I have to disagree. It should not be a consequential or self-conscious act if you aren't using your real identity. (If you are, the expectation that you should be very careful with how you participate remains unchanged. This isn't LinkedIn and it shouldn't be trying to be.)

Commenters on the GitHub issue have put it better than I can:

An average user absolutely benefits from being able to see who voted on a post or comment and what their vote was. A person noticing that someone is actively down voting their content in a deliberate way empowers the user to have it dealt with. Mods might not [cue in to] that kind of targeted harassment.

Your vote isn't private in either case regardless. At most you need to know someone's birthday, first name, and last name to find someone's voting record in America (might depend state by state). Someone willing to set up a Lemmy instance to see your votes is also capable of then setting up bots to specifically target you with down votes, which is the more egregious of the two actions.

People who use the fediverse need to get used to the fact that things are not private here, that's the point of interoperability, trying to convince them that they have fake privacy is just going to make them feel self entitled and violated when they learn that nothing here is really private, which shouldn't really be expected as it is a public and decentralized forum.

I don't think users are done any favours by pretending they are private as bad actors can already do whatever odious crap they want to and it leads people into a false sense of security. For example someone liking controversial content on an account which can be traced to an identity they may need to keep separate is already taking a massive risk under a false assumption of privacy.