Hide table of contents

This post is co-authored by Jess Whittlestone and Michelle Hutchinson

In a previous post, Jess talked about how our scepticism as a community might be threatening our ability to be supportive of one another. In this post, Jess and I would like to explore some concrete ways in which we can all be more supportively sceptical.

Supportive scepticism requires a conscious effort

First, let’s just highlight that for most of us, it’s incredibly difficult to take criticism well - and that’s natural. In his best-selling book “How to Win Friends and Influence People”, Dale Carnegie goes so far as to advise that you should never criticise anyone. Criticism puts people on the defensive, he claims, so they’ll rarely take it on board or learn from it - most likely, it will just lower their opinion of you. But most of us care about more than simply winning friends and influencing people - we want to try and figure out true answers to difficult questions.

This suggests that we face a tough challenge if we want to both give critical feedback and remain motivated and caring towards each other. Although we’re a community of people who value feedback and scepticism more than average, it’s important to acknowledge that we’re all human, and all likely to find critical feedback difficult. It’s not just a matter of “not being mean” - we’re all going to need to make a conscious and deliberate effort to think about how the things we say might come across and affect others. This isn’t just a problem we can brush off - it needs a lot of thought and effort if everyone is going to feel positive and supported.

Why are you giving feedback?

Before giving someone a piece of feedback, you might want to think about a couple of questions: How are you feeling? Why are you giving this feedback, and why now? What is the best possible outcome?

This post has a nice breakdown of the various different reasons we might give critical feedback:

  • To help someone improve
  • To see a change we would like
  • To further the discussion
  • To hurt someone
  • To vent frustrations
  • To boost our ego

Hopefully the latter three don’t occur too much in this community, but it’s worth being aware that sometimes we might say something critical for bad reasons. So it may be helpful to check whether the reason you’re commenting is one of the first three. If it isn’t, it’s probably best to refrain from saying anything.

One way to check that you’re not motivated by the latter reasons could be to consider how you’re feeling as you comment - whether you feel angry or hurt, for example. As well as knowing that you’re motivated by one of the first three reasons, it’s often useful to know which, because it means you can focus on framing your criticism in the best way to achieve that goal.

Consider context

Some situations are more appropriate for critical feedback than others. It’s important to consider the content of the post, who the poster is, who they are addressing and where they are posting it, for example.

Suppose someone posts a ‘feel good’ article about some way in which the world is getting better on their own facebook wall. They may well find it demotivating to have the article criticised, and the benefit of criticism in this situation is also probably quite minimal. This seems quite different to a case in which someone posts an idea they had in a group, asking people what they think of it - where critical feedback is much more appropriate.

Timing can also be important: is the feedback about a project which the person is about to embark on, such that they can take your feedback into account? This seems an ideal time to provide feedback. On the other hand, giving someone feedback when they’ve already completed a project and can’t act on it risks just making them feel bad.

If you’re in an ongoing conversation, it might be useful to take note of how the person is responding to your comments. If they appear to be at all upset or defensive, it is likely to be better to step back from criticising them. Another piece of advice Carnegie has is taking care to avoid arguments. He points out that they often simply cement each side’s views, and leave both parties feeling worse about each other. If a discussion seems to be taking that turn, it might be better simply to listen to the other side, and be grateful for their input, rather than continuing to put forward your view. If you find that too difficult, then the best move is probably to bow out entirely.

The importance of positive feedback

Whether you’re giving feedback on an idea, a blog post or a whole project, the person you’re talking to has likely put in effort and good work towards a valuable goal. Highlighting what you think they’ve done well, what is valuable and what you admire in what they have done is likely to make them feel more motivated and more likely to take improvements you suggest on board. Likewise, people tend to feel motivated by acknowledgement of and gratitude for their hard work. We’re so lucky to be part of a community where there are so many nice people working so hard to help others, which thankfully makes it easy to find positive things to say under most circumstances (though unfortunately it’s also really easy to forget to look for them!).

