Thanks for asking this question, and I look forward to answering it and diving in. I do agree that it’s important to get that nuance.
However, there is one problem I must first correct:
with me having added percentage of survey respondents who favored each cause as a percentage of the total number of respondents who answered the question
You can’t add the percentages like you did, because people were allowed to indicate support for more than one cause. So the 579 people who supported poverty and the 422 people who supported metacharity are not 1001 distinct people. (In fact, they were 683 distinct people.)
I don’t remember the exact phrasing of the question, but because of the flexibility you note, I would find it easy to answer something like “In what cause areas have you supported or are you planning to support interventions this year?” than “In what cause areas do you support interventions?”
The overlap between the answers is not a great problem for me as respondent, since I can select several, but some of the options will become less meaningful if some people perceive the hypernym relationships differently. Maybe these answers can be instrumented via monetary donations or direct work for charities in the areas. (I hope I’m using that word right. ^^) That strategy will probably become more promising once Open Phil has put out concrete recommendations.
I don’t remember the exact phrasing of the question, but because of the flexibility you note, I would find it easy to answer something like “In what cause areas have you supported or are you planning to support interventions this year?” than “In what cause areas do you support interventions?”
The precise wording was “Which of the following causes do you think you should devote resources to?”
[Important meta point]
Thanks for asking this question, and I look forward to answering it and diving in. I do agree that it’s important to get that nuance.
However, there is one problem I must first correct:
You can’t add the percentages like you did, because people were allowed to indicate support for more than one cause. So the 579 people who supported poverty and the 422 people who supported metacharity are not 1001 distinct people. (In fact, they were 683 distinct people.)
I missed that. Thanks. I’ll fix it.
If there were 813 EAs in total, that also puts those who prioritize animal rights at 36.4%. Here’s the breakdown:
Poverty: 71.22%; Metacharity: 51.91%; Rationality: 50.55%; Prioritization: 42.44%; AI Risk: 40.84%; Environmentalism: 38.99%; X-Risk: 37.02%; Animals: 36.41%; Politics: 35.79%; Far Future: 28.66%;
I don’t remember the exact phrasing of the question, but because of the flexibility you note, I would find it easy to answer something like “In what cause areas have you supported or are you planning to support interventions this year?” than “In what cause areas do you support interventions?”
The overlap between the answers is not a great problem for me as respondent, since I can select several, but some of the options will become less meaningful if some people perceive the hypernym relationships differently. Maybe these answers can be instrumented via monetary donations or direct work for charities in the areas. (I hope I’m using that word right. ^^) That strategy will probably become more promising once Open Phil has put out concrete recommendations.
The precise wording was “Which of the following causes do you think you should devote resources to?”