By age
The same questions aren't used at every age. Who answers, the item format, and how firmly to read the result all differ by age band.
Ages 3-4 - when to hold the result most loosely
Results vary most at this age. What drives the variation, and what you can genuinely see at this stage.
Ages 5-7 - when children start answering for themselves
Self-regulation grows fast and peer relationships take hold. A change from a few months ago usually means growth, not a different personality.
Ages 8-10 - when a result becomes a usable tool
Results are comparatively stable and a child can use them as material for self-understanding. The risk of labelling grows in the same measure.