I don’t think it’s clear if we should insist on anything in particular, but I don’t see it as a no-brainer.
“Getting prominent politicians and scientists like Stephen Hawking or Bill Gates to affiliate...”
People respect and are impressed by those who are making big sacrifices for the benefit of others. Note how much attention and respect from important figures Toby Ord has gotten by pledging a large fraction of his income. Also the kudos generally given to doctors, soldiers, firefighters, Mother Teresa, etc.
“But there’s a cap—once you give away about half or three-quarters of your funds, you will run out. Whereas how effectively or cleverly we can donate has no obvious upper bound. If we gather greater insights, we can always start newer and better projects.”
We might be able to get people to give 50 times more than they do now (an average of 1% to 50% of income). Do you think we can persuade many people, who wouldn’t be motivated to give more, to give to a charity that is, ex ante, 50x better than they do now on average (keeping in mind the mean of a log-normal distribution is already much higher than the median due to the right tail)?
“As any philosophical movement gains widespread support, its idea gets watered down.”
This seems like an argument in favour of very high expectations to start with, knowing it will be diluted later anyway on as more people get involved.
“The whole idea of extreme self-sacrifice has pretty mixed effects”
A standard doesn’t have to and shouldn’t embody extreme self-sacrifice, it can just ask for something like 10%, which is not extreme—indeed it used to be the norm.
“Which is too bad, because the theology of liberal Protestantism is pretty admirable. Openness to the validity of other traditions, respect for doubters and for skeptical thinkers, acceptance of the findings of science, pro-environmentalism – if I had to pick a church off a sheet of paper, I’d choose a liberal denomination like the United Church of Christ or the Episcopalians any day. But their openness and refusal to be exclusive – to demand standards for belonging – is also their downfall. By agreeing not to erect any high threshold for belonging, the liberal Protestant churches make their boundaries so porous that everything of substance leaks out, mingling with the secular culture around them.
So what if liberal Protestants kept their open-minded, tolerant theology, but started being strict about it – kicking people out for not showing up, or for not volunteering enough? Liberals have historically been wary of authority and its abuses, and so are hesitant about being strict. But strictness matters, if for no other reason because conservatives are so good at it: most of the strict, costly requirements for belonging to Christian churches in American today have to do with believing theologies that contradict science, or see non-Christians as damned. What if liberal Protestantism flexed its muscle, stood up straight, and demanded its own standards of commitment – to service of God and other people, to the dignity of women, and to radical environmental protection? Parishioners would have to make real sacrifices in these areas, or they’d risk exclusion. They couldn’t just talk the talk. By being strict about the important things, could liberal Protestant churches make their followers walk the walk of their faith – and save their denominations in the process?”
Do you think we can persuade many people, who wouldn’t be motivated to give more, to give to a charity that is, ex ante, 50x better than they do now on average (keeping in mind the mean of a log-normal distribution is already quite high due to the right tail)?
This is a backwards interpretation of the dynamics of log-normal distributions.
The (rough) equivalent operation of moving everyone’s donations from 1% to 50% would be moving everyone’s donations from the (dollar-weighted) mean charity to the best charity. Although (as you noted) the heavier tail of a log-normal distribution means that the sample mean is higher relative to the mode or median, it has an even stronger effect on the sample maximum.
This means that overall, a lognormal has a higher, not lower, maximum:mean ratio for a fixed, say, median and standard deviation, compared to a thinner-tailed distribution like the normal. For instance, in numerical simulations I just ran with 100 samples from a log-normal and normal distribution, both with median 2 and variance approximately 4.6, the average ratio of sample maximum to sample mean was 5.5 for the log-normal and 3.7 for the normal.
Yes, but ex ante. The higher up the distribution the harder they will get to identify because ‘if it’s transformative it’s not measurable; if it’s measurable it’s not transformative’. The weakness of the measurements mean you’re going to be hit with a lot of regression to the mean.
Also that stuff is likely to be weird (must be extreme on neglectedness if it’s so important and still useful to put more money in to), and so just as it’s hard to get someone to give enormous amounts, it will probably also be hard to move donations there.
I’m not talking about in-practice difficulties like convincing people to donate. I’m just talking about statistics.