The popular management handbook ‘The One Minute Manager’ highlights the importance of using praise to encourage people, not just highlighting flaws in people’s work. It contains a favourite sentiment of Michelle’s: “Here we put the accent on the positive. We catch people doing something right.” It’s so much easier to produce work and to share your thoughts when you feel they’ll be appreciated and useful than if you expect only negative reactions to them.

It might be particularly good to start and end your comment by telling the person what you like and admire about their post. This is particularly the case if you think the overall idea is a good one, though you think there are particular refinements that could be made. That way the person has an overall positive feeling toward your comment, which will hopefully leave them feeling enthused to carry out the specific suggestion you had.

An underlying theme here is to cultivate an attitude of appreciation for others. Being grateful for the hard work people put in and respectful of the good work they are doing, whether that’s in writing a blog post or carrying out a project, allows us to discuss with them ways in which they might be able to do even better without putting them on the defensive. Since it’s well established that being grateful makes you happier, this should benefit you, too!

Highlighting areas of agreement

Highlighting areas of agreement between you and the person you’re addressing could also work well to improve a discussion. Adam Grant writes in Give and Take about the fact that ‘rare commonalities’ can bring people together: you naturally feel close to someone who has something in common with you, and the more so the more rare the thing you have in common is.

Highlighting the assumptions you share with someone you’re talking to could make you feel more like you’re engaging in a collaborative search for truth rather than arguing. Similarly, if you’re suggesting some improvement, frankly acknowledging that it’s something you could also be doing better can help you connect, and also prevent you from coming across as patronising - thereby again cultivating a feeling of shared pursuit for knowledge.

Focusing on actions rather than people

Feedback about character traits, or general tendencies that people have, can often feel like personal attacks. For that reason, it can be better when suggesting improvements to focus on particular actions, rather than on people’s dispositions. For example, it might be more constructive to say ‘claim X in this argument seems overconfident’ rather than ‘you seem overconfident’. While the latter feels like a criticism of you as a person, the former is more like a suggestion that you rephrase a particular claim you made.

One complication of this might be when criticising projects people feel extremely invested in. When we invest a lot of time and energy in a project, we often end up investing a large part of ourselves in them too. This means an attack on that project can end up feeling like an attack on your identity. It’s worth bearing this in mind when we give feedback on a specific project we know someone is somewhat invested in. In this cases, we probably need to make even more effort to give positive encouragement as well as constructive feedback, and to be extra conscious of the language and tone we use.

Being specific and constructive

The more specific and constructive a piece of advice is, the easier it is to act on, and the easier it is to feel positive about. In an ideal world, we would each have the time and motivation to look into every piece of feedback we got, and work out how we could act on it. In practice, frequently we don’t. Making a comment actionable could ensure it gets acted on rather than forgotten about.

Framing is also important - framing a comment as a way of making something even better than it already is, is likely to empower the person to want to improve their post. On the other hand, saying that a particular feature is bad runs the risk of creating an ugh field for that person around that area.

It might also be helpful to think about what the other person’s goals are, in order that you can frame your comments to align with those goals. That can help prevent you from talking past each other, and is extra helpful to them because it makes clear to them how they might be able to better achieve their goals.

Hedging and warmth

Using language that conveys warmth, support and humility may be easier said than done, but can make a huge difference. It’s difficult to know how other people will take what you write. A few things that might help:

Using ‘adding strength words’ (like ‘definitely’, ‘absolutely’, ‘extremely’) when agreeing with someone or giving positive comments, but not when criticising. People often take criticism strongly to heart. For that reason, you might, for example, want to say something like ‘this doesn’t seem quite true’ rather than ‘this is totally untrue’, since the former conveys your message while dealing less of an emotional punch.

Using exclamation marks. I was once with someone who texted two people to say he was going to be late to see them. One replied ‘Fine’ and the other ‘Fine!’. The sense we both got was that the former of these might be rather irritated by the lateness, but the latter wasn’t - which even at the time seemed pretty ridiculous!

Hedge: make clear that what you are saying is simply your opinion on the matter. Err on the side of saying ‘it might be that...’, ‘my intuition is that...’ rather than ‘this is wrong’. Online, simply stating something can seem aggressive when that was in no way the intent.

