A commitment to rigor
Measure. Document. Publish.
Many initiatives claim to improve school climate. Few are willing to measure their impact. dialog’ has chosen a different path: every rollout includes an evaluation protocol combining questionnaires, observation tools, and analysis. At the same time, we are developing our own measurement framework, the ISR-dialog’ barometer.
Field results to date
students involved in field research in one year.
2 000+ people reached through dialog’ in just one year.
average satisfaction rate
95 % adoption by educational teams
90 % of surveyed players — young people and adults — want to experience dialog’ again, a strong sign of engagement across schools, families, and organizations
partner school · through programs (HEC Paris Young Innovators) or directly participating in deployment
A growing network of trained practitioners is building the capacity to bring dialog’ to more schools and organizations : 100+
certified dialog’ Animators, Masters or Expert Leads
users of the Savoir-Relier methodology over 15 years with average satifaction rate of 95 %
Measuring impact at every level
Empathy and social cohesion in primary school
In parallel, a research-based protocol has been designed to assess the impact of dialog’ in elementary schools, focusing on classroom cohesion, psychosocial skills, relationship climate, and students’ ability to self-regulate tensions. The assessment is conducted by researchers (HEC Paris and University of Lorraine, using a protocol that combines quantitative and qualitative measures and aligns with the evaluation approach recommended by the French Ministry of Education.
The studies test whether regular use of dialog’ improves five measurable dimensions: 1. Belonging & group cohesion — stronger sense of belonging, inclusion, and connection within the group 2. Listening & openness — more active listening and less judgment of others. 3. Emotional skills — richer emotional vocabulary and better emotion regulation. 4. Self-confidence & participation — greater confidence to speak up, including among quieter students. 5. Learning together — stronger group dynamics and greater enjoyment of learning with others over time.
Four tempos — T0 before the first session, then 1, 3 and 6 sessions (one school year).
Three complementary sources — students complete a questionnaire about their own experience and relationships within the group; teachers complete a separate questionnaire assessing changes they observe in students and classroom dynamics; and trained observers use a structured observation grid to track concrete behaviors, including interruptions, spontaneous rephrasing, emotional vocabulary, conflicts arising and being resolved, and the need for adult intervention.
Trois indices synthétiques — Cohésion Sociale (ICS), Qualité Relationnelle (IQR) et Engagement (IE), avec des seuils cibles explicites à T3.
Three composite indicators — Social Cohesion (SCI), Relationship Quality (RQI), and Engagement (EI) — are used to measure progress over time, with clearly defined target thresholds at T3, the third measurement point, to assess whether observed changes are sustained.
School-level indicators — reported conflicts, verbal incidents, temporary exclusions, and family feedback — are tracked over time to identify broader changes in school climate and compare them with findings from student questionnaires, teacher assessments, and structured observations. Together, these measures provide a more complete picture of how relationship dynamics evolve across the school.
Research rigor — the study protocol is preregistered before data analysis begins. Parental consent and school team approval are obtained in advance. Comparison groups are constructed to have equivalent baseline mathematics achievement, reducing pre-existing differences between groups. Students discover the games only when the study begins, limiting anticipation and expectation effects.
We welcome new schools in our research ↓
May–June 2026 · Testing dialog’ in elementary schools
Test in 2nd grade
and 4th–5th grades
Two sessions were conducted with the same 2nd-grade class (ages 7–8), involving 21 students in the first session and 20 in the second. Two facilitators each worked with a group of four students. Between sessions, the number of emotions and the way students guessed the storyteller’s emotion were deliberately adjusted to test which format worked best for this age group.
In 4th–5th grade, sessions conducted in two classes of 32 students provided an opportunity to observe how students with more disruptive or impulsive behaviors engaged with dialog’. Early observations suggest that the structured format may help these students regulate participation, express emotions more constructively, and engage more calmly with the group. These preliminary findings require further testing.
100 % of students say they enjoyed the game and want to play again. What they value most: getting to know their classmates better, sharing their opinions and asking questions, listening to one another, and finding answers together.
Preliminary observations from these pilot sessions suggest that group composition may influence the quality of interaction. Groups that mixed girls and boys and balanced more and less vocal students appeared to show more attentive listening and calmer exchanges. These early findings remain to be tested on a larger scale.
