TL;DR: I am looking for funding of around $1,000 to complete an effective altruism donations experiment with real money. I think this experiment has implications on how to best spread EA ideas. I am posting on the EA forum to get feedback on the idea and to discover potential funding opportunities.
As far as I'm aware, not much research has been done on the psychology of effective altruism. A lot of psychology literature focuses on the psychological factors behind charitable giving (see Small et al. - Sympathy and callousness: The impact of deliberative thought on donations to identifiable and statistical victims as an example). I am interested in understanding how exposure to effective altruist thinking impacts actual donations in a psychological experiment. This would be completed as my master's dissertation at the London School of Economics.
Research Implications
I believe the results of this experiment may have implications on how to do the most good. In Small et al.'s experiment among others, findings show that being made to think analytically about donations reduces donation amounts. If this result holds when considering charity cost effectiveness, this means that effective giving could backfire. If this is true, I think it is important to be aware of this when spreading effective altruist ideas, especially to non-typical effective altruist audiences.
Experimental Design
This is the rough idea for the experiment, however the practicalities (including if any moderating/mediating variables are included) are likely to change through fine tuning with my supervisor.
The experiment would follow a 2x2 design:
- The first factor would be effectiveness, each participant would receive either a cost ineffective charity or a cost effective charity, although effectiveness information would not be present.
- The second factor would be an EA intervention, half of the participants would read an effective altruist vignette discussing cost effectiveness, and the other half would receive no intervention.
| Effective | Ineffective | |
| No Intervention | "Any money that you donate will go to the Against Malaria Foundation. Around 625,000 people die each year from Malaria and 70% of them are children under 5. Your donation will be used to provide life saving bed nets." | An ineffective charity would be chosen which provides life saving medical treatment - TBC. |
| Intervention | EA vignette read initially "Any money that you donate will go to the Against Malaria Foundation. Around 625,000 people die each year from Malaria and 70% of them are children under 5. Your donation will be used to provide life saving bed nets." | EA vignette read initially An ineffective charity would be chosen which provides life saving medical treatment - TBC. |
Effective altruist vignette:
"The researchers would like to bring your attention to the effective altruism movement. Effective altruists use evidence to find the best ways of doing good, and try to do the most good that they can.
Effective altruists are particularly concerned with charitable donations, and claim that we should donate to cost effective charities which save the most lives per dollar. For example, providing malaria bed nets is one of the most effective causes, with an estimated cost of $4,500 to save one life. 'Cause here' is an example of an ineffective cause, with an estimated cost of 'cost here' to save one life."
Prediction
Two hypotheses would be tested:
- Hypothesis 1: Considering the cost effectiveness of giving should reduce giving to an ineffective charity
- Hypothesis 2: Considering the cost effectiveness of giving should have no effect on giving to an effective charity.
These predictions follow from psychological literature on the effect that analytical thinking has on donations (see Small et al.). Of course, it is quite possible that the experimental result differs from the hypotheses, which would nevertheless be useful information.
Procedure
150 US participants would be recruited from Prolific (preliminary power calculations predict that n=150 would be more than sufficient to detect the same effect size as found in Small et al.). Each participant would be paid $2 for participating in a short unrelated survey. Upon completion, participants will be given a $5 bonus - they will be told that researchers are raising funds for charity, and they will be given of the option of donating a proportion of their bonus to the charity in their experimental condition.
What I Need
Psychological experiments on donations work best with real money. I doubt I would observe an effect measuring donation intentions given the intention behaviour gap. I am therefore seeking funding of~$1,000 to complete this experiment with real money. To reiterate, I think the results of this experiment could have implications on how best to spread EA ideas, and therefore the small cost of this experiment is more than worth it on balance. As such, I am seeking recommendations for small grants to apply for. I am considering the EA infrastructure fund but I'm not quite sure that this fits.
I am also seeking thoughts and feedback on this idea. I'm an inexperienced researcher, so I'm quite aware that the structure of the experiment could be improved. I'm also interested to know if you disagree with my predictions or if you think the implications aren't as meaningful as I propose. I may also have missed some previous discussions on the same problem.
Finally, I am aware of the limitation that I'm only really testing one shot, one option donation decisions, and not donation reallocation. However, I propose that the experiment is still meaningful as non-typical EA audiences are less likely to spend time or effort researching charity effectiveness and reallocating donations. Exposure to EA ideas may therefore mean that:
- Present non-effective donations are reduced or cancelled without being reallocated to a more effective cause.
- Future ad-hoc giving opportunities are passed up on due to being 'cost ineffective', but money that would have been given is not allocated to a more effective cause.
Donation allocation could be tested by offering the choice between two charities in another experiment - however this may require more funding. I am, of course, also interested in hearing your thoughts on this.
See these papers on EA and effective charitable giving by Paul Bloom and Lucius Caviola.
You might find it tricky to fill these in. In general cost estimates for less effective charities are, when they exist at all, much less developed and much lower quality, because it's laborious to develop an accurate estimate and there's not much demand for precision once something is unlikely to be a top charity.
The nature of the effective altruist project is mostly to distinguish between "known to be effective" and "not known to be effective", and there's rarely any appetite for going the extra mile to "known not to be effective".
(I do expect there will be something you can use here, with at least a rough estimate, but don't expect the same rigour as you'll see with the effective charities.)
Some feedback on your vignette—I can imagine a confounding effect from feeling like 'a group of people wants to influence my decision-making' or something similar. A purer form of your experiment might just include the cost-effectiveness numbers for both charities.