One quibble in two parts on the methodology: the use of a vocabulary test to measure intelligence. First, vocabulary is in part a measure of education, rather that raw ability. Second, verbal ability is semi-independent of mathematical ability, and I suspect math may be more important. However...
The last graph results looks STRONGLY suggestive of a catastrophe theory transition; specifically, a Cusp Catastrophe.
Using the labels on the axis as depicted, variable x indicates acceptance versus doubt of evolution (positive = doubt evolution, negative = accept evolution), and variable u indicates verbal level (positive = weak verbal, negative = strong verbal; of potential interest for further investigations, children tend to move from positive to negative as they age). Variable v in this study indicates degree of acceptance of the Bible (negative = Inerrant, positive = Fables). I suspect the true underlying variable v may more accurately be a measure for relative degree of acceptance of the importance of the Bible as a tool for thinking about the universe, in contrast with Science and Evidence. (The math-versus-verbal balance may take part.)
As v values become increasingly negative, the x versus u view will become dominated by belief of the biblical account. People will be on the "top" surface; as verbal sophistication increases, so does their skepticism of evolution. Contrariwise, highly positive v values lead to dominance of the evolutionary view, and location on the lower surface. As verbal sophistication increases, the x versus u view shows a decrease in skepticism about evolution results.
Considering the x versus v view for strongly negative u (the verbally sophisticated), we have two stable surfaces and an unstable "double bind" intermediary. The lower surface represents acceptance of the evolutionary account; as acceptance of the Bible's importance increases, it becomes harder to retain the evolutionary view. Contrariwise, the upper surface represents belief in a literalist biblical account; as science exposure increases, it becomes more and more difficult for the verbally sophisticated to remain with Biblical belief. The moderate "Bible Inspired" position may correspond to the unstable double bind surface.
I would suggest that it is not an inborn human intuition that causes the initial bias, but rather a universal social bias (due to the evolutionary benefit of religion for organizing societies; cf. Wilson and Diamond), and conveyed to new members (children) very early in the socialization component of the education process. I further suspect that the other correlations of intelligence to social group membership are dominantly the result of the correlation of group membership to state on the v axis.