Multisport Diagnostics Lab for Children
Multisport Diagnostics Lab for Children
Our model has been validated on 300+ children: reliability 0.87–0.89, accuracy 85%. But the real exam for any diagnostic happens outside the lab: did the child join a club, enjoy it, and stay six months later? Only you can answer that — which is why we invite families to take part in the research.
Important: nothing changes for the child. The diagnostic is the same for everyone. Participation means a few short questionnaires for the parent — all optional, and only after you receive the report. Any of them can be skipped.
85% is accuracy "on yesterday's data": the children we assessed in 2024–2025. To stay accurate going forward, the model needs real-life outcomes.
Your "matched / didn't match" answer for each of the 4 components shows where the model misses for children like yours — and what to fix first.
Whether the child enrolled and stayed in sport is an external criterion — the strictest quality test of any diagnostic. It is impossible to obtain in a lab.
Children aged 6–11 change fast. A repeat diagnostic a year later shows what in the profile is stable and what is developing — without it you cannot tell a test error from growing up.
Every step is voluntary and any can be skipped. Even a single answer from you is a unit of scientific data.
Age, sport experience, current clubs, the family's expectations. This is context: it helps distinguish "the model was wrong" from "there simply is no such club in town".
Was it interesting for the child, was everything clear, would they do it again. The diagnostic must stay a game — we monitor that.
For each of the 4 report sections: matched, partly, or not — plus a free comment. The most scientifically valuable questionnaire of all.
Did the child join a club, was it a recommended one, do they enjoy it. Three months is the typical first-adaptation period.
Stayed, switched sports or quit — and why. Optionally, a short coach's assessment as a second source.
The child goes on the "journeys with Leos" again, and you see the profile dynamics: what grew, what stayed stable. Only if the child wants to.
Participants get a repeat diagnostic a year later and see how the profile changed: what grew, what is stable.
"Didn't match" is not a complaint — it's data. We will see where the model misses for similar children and fix it; your family's next report will be more accurate.
The model version is stated in the report; how and why the model changed is published openly in Scientific Materials. Once a year — a scientific paper on the anonymized cohort data.
Consent to participate in the research is a separate document. Declining changes nothing: you receive the full report in any case.
No explanation needed. Previously collected answers are deleted at your request.
Data is pseudonymized for analysis; publications contain only aggregated figures ("72% of children…"). No child is identifiable. Data is not shared with third parties and is not used for advertising.
The repeat diagnostic happens only if the child wants it. The methods are built as a game and children usually ask for "more" — but any "I don't want to" is respected without discussion.
The surveys contain no medical questions and do not replace a doctor's care.
It will. The turning point in diagnostic-model statistics comes at dozens of observations: 25–30 families with 3-month answers — and we can already see a systematic error in a specific indicator. Your answer may be exactly that one.
It means the test is honest. No instrument in the world gives 100% (ours gives 85%, and we say so openly). The difference between science and advertising is that science looks for its own errors and fixes them. Your "didn't match" is the most valuable answer of all.
The diagnostic is free for research program participants. Questions about the research — athleos57@gmail.com.