Rethink Priorities' analysis of the 2019 EA survey concluded that 42% of EAs changed their cause area after joining the movement, 57% of change was away from global poverty, and 54% towards long term future / catastrophic and existential risk reduction.

Rethink Priorities, and Faunalytics also have much content on how to do effective animal advocacy, which would likely be useful for your purposes.

This is probably not the extent of research that Rethink Priorities has on this issue, but it's what I could remember reading about.

Do we have strong evidence that "average donors" even have "cause areas," as an accurate/descriptively useful mapping of how they understand the world? My young and pre-EA self feels so distant from me that it's barely worth mentioning, but I vaguely recall that teenage me donated to things as disparate as earthquake relief in Sichuan, local beggars, LGBT stuff and probably something something climate change. 

I don't think I ever consciously considered until several years later how dumb it was to a) donate to multiple things at the tiny amounts I was donating at the time and b) to have multiple cause areas of very varying cost-effectiveness and theories of change.

Thank you for posting this question. It has spurred me to consider taking some action; I am now interested in creating a survey on this topic, and then submitting it to my college.  

I am interested in advice on this idea in general (once each section has been read) and on each of the individual sections listed below (execution details, hanging variables, proposals for survey questions, narrowing / expanding the scope, redefining the objectives, etc...) .

Should you find that this has the potential to be effective or informative, I am interested in receiving help and discussing the content/execution of the survey. Should you think this is a waste of effort, or should you have any criticisms, please notify me, I would greatly appreciate the feedback. 

Objective and Measurements

The objectives  (in order) of this survey are to evaluate (1) which issues people believe are worth donating to, coming into the survey, (2) what people believe it would take for them to change what they believe is worth donating to, (3) how people evaluate EA affiliated organizations' cause-area selections, and (4) how people change their evaluations of which issues are worth donating to once they are exposed to EA affiliated organizations' measurements / reasoning of the importance of the cause-areas (e.g. scale, neglectedness, solvability ratings for a cause or metrics attached to a problem area) .

Questions by Objective

These are initial conceptualizations of questions  for each of the objectives listed above. Some questions depend on previously asked questions, so there are multiple sequences of questions that can be used. There are many other ways this survey can be created, so please do not think that I am set on these questions, or even on this particular ordering or set of objectives. 

(1) which issues people believe are worth donating to, coming into the survey

• What issues are worth donating money to?
• Please list the top five issues you think are worth donating money to.
• If you had money to donate, which issues would you donate to?
• Please list the top five issues you think are worth donating money to. Specify why each is important to you personally and to humanity as a whole.
• Please list the top five issues you think are worth donating money to. Specify why these are the top five issues humanity should focus on solving.
• (follow up): Please list some other issues that are maybe slightly less pressing, but are still important.
• You have $[quantity of funds] you must donate to some issue/s:  What causes would you donate the money to? How much would you donate to each cause?

(2) what people believe it would take for them to change what they believe is worth donating to

• What would it take for you to consider [iterate through EA affiliated organizations' cause-areas or a subset of causes areas, where the subset could differ between surveys] to be in your (top five) or (selection) of causes worth donating to?
• Why isn't [iterate through EA affiliated organizations' cause-areas or a subset of causes areas, where the subset could differ between surveys] in your list of causes worth donating to?

(3) how people evaluate EA affiliated organizations' cause-area selections

[iterate through EA affiliated organizations' cause-areas or a subset of causes areas, where the subset could differ between surveys] 

•  Which of these causes shouldn't be considered causes? Why? 
•  What would you change about this list? Why?
• Please select the elements of this list that should be removed, and add the ones that you think are missing. Please provide some reasoning behind your decisions.

(4) how people change their evaluations of which issues are worth donating to once they are exposed to EA affiliated organizations' measurements / reasoning of the importance of the cause-areas

x = [iterate through EA affiliated organizations' cause-areas or a subset of causes areas, where the subset could differ between surveys] 

y = [a quantitative negative or positive outcome]

• What if you knew working on [x] would/might result in [y]; does this change your cause-area selection?
• Quantitative statement about the pressingness of [x]; would you now consider donating to this? 
• Quantitative statement about the pressingness of [x];  do you think this is pressing enough to replace one of your original cause-areas to donate to?

Implementation

Given my circumstances, I am 75% confident that I will be able to submit a survey on this topic to at least two of the following departments [Economics, Neuroscience, Computer Science, Mathematics, Sociology, Anthropology, Philosophy] in my college. Additionally, I am somewhat confident that I will be able to convince EA groups at two other colleges in the area to host this survey in at least one department. Finally, there is a small chance that my non-EA affiliated friends will be able to host this survey and, should the survey be formulated well, a small chance that some members of this forum will take the survey or consider spreading it. 

Across all of these, I estimate that I can get at least 50 survey responses. 

Some Final Remarks

I can envision this survey beginning with a scenario where people divide 1,000,000 hypothetical dollars between cause-areas they personally find important. Then they'd choose how to reallocate funds in subsequent questions that pertain to different objectives (e.g. after reading about the reasoning EA affiliated organization use to select cause-areas, they'd have the option to reallocate money to another cause). A situation where the participant reallocates hypothetical money  to donate upon learning new information could be a short survey-game, similar to Explorable Explanations , but perhaps of a slightly shorter duration. 

Thank you for taking a look at this; I will wait for some feedback before taking more serious steps to conduct this survey (I will still be thinking about question phrasing, objectives, and implementation details). 

There are some studies suggesting people sometimes donate to less effective charities even when informed that other charities are more effective. E.g. this paper found that people prefer to donate to cancer research even when told that arthritis research is more effective. We made similar findings in this paper. 

These papers just ask one-off questions, though - they don't concern whether sustained persuasion would cause people to change cause area. But they do indicate that preferences for particular cause areas often override effectiveness information.

At Animal Advocacy Careers, we ran two concurrent studies of our one-to-one careers advice calls and our online course. We had a measure in there for cause prioritisation (a subcomponent of our overall "attitudes" metric), amongst many other outcome measures. Both interventions devoted at least some effort to encouraging (some) people to shift cause prioritisation.

We found evidence that the interventions each had significant effects on some outcomes (e.g. career plans, "career-related behaviours,"), but neither had significant effects on attitudes. In fact, there was some somewhat concerning evidence of a backfire effect on the cause prioritisation question, although this seems to be reduced to nothing in some of the sensitivity analyses.

So in short, we found that our intervention failed to persuade people to change cause areas despite being effective at some of the other things we tried. Ofc, this could be a reflection of our interventions. But it's at least weak evidence that persuading people to alter their cause prioritisation is difficult in general.

I've finished the write-up of this but am waiting on some additional feedback and we haven't published it yet. Feel free to email me ([email protected]) if reviewing the current draft would be helpful.