So the more strongly people trust science, the more value there will be to con men in passing themselves off as scientists. We can mitigate that to some extent with scientific reputation and whatnot, but I don't think we have trust fully figured out in general, as there's always the bootstrapping problem.
In reality, an anti depressant for example, probably works out to something like: it works about 50% of the time, with a bit of a lean over 50%, but if we calculate in spontaneous remission, placebo effect, and some other variable, now it looks like it is either a net zero effect, or in many cases, like AA meetings, statistically detrimental. But the general public will need a binary answer.
And I consider myself part of this general public, I just know there is more than meets the eye on anything, and in general I feel correlation and causation more often then not don't agree.