Nitpick, but it's DSM 5; roman numerals were dropped after DSM IV.
> I do not know if the experts are not aware of the changes or don't agree with them
Almost certainly the latter. Lots of DSM V changes were, and are, controversial (even among people who acknowledge the existence of the problems the changes were intended to fix, the aisle solutions chosen are sometimes controversial.)
Its more of a Pavlovian reinforcement than reputation. You see what got you lots of points (what people liked) and you're more apt to do that in the future.
By keeping everyone else blind to someone's scores, you don't have as much popularity bias creeping in.
For example, as ternary predicates:
1: water is wet
2: identity_1 corroborates 1
3: 2 by_way_of blind_faith
4: identity_2 denies 1
5: 4 by_way_of blind_faith
etc.
You can go as far with this as you want, and the consumer gets to decide what establishes or hurts credibility, and which statements of fact corroborated by what methods and by whom will either count toward or against the credibility of a specific form of statement on a specific topic by a specific set of identities.
For example, you could have predicates describing conflicts of interest, and identify the relevant conflicts of interest by querying for interests in the predicate in question to find what people you trust will say about the interest of individuals and organizations in corroborating or denying a particular claim.
You could also have predicates which show ways in which a statement is controversial in its general form, but is uncontroversial when refined. For example: liquid water can not be wetted with liquid water.
Sure, if those with higher reputation have their claims promoted above others with poorer reputations.
Long term the goal is to have an open reputation score for every fact checker. I'm currently experimenting with different implementations of this, and seeing good promise.
Working on this over the last few years though, I've personally fact checked over 500 assertions.
Also, on first glance the app looks great, well-polished, and the UX of reading an article and looking at sources for claim is nice :). Thanks for working on and sharing this!
Specifically, let's say you support belief A, and I want to add a whole bunch of facts that support belief B. You're still the gatekeeper to decide whether I get to be a contributor. There's no way for someone to know that you're going to approve equally-qualified contributors on different sides of an issue, so they're always going to wonder whether there are voices or facts being omitted.
(But if that's just the plan for beta, that's fine.)
Long term every review, even those that are flagged, will be accessible. Albeit, those that are flagged will be in a separate section, but the point is they will always be accessible, so that you never have to trust me or the community, you can always review for yourself.
Again, not for the omission of facts, or the omission of counter-evidence.
E.g, suppose you only let in contributors who agree with you. An article is posts with "Fact A," which you happen to agree with. A bunch of your hand-picked contributors put up evidence to support "Fact A." No one ever posts the evidence debunking "Fact A."
So there's no way to "verify" anything, because we can't see what's not there. Most readers will be satisfied that "Fact A" is well-sourced.
So it's not enough to say that the reader can verify everything and not need to trust you. I mean, they could do their own research to Google the evidence debunking Fact A, but they could do that anyway and the platform doesn't support them.