Emotion definitions provide useful support — 15 out of 20 students used them to better understand the emotions in the game and expand their emotional vocabulary, suggesting that definitions may help children identify and talk about emotions more precisely.
Anti-dialogue tokens are visual cues used during the game to help children recognize behaviors that can get in the way of dialogue — such as interrupting, not listening, or judging others. They are best introduced gradually, one behavior at a time. During the pilot sessions, children needed time and guidance to recognize these behaviors as they occurred.
The points didn’t seem to matter — none of the students appeared to focus on scoring during the pilot sessions. Their engagement seemed to come instead from the pleasure of telling their own stories, listening to others, and discovering more about their classmates.
One revealing moment. In a tense group of four boys, one student asked to play again and joined a mixed group. His behavior was noticeably different: he followed the rules and gave his classmates space to speak. Hesitant to reveal too much, he struggled to structure his story. Another student spontaneously helped him clarify it — listening carefully and supporting him without speaking for him.
« She didn’t come home talking about a game she had won. She came home talking about someone she had listened to. »
— Parent of a school 2nd grader
Pilot Experiment in high school
A promising pilot in high school
The first comparative study of dialog’ in a high school setting involved four 10th-grade classes and 104 students, with approval from school leadership, teachers, and parents. Students took part in three conditions: playing dialog’ before completing 26 math problems; completing the math problems before playing; or completing two consecutive sets of math problems, reflecting a more typical classroom sequence in France.
Preliminary observation: in this initial study, students who played dialog’ before the Kangourou math assessment showed higher average performance than students in the other conditions. They also reported greater enjoyment of learning and solving problems together. Given the small scale of the study, these findings should be considered exploratory and do not establish that dialog’ caused the differences observed.
They provide a hypothesis worth testing: whether a relationship-based experience before a learning task may influence students’ engagement, enjoyment of learning together, and potentially their performance. This hypothesis is being investigated on a larger scale in 2026–2027.
Fall 2026 · scaling
700+ students. 30+ classes. A controled study.
Fall 2026 marks a major step: the first controlled scientific study of dialog’, led by HEC Paris researcher Anne-Laure Sellier in collaboration with Craig L. Anderson, and supported by a scientific consortium bringing together FHU VEAVE / Necker Hospital AP-HP, Lille University Hospital, and the University of Lorraine (LISEC, UR 2310).
Research sites — the main high school study is being conducted at Rosa Parks High School (20 10th-grade classes) and Marcel Pagnol High School (10 10th-grade classes). The research program also extends to Jules Verne Middle School in Les Mureaux, within the Versailles education district.
Participants — 700+ 10th-grade students across 30+ classes, with experimental and control groups and 345 students over two levels (7th and 8th graders).
School teams — 30+ teachers trained to facilitate dialog’ independently, with school leadership teams and student support staff actively involved in the research program.
Outcomes measured — trust, emotion regulation, cooperation, violence prevention, relationship skills, and learning outcomes.
Research instruments — Qualtrics questionnaires capture students’ and teachers’ reported experiences, including trust, emotions, cooperation, relationship skills, and learning; the ISR-dialog’ barometer tracks changes in relationship climate and group cohesion; structured observation tools record concrete behaviors during sessions, such as listening, participation, interruptions, and interactions; and monthly indicators track how these measures evolve over time.
Research rigor — preregistered study, parental consent and school team approval obtained in advance, groups matched for mathematics achievement, and games introduced only at the start of the study to limit anticipation effects.
Measuring impact at every level
Measuring relationship climate in middle and high schools
Developed through 18 months of field testing, the Savoir-Relier Index (ISR-dialog’) is a psychometric instrument designed to measure relationship skills at the student, classroom, and school levels. It provides an overall score out of 100 and four component scores, each measured through five items on a 7-point Likert scale.
I1 — Genuine · Know yourself —Relationship with self — authenticity, self-awareness, confidence, and self-esteem.
I2 — Generous · Recognizing others — listening, attention, suspending judgment, clarification, and discernment.
I3 — Generative · Acting and cooperating — cooperation, autonomy, and co-creating solutions.
I4 — Connecting · Building bonds — dialog’s cross-cutting pillar: shared meaning, sense of belonging, and psychological safety.
