You may find these resources of interest: in a recent meta-analysis of correlations between a few versions of CogAT and various instruments, r's with intelligence tests were mostly between 0.54 and 0.79, with the highest average values (0.84) in technical manuals (data and analysis generated by the publisher). Along those lines, another significant finding was that the highest numbers came from studies authoried by the test author. (Original source is behind a paywall, but you can see these numbers in the abstract. Limited free access on researchgate.)
(
https://journals.sagepub.com/doi/10.1177/00169862241285593)
(
https://www.researchgate.net/public...Validity_of_the_Cognitive_Abilities_Test)
Somewhat outdated in terms of the actual instruments (some have been revised since) and data catchment (obviouslly there's been a decade-plus of research done since then), but still relevant in terms of the discussion, here's a John Wasserman moderately-technical presentation on some influences on the applicability of various tests (including CogAT7) in gifted identification, including the intent to measure g, as well as validity via special group studies/correlations with gold-standard instruments. Note there are no WISC-IV comparison numbers for the CogAT7 specifically in the gifted population.
(
https://johndwasserman.com/index_ht...earners%20SMU%20Presentation%20Final.pdf)