Being vulnerable: If a commenter takes a dominant tone, the person receiving the feedback is likely to feel put-down and defensive. They might react better to a tone which made clear that you have similar difficulties to theirs. For example, you might want to frame your suggestion as something that has worked for you, when you faced similar difficulties. Or, you might want to frame it as something the person may have already thought about: rather than saying ‘you should have done x rather than y’, you might ask ‘I was wondering why you decided on x rather than y?’.

Remember, it’s a person you’re talking to

When commenting on blog posts or having conversations with online strangers, it seems rather easy to forget that there’s a real person sat at a computer receiving your feedback. That person, like the rest of us, struggles to stay motivated and happy. Like the rest of us, they don’t have the time or skill to do everything perfectly. Remembering that could help us to remain friendly, constructive and positive in our feedback, even when we are short on time and irritated by the imperfections in a post.


We thought it might be helpful to provide what seems to us like a great example of supportive scepticism. The following is a quoted email:

Some feedback attached. It sounds like things are going very well, apart from [feature eggplant]. Congratulations! One thing I'd encourage you to do with these reports is, in the very first section, list something like:

A. Top 3 achievements over last 6 months (this is for qualitative stuff beyond members increases) (I'm thinking just like 3 bullet points with a sentence or two)

B. Biggest problem you have right now

This would give us an opportunity to zero in on things you might need advice on. Anyway, glad it's going well! Thanks for your hard work.

This acknowledges and is grateful for the work put in. The person makes clear that they are happy with the overall progress, and that they are enthusiastic about that (exclamation marks!). This made it really motivating to work on the changes suggested. Acknowledging that [feature eggplant] was less than ideal indicates that the overall view was genuine rather being fakely positive. The suggestion in the body of this email (as with the attached feedback) is very specific and actionable, making it easy to implement, and it explains the benefit of the suggestion. The whole inspires a spirit optimism and collaboration.

Where to go from here?

It might be useful, going forward, for us to communicate to each other a bit about how we experience feedback. Someone might find it useful to hear that you found a particular piece of feedback they gave you upsetting or stressful. Even better might be to convey to someone when a piece of feedback they gave you was really helpful to you. You might also like to follow up on feedback you’ve given others, by asking them whether they found it helpful, and how it could have been more so. Hopefully we can work together to be a really supportive (as well as sceptical!) community.

Further reading:

  • How to Win Friends and Influence People
  • The One Minute Manager
  • Give and Take
Sorted by Click to highlight new comments since:

Thanks, Michelle!

How to Win Friends and Influence People

I have not read the book, but this seems like a good summary.

Dennett has a nice article on this for written criticism: http://www.brainpickings.org/2014/03/28/daniel-dennett-rapoport-rules-criticism/

Here's his checklist:

How to compose a successful critical commentary:

  1. You should attempt to re-express your target’s position so clearly, vividly, and fairly that your target says, “Thanks, I wish I’d thought of putting it that way.
  2. You should list any points of agreement (especially if they are not matters of general or widespread agreement).
  3. You should mention anything you have learned from your target.
  4. Only then are you permitted to say so much as a word of rebuttal or criticism.


An advantage to steel-manning an opponent (arguing against the best version of their argument) is that you get to see if they agree with your steel-man. This leads to many possible outcomes, and almost all are good for information within the debate. If the person disagrees with your steel man, they may rephrase their argument in a stronger way than your steel man, which may convince you of their position and cause you to change your mind. If they agree, you know exactly what you have to argue against.

There is some nuance to this which is very useful in social situations. For example, if someone immediately agrees with a steel-man, in a rather suspicious way, you may be able to detect if the person is using sophistry to convince you of something. Some people just speak non-sense while trying to signal intelligence or group loyalty, and you can accidentally create a coherent argument where there was none initially.

To summarize, steel-man arguments often lead to good outcomes: (1) You may change your mind for good reasons (2) You can find a clear path to argue with minimal semantic issues and confrontation (3) You may identify motivated reasoners faster

Thanks Ben - great guidelines!

