Research articles are rated in the following way:
Meta analysis and Systematic Reviews
Studies of unselected groups.
In order to be awarded 4 stars, studies must have sampled from the general population before determining who belong to the "gifted" group. These studies are therefore able to say something about the entire gifted population. Studies that have casted a wide enough "net" in pre-selecting from the general population (such as the SMPY) also achieve this rating.
Studies of selected groups and literature reviews
Studies of selected groups might not be representative of the entire "gifted" population, because they have sampled from a selected group; for example Mensa members, children enrolled in special "gifted" classes, etc....
Literature reviews at this level may have been selective and/or non-systematic in nature.
Surveys, qualitative studies and case reports.
These studies have actually examined subjects, but have not done so in a methodologically sound way to be awarded more stars.
Expert opinion / Other.