Teachers are part of the measurement process, not outside it. The barometer includes a teacher component, ISR-E: a 12-item perception questionnaire mirroring the four ISR-dialog’ components, together with eight factual, observable indicators (ISR-F), such as time spent on classroom management or the number of lessons taught in small-group configurations.
Five measurement points throughout the year — T1 in September (baseline), T2 in November, T3 in January, T4 in March–April, and T5 in June (annual assessment and T1–T5 comparison). These points are aligned with the school year’s actual milestones, including report periods, class councils, and meetings with families.
Our ambition — for every participating classroom or school to become a measurement point for relationship climate, with longitudinal tracking over time.
students involved
partnering High Schools
trained professors and animators
participating French regions
On example of field action : the young innovators program
Innovation as a driver for teaching & learning
As part of HEC Paris’s Young Innovators program, dialog’ prepares high school students from underserved backgrounds for the collaborative demands of an innovation sprint. Before tackling the challenge, students experience dialog’ to practice speaking up, listening, understanding different perspectives, giving and receiving feedback, and making decisions together.
They then put these relationship skills into action as they define a problem, develop and iterate on solutions, and pitch their project to a jury. In this context, academic achievement alone is not what matters: the ability to contribute, listen, collaborate, and build on others’ ideas can make a real difference — skills that also matter when choosing a path, applying for opportunities, or interviewing.
« Before playing dialog’, some students in the same group were not speaking to one another. One student even asked to move to another group. During the game, they began sharing personal stories and listening to one another. By the end of the session, she asked to stay with the group. »
— One Professor ,Young Innovators program
The student’s academic level did not change in an hour. Her place within the group did. That is exactly the kind of change we measure.
What we will aim to measure — 2026-2027 cohort
Self-confidence and relationship safety (Rosenberg scale) — feeling confident to speak up, be yourself, and engage with others without fear of judgment.
Cohesion et cooperation — feeling part of the group, working together, and building solutions collectively.
Public speaking and authenticity — feeling safe to speak openly, be yourself, and express different perspectives without fear of judgment.
Entrepreneurial mindset — taking initiative, turning ideas into action, persevering through challenges, and adapting when things do not go as planned.
Enjoyment of learning — in the pilot study, groups that played dialog’ performed better and reported greater enjoyment while completing logic exercises.
Three measurement points — T0, T1, T2
One composite measure — Savoir-Relier Index, scored out of 210 across the relationship and collaborative skills assessed. Two groups compared over time — students who experience dialog’ and control-group students who do not, both measured at the same points to compare how their scores evolve. When a young person discovers that they can express and defend an idea and work effectively with others, they discover more than a skill: they discover a possibility. Measuring these changes helps us assess whether dialog’ can strengthen some of the capabilities that contribute to equal opportunity.
Take part in the measurement
Schools, education districts, local communities, and other participating organizations: every classroom using dialog’ can become a measurement point. We provide the questionnaires and observation tools, then share the results with your team.
Every classroom can participate in the research
From field testing to scientific evidence — 1000+ students tested, 97% satisfaction, a controlled scientific study with HEC Paris and FHU VEAVE involving 700+ students starting in fall 2026, and a relationship climate barometer for students and teachers.
Publications et calendrier scientifique
EN VEILLE Les prochaines étapes
Septembre 2026
lancement : constitution des groupes contrôle et expérimental, sessions dialog’ in situ sur deux sites.
Octobre - décembre 2026
questionnaires post-protocole, baromètre, analyse comparative, suivi mensuel des indicateurs.
Janvier 2027
rapport intermédiaire sur 700+ élèves, premiers résultats publiés et partagés ; réplications sur d’autres terrains (Jeunes Innovateurs, primaire).
2027-2030
deux contrats doctoraux (sciences comportementales, sciences de l’éducation) accueillis par l’école doctorale SLTC (Université de Lorraine).
2028-2029
publications scientifiques HEC Paris × FHU VEAVE, plaidoyer national, ouverture internationale.
Objectifs institutionnels — produire un rapport scientifique publiable permettant le référencement national, contribuer au référencement de dialog’ dans le programme pHARe, et ouvrir la voie à un déploiement massif : 500 établissements à l’horizon 2028.
Rigor matters, too
What we don’t claim to know
Scientific rigor also means recognizing what we do not yet know. Pilot results remain preliminary and require replication. Evidence will come from multiple sources — students, teachers, observed behaviors, and school data — not from any single measure.