Can you point to actual parameters for a toy model where changing the distribution from normal to log-normal (holding median and variance constant) decreases the benefits you can get from convincing people to switch charities? The model can include things like regression to the mean and weirdness penalties. My intuition is that the parameter space of such models is very small, if it exists at all.
If we thought that the charity they were switching to were only at the 95th percentile, it could be worse in a log-normal case than a normal case (indeed it could be worse than not getting them to switch).
However that would be an unusual belief for us. More reasonably we might think it were uniformly drawn from the top ten percent (or something). And then log-normal is again much better than normal. I agree with the thrust of your intuition.
Yeah, in GiveWell classic, you’re generally going to estimate that a high-impact charity is on the 95th percentile but with uncertainty around that. Which is in-between the cases that you describe.
I can imagine that knowing that something is on the 95th percentile with high certainty might be worse than guessing that something is on the 95th percentile with high uncertainty, if you have a prior that is some mixture of log-normal, normal and power-law. That’s what we’d have to show to really question the classic GiveWell model.
I’m sure you’re right about the math, but I am concerned with the in-practice difficulties.
My point about the mean was simply that one shouldn’t compare the max-median on the log-normal—which would be natural if you visualise where ‘typical donations’ that you see go—but rather the max-mean, which is a less extreme ratio. I wasn’t drawing a contrast with a normal or any other distribution.
I’m also not sure about the answer to my question to Ryan—maybe the effectiveness is still the better focus, but I’m prompting him with possible counterarguments.
Ah, from your first comment it sounded like you were comparing the mean of the log-normal to the mean of a less-skewed distribution, rather than to the median of the log-normal. That sounds more reasonable.
In practice, isn’t it relatively easy to identify whether someone is already up the right tail in their giving? I don’t recall struggling with this in initial conversations. You can just ask whether they give abroad for instance, which will seemingly get you most of the way since most people don’t (i.e. it is true that those who do vastly skew the mean, but you can just exclude almost all of them ex ante...).
What you’re saying might be appropriate for mass marketing I suppose where you can’t cut off the right tail.
Sorry, pretty unclear post on my part. Owen basically got it right though; if we’re talking practically rather than theoretically, you don’t have to decide to always focus on effectiveness or always focus on quantity. You can choose, and your choice can be influenced by the information you have about your audience.
Since most individual people are around the median/mode and a long way below the mean, for most individual people talking about effectiveness is correct. There are a few exceptions to this (those ‘up the right tail’), and then you can talk about amount...or just accept that you aren’t going to achieve that much here and find more people where you can talk about effectiveness!
This is obviously dependent on how much ability you have to discriminate based on your audience, which in turn depends on context, hence my ‘mass marketing’ point.
I think it’s the reverse—if you exclude the people who already do give effectively then you’ve brought the mean of those who remain down closer to the vicinity of the median.
Another speculative argument in favor of big asks: most charity seems to focus on making small asks, because people think getting people on the first step towards making a difference is the crucial bottleneck (e.g. see this guy), so the space of ‘making big asks’ is neglected. This means it’s unusually effective to work in this space, even if you appeal to many fewer people.
This seems to be one of the main reasons GWWC has been much more effective than ordinary fundraising techniques.
The downside is that we’re concerned with total scale as well as cost-effectiveness, and a ‘big ask’ approach probably has less total growth potential in the long-run.
The GWWC pledge isn’t really an ‘ask’ - people may make particular donations because they’re asked to, but no one commits to donating 10% of their income every year until they retire because someone asked them to. Instead they make this commitment because they want to do it anyway, and the pledge provides a way for them to declare this publicly to influence others. So it would be interesting to find examples of more typical big asks working—eg. fundraising teams which highball potential donors. Does anyone know of these?
This would, if true, imply that GWWC does not actually result in any initial donations on the behalf of its members. It might still result in more donations on behalf of non-members.
Ah sure, but I’m saying that no one gives this sort of money just because they’ve been asked to—it’s too large and long-lasting a commitment, and being asked is not a powerful enough reason or prompt. Asking them to sign the GWWC pledge may prompt them to make this public declaration, but only if they were already happy to give that sort of money.
In my experience, some of the people I’ve asked to take the pledge would have donated 10% “eventually”, but the pledge actually made them follow through for at least that year in particular where I’m confident they otherwise would not.
