Seeking feedback on Cause Prioritization Platform

post by adamkruger · 2019-08-12T22:06:10.320Z · score: 6 (6 votes) · EA · GW · 4 comments

This is a link post for https://listofproblems.bubbleapps.io/

Hey there,

Thanks for opening this post - glad it caught your eye. :)

I created a Cause Prioritization website with the purpose of creating a space for all EAs (and people in general) to share their views on which problems/causes should be top priorities.

https://listofproblems.bubbleapps.io/

The goal is that the platform would be able to help us better prioritize our efforts, and would serve as a forum similar to this one, but focused specifically on cause prioritization.

The website is still in its very early stages, so I'd love to hear any ideas, suggestions, or feedback you have for improving it.

Thanks again!

Adam


4 comments

Comments sorted by top scores.

comment by saulius · 2019-08-13T17:02:19.092Z · score: 27 (10 votes) · EA · GW

Some quick thoughts:

  1. There is no way to say that I think the problem is not big/neglected/solvable. Hence ratings provide little information. It could be that only 1 in 10 people think that a problem is solvable, but if enough people think about voting on the cause, it will have a higher solvability rating than a problem that is in a less prominent place within the website. And if you sort problems by "most attention required", then problems that received most votes initially will continue receiving more votes. Instead of a checkbox, I would put a slider so we could vote how solvable a problem is.

  2. People who know very little about problems can vote, and their votes will weight the same as people who know a lot about arguments for and against the issue. If someone mentions the website in some animal discussion group, you might get many people voting favourably for animal causes and skew the results. Hence, as it is, it's currently more of a popularity contest rather than something that could be helpful prioritise. And the EA survey might be better for gauging popularity of various causes.

  3. Scale-neglectedness-solvability framework needs to be applied with care [EA · GW] and not all the readers will do that. Maybe a different phrasing of questions would help. E.g. instead of "think this is a solvable problem" you could have something like "if we doubled the resources dedicated to solving this problem, what fraction of the problem would we expect to solve?" (taken from here). Also, it's unclear whether this is the best framework for prioritizing causes, it's been criticised. And we already have ratings for scale-neglectedness-tractability at the 80,000 hours website, I'm not sure this will give more accurate ratings.

I like how the website looks, how it's organized, and how fast it is though. I think that it could be a good idea to have a website for cause prioritization discussions. Although I think there would be a big chance that it wouldn't become popular, even if you do it well. One way to make such website that I imagine could maybe be useful is if you could post various arguments for and against a cause area, and other people could upvote/downvote/comment on these arguments, or propose different ways to phrase them. I remember that a couple of years ago some Scandinavian EA wanted to make a platform that was similar to this but I can't find it now.

Also, in case you didn't know, there are two wikis about cause prioritisation: http://priority.wiki/ and https://causeprioritization.org/ Maybe information in them could be used in some way.

comment by adamkruger · 2019-09-02T18:43:04.092Z · score: 1 (1 votes) · EA · GW

Thanks so much for your thoughtful feedback - I really appreciate it.

A few responses:

  • Don't you think there will be at least some people who view certain problems as not big, neglected, and/or solvable? I was thinking it'd be interesting to see which problems received the most "votes" across each of these 3 dimensions
  • I definitely agree with your concern that problems that receive the most votes initially may continue to receive more votes. Would your suggestion for addressing this be to remove the ability to sort by "most attention required," or do you have any other ideas?
  • I thought about doing a slider but felt a checkbox required less thinking/effort, and I thought that on aggregate there was still value in seeing which problems were "checked" off the most
  • To your point about the weighting of votes based on people's backgrounds, doesn't this come down to the pros and cons of a democratic voting system?
  • I completely agree with you that the phrasing of the statements needs to be used with care. My thinking though was that a statement like "think this is a solvable problem" would be easier to understand and therefore more likely to get a response than a statement like "if we doubled the resources dedicated to solving this problem, what fraction of the problem would we expect to solve?". Thoughts?
  • If you happen to find the Scandinavian EA, would you mind connecting me with them please?
  • Finally, thanks for those two wiki links - they were very useful!

Adam

comment by Denis Drescher (Telofy) · 2019-08-14T16:03:45.150Z · score: 3 (2 votes) · EA · GW

Cool idea!

A few things I just noticed:

  1. It would be nice if the reordering happened only on reload and not immediately. I sometimes ticked an “importance” box fully intent to also tick another one, but then the block jumped away, and I’d have to go looking for it again.
  2. I second Saulius’s feedback: I also feel like it’d be easier for me to rate things on a three- or five-point scale or even with a slider. As it is, I’m tempted to tick almost all the boxes on all problems.
  3. I really like your initial selection. I had a hard time voting on animal welfare because my assessments of farmed and wild animal welfare are so different. Then I noticed that you had anticipated my problem and added the subsections!
  4. The term tractable seems more intuitive to me than solvable, since the second can be misunderstood as whether the problem can be fully solved as opposed to the definition that Saulius cites.
  5. I’d like to make a distinction (and maybe a granular one) between whether I think a problem is not important/neglected/tractable or whether I have no opinion on it. E.g., I currently don’t know whether the Pain Gap is tractably addressable, but not ticking the box feels as if I were saying that it’s not tractable.
  6. Not showing other people’s ratings could help avoid anchoring people.
comment by adamkruger · 2019-09-02T18:57:22.213Z · score: 1 (1 votes) · EA · GW

Thanks so much for your feedback Denis!

Here are my responses:

1. I definitely agree - I tried fixing this but had some difficulty. It may be doable, I'll have to try again haha. Good news though is that if there are more people using the platforms, there will likely be more of a delta in scoring between the different problem areas, therefore blocks will jump/reorder less often.

2. Similar to my reply to Saulius above on this, do you think a three or five point scale would be viewed as more effort / thinking required for some people? Maybe it's actually a good thing if it requires more thinking, but just wanted to get your perspective on this

3. Thanks! The goal is to have sub-sections, sub-sub sections, etc. That's what I find most excited about this concept - the ability to break down these problems further and further to identify where we should focus our efforts within each umbrella :)

4. Interesting, for me it was the opposite haha. Before learning about EA, I wasn't really familiar with the term tractable. I'll definitely change this if I get more feedback like yours.

5. I completely understand your perspective on this. I thought quite a bit about the construct for this, and wanted to keep it as simple as possible - which meant limited the number of options available to select. My thought was that if someone is on the fence about whether or not to select a box, and some of those people end up choosing not to select it, that's probably an indication that it's a less important issue than one that seems unequivocally large, neglected or solvable. Is that fair?

6. What if I only showed other people's ratings after your voted yourself? Would that help?

Thanks again!