(Cross-posted on The Life You Can Save blog and the Intentional Insights blog)
“If others have half the experience I had today, they will be completely changed.” That’s what a participant told me after the end of a Giving Game I facilitated that aimed to engage people’s emotions in making a decision about where to donate.
Giving Games are workshop-style events aimed to advance philanthropic education. Participants learn about a couple of pre-selected charities, think about and discuss their personal values and reasons for giving, consider what methods and metrics they should use to select a charity, and then vote on what charity will get a real donation. The donation is sponsored by an outside party, typically The Life You Can Save, which donates $10 per participant to the charity that wins the vote.
I led a Giving Game for several core organizers of Columbus Rationality, a group that popularizes strategies on improving thinking, feeling, and behavior patterns to reach one’s goals. Columbus Rationality is a branch of the broader Humanist Community of Central Ohio, a community for skeptics and secular humanists in Columbus, OH with over 700 members on its meetup page. This Giving Game served as one aspect of a collaborative effort between Intentional Insights and The Life You Can Save to spread effective giving strategies to the skeptic and secular communities.
The large majority of Giving Games compare two types of charities. They compare charities working in developing countries that include The Life You Can Save and GiveWell-endorsed charities that are highly effective in helping save lives and advance human flourishing, such as the Against Malaria Foundation (AMF), to other charities that are less effective on this metric. The purpose of this type of Giving Game is to encourage people already interested in giving to charities working in developing countries to give to more effective charities. Another type of Giving Game compares a highly effective international charity to a local food bank or a similar local nonprofit. This type of Giving Game addresses the tendency of people to prefer to give to those they can see immediately around them, as opposed to those far away, known as the “drowning child problem.” These two types of Giving Games address important issues, and help people recognize the value of reflecting on the social impact of charities when considering their giving.
Yet discussing a local food bank does not arouse high emotions, and it is emotions that tend to drive our underlying decision making. Moreover, strong emotions cause the formation of much longer-lasting and easily-accessible memories. Thus, I decided to choose a much more emotional charity to have a stronger impact on the participants involved, especially considering the skeptic-oriented perspective of the audience.
Choices for Victims of Domestic Violence serves the Columbus, OH area. It provides a shelter for victims of domestic abuse, and a hotline they can call for help, or anyone else can call for help if they suspect anyone they know suffers from domestic abuse. This charity was aimed to address at once both the drowning child problem and to provide a strong and visceral emotional appeal. AMF served as the other charity.
As anticipated, the Giving Game itself was a powerful experience. Choices appealed to people’s emotional desire to help save victims of domestic violence. This was especially so since they could identify with victims of domestic violence, since this problem impacts the kind of middle-class people to which most participants in the Giving Game belonged – unlike a local food bank, which addresses the needs of the poor. Moreover, Choices provided the audience with a hotline they can use to call if they saw someone they know suffering from domestic abuse. It also helped improve the community around themselves. AMF, however, provided a much more impactful and cost-effective way of saving people’s lives and contributing to human flourishing.
It was an intense discussion, with much back-and-forth. People realized their own intuitive biases of valuing lives near to them higher than those far away, and valuing protecting victims of domestic abuse who they could identify with over victims of malaria. They had strong emotional experiences when they made these realizations, and weren’t sure whether to go with their heads in giving to the most impactful charity or with their cached patterns and intuitive beliefs.
I wasn’t sure until the end which way the vote would go. We even discussed the possibility of giving our own money to whatever charity ended up losing. We voted, and AMF did finally win out. After the vote, one of the participants told me that “I feel like I need to take a shower,” and expressed a strong desire to volunteer for Choices. I encouraged people to give to Choices if they feel it’s important as well.
Now, some of you might feel that this was a risky move, namely pitting a highly effective charity against an emotionally appealing one like Choices. Yet think about the impact! In this Giving Game, although AMF won, people made a commitment to Choices too. So if Choices won, it is very highly likely they would give to AMF as well. Yet more than this short-term result, people are likely to be much more powerfully impacted by a Giving Game that arouses their emotions. They would be strongly moved to explore effective giving if made to face their biases. So even if AMF lost, the cause of effective giving would have won, due to the tapping of strong emotions.
Based on my experience, I encourage you to inject emotional appeal into your Giving Games, whether targeting skeptics or just a broad audience, as a means of spreading effective giving ideas effectively.
P.S. This article is part of the EA Marketing Resource Bank project lead by Intentional Insights and the Local Effective Altruism Network, with support from The Life You Can Save.
Thanks Gleb. As you know I'm interested in this question:
A few different arguments seem to be in play here so I wonder if you can clarify your rationale for me?
It seems like the claim here is that both charities gain some support. Maybe that's so. But I still wonder if AMF (and effective giving in general) would win more support if pitched against a charity with less visceral emotional appeal (the kind people might feel compelled not to betray)?
There are a few different possible arguments I can imagine you might be making here, but I'm not sure which ones you are making.