Some would have done so anyway, but I think the example set by hundreds of others, including some they know personally, normalises giving a large amount and makes them more likely to copy. Also making a public declaration makes people more likely to follow through.
We also obviously make arguments in favour of doing so. It rarely convinces people immediately but it contributes to moving them in that direction.
“People respect and are impressed by those who are making big sacrifices for the benefit of others. Note how much attention and respect from important figures Toby Ord has gotten by pledging a large fraction of his income. Also the kudos generally given to doctors, soldiers, firefighters, Mother Teresa, etc.”
Attention and admiration, absolutely. But how much copying? I would expect people to be most drawn to people who seem to be only ‘one step’ more self-sacrificing than they are, in a conceptual framework that I suppose is analogous to the idea of inferential distance. For instance, I think Toby benefits in this from having an ‘ordinary’ income. Anything further away than that rapidly becomes too weird to be taking seriously. Note Jeff Kaufman has written about pretty much exactly this:
One weak support for the idea that these examples are sufficiently far away for most people to dismiss without much thought is the very fact media happily talks about them; this would make them less challenging and cognitive-dissonance inducing, so others’ dominant impressions on reading are to be impressed rather than unsettled.
The fact that each person is less impacted has to be weighed against the fact that about 100,000-1,000,000 times as many people have heard about what Toby is doing than someone who just privately gave a few percent of income.
Of course we need examples of more typical people to refer to as well.
No way, the main media interest was in the extremeness of the amount. The press wouldn’t care about or cover Giving What We Can if the ask was 1% and Toby was giving 1%.
Most of GWWC’s exposure is through word of mouth. This is even truer for the exposure that matters, and that leads to people signing up, much of which is through meeting Giving What We Can members.
The idea that 99.99% of the exposure of Giving What We Can would have disappeared if they couldn’t focus on Toby’s generosity could only result from very unclear thinking. The press would’ve just looked different. Trivially, other members would have given a large fraction even if Toby hadn’t told them to, and press could’ve focussed on that. The example is The Life You Can Save. Granted, it had Peter Singer, but I hope that we’re not going to have the discussion that without Peter, they could’ve still aroused at least 0.1% of the same press. I mean, EA Melbourne has been able to give talks, go on community radio et cetera without citing the 10% figure...
Like, I know things like press can have weird power law distributions but we’re really short-changing the exposure of the rest of GWWC’s message—the parts that are about quality of donation—if we say that 100,000 of its exposure is just because of the quality.
So although the benefits for exposure of of giving 10% are presumably there, and they might matter a lot, they’re like 2-10x.
“The idea that 99.99% of the exposure of Giving What We Can would have disappeared if they couldn’t focus on Toby’s generosity could only result from very unclear thinking.”
Yes, but this is what you said: “Although Toby’s story is only about 2-10x more widely known compared to if he gave 1% and still founded GWWC.”
You may think GWWC would have similar numbers of members today because of how it grows person-to-person, but his story would be much less known because there would basically be no story to report on. Someone gives 1% of charity and encourages others to do the same? It’s not newsworthy. As it was, it got to the most read stories on BBC News and other similar outlets and literally millions (maybe tens of millions) of people heard about him.
Furthermore, I think that media exposure was necessary to turn Giving What We Can from just a group of friends in Oxford into the going concern it is today. That story isn’t as valuable today, but it was the main asset we had in the early days.
Haha Rob it’s Christmas, can’t we stop fighting? Because I’m right, and you should convert to my point of view. :|
But seriously, you’re saying that if Toby had given 1% instead of 50%, then rather than 10 million people knowing his story, rather than 10 million people knowing about him, only 10-100 people would? That simply not reasonable. Without press, even if each of the handful of academics who signed up for GWWC had mentioned it in their opening lectures that semester, you would already have a thousand people who had heard the story.
Haha, let’s split the difference. Maybe 100,000 people would have heard of Toby, so 1-2 orders of magnitude? I think that’s enough for me to make my point that it could be better overall, even if each person takes the example less seriously. :P
I don’t think it’s clear if we should insist on anything in particular, but I don’t see it as a no-brainer.
“Getting prominent politicians and scientists like Stephen Hawking or Bill Gates to affiliate...”
People respect and are impressed by those who are making big sacrifices for the benefit of others. Note how much attention and respect from important figures Toby Ord has gotten by pledging a large fraction of his income. Also the kudos generally given to doctors, soldiers, firefighters, Mother Teresa, etc.
