I can't speak for JamesBarney's reasons, but agree with them that it's probably minimal.

Facebook and Twitter have better funded research groups focusing just on the sub-task of knowing what my current state is (a necessary part of being able to tell if you've changed my preferences), and for advertising purposes I am categorised as being into nouns I don't recognise, foods and sports which repulse me, and so on.

I expect the three letter agencies to mix me up with the horror film director I share a name with, and also to confuse me with three pet dogs and two pet cats.

Because there's been lots of research on how much advertising affects election outcomes and it's minimal.( At the federal/state level, in smaller races where voters lack information about either candidate it can be slightly more influential)

I can't imagine a Russian internet trolls with much a smaller reach and lower budget are going to be that much effective than best political ad minds of America at changing our opinions. (For Russian governmental organizational competency see Ukraine)

And then anecdotally I could have predicted every person's i knows vote a year before based on how they felt about woke culture and their party identification.