I can see that people might be more affected by a GG that arouses their emotions, but why is this necessarily a good thing? Presenting people with a heartwrenching choice may affect them, but it's not clear to me why it would make them more pro one of the choices (or one particular way of making choices).
Perhaps your thought is that arousing people's emotions will make them think more about charity choices in general, and this is a desirable outcome. But I'm not sure why we should suppose that the result of their thought will come down on the right side. If presented with a really emotionally compelling ineffective charity, why might they not simply become resolved against effective giving (or against comparing charities in general?).
Relatedly, I think I can see why we might think it's good to get people to "face their biases" if the result is that at the end they overcome their biases. But why assume that they will? Why might they not simply feel the pull of the emotionally engaging charity and just go with it, deciding that absolutely the right thing to do in general is to give to the local charity where they know the people and have built connections and seen the person in need first hand and definitely not to impartially decide that they should send the money to Africa?
Moving away from the abstract principles into the realm of concrete examples this just seems straightforward. It's easy to imagine cases where if you want to persuade people of the benefits of X versus Y, you'd do better to set up X in contrast to a very poor unengaging Y, and that if you contrast X with a highly emotionally appealing Y, X will lose out.
I don't think that it's always worse to contrast X with a highly emotionally engaging, and appealing, alternative, but it still seems somewhat counterintuitive to me, so I'd be interested if you can explain further.
David, let me make sure I see what seems counterintuitive to you.
You are wondering whether setting up an effective charity X against a not effective not-emotionally-engaging Y might be better to persuade people to give to X, rather than the scenario I did, of an effective charity X against a not effective but highly emotionally engaging Z, which might result in people giving to Z instead.
I hope I understood your point correctly.
The reason I think X vs. Z is optimal, rather than X vs. Y, is that in the real world, people generally don't make a trade-off between X and Y. In any situation where people are giving significant money to a charity, they are by definition already having a significant emotional bond with that charity. In other words, in the real world, it will always be a trade-off between X and Z.
Thus, by setting up X vs. Y, we would, at best, influence some non-committed people to give to X, and at worst, we would get no counterfactual giving to X at all.
In order to get people to actually change their minds, they need to have an emotionally engaging experience, where they get to truly face their biases, and avoid the opportunity to flinch away. Truly shifting people's giving, those who are already committed donors, is quite hard, and doing so in a workshop setting requires engaging their emotions well.
Now, let's get concrete and specific. Having led this Giving Game I described above, there was no one who was not moved to give more to both charities. Heck, I was moved to give to Choices myself emotionally, but chose not to due to EA reasons. So having experienced this on the ground level, I have a high probability estimate that other GG will go in a similar fashion.
Is it possible that some people might become more committed to Z? I can't say it's not impossible, just quite improbable. Moreover, we have to remember that it's not like X and Z had an equal estimate in people's minds beforehand. People are already predisposed to favor Z. In this case, there were people in the room who knew and in fact already donated to Choices for Domestic Violence, and the donor actually chose to vote for AMF instead of Choices.
Of course, Intentional Insights will try more of these, and see how they go :-)
An additional point particular of relevance to a skeptic/secular audience. Skeptics are much more likely to be, well, skeptical of an X vs. Y comparison, they will be much more likely than, say, students to see Y as a strawman. We don't want them to have an experience that turns them off from the whole effective giving project.
Moreover, an emotionally engaging experience will be remembered much better down the road, rather than a dry educational one. There's quite a lot of research on this, and so it would be way more effective to have an emotionally engaging GG.
I think the biggest issue is that you shouldn't pit an effective charity against (only) a strawman. There should be at least one other choice that's compelling for emotional/effectiveness/etc based reasons. It's fine to include a strawman in a 3+ charity GG and could lead to some useful outcomes. For instance, if you had college students choose between AMF/not EA but still compelling charity/their school, they'd probably see their school as a strawman. So this setup can get people who might counterfactually give to their school without thinking to instead think: "there are reasons to give to my school, but my charitable dollars can do way more good elsewhere." Not only do people explicitly reject a default giving option, they also do so based on comparative impact which is how we want them to make future giving decisions.
My hunch is that the real value of an emotionally compelling charity is in the intensity of the experience it can create. This provides the opportunity to make really valuable "post-game asks" like getting them to sign up for the newsletter of an effective charity or charity evaluator.
That's pretty much it.
I don't entirely agree. I think many people give to [generic charity] because they want to do some good and this charity seems to do a good job but without any particular knowledge/interest/affiliation about/in/for the charity. Perhaps the disagreement about this is somewhat merely a verbal disagreement about "significant money" and "significant emotional bond."
But I do think that there is clearly a difference between just any old charity which a person finds appealing and has a bond with and donates a significant amount of money to and a charity which is maximally elicits deontological trumping responses. I think domestic violence shelters are pretty close to the latter and more than most charities it would feel like a taboo to oppose or, indeed, taboo to apply instrumental, rational (i.e. cold, calculating) cost-benefit analysis to at all.
