Research methodology
Data, sources and methodology.
Huddle separates verified context, owner-confirmed project facts, prototype evidence, assumptions and future outcome claims so readers can see what is known and what is still being tested.
Method
Every study should remove a named uncertainty.
A useful research plan begins with a statement that could be wrong, explains why it matters and defines what evidence would support, narrow or reject it. Activities are selected because they inform a real product, safety, operating or funding decision.
The public record should distinguish external context evidence, qualitative learning, prototype observations and measured outcomes. These forms of evidence answer different questions and cannot be substituted for one another.
Evidence record
A finding is useful only when its source and limits travel with it.
Published findings should identify the question, method, population, geography, period, sample, source, responsible owner and review date. Quotations and images require permission; small groups and sensitive contexts require additional privacy review.
Negative and inconclusive results must be recorded when they change the direction. Huddle should revise, pause or stop a concept when evidence does not support it, rather than converting uncertainty into marketing language.
Study lifecycle
A study begins with a decision, not with a questionnaire.
Before recruitment, the study brief should name the uncertainty, the decision it informs, the people whose perspective is missing and the evidence that would support, narrow or reject the assumption. The method follows that question: interview, observation, diary, comprehension test, prototype task or a bounded field trial.
The brief should also define inclusion and exclusion criteria, recruitment route, sample limits, responsible owner, review date and stop conditions. A larger sample does not repair a vague question, and a compelling quotation does not become a measured outcome.
Participant and data care
The minimum useful evidence should require the minimum necessary data.
Each protocol should explain what is collected, why it is needed, where it is stored, who can access it, how long it is kept and how withdrawal affects existing records. Identifiers should be separated from research material whenever possible, and sensitive family or child information should not be collected merely because it might become useful later.
Children are beneficiaries of the work, not a convenient source of data. Any direct involvement would require a separately approved safeguarding approach, age-appropriate information, guardian responsibility and a clear route for the child to decline or stop without pressure.
Analysis and publication
Findings must show what was learned, how certain it is and what changes next.
Analysis should preserve disagreement, outliers and missing perspectives rather than forcing one clean story. Where useful, qualitative themes may be compared with task completion, comprehension or operational records, but correlation must not be presented as causation and prototype behaviour must not be presented as real-world impact.
A public update should state the question, method, sample, period, limitations, conflicts, result and decision. Material changes, corrections and negative findings belong in the record. Publication requires claims review and a privacy check before participant quotations, images or small-group details are used.