The author doesn’t address predictive power. She discusses social science prediction compatibility with mathematical formulas.

The author doesn’t compare mathematical importance overall. She only speaks on mathematics as a means of prediction; we don’t know her opinion on other ways that math could factor into these sciences.

We don’t know anything about a need for improvement. We know some social scientists want the power to predict accurately, but that’s all.

This accurately paraphrases the conclusion. The author says “this” (reducing phenomena in the social sciences to mathematical expressions) “would be a mistake” (should not happen).

This is an inaccurate interpretation of context. We’re told that prediction is the hallmark of natural sciences; a trademark isn’t the same as being causal.