Consider the following comparison: 'Your significant other/child will suffer a painful [gruesomely described] death unless you pay $1000, or you could use the $1000 to save X-many distant others from equivalently bad deaths. Which do you choose?' ^That seems emotionally evocative and challenging people's biases, but unlikely to encourage people to favour effective giving.
It seems plausible to me that people would "face their biases" presented with a more moderately appealing charity, whereas when presented with a maximally attractive charity they may be especially likely to "flinch away" and simply refuse to bite the bullet that helping many distant impoverished people is worth taking money away from female victims of domestic abuse.
Some more money going to an effective charity is a good thing, but you've also talked about the benefits of getting people to accept 'effective giving' as a process. If people just decide to give more to both charities because 'both are good' plausibly they haven't really adopted effective giving.
No doubt more emotionally engaging things are remembered more, all else being equal. I'm not aware of any research which would settle the question of whether more emotionally engaging experiences (by opposing an effective charity to a particularly emotionally compelling ineffective one) would be good. If people are emotionally engaged by how awful it is to be asked to consider whether domestic abuse victims deserve money (asked only implicitly of c.) and remember the experience deeply, this may be no good thing.
It sounds like we are on the same page about the benefits of an emotionally engaging experience for the goal of moving people and having them remember the experience in the long term. So the only point of disagreement is whether the experience of the GG itself is worthwhile.
I see your point. I think this is a matter where we need to experiment and learn. I shared the experience of my experiment, which pretty clearly moved people strongly and causes them to be quite engaged with AMF. Intentional Insights will do other experiments and see what happens in future cases of emotionally intense charity comparisons.
My intuition is that we will have positive outcomes. Jon Behar's comments here suggests he shares that intuitive sense. My guesstimate is that emotionally engaging experiences will be most powerful for changing long-term giving, and I am glad that Jon is studying this question.
However, I am happy to update - the goal is to get people to give effectively in the long run, after all :-) So if future experiments go differently, then we will change course.
If people give to an effective charity as well as the emotionally engaging charity, they by definition implicitly accepted the value of effective giving to some extent - they have crossed some of the inference gap. So giving to both I perceive as a highly beneficial outcome, since this results in them shifting their giving to give at least something to effective charities. These are the kind of slow steps and behavior changes that will lead to big shifts of giving in the future.
That’s an interesting move. I’m looking forward to hearing more reports on successes and failures of different formats of giving games now that TLYCS is focusing on them so much.
How did you solve the problem of the information dissemination? In our case we had DMI compete with IPA, and the discussion became very moderation heavy because those of us who had read all the material on the charities, the organizers, had to dispense facts for most of the time, which left some but not much time for discussion. (DMI won, even though it was after the latest GiveWell update.)
Quick update on where things stand w/ our ability to analyze the relative efficacy of different sorts of Giving Games…
I'm working on a spreadsheet that will basically take all the data we collect about each game and put it into a pivot table. Then users will be able to look at whatever cuts of data they want: all data for games involving AMF, all data for games including a local option, games w/ 2 charities vs. games with 3, games in US vs. games in UK, etc. I just completed a very preliminary prototype, but there's still a bunch of work to do to get it operating smoothly and back populate some of our old data. Eventually I'd like to make this a shared resource that GG facilitators and researchers can access.
Awesome! I’m looking forward to reading/faceting that.
Hi Jon Behar: I read that there might be an academic paper looking at the impact of giving games in the works, could you confirm this? Would be interested in reading if so.
We've run a lab experiment where we tested the effects of giving people money to donate to their choice of two charities featured in a TED talk all subjects watched. It's described here, my collaborator is working on writing it up for publication. Takeaway is that the "Giving Game Treatment" saw huge (2-4x) increases in amount of their own money people donated, % of people who signed up for at least one charity newsletter, and total # of subscriptions relative to the control group.
Next thing we want to test is whether we can create lasting changes in attitudes and behaviors. Not sure yet whether that will be a formal (intended for academic publication) study or not.
Btw, while not directly related to the efficacy of Giving Games, there are also a couple of academic studies likely to be published that have used the Giving Game model to test hypotheses by varying the information provided to different treatments. Some U Chicago researchers did this to extend their research on the "other minds problem" and I'm working with someone at U of Hamburg to expand the literature on how "social information" affects giving.
On the meta-issue, I agree, it will be interesting to see different results. Charity Science recently ran a skeptic-oriented Giving Game, they will write it up soon.
For the information dissemination, we started with a brief presentation on the two charities, for AMF using TLYCS' materials and for the local charity based on research done by a couple of members. That way, everyone had about the same amount of material, and we didn't focus too much time on the details. We instead focused on the higher-level reasons for giving and priorities - things such as "do you give locally but less impactfully" vs. "do you give globally but more impactfully." That worked out well for us, and really got the participants well engaged with emotion-laden topics.