“But there’s a cap—once you give away about half or three-quarters of your funds, you will run out. Whereas how effectively or cleverly we can donate has no obvious upper bound. If we gather greater insights, we can always start newer and better projects.”
We might be able to get people to give 50 times more than they do now (an average of 1% to 50% of income). Do you think we can persuade many people, who wouldn’t be motivated to give more, to give to a charity that is, ex ante, 50x better than they do now on average (keeping in mind the mean of a log-normal distribution is already much higher than the median due to the right tail)?
“As any philosophical movement gains widespread support, its idea gets watered down.”
This seems like an argument in favour of very high expectations to start with, knowing it will be diluted later anyway on as more people get involved.
“The whole idea of extreme self-sacrifice has pretty mixed effects”
A standard doesn’t have to and shouldn’t embody extreme self-sacrifice, it can just ask for something like 10%, which is not extreme—indeed it used to be the norm.
Strong rules can make for stronger communities: http://slatestarcodex.com/2014/12/24/there-are-rules-here/
Also note the empirical regularity that churches that place high demands on their members tend to last longer and have more internal cooperation (e.g. http://ccr.sagepub.com/content/37/2/211.abstract).
Quote relating to this:
“Which is too bad, because the theology of liberal Protestantism is pretty admirable. Openness to the validity of other traditions, respect for doubters and for skeptical thinkers, acceptance of the findings of science, pro-environmentalism – if I had to pick a church off a sheet of paper, I’d choose a liberal denomination like the United Church of Christ or the Episcopalians any day. But their openness and refusal to be exclusive – to demand standards for belonging – is also their downfall. By agreeing not to erect any high threshold for belonging, the liberal Protestant churches make their boundaries so porous that everything of substance leaks out, mingling with the secular culture around them.
So what if liberal Protestants kept their open-minded, tolerant theology, but started being strict about it – kicking people out for not showing up, or for not volunteering enough? Liberals have historically been wary of authority and its abuses, and so are hesitant about being strict. But strictness matters, if for no other reason because conservatives are so good at it: most of the strict, costly requirements for belonging to Christian churches in American today have to do with believing theologies that contradict science, or see non-Christians as damned. What if liberal Protestantism flexed its muscle, stood up straight, and demanded its own standards of commitment – to service of God and other people, to the dignity of women, and to radical environmental protection? Parishioners would have to make real sacrifices in these areas, or they’d risk exclusion. They couldn’t just talk the talk. By being strict about the important things, could liberal Protestant churches make their followers walk the walk of their faith – and save their denominations in the process?”
http://www.patheos.com/blogs/scienceonreligion/2013/07/why-is-liberal-protestantism-dying-anyway/
This is a backwards interpretation of the dynamics of log-normal distributions.
The (rough) equivalent operation of moving everyone’s donations from 1% to 50% would be moving everyone’s donations from the (dollar-weighted) mean charity to the best charity. Although (as you noted) the heavier tail of a log-normal distribution means that the sample mean is higher relative to the mode or median, it has an even stronger effect on the sample maximum.
This means that overall, a lognormal has a higher, not lower, maximum:mean ratio for a fixed, say, median and standard deviation, compared to a thinner-tailed distribution like the normal. For instance, in numerical simulations I just ran with 100 samples from a log-normal and normal distribution, both with median 2 and variance approximately 4.6, the average ratio of sample maximum to sample mean was 5.5 for the log-normal and 3.7 for the normal.
Yes, but ex ante. The higher up the distribution the harder they will get to identify because ‘if it’s transformative it’s not measurable; if it’s measurable it’s not transformative’. The weakness of the measurements mean you’re going to be hit with a lot of regression to the mean.
Also that stuff is likely to be weird (must be extreme on neglectedness if it’s so important and still useful to put more money in to), and so just as it’s hard to get someone to give enormous amounts, it will probably also be hard to move donations there.
I’m not talking about in-practice difficulties like convincing people to donate. I’m just talking about statistics.
Can you point to actual parameters for a toy model where changing the distribution from normal to log-normal (holding median and variance constant) decreases the benefits you can get from convincing people to switch charities? The model can include things like regression to the mean and weirdness penalties. My intuition is that the parameter space of such models is very small, if it exists at all.
