Diversity and team performance: What the research says

(Photo of group of people doing a hard thing from Wikimedia user Rizimid, CC BY-SA 3.0.)

This is an extended version (more info, more sources) version of the talk I gave at EA Global San Francisco 2017. The other talk I gave, on extinction events, is  here. Some more EA-focused pieces on diversity, which I’ve read but which were assembled by the indomitable Julia Wise, are:

Effective altruism means effective inclusion

Making EA groups more welcoming

EA Diversity: Unpacking Pandora’s Box

Keeping the EA Movement welcoming

How can we integrate diversity, equity, and inclusion into the animal welfare movement?

Pitfalls in diversity outreach

There are moral, social, etc. reasons to care about diversity, all of which are valuable. I’m only going to look at one aspect, which is performance outcomes. The information I’m drawing from here are primarily meta-studies and experiments in a business context.

Diversity here mostly means demographic diversity (culture, age, gender, race) as well as informational diversity – educational background, for instance. As you might imagine, each of these has different impacts on team performance, but if we treat them as facets of the same thing (“diversity”), some interesting things fall out.

(Types of diversity which, as far as I’m aware, these studies largely didn’t cover: class/wealth, sexual orientation, non-cis genders, disability, most personality traits, communication style, etc.)

Studies don’t show that diversity has an overall clear effect, positive or negative, on the performance of teams or groups of people. (1) (2) The same may also be true on an organizational level. (3)

If we look at this further, we can decompose it into two effects (one where diversity has a neutral or negative impact on performance, and one where it has a mostly positive impact): (4) (3)

Social categorization

This is the human tendency to have an ingroup / outgroup mindset. People like their ingroup more. It’s an “us and them” mentality and it’s often totally unconscious. When diversity interacts with this, the effects are often – though not always – negative.

Diverse teams tend to have:

  • Lower feelings of group cohesion / identification with group
  • Worse communication (3)
  • More conflict (of productive but also non-productive varieties) (also the perception of more conflict) (5)
  • Biases

A silver lining: One of these ingrouping biases is the expectation that people more similar to us will also think more like us. Diversity clues us into diversity of opinions. (6) This gets us into:

Information processing 

Creative, intellectual work. (7) Diversity’s effects here are generally positive. Diverse teams are better at:

  • Creativity (8) (2)
  • Innovation (9)
  • Problem solving. Gender diversity is possibly more correlated than individual intelligence of group members. (Though this conclusion might have failed to replicate. Taymon Beal kindly brought this to my attention after the talk, and I’ve only skimmed it since, but you may want to look into it.) One of the factors implicated is more equal distribution of who’s talking (observed in ethnically diverse groups as well.) (10)
  • Decision-making (15)

Diverse teams are more likely to discuss alternate ideas, look at data, and question their own beliefs.

This loosely maps onto the “explore / exploit” or “divergent / convergent” processes for projects. (2)

    1. Information processing effects benefit divergent / explore processes.
    2. Social categorization harms convergent / exploit processes.

If your group is just trying to get a job done and doesn’t have to think much about it, that’s when group cohesiveness and communication are most important, and diversity is less likely to help and may even harm performance. If your group has to solve problems, innovate, or analyze data, diversity will give you an edge.

How do we get less of the bad thing? Teams work together better when you can take away harmful effects from social categorization. Some things that help:

    1. The more balanced a team is along some axis of diversity, the less likely you are to see negative effects on performance. (12) (7) Having one woman on your ten-person research team might not do much to help and might trigger social categorization. If you have five women, you’re more likely to see benefits.
    2. Remote teams are less biased (w/r/t gender). Online teams will be less prone to gender bias.
    3. Time. Obvious diversity becomes less salient to a group’s work over time, and diverse teams end up outperforming non-diverse teams. (13) (6) Recognition of less-obvious cognitive differences (e.g. personality and educational diversity) increases over time. As we might hope, the longer a group works together, the less surface-level differences matter.

This article has some ideas on minimizing problems from language fluency, and also for making globally dispersed teams work together better.

How do we get more of the good thing? Diversity is a resource – more information and cognitive tendencies. Having diversity is a first step. How do we get more out of it?

    1. At least for age and educational diversity, high need for cognition. This is the drive of individual members to find information and think about things. (It’s not the same as, or especially correlated to, either IQ or openness to experience (1)).

Harvard Business Review suggests that diversity triggers people to stop and explain their thinking more. We’re biased towards liking and not analyzing things we feel more comfortable with – the “fluency heuristic.” (14) This is uncomfortable work, but if people enjoy doing it, they’re more likely to do it, and get more out of diversity.

But need for cognition is also linked with doing less social categorization at all, so maybe diverse groups with high levels of this just get along better or are more pleasant for all parties. Either way, a group of people who really enjoy analyzing and solving problems are likely to get more out of diversity.

2) A positive diversity mindset. This means that team members have an accurate understanding of potential positive effects from diversity in the context of their work. (4) If you’re working in a charity, you might think that the group you might assign to brainstorming new ways to reach donors might benefit from diversity more than the group assigned to fix your website. That’s probably true. But that’s especially true if they understand how diversity will help them in particular. You could perhaps have your team brainstorm ideas, or look up how diversity affects your particular task. (I was able to find results quickly for diversity in fundraising, diversity in research, diversity in volunteer outreach… so there are resources out there.)

