Pitfalls in Diversity Outreach
By Kelsey Piper @ 2015-08-18T21:52 (+39)
(This is an adaptation of a post on my blog.)
EA is one of several movements I have seen which have tried to address the problem of a lack of diversity, either demographic diversity (i.e. too many men, too many white people) or ideological diversity (too many programmers, too many technolibertarians). The first category of efforts are often ugly and counterproductive, and in other movements I've witnessed the debate itself scaring members of the targeted demographic away. Without rehashing the debate on the merits of diversity, given that a person has decided they want to increase movement diversity, here are some common failure modes I have witnessed:
- exaggerating the demographics to make your point, in a way that suggests that there are no members of the target demographic currently or at least none whose contributions are meaningful.
useful ways of making this criticism: “we’re 70% male. why is that?” "we want more people from developing countries." "the majority of EAs are just out of college."
ways of making this criticism which have frustrated me: “why aren’t women involved in this movement?” "EAs are from all over the world - well, if the only countries in the world are America, Britain, and Australia" "if EA wants to appeal to anyone who isn't an autistic white nerd...
- suggesting that the movement needs members of the target demographic by appealing to sexist/racist/offensive stereotypes (”we’re not warm and empathetic enough. that’s why we need more women in the movement”. "the reason we don't have enough black people is because we're too intellectual and data-focused.")
- suggesting that the movement recruit by appealing to offensive stereotypes (“if we want more non-white people we need spicier food and fewer long position papers”)
- tokenizing the members of the target demographic who you do have ( “hey, will you be the organization president? you don’t have to do any work but we need a black person on the leadership board” )
- not asking people in the groups you're trying to reach what they think or recommend ( “As a man, I’m concerned that we have too many men, and here’s how I think we should go about fixing it”)
- treating the members of the targeted demographic who you do have like they’re zoo animals ( “A woman interested in Our Movement? Cool! Those are so rare, you know. But we’re getting more of them. Look, over there - that’s a new one.” )
undefined @ 2015-08-18T22:23 (+9)
It seems to me that often conversations about diversity outreach are offputting not because anyone accidentally says something wrong but because lots of people sincerely think diversity outreach is a bad idea, and argue so.
For example, in a recent discussion someone said "for example, here's how I think we could do outreach to religious people." Someone else said "honestly, I think we shouldn't do outreach to religious people; religiosity is incompatible with effective altruism". The first person said "some religious people have the same values as we do, and don't take that Biblical law stuff seriously". The second person said "and some of them have horribly divergent sexist Stone Age values and welcoming them would destroy everything that makes EA appealing".
This conversation, I suspect, was not encouraging to religious would-be EAs. But I don't think the first speaker did anything wrong (and the second speaker did everything right, given his/her values). Should there be an expectation that we not publicly argue when someone talks about the merits of diversity, lest our public disagreement make diversity impossible to achieve even if the community ends up concluding it is beneficial?
undefined @ 2015-08-19T08:57 (+6)
I think anybody wanting to raise a potentially divisive or negative discussion should think carefully about how likely a given discussion is to be self-defeating, or to yield negative results that outweigh the benefits.
The setting matters a lot to this: if you post on Facebook, the discussion gets published in lots of people's feeds in a manner that posters don't control (I find 'likes' on comments I make in the EA FB group from friends I know are not members of that group). Also, the FB policy of only allowing 'upvoting' means that the degree to which people's statements are well or badly received is not well reflected. Finally the listing of threads by order of most recent comment keeps pile-ons in the current discussion.
(This also creates an important asymmetry: those who don't care about the discussion being damaging are more likely to continue it, while those who disagree might avoid voicing their disagreement in the hopes that the thread will die away.)
This forum doesn't suffer any of those drawbacks, so I believe it is a better arena for raising these issues for discussion if you reasonably believe there is something important at stake.