This is nice and practical - it's good that it focusses on specific behaviours that people can practice rather than saying anything that could come across as "you're alienating people and you should feel bad".

One thing I'd add to this is to try to debate less and be curious more. Often discussions can turn into person A defending one position and person B rebutting this position and defending their own. I've found that it is often more helpful for both people to collaborate on analysing different models of the world in a curious way. Person A proposes a model of how the world works and person B the starts trying to understand person A's model - what its assumptions are, where it applies, and where it doesn't. They can then contrast this with other models of the world and try to work together to find out which is best. If you want to get really into it, drawing diagrams can help both because it helps you think and because it increases the sense that you are working together on a problem, rather than arguing against one another. But it doesn't have to be this formal - it could just be a friendly discussion about the strengths and weaknesses of different ideas.

On a related note, I think it's important to realise that people don't always believe the positions they're arguing for. I'll often tell my friends an idea because I'm interested in working out it's strengths, weaknesses, and implications. If they're dismissive and try to argue against it I feel that they're missing the point - it would be more helpful to explore the idea's strengths and weaknesses together rather than turning it into a debate. This would help us to be more accepting of new ideas that don't come from the usual EA sources.

Great comment! :)

I really like that framing of trying to be more curious and less debating. As you say, it's really useful to investigate new ideas, even if you don't agree with them!

Great article! I really think it sets an excellent standard for us to all aspire to and work at.

I recently discovered a downside to hedging and being vulnerable in a public forum. In responding to a blog post I made, a commenter claimed I had an important figure wrong by 2-4 fold. He'd used an adjustment that I thought was not applicable and rather than say 'no your calculation is wrong' I hedged by saying I wasn't an expert in field x and could be explain what calculation he'd used and why. Unfortunately that exchange was used by a critic of EA as proof of my - and the movement's a 'ignorance'.

The moral of this story is, I suppose, we risk vulnerability to unsupportive critics by taking these standards. I definitely think it's a worthwhile trade, and one that will lead to better discussion and a more productive community.

Great comment! Coming from a background of more mainstream altruistic circles, I've found appearing confident and avoiding hedging is an important aspect of leadership and inspiring others to join me in my goals.

I think one resolution for this trade-off is to adjust based on the subject matter. When making certain basic claims (e.g. the importance of the far future, giving a public speech about effective altruism), we might want to err on the side of confidence, but when making more tenuous claims (e.g. regarding the effectiveness of Against Malaria Foundation versus the effectiveness of GiveDirectly), we might want to err on the side of hedging.

Another resolution is to try to come off as both confident and hedging. This is easier in person when we can communicate confidence with our tone and body language, but hedge with our language. It seems difficult to pull off online.

I'd be interested in hearing other ways to best handle this tradeoff.

I'd accept there's some tradeoff here, but I'd hope it's possible to defend your reasoning while being sufficiently supportive.

Some of the comments on the forum don't seem to have any purpose but to make the commenter sound smart/morally superior. I try to only post comments that are useful, not just ones where I can prove someone wrong. It's actually hard not to make unhelpful “you're wrong” type of comments, as per human nature, but I think I have a good system to avoid them: I always write any comment I make in my word processor first, because it's easier to edit, and then cut and paste it into the comment form. If I'm not sure whether my comment will be really useful (despite how right I think it is), I'll just leave it in my word processor for a day or two, occasionally referring back to it. I'll often decide to seriously edit the comment to make it more benign/useful, but often I end up just dropping it altogether if I think it may not be that constructive. Replying to something as soon as you read it can lead to nonproductive comments – often routed in emotion – that come across as showing someone up.

I quite appreciated both this post and your comment. (Which I've stumbled upon 6 years late due to this post.)

I was going to comment something related. Specifically, when I read:

This post has a nice breakdown of the various different reasons we might give critical feedback:
To help someone improve
To see a change we would like
To further the discussion
To hurt someone
To vent frustrations
To boost our ego

...it occurred to me that I think an additional, common reason for giving critical feedback is to signal one's knowledge or intelligence. (I'd say that that's related but not identical to boosting our ego.)

