Or: UI musings for the Voting Phase of the 2018 Review. This post outlines the current plans, and asks a couple concrete questions at the end.


Range Voting, which is converted into Quadratic Voting

The current plan for the Voting Phase, after some discussion on this thread, is for most people to start with a range-voting system, which is converted* into a set of quadratic votes that people can then fine tune.

* the precise conversion method is still up in the air. More comments and suggestions appreciated.

The range-voting part will basically be a "-1 to 3" scale, with the points corresponding to english-labels of "No", "Neutral", "Decent", "Important", and "Crucial". 

(This is a bit asymmetric between downvotes and upvotes, but you can add stronger downvotes in the quadratic step. Our experience testing voting systems was that only a small number of posts were things we expected anyone to downvote much. So it didn't seem useful to heavily emphasize downvotes in the first-pass-phase)

If you want to skip the range voting and go right to the quadratic voting, you can do that.

What common comments do you expect people to make?

You're also encouraged to leave comments on each post, if there's anything you haven't yet said about it in the Nomination or Review Phases. The comments can be marked private (if so, only LW team members will see them).

But there are a couple types of comments I thought might be common enough to warrant some kind of... standardizing. 

For the most part, I expect "how useful is this post?" and "how epistemically sound is this post" to blur together (for good or for ill – this may just be halo effect, but I think it's hard to avoid it). But I at least wanted people to have the opportunity to say "I rate this post overall a 2-out-of-3 stars, but it had particularly good, or bad, epistemics, compared to other posts I gave 2-stars." Or something like that.

I was thinking of including a few checkboxes for "common, default comments" that people could leave, so that certain types of feedback could be more easily aggregated.

If you have time, it'd be helpful if you looked over the posts on the /reviews page and think about what sort of short, simple comments you might have wanted to leave on them (either positive or negative ones), especially if you think others might want to leave similar comments.

(Ideally post each one as a top-level comment here, so that it can get voted on or discussed independently)

(This is somewhat similar to my previous thread discussing "Reacts", although I think it's worth asking the question separately here, where there's an immediate concrete task to use them for)


5 comments, sorted by Click to highlight new comments since: Today at 10:15 AM
New Comment

"This post makes claims that I'd like to see better justified"

(Phrasing could be dialed up or down in how 'accusatory' it is. I think I'd like the phrasing to work equally well for a post you think was bad, as for a post that you think was overall good but just didn't justify everything.)

An alternate version of this whole thing might just be to basically have an "epistemic upvote/downvote" which is separate from the overall score. I found myself having some posts where I wanted to be: 

"Okay, this post wasn't actually that important, but I was pretty impressed with it's rigor and/or thoughtfulness",

and other posts where I wanted to be like: 

"This post was quite important but also makes some iffy claims."

"I'd like this post's claims to be made more explicit." (There's a few different plausible wordings for this, such as "I'd like this post to draw a more explicit model")

"I appreciated this post's rigor"

Quadratic Voting is a very bad idea. Score Voting (aka Range Voting) or STAR Voting are better.


New to LessWrong?