Mean: the average answer across everyone who responded, on the answer scale. On this survey a lower mean means people mostly chose the favourable end.

SD (standard deviation): how spread out the answers are. Small means people answered similarly; large means opinions varied. Around 1 or less is typical here.

Item-rest correlation: how well this question moves in step with the rest of the question bank. 0.30 or higher is healthy. Near zero (or negative) means the question is measuring something different from the others.

Loading (factor loading): how strongly the question belongs to its category. 0.70 or more is preferred, 0.50 acceptable. Below 0.40 is weak, and a negative number means it moves the opposite way to its category, which usually points to reverse wording.

Variance: the share of a question's answers explained by its category. 0.50 or more is desirable; under 0.40 means the category does not represent the question well.

p value: the chance the result is a fluke of this sample. Below .05 counts as statistically significant. "<.001" means less than 1 in 1,000.

Cronbach alpha: how consistently the whole question set measures together, from 0 to 1. 0.70 acceptable, 0.80 good, 0.90 or higher excellent. This round scored 0.98.

KMO (sampling adequacy): whether the data suits this kind of analysis. Above 0.60 acceptable, above 0.80 good.

Bartlett's test: checks the questions relate to each other enough to analyse as groups. It passes when p is below .05.

Chi-square: the overall gap between the category model and the real answers, judged together with the measures below rather than alone. Smaller is better.

CFI and TLI: how much better the category model fits than assuming no structure at all, from 0 to 1. 0.90 acceptable, 0.95 or higher preferred.

RMSEA: the model's error per question. Below 0.08 acceptable, below 0.06 good. The bracket after it is the 90 percent confidence range.

Verdict pills: the report text gives the cut-offs (CFI and TLI 0.90 acceptable and 0.95 preferred, RMSEA below 0.08 acceptable and 0.06 preferred), but each category's verdict is awaiting confirmation from the research team; Trauma is the one fit marked red in the report.

Low (red) and Amber: exactly the rows marked red or amber in the research report, read directly from the coloured file (the exact amber rule is with the research team to confirm). Strong: unmarked with loading 0.70 or more (the preferred level). Good: unmarked with loading 0.50 to 0.69 (the minimum acceptable level). Red and amber questions are coloured; validated questions stay plain with their pill.

Category bar and colour: one bar per category showing green (validated), amber and red segments, with the count beside it, for example 6/6 or 2/6. The header shows the overall status: green when more than two thirds of its questions are green, red when at least half are red, amber in between.