If I'm being honest, I notice I'm sometimes motivated to comment on the forum for this reason. I don't think that that's inherently bad, and I think there is an upside: the desire to signal that I know stuff motivates me to post and comment more on the forum, which (I hope) actually does benefit others in some ways. But that motivation could also lead to making comments that aren't actually net positive (once you factor in how the comments make other people feel), or making comments in a way that's suboptimal (e.g., with the wrong sort of tone).

So I try to pay attention to whether that motivation is playing a major role in my decision-making. And, when it seems like it might be, I try to increase how carefully I assess whether it's worth me making the comment, and whether I should adjust how I phrase it.

(That said, a norm promoting that sort of approach could conceivably go too far, in a way that slows people down and reduces discussion to a greater extent than is worthwhile.)

I like this idea, and have done it before, but it is good if the process can be sped up. Being more responsive increases the likelihood that the useful things you post will get read by those you are responding to. Some forums boot people for not explaining their arguments fast enough.

Yeah, I often use a heuristic of "if you're not sure if you should post this, don't."

Jess & Michelle: thanks for this excellent post. Three remarks I'd like to add:

1. We all need support, but individuals vary considerably in the kind of support they need in order to flourish. A kind of support that works well for one person might feel patronising, frustrating or stifling to another, or cold, distant and uncaring to a third. To be effectively supportive, we must be sensitive to individual needs.

2. Being supportive is difficult, so individuals in the community should help others support them. If the support you're getting from the community is suboptimal, it's unlikely that other individuals are entirely to blame.

3. As a community, we should create an atmosphere where it's easy for people to ask for more or different kinds of support when they need to. Admitting vulnerability and requesting support is a sign of strength and maturity, not weakness, so we should praise, encourage and reward those who do this.

Thanks Peter, great points!

Michelle and Jess, thank you so much for continuing to write and think about these problems!

I've noticed recently that a lot of successful communities seem to succeed because people want to be in them for their own sake, even leaving aside the stated goals of the community. This kind of hard thinking about our social norms really help to push effective altruism towards that goal.

It might be particularly good to start and end your comment by telling the person what you like and admire about their post. This is particularly the case if you think the overall idea is a good one, though you think there are particular refinements that could be made. That way the person has an overall positive feeling toward your comment, which will hopefully leave them feeling enthused to carry out the specific suggestion you had.

If you need a handy name for this technique, I've heard it called the "compliment sandwich" :)

I've discovered with these that it feels really important to get the compliments right--not just "great post!" or something but taking the time to put together something specific, thoughtful, and relevant. (For instance, see my attempt above.) Otherwise, if the person you're talking to is familiar with the compliment sandwich, it can come off as slightly insincere.

Of course, this is hard, but the skill of giving compliments that are sincere, topical, and meaningful will take you very far. And as I mentioned above, I think they also have huge benefits for EA culture at large, in terms of making this a community that people want to participate in.

Highlighting areas of agreement

I think another (underrated) benefit of this, is that it stops the other person from rounding off what you're saying to the closest argument they're used to hearing. For instance, if I were trying to persuade a friend that, say, global poverty causes were probably higher-impact than environmental ones, I would be very careful to say that I agreed that anthropogenic climate change was a serious problem, etc., so that the friend didn't get distracted and start hashing out carefully-rehearsed anti-climate-denial arguments.

As another tip--in informal communication, I'm a big booster of using smiley faces [:)] when I say something that could be read ambiguously. For instance, if I say something that could be read sarcastically, I'll often end with a smiley to make sure it's not taken the wrong way.

I definitely agree it seems important to create a community atmosphere that's appealing in itself, tricky as that is likely to be. As you say, the compliments need to not just be sincere, but the kind that come across as sincere. Being really specific seems like a good way to do that. It might also be helpful not to think of it as a 'compliment sandwich', which sounds kind of insincere, but rather as showing appreciation for the person. Putting yourself in the other's shoes, and thinking about the fact you would like to be appreciated, and then trying to do the same, might help get into a frame of mind that will make showing true appreciation easier. I'm actually not such a fan of smilies - I think they can come across as patronising, or as a bit fake - like you're saying something mean and then trying to nullify the offence. But I imagine it just really depends on the person!