If we thought that the charity they were switching to were only at the 95th percentile, it could be worse in a log-normal case than a normal case (indeed it could be worse than not getting them to switch).
However that would be an unusual belief for us. More reasonably we might think it were uniformly drawn from the top ten percent (or something). And then log-normal is again much better than normal. I agree with the thrust of your intuition.
Yeah, in GiveWell classic, you’re generally going to estimate that a high-impact charity is on the 95th percentile but with uncertainty around that. Which is in-between the cases that you describe.
I can imagine that knowing that something is on the 95th percentile with high certainty might be worse than guessing that something is on the 95th percentile with high uncertainty, if you have a prior that is some mixture of log-normal, normal and power-law. That’s what we’d have to show to really question the classic GiveWell model.
I’m sure you’re right about the math, but I am concerned with the in-practice difficulties.
My point about the mean was simply that one shouldn’t compare the max-median on the log-normal—which would be natural if you visualise where ‘typical donations’ that you see go—but rather the max-mean, which is a less extreme ratio. I wasn’t drawing a contrast with a normal or any other distribution.
I’m also not sure about the answer to my question to Ryan—maybe the effectiveness is still the better focus, but I’m prompting him with possible counterarguments.
Ah, from your first comment it sounded like you were comparing the mean of the log-normal to the mean of a less-skewed distribution, rather than to the median of the log-normal. That sounds more reasonable.
Yep my bad making the original comment ambiguous.
In practice, isn’t it relatively easy to identify whether someone is already up the right tail in their giving? I don’t recall struggling with this in initial conversations. You can just ask whether they give abroad for instance, which will seemingly get you most of the way since most people don’t (i.e. it is true that those who do vastly skew the mean, but you can just exclude almost all of them ex ante...).
What you’re saying might be appropriate for mass marketing I suppose where you can’t cut off the right tail.
Sorry Alex, I can’t quite follow what you’re arguing here. Are you saying you can just focus on people who already give fairly effectively?
Sorry, pretty unclear post on my part. Owen basically got it right though; if we’re talking practically rather than theoretically, you don’t have to decide to always focus on effectiveness or always focus on quantity. You can choose, and your choice can be influenced by the information you have about your audience.
Since most individual people are around the median/mode and a long way below the mean, for most individual people talking about effectiveness is correct. There are a few exceptions to this (those ‘up the right tail’), and then you can talk about amount...or just accept that you aren’t going to achieve that much here and find more people where you can talk about effectiveness!
This is obviously dependent on how much ability you have to discriminate based on your audience, which in turn depends on context, hence my ‘mass marketing’ point.
I think it’s the reverse—if you exclude the people who already do give effectively then you’ve brought the mean of those who remain down closer to the vicinity of the median.
Another speculative argument in favor of big asks: most charity seems to focus on making small asks, because people think getting people on the first step towards making a difference is the crucial bottleneck (e.g. see this guy), so the space of ‘making big asks’ is neglected. This means it’s unusually effective to work in this space, even if you appeal to many fewer people.
This seems to be one of the main reasons GWWC has been much more effective than ordinary fundraising techniques.
The downside is that we’re concerned with total scale as well as cost-effectiveness, and a ‘big ask’ approach probably has less total growth potential in the long-run.
The GWWC pledge isn’t really an ‘ask’ - people may make particular donations because they’re asked to, but no one commits to donating 10% of their income every year until they retire because someone asked them to. Instead they make this commitment because they want to do it anyway, and the pledge provides a way for them to declare this publicly to influence others. So it would be interesting to find examples of more typical big asks working—eg. fundraising teams which highball potential donors. Does anyone know of these?
This would, if true, imply that GWWC does not actually result in any initial donations on the behalf of its members. It might still result in more donations on behalf of non-members.
“but no one commits to donating 10% of their income every year until they retire because someone asked them to”
We do in our outreach efforts!
Ah sure, but I’m saying that no one gives this sort of money just because they’ve been asked to—it’s too large and long-lasting a commitment, and being asked is not a powerful enough reason or prompt. Asking them to sign the GWWC pledge may prompt them to make this public declaration, but only if they were already happy to give that sort of money.
