I'm trying to figure out how realistic the mentioned concept of a "window of comprehension" actually is. This [paper? article? post?] only has one external reference - to another document from the same group[1]. That short document, in turn, makes unsupported claims that "Studies have shown that there is a “window of comprehension” of about 30 IQ points." A quick Google search didn't provide anything further in suppo…
The footnote in the post also references this paper for the claim- Intelligence and personal influence in groups: Four nonlinear models. Simonton, Dean K. Psychological Review, Vol 92(4), Oct 1985, 532-547 http://psycnet.apa.org/psycinfo/1986-14442-001
"Describes 4 progressive models that provide a conceptual basis for a curvilinear relation between intelligence and an individual's influence over other group members.
"Model 1, the intellectual superiority model, by assuming that influence is a function of percentile placement in intelligence, predicts that beyond an IQ of about 120 intelligence bears a negligible connection with influence.
"Model 2, the comprehension factor, adds the consideration of the degree of comprehension by potential followers, yielding a nonmonotonic function with a predicted peak IQ of about 108, approximately 0.5 standard deviations (SD) above the mean.
"Model 3 incorporates the criticism factor that acknowledges a group member's vulnerability to intellectual superiors, and thereby predicts a 2nd nonmonotonic function with an optimal IQ of about 119 (or 1.2 SD above the mean).
"Model 4, the intellectual stratification model, expands on the fact that the mean group IQ varies across different groups and, consequently, predicts a high correlation between the group mean IQ and the IQ of its most influential member, with a leader–follower gap of between 8 and 20 points, depending on the submodel.
"It is surmised that other factors besides intelligence must participate in the determination of who becomes the leader of a group. Even if intelligence is not the exclusive determinant, the models strongly imply that published validity coefficients are biased downward."
I don't see any evidence for the existence (beyond the assumptions of "a model") of a "window".