Kelsey Piper @ 2015-08-21T04:47 (+3)
I really agree here - other factors that make Facebook conversations particularly inflammatory include Facebook's lack of threading, so you can't easily see who a person is responding to and if the tone of the response is appropriate to the original post, the way Facebook comment threads rapidly stack up with hundreds of comments, some only tangentially related to the original post, and the wide variance in moderation schemes. I've been disillusioned by some of the conversations on Facebook, but this comment made me more optimistic that is a platform issue, not a problem with open discussion of EA concerns.
undefined @ 2015-08-22T01:24 (+2)
Empirically, discussions of diversity here do seem to be doing a lot better than the ones on the FB group. (I'm thinking particularly of this thread and AGB's post from a while ago.)
undefined @ 2017-10-29T16:26 (+4)
Still think so, in light of the heated discussion in the comments at http://effective-altruism.com/ea/1g3/why_how_to_make_progress_on_diversity_inclusion/ ?
undefined @ 2015-08-18T23:18 (+2)
For the record, this happens enough on, e.g., social media without headway being made in the debate I don't believe the most influential or well-connected effective altruists pay mind to these discussions, or expect them to make new insights. Since, e.g., CEA or EA Global conferences are unlikely to discuss effective altruism and diversity issues in a damaging way, I doubt these conversations will cause outsized damage to the community's potential at large. I do believe it would pay off, though, to be as courteous as possible in these cases, even if people disagree with each other. This is not likely an enforceable policy, though. For us, Rivka, I believe the best we can do is that when we witness or participate in these conversations, we translate opposing viewpoints into charitable interpretations of the other side on behalf of debaters, steelman positions when possible, and promote civility as much as possible. Seeing there are other stewards in effective altruism who are just as willing not to accommodate, but see eye to eye and empathize with other perspectives, as others are to bash the worldviews of would-be effective altruists, is the best we can do to show a divisive opinion only represents one individual, and doesn't imply what effective altruism as a whole thinks.
There have been debates on purely intellectual grounds about how if one doesn't favor a specific cause, acting on another because of a difference of value constitutes a monstrous failure to do good. Effective altruism has and is surviving and becoming more robust against such debates which would otherwise threaten to break effective altruism as a union and an alliance. Sociopolitical movements have thrived on pluralism and diversity. Just look at how successful, e.g., Obama's presidential (re-)election campaign went, relative to his Republican rivals, who didn't as much make explicit how much they would try to welcome others regardless of their creed, disposition, or ethnicity. Effective altruism can be that too.
On the object-level subject of religion and effective altruism, here are my takes:
Religiosity is compatible with effective altruism. Effective altruism is about practical prescriptions or recommendations for ethical behavior, controlling for the vast majority of presently popular ethical worldviews. This includes religion. While effective altruism does borrow assessment tools from utilitarianism, and thus effective altruism has a large contigent of secularists, avowed atheists, other non-religious persons, and utilitarians, effective altruism borrows just as many tools from social science and normative rationality, which have little to do with philosophical precepts for ethics. What effective altruism borrows from pure philosophy is mostly just that: tools for thinking, not absolutes which would preclude religion.
A religious person with "horribly divergent sexist Stone Age values" may likely have those values on the basis of religious arguments and evidence, which effective altruism doesn't consider. There are just as many, if not more, religious people on Earth who consider lessons taken science, philosophy, politics, and other ideologic frameworks as think lessons for how to behave in the world can only come from their religion. The former group will naturally accommodate effective altruism in their lives, not by pulling effective altruism in the direction of their religious doctrine, but by perceiving how the imperative of effective action is consistent with the moral intuitions they already hold, as the rest of us currently do. I expect belligerent religious persons will be turned of from effective altruism. Not because those of us already here would be pernicious, or unwelcoming, or uncharitable, but because we simply ask them questions to demonstrate the effectiveness of adopting their methods and values. I believe in being unable to answer such questions without relgious evidence, they would find effective altruism isn't compatible with their particular interpretation of their own religion. Thus, the sort of religious persons hardline atheists or secularists wouldn't want in effective altruism are also the sort of religious persons who will self-select themselves to not join. Thus, I think it's unnecessary and ill-advised for any effective altruist to bash religious peoples' general involvement with effective altruism.