Nice article! I also think all of these factors are important, not only to build to strengthen our ability to be supportive of another, but also because they simply make us better decision-makers. Here's an interesting experiment on the "collective intelligence" of groups and what factors contribute to it:

Psychologists have repeatedly shown that a single statistical factor—often called “general intelligence”—emerges from the correlations among people’s performance on a wide variety of cognitive tasks. But no one has systematically examined whether a similar kind of “collective intelligence” exists for groups of people. In two studies with 699 people, working in groups of two to five, we find converging evidence of a general collective intelligence factor that explains a group’s performance on a wide variety of tasks. This “c factor” is not strongly correlated with the average or maximum individual intelligence of group members but is correlated with the average social sensitivity of group members, the equality in distribution of conversational turn-taking, and the proportion of females in the group.

My emphasis.

Here's the article (Science) and here's a test of social sensitivity.

Thanks for the post! I really liked it, and it helped me further internalize a lot of what I had read elsewhere (e.g., How to Win Friends and Influence People). I definitely need to think about that again and figure out how to actually integrate it into my list.

Maybe a mandatory criticism checklist before I criticize anyone, with a negative HabitRPG habit if I catch myself having criticized without going through the checklist? Or just a negative HabitRPG habit for all criticism? (I like to think of negative HabitRPG habits as things which are unvirtuous (figuratively soul damaging), and I think giving criticism is that, regardless of how needed it is.)

I'm also reminded of Slate Star Codex's comment policy: "There is an ancient Sufi saying beloved of the Buddha, which like a surprising number of ancient Sufi sayings beloved of the Buddha, originates from a book of preachy Victorian poetry. It goes: 'Before you speak, let your words pass through three gates; At the first gate, ask yourself, is is true? At the second gate ask, is it necessary? At the third gate ask, is it kind?' Slate Star Codex has lower standards than either ancient Sufis or preachy Victorians, and so we only require you to pass at least two of those three gates."

Err ... I like your article. Thanks for the hard work. :)

The above was my main reaction to reading this! Then I thought my comment would be more apposite if it also contained a grain of criticism, so I started looking for something you might have improved.

The only thing I have is that perhaps it would help to lead with a paragraph summarising the recommendations, so that people who don't read the whole thing still get some of the value, and people who do are more likely to remember it.

I think people perceive differences in their treatment more than they perceive the absolute harshness or kindness of their treatment. If everyone receives harsh criticism from their teacher or boss, in my experience this results in little resentment, and it makes the praise that much better. When I did ballet, the criticism was almost comically harsh. The teacher would say our dancing was "ugly" or "stupid" and that she "hated" it. Sometimes she would just laugh at us. Praise was very rare and very mild ("not half-bad"). But it was the same for everyone, so almost all of us just reacted with amusement at the harshness, and I think we did become much better dancers as a result. Contrast that with the way girls treated one another: some were popular, some were the target of cruel remarks, and in general it was very unequal. This led to a lot of unhappiness, of course.

I'm not saying that over-the-top cruelty is the best way to deliver criticism, but I think people perceive who's popular and who's not, even if the criticism is implicit. (Keep in mind that "implicit" sometimes just means passive-aggressive, which can be even more destructive.)

Also keep in mind that people perceive differences in praise. Sometimes the EA community feels a bit clique-ish. For example, it feels kind of hurtful when someone will post statuses naming their favorite people in the EA community.

So the lesson is: don't be driven by grudges or favoritism. (I know this is one thing I need to be better about myself.)

For example, it feels kind of hurtful when someone will post statuses naming their favorite people in the EA community.

Yikes, that does sound obnoxious! It would be good if we could cultivate a complimentary atmosphere without causing these effects...

Are there types of praise that are more or less likely to trigger this reaction? Thinking about this, my intuition suggests that it's much more of a problem when you praise people and less of a problem when you praise specific actions. Does that seem right?

(As a side bonus, praising specific actions instead of people generally also helps with growth mindset.)

Curated and popular this week
Relevant opportunities