What science can do is provide estimates with error bounds. Sometimes the error bounds are extremely large during a developing situation, and people should absolutely be aware of that. There's a major discussion to be had about risk/reward tradeoffs in the face of high uncertainty. (E.g., the relative cost of mandating face coverings is pretty low, even if the science is temporarily uncertain.)

However it becomes impossible to have such discussions when a large contingent of the population are ignoring all of the scientific data, and instead basing their opinions on motivated reasoning, faith, or political preferences.

I don't disagree with any of that. Indeed, there are many policies and behaviors that probably fall into the "better safe than sorry" category--given how much noise and uncertainty there is in the data--although even possibly reasonable people will vehemently disagree about what precise policies/behaviors that applies to. (As you see from voting patterns and discussions here.) And, yes, I also agree that face coverings under at least many circumstances fall into that category.

At the same time, there are different tradeoffs because different people (and governments) are approaching risks and rewards differently even if the facts were known with high certainty.

This is perfectly what I wanted to say, but was incapable of saying. Thank you.