The numbers on how useful things are seem quite low to me. What did you write as the ends of the scale? I’m thinking in terms of net promoter scores where anything below a 9 or a 10 is considered neutral or bad.
Can you explain Hamming circles? I couldn’t find out how they worked even after a quick google.
Did you ask people if there was anything they wanted to do on the weekend but didn’t do? I’d be curious to see if people came up with anything.
Perhaps a little more discussion about actual causes e.g. what should be done about AI
In addition to the 1-1s(which were really useful), the could have been some more structured socialising (eg having people come together in groups of 3 and talk about what they care about, what their problems are, etc and switch groups in eg 5 minute intervals)
Maybe one minute intros could be on a post it on a wall so that people coming a day late still know who does what.
We also had some event ideas ourselves that we decided not to include in the weekend:
Community health session (a discussion of ways to improve the health of the EA community, be that through more diversity, more mutual support etc.)
Self-care for the altruistic (discussion and support session)
Discussions/anti-debate on cause prioritisation
Watching a video of a talk together and then discussing it (vote on a talk to watch during the event)
Values session: discussing why we have certain moral values
Problem-solving circle (participants raise problems they have and others try to help them)
The Humane League Work Party (Encouraging companies to commit to improving conditions on farms—writing letters & emails, phone calls, petitions. Would have been an option throughout the event.)
Gathering and interpreting data relating to the impact of charities for SoGive
2) I will ask the person who ran Hamming circles to explain it, I only participated in half of the event myself. You can get some idea from the slides (slides 21- 28).
1) >”The numbers on how useful things are seem quite low to me...”
On the scale 1 was “Useless” and 10 was “Life-transforming”. But just before asking for feedback, I made a change in the slides and added this meaning to the ratings of the events:
“3 - £100, 5 - £1,000, 8 - £10,000, 10 - £100,000 (e. g. career change)”
I explained it to people as well. This was… not smart. Because of this, some respondents gave low scores to all the events. E.g. someone said that the weekend was “Far more valuable (10-30x the counterfactual)” but did not gave any event a rating that is higher than 4. Others ignored the point and gave high ratings for all events.
That’s why I weighted and normalised the ratings. If someone said that the weekend was “Vastly more valuable (>30x counterfactual)”, I multiplied all their ratings by a constant so that their highest rating would be 10. If they rated the weekend as “Far more valuable (10-30x the counterfactual)”, I multiplied all their ratings so that the highest rating would be 9. 8 for “Much more valuable”, 7 for “Somewhat more valuable”, and 6 for “About as valuable”.
Yeah, I thought the ends of the scales might have been more extreme than we’d normally use. It’s probably quite hard to get people to sensibly answer unfamiliar, tricky questions.
Thanks for writing this up. Three questions
The numbers on how useful things are seem quite low to me. What did you write as the ends of the scale? I’m thinking in terms of net promoter scores where anything below a 9 or a 10 is considered neutral or bad.
Can you explain Hamming circles? I couldn’t find out how they worked even after a quick google.
Did you ask people if there was anything they wanted to do on the weekend but didn’t do? I’d be curious to see if people came up with anything.
3) Because of your question I added a link to the summary of the responses to the feedback form to the article. Here are responses I see that suggest different activities:
Perhaps a little more discussion about actual causes e.g. what should be done about AI
In addition to the 1-1s(which were really useful), the could have been some more structured socialising (eg having people come together in groups of 3 and talk about what they care about, what their problems are, etc and switch groups in eg 5 minute intervals)
Maybe one minute intros could be on a post it on a wall so that people coming a day late still know who does what.
We also had some event ideas ourselves that we decided not to include in the weekend:
Community health session (a discussion of ways to improve the health of the EA community, be that through more diversity, more mutual support etc.)
Self-care for the altruistic (discussion and support session)
Discussions/anti-debate on cause prioritisation
Watching a video of a talk together and then discussing it (vote on a talk to watch during the event)
Values session: discussing why we have certain moral values
Problem-solving circle (participants raise problems they have and others try to help them)
The Humane League Work Party (Encouraging companies to commit to improving conditions on farms—writing letters & emails, phone calls, petitions. Would have been an option throughout the event.)
Gathering and interpreting data relating to the impact of charities for SoGive
2) I will ask the person who ran Hamming circles to explain it, I only participated in half of the event myself. You can get some idea from the slides (slides 21- 28).
1) >”The numbers on how useful things are seem quite low to me...”
On the scale 1 was “Useless” and 10 was “Life-transforming”. But just before asking for feedback, I made a change in the slides and added this meaning to the ratings of the events:
I explained it to people as well. This was… not smart. Because of this, some respondents gave low scores to all the events. E.g. someone said that the weekend was “Far more valuable (10-30x the counterfactual)” but did not gave any event a rating that is higher than 4. Others ignored the point and gave high ratings for all events.
That’s why I weighted and normalised the ratings. If someone said that the weekend was “Vastly more valuable (>30x counterfactual)”, I multiplied all their ratings by a constant so that their highest rating would be 10. If they rated the weekend as “Far more valuable (10-30x the counterfactual)”, I multiplied all their ratings so that the highest rating would be 9. 8 for “Much more valuable”, 7 for “Somewhat more valuable”, and 6 for “About as valuable”.
Yeah, I thought the ends of the scales might have been more extreme than we’d normally use. It’s probably quite hard to get people to sensibly answer unfamiliar, tricky questions.