The Effective Altruism Survey results confirm effective altruists are by and large a liberal and pluralist bunch, who personally and politically, are likely to reject advances of allegedly "horribly divergent sexist Stone Age values", all else equal. This and the above consideration ensure it's unlikely the most belligerent religious persons would become even a significant minority of effective altruism, let alone a major force or any other belligerent class of persons.
There are religious discussion groups for effective altruism, such as the Christian effective altruism group on Facebook. It's got over 100 people, if I recall correctly. I think engaging more religious effective altruists is best left to organizations who explicitly have a goal of growing the community, such as Giving What We Can, and the Centre for Effective Altruism, and the general community of already religious effective altruists. The rest of us simply don't have the nuanced take or comparative advantage of assessing the issue of religion and effective altruism, without also shooting ourselves in the foot.
undefined @ 2015-08-19T16:24 (+6)
One worry I've had for a while is that efforts to increase demographic diversity, at least to groups that face particular struggles, could be inherently futile. At the risk of lapsing into postmodern gobbledygook, I think it can be difficult for people who face direct, personal struggles to embrace the universality and objectivity EA aims for. One thing I think about is Mullainathan and Shafir's Scarcity, which talks about the degree to which poverty frames and distorts consumer behavior, and I think struggles that various marginalized groups face may have similar effects. It's hard to help others when you're actively fighting for yourself.
(Anecdotally, on a personal level, as a gay man with an upper middle class upbringing, I find it easy to subscribe to EA, but I don't think I would have when I was closeted and mired in anxiety and self-doubt.)
undefined @ 2015-09-25T04:07 (+3)
Maslow's hierarchy of needs idea points to the same conclusion if you figure EA is a form of self-actualization. I suspect some EAs would be more effective in the long run if they focused on EA less and focused on meeting prerequisite needs in the hierarchy more.
undefined @ 2015-08-19T08:49 (+6)
Those quotes aren't real quotes right? I recognize the offensive one about 'autistic white nerds' from a wierd article by one of the Vox founders but I would have bet against much of the others being said by someone?
Kelsey Piper @ 2015-08-19T21:05 (+4)
No, sorry, they are not. And not all of these are pitfalls I've witnessed specifically in EA outreach - the atheist/skeptic community and campus conservative/libertarian groups are where I watched a lot of these mistakes get made.
undefined @ 2015-08-18T22:48 (+4)
Upovted. When I originally read this on your Tumblr, I legitimately learned things, as I've made the mistake of saying such things in the past, and also observing others doing the same without speaking up about it as a failing tactic, and I agree with you using such tactics in the past and present largely is and was a mistake. This comment is also a signal to cross-post well-received posts from your Tumblr to this Forum, as I don't use Tumblr, as I suspect other effective altruists who find value in your writing. Cross-posting here allows for a greater dissemination of important considerations, and also faciliates more comments/feedback.
My suspicion is that the best way to increase demographic diversity is through project partnerships with groups that are skewed towards people we lack; collaborating with a local racial justice group on a fundraiser for GiveDirectly, for example, or with an environmental group on a cost-benefit analysis of different forms of climate change intervention advocacy. Partnerships for specific projects reduce the risk of value drift and, if the groups turn out to be incompatible, make it easy to part on good terms and maybe win over a few members who are intrigued by the EA approach.
This seems at least fertile ground for running charity experiments. I'm friends with everyone at Charity Science, so I will talk to them about running something like this in the future. If not, I suggest we talk to either Giving What We Can or The Life You Can Save to help push for local EA chapters to try projects like this with local organizations. I think that would have the added benefit of EAs at more isolated chapters build a sense of community, which, in my experience, is crucial for getting people more involved with effective altruism.
undefined @ 2015-08-19T00:17 (+1)
Another reasonably common failure mode seems to be claiming that the under-represented minority is being oppressed or victimized. This apparent attempt to diagnose the problem simply deters members of the minority from applying, making the problem worse.