesion, and mixed groups of unfamiliar students consistently showed lower energy and poorer communication.
Phase D · Weeks 11–13
It reaches their plates
Beyaynetu Level 3 arrived and revived the flagging groups. But the breakthrough came in week 11 at School 5: asked what they had learned, students described — unprompted — reorganising their actual daily meals at home around the game's rules. School 4 echoed it.
Week 13 was emotionally charged across all six schools. Students were genuinely sad it was ending, and named what they would miss — a safe, highly interactive space where complex material arrived without a lecture attached.
Final sessions, week 13. Active gameplay was concluded exactly at thirteen weeks so students could focus on final examinations.
Quantitative analysis
Attendance and retention
Target: 50 students per session per school. Week 1 cleared 50% everywhere, several above 60% — launch novelty and tournament framing doing their work. Retention is where the schools separated.
64.9%School 4Mid-range · one sharp single-week dip
64.2%School 1Mid-range · gradual decline
58.5%School 5Mid-range · flat trend
40.6%School 3Lowest retention · steady decline
Result 01Rewards supported attendance where school-level engagement was already in place
Observed
Comparing weeks 1–7 against weeks 8–13 showed a drop in four of the six schools. The repetition of the same core mechanics had worn the novelty off.
Evidence
After introducing tangible rewards, School 6 climbed from 67.4% to 80% and School 5 stabilised. School 3 and School 2 barely moved.
Design decision
Incentives sustain motivation but cannot overcome deep institutional or scheduling barriers. The fix for a school like School 3 is structural, not material.
The reward system
A weekly data or voice package went to each day's top scorer, every week, for the rest of the intervention. At the midpoint we asked students what they actually valued, and added EarPods and club jerseys for consistent attendance and high scores.
The effect was immediate: renewed interest, steadier attendance, and harder play. The honest trade-off — it shifted motivation from intrinsic to extrinsic.
Students choosing their own rewards. Asking them what they wanted, rather than guessing, is why the incentives landed.
Qualitative analysis
Learning, and evidence of behavioural change
Indicators of mastery
Self-correction — at School 6, students caught their own mistakes and immediately recalled the right food group.
Peer instruction — at School 3, students explained the rules and the FBDG concepts to newcomers without any facilitator prompting.
Strategic reasoning — at School 2, avoiding street-food spaces in Digis to manage penalty risk.
Comprehension built cumulatively across the thirteen weeks, which cut both ways: students who missed foundational sessions visibly struggled and felt isolated when grouped with peers who had played continuously.
By weeks 12 and 13, unprompted
The games taught us about food, and I actually changed my own eating habits because of it.School 4
The games helped me understand how to organize my diet daily to live healthy.School 6
I want to focus on my meals more going forward.School 1
Group dynamics — a caveat
Session quality depended heavily on social stability. Merging unfamiliar students dropped gameplay quality — hardest in Gursha, which needs strong intra-team communication. Fast, competitive players sometimes intimidated quieter ones into disengaging, so Game Masters stepped in to keep participation fair.
What broke
Four operational challenges
Running a longitudinal study across six institutions produced friction that had nothing to do with the games. The team logged all of it.
Fourteen weeks became thirteenSevere disruption to the Ethiopian school calendar compressed the schedule, which forced two or more sessions into a single day. The result was facilitator burnout and less individual attention per student. The team recruited and trained an additional batch of Game Masters mid-intervention to redistribute the load and restore session quality.
Access restricted at School 6A misunderstanding with the school director produced a mandate barring male Game Masters from the compound. The five female Game Masters absorbed the entire workload for that school, doubling their group-management responsibilities and delivering every scheduled session without skipping a week. That school went on to record the highest retention in the study.
Focal person engagement varied widelyThe gap between School 3's 40.6% and School 6's 73.2% tracked focal person involvement more closely than any other variable.
Final exams closed the windowRather than push into the examination period, the team concluded active gameplay at exactly thirteen weeks — meeting the data collection and educational goals just before students needed full academic focus.
How it was managed
Weekly review and planning meetings ran between the research and project teams for all thirteen weeks — upgrading gameplay week to week, catching facilitator fatigue early, and replanning logistics against whatever the school calendar had just done. Staying responsive rather than rigid is the only reason the compression was survivable.
Closure · 6 July 2026
Handing it over
The closing event brought together the Ministry of Agriculture, EPHI, WUR, school directors and focal persons from all 12 schools — intervention and control — along with top-performing students and their families.
Stakeholders were walked through the full evolution — from the adolescents' first sketches to the finished games.
The adults had to play
Some school directors and focal persons had never quite grasped why board games belonged in a school. So the event sat them down at the tables. Playing the games — and finding they demand real cognitive and nutritional focus — closed the understanding gap faster than any presentation had.
Top-scoring adolescents then took the stage to describe their own behavioural shifts. Unscripted, from the students themselves, in front of the people who decide whether this continues.
Sustainability by distribution
Physical game sets went to all 12 schools — control schools included, each package with full rules so those students can learn independently. Intervention students already have the mechanics mastered well enough to form clubs and teach their peers.
That is what converts a research instrument into a community asset: the games stay in the schools, and they no longer need a facilitator to be played.
Officials, focal persons and families at the tables. Top-performing students were formally awarded their prizes.
What's next
Scaling past the six schools
The final phase takes the methodology national through social media — framing healthy eating as a level-up, not a chore — and offers donors and institutional funders a validated, evaluation-backed model ready to scale.
TikTok
The hype engine
High-energy, trend-driven content aimed at the younger demographic — the same audience the games were designed for.
Instagram
The community hub
Young adults and parents. Lifestyle integration and community building rather than instruction.
LinkedIn
The authority
Data-driven posts for policymakers and educators, establishing the scientific and institutional credibility of the approach.
A 30-day release in three phases
Days 30–21 · Spark curiosity Teaser videos with a deliberate small error in the artwork, so that corrections in the comments push reach.
Days 20–11 · Bust the myths Partnerships with nutrition creators showing that eating well is not expensive — the exact misconception the co-creation workshops uncovered.
Days 10–1 · Influencer launch Game streamers play on their own channels, with early-access codes for people who engage.
Inviting people in rather than lecturing them
Health advice on social media gets ignored because it arrives as a one-way instruction — the same failure mode the classroom lecture hit in Round 1 of co-creation. The campaign target is 25% of viewers actively engaging.
The tactics follow from that: posting a common unhealthy habit and asking which one people struggle with, using Instagram's "Add Yours" to collect real meals, and posing genuinely contested questions — who is responsible for teaching children about nutrition, the school or the parents?
Tracking focuses on shares as the measure of reach, and on the ratio of engagement to views.
Having deployed the validated game across thirteen weeks, we have handed the physical games to the schools for sustainable use and are launching a digital campaign to scale this nutritional impact nationwide.
Project conclusion, July 2026