In my experience, some of the people I’ve asked to take the pledge would have donated 10% “eventually”, but the pledge actually made them follow through for at least that year in particular where I’m confident they otherwise would not.
Some would have done so anyway, but I think the example set by hundreds of others, including some they know personally, normalises giving a large amount and makes them more likely to copy. Also making a public declaration makes people more likely to follow through.
We also obviously make arguments in favour of doing so. It rarely convinces people immediately but it contributes to moving them in that direction.
“People respect and are impressed by those who are making big sacrifices for the benefit of others. Note how much attention and respect from important figures Toby Ord has gotten by pledging a large fraction of his income. Also the kudos generally given to doctors, soldiers, firefighters, Mother Teresa, etc.”
Attention and admiration, absolutely. But how much copying? I would expect people to be most drawn to people who seem to be only ‘one step’ more self-sacrificing than they are, in a conceptual framework that I suppose is analogous to the idea of inferential distance. For instance, I think Toby benefits in this from having an ‘ordinary’ income. Anything further away than that rapidly becomes too weird to be taking seriously. Note Jeff Kaufman has written about pretty much exactly this:
http://www.jefftk.com/p/optimizing-looks-weird
One weak support for the idea that these examples are sufficiently far away for most people to dismiss without much thought is the very fact media happily talks about them; this would make them less challenging and cognitive-dissonance inducing, so others’ dominant impressions on reading are to be impressed rather than unsettled.
The fact that each person is less impacted has to be weighed against the fact that about 100,000-1,000,000 times as many people have heard about what Toby is doing than someone who just privately gave a few percent of income.
Of course we need examples of more typical people to refer to as well.
Although Toby’s story is only about 2-10x more widely known compared to if he gave 1% and still founded GWWC.
No way, the main media interest was in the extremeness of the amount. The press wouldn’t care about or cover Giving What We Can if the ask was 1% and Toby was giving 1%.
Most of GWWC’s exposure is through word of mouth. This is even truer for the exposure that matters, and that leads to people signing up, much of which is through meeting Giving What We Can members.
The idea that 99.99% of the exposure of Giving What We Can would have disappeared if they couldn’t focus on Toby’s generosity could only result from very unclear thinking. The press would’ve just looked different. Trivially, other members would have given a large fraction even if Toby hadn’t told them to, and press could’ve focussed on that. The example is The Life You Can Save. Granted, it had Peter Singer, but I hope that we’re not going to have the discussion that without Peter, they could’ve still aroused at least 0.1% of the same press. I mean, EA Melbourne has been able to give talks, go on community radio et cetera without citing the 10% figure...
Like, I know things like press can have weird power law distributions but we’re really short-changing the exposure of the rest of GWWC’s message—the parts that are about quality of donation—if we say that 100,000 of its exposure is just because of the quality.
So although the benefits for exposure of of giving 10% are presumably there, and they might matter a lot, they’re like 2-10x.
“The idea that 99.99% of the exposure of Giving What We Can would have disappeared if they couldn’t focus on Toby’s generosity could only result from very unclear thinking.”
Yes, but this is what you said: “Although Toby’s story is only about 2-10x more widely known compared to if he gave 1% and still founded GWWC.”
You may think GWWC would have similar numbers of members today because of how it grows person-to-person, but his story would be much less known because there would basically be no story to report on. Someone gives 1% of charity and encourages others to do the same? It’s not newsworthy. As it was, it got to the most read stories on BBC News and other similar outlets and literally millions (maybe tens of millions) of people heard about him.
Furthermore, I think that media exposure was necessary to turn Giving What We Can from just a group of friends in Oxford into the going concern it is today. That story isn’t as valuable today, but it was the main asset we had in the early days.
Haha Rob it’s Christmas, can’t we stop fighting? Because I’m right, and you should convert to my point of view. :|
But seriously, you’re saying that if Toby had given 1% instead of 50%, then rather than 10 million people knowing his story, rather than 10 million people knowing about him, only 10-100 people would? That simply not reasonable. Without press, even if each of the handful of academics who signed up for GWWC had mentioned it in their opening lectures that semester, you would already have a thousand people who had heard the story.
Haha, let’s split the difference. Maybe 100,000 people would have heard of Toby, so 1-2 orders of magnitude? I think that’s enough for me to make my point that it could be better overall, even if each person takes the example less seriously. :P