High Confidence

Confidence describes how strong the evidence for a finding is. High means the finding is anchored in a primary statutory source or corroborated by at least two independent official sources. Medium marks credible official evidence that does not itself provide the primary statutory anchor. Low marks evidence that remains indirect, uncertain, outdated or in unresolved conflict.

Gap identifies a requested datum that is absent rather than assigning it an evidence-strength grade, and Gap/partial one that is present but incomplete. Neither says why: the evidence-status chip beside it carries that, and the two readings are deliberately separate — a datum can be missing because the authority does not publish it, which is a finding, or because we have not found it, which is not. Not graded means a grade does not apply, including analytical summaries and observations that are structurally not applicable.

Study-derived Derivation

Derivation identifies whether the figure is stated by the source or calculated by the study. Study-derived means we computed it from figures the sources do state, most often time-to-licence where education and practice are added together. Study-characterised means we compressed a longer description into a short cell. Values with no chip here are stated by the source. This is not a weakness rating: a Study-derived value can rest on impeccable statutory sources.

Primary-backed Evidence status

Evidence status identifies what kind of source stands behind the observation. Primary-backed means it rests on statutory evidence. Primary + secondary corroborated means primary and independent comparative evidence both cover the observation and agree. Secondary-only — primary pending means only comparative evidence is available; it is a research target and is not publishable as a finding until a primary source is found or the observation is restated as an explicit gap. No statutory instrument means the state does not regulate the matter, so the voluntary body is the authoritative publisher of its own arrangements rather than a missing statutory source. No statutory instrument — primary-confirmed is the same fact of absence, but affirmatively established by an authoritative source rather than asserted by a voluntary body, and is the stronger of the two.

Not published — primary consulted means the authoritative source has been consulted and does not publish the requested datum. That is a finding about the body's publication practice rather than a gap in the research, and the next step is a direct enquiry to it. No primary located (Phase 3 target) means desk research did not locate a primary source, so the next step is further search and, if necessary, a Phase 3 enquiry. Neither should be read as evidence supporting a substantive value, and they are not interchangeable: the first describes what the authority publishes, the second describes what desk research could reach.

Graded observations

The figure above the table is the share of that country and area's observations carrying an evidence grade of High, Medium or Low. Observations recorded as a gap, and analytical or structurally not-applicable rows, are counted in the denominator, not excluded from it — a jurisdiction whose fees and CPD load are unpublished genuinely has less established record than one whose are, and that is what the figure is for. The composition is printed beneath it so the percentage can be checked against the table rather than taken on trust. It ranges widely: at the low end are jurisdictions such as Albania, where the Order law is not in force and almost every observation is a gap; at the high end, jurisdictions whose access rules, fees, CPD and register are all published. The current range can be read off the dashboard itself, and this note deliberately does not restate it — an earlier version quoted a fixed floor of 0% for Albania, which ceased to be true as soon as the first Albanian observation was graded, and nothing in the pipeline noticed.

Reading the source column

Where a primary source stands behind the value, the source column is labelled Primary. Where the evidence status reads Not published — primary consulted it is labelled Consulted instead, because the source was read and does not support the value — it establishes that the authority publishes nothing on the point. Calling it Primary there would claim support that the source does not give.

Two counts, both correct

This dashboard counts observations: one row per country, canonical indicator and qualifier. The written report counts authored dossier rows, which is the smaller figure, because a few dossier rows carry two distinct concepts and are split when published — title and practice protection, temporary services and establishment, and Area 2 application volume and qualification origins. Neither figure is wrong; they answer different questions, and a total here will not match a total in the report.