The term "diversity" is being used in very different ways. The article you posted drfines it based on skill and knowledge:

> By diversity, we mean differences in how agents build problem representations which allow them to access various solutions. We deploy the concept of diversity along two dimensions: knowledge amplitude, which accounts for the level of available knowledge allowing access to poorer or richer problem representations (compared with complete problem representations), and knowledge variety, which pertains to the differences in the constituents of agents’ representations.

So it just further demonstrates the logical gap between A=B and C=D, since the studies didn't even show A=B, they showed E=F