Every organisation that publishes a psychological tool makes choices about language. Too much caution can hide useful work. Too much certainty can turn a limited finding into a claim the evidence cannot carry. Foundation Nova Diem uses a clear standard between those extremes.
Say what Lux does well, in plain and confident language, and never say more than the evidence supports.
Confidence first, not confidence instead of rigour
Foundation Nova Diem exists to build psychological tools that start from strengths and patterns, not deficits and categories. Lux gives people language for how they think, relate, respond and grow. We can say that plainly because it describes the assessment’s purpose and output.
But confidence about Lux’s purpose and design does not turn active evidence development into completed Lux-specific validation results. That is the boundary this claims standard sets.
What held to evidence means day to day
In practice, this standard is supported by concrete engineering controls and publishing rules:
- Every scoring pathway is versioned. If a result was produced by an earlier version of the engine, that remains traceable rather than being silently absorbed into the current version.
- Scoring changes are checked with committed fixtures and regression tests. Selected outputs are compared with defined reference values during development; this is not a claim that every output is independently revalidated before every report.
- Public claims should identify the evidence that supports them. A source-bank reliability figure should be described as source-bank internal-consistency evidence, not stretched into a broader claim about validity, outcomes or fitness for a use case it has not been tested against.
- We describe evidence development as work in motion. Test–retest evaluation, for example, is part of the active programme and will assess score stability as repeat-response data becomes available. We do not present the study or its result as complete before it is.
Why the distinction matters
A figure without its scope can invite a reader to assume more than a study showed. We publish the scope with the figure. The inherited internal-consistency range, for example, belongs to the source/reference item banks and dataset. It is not presented as completed Lux-specific validation.
This makes each claim easier to trace, question and correct. It also makes the limits of psychological insight easier to inspect. The current evidence position records what we can say today, what remains under study and what we do not claim.