You make a valid point. I think the issue is the testing
   methodologies and the nature of generalized abstract reasoning
   as considered by those testing methodologies.

   What might be considered "generalized abstract reasoning" might
   be better described as visual-spatial processing ability.