Finally, diversity’s effect size on innovation isn’t huge. It’s smaller than the effect size support for innovation, external and internal communication, vision, task orientation, and cohesion – all these things you might correctly expect correlate with diversity more than diversity (15). That said, I think a lot of people [at EA Global] want to do these creative, innovative, problem-solving things – convince other people to change lives, change the world, stop robots from destroying the earth. All of these are really important and really hard, and we need any advantage we can get.

  1. Work Group Diversity
  2. Understanding the effects of cultural diversity in teams: A meta-analysis of research on multicultural work groups
  3. The effects of diversity on business performance: Report of the diversity research network
  4. Diversity mindsets and the performance of diverse teams
  5. The biases that punish racially diverse teams
  6. Time, Teams, and Task Performance
  7. Role of gender in team collaboration and performance
  8. Team-level predictors of innovation at work: A comprehensive meta-analysis spanning three decades of research
  9. Why diverse teams are smarter
  10. Evidence of a collective intelligence factor in the performance of human groups
  11. When and how diversity benefits teams: The importance of team members’ need for cognition
  12. Diverse backgrounds and personalities can strengthen groups
  13. The influence of ethnic diversity on leadership, group process, and performance: an examination of learning teams
  14. Diverse teams feel less comfortable – and that’s why they perform better
  15. Team-level predictors of innovation at work: A comprehensive meta-analysis spanning three decades of research

Fictional body language

Here’s something weird.

A common piece of advice for fiction writers is to “show, not tell” a character’s emotions. It’s not bad advice. It means that when you want to convey an emotional impression, describe the physical characteristics instead.

The usual result of applying this advice is that instead of a page of “Alice said nervously” or “Bob was confused”, you get a vivid page of action: “Alice stuttered, rubbing at her temples with a shaking hand,” or “Bob blinked and arched his eyebrows.”

The second thing is certainly better than the first thing. But a strange thing happens when the emotional valence isn’t easily replaced with an easily-described bit of body language. Characters in these books whose authors follow this advice seem to be doing a lot more yawning, trembling, sighing, emotional swallowing, groaning, and nodding than I or anyone I talk to does in real life.

It gets even stranger. These characters bat their lashes, or grip things so tightly their knuckles go white, or grit their teeth, or their mouths go dry. I variously either don’t think I do those, or wouldn’t notice someone else doing it.

Blushing is a very good example, for me. Because I read books, I knew enough that I could describe a character blushing in my own writing, and the circumstances in which it would happen, and what it looked like. I don’t think I’d actually noticed anyone blush in real life. A couple months after this first occurred to me, a friend happened to point out that another friend was blushing, and I was like, oh, alright, that is what’s going on, I guess this is a thing after all. But I wouldn’t have known before.

To me, it was like a piece of fictional body language we’ve all implicitly agreed represents “the thing your body does when you’re embarrassed or flattered or lovestruck.” I know there’s a particular feeling there, which I could attach to the foreign physical motion, and let the blushing description conjure it up. It didn’t seem any weirder than a book having elves.

(Brienne has written about how writing fiction, and reading about writing fiction, has helped her get better at interpreting emotions from physical cues. They certainly are often real physical cues – I just think the points where this breaks down are interesting.)


There’s another case where humans are innovatively trying to solve the problem of representing feelings in a written medium, which is casual messaging. It’s a constantly evolving blend of your best descriptive words, verbs, emoticons, emojis, and now stickers and gifs and whatever else your platform supports. Let’s draw your attention to the humble emoticon, a marvel of written language. A handful of typographic characters represent a human face – something millions of years of evolution have fine-tuned our brains to interpret precisely.

(In some cases, these are pretty accurate: :) and ^_^ represent more similar things than :) and ;), even though ^_^ doesn’t even have the classic turned-up mouth of representation smiles. Body language: it works!)




Now let’s consider this familiar face:


And think of the context in which it’s normally found. If someone was talking to you in person and told a joke, or made a sarcastic comment, and then stuck their tongue out, you’d be puzzled! Especially if they kept doing it! Despite being a clear representation of a human face, that expression only makes sense in a written medium.

I understand why something like :P needs to exist: If someone makes a joke at you in meatspace, how do you tell it’s a joke? Tone of voice, small facial expressions, the way they look at you, perhaps? All of those things are hard to convey in character form. A stuck-out tongue isn’t, and we know what it means.

The ;) and :D emojis translate to meatspace a little better, maybe. Still, what’s the last time someone winked slyly at you in person?

You certainly can communicate complex things by using your words [CITATION NEEDED], but especially when in casual conversations, it’s nice to have expressive shortcuts. I wrote a bit ago:

Facebook Messenger’s addition of choosing chat colors and customizing the default emoji has, to me, made a weirdly big difference to what it feels like to use them. I think (at least with online messaging platforms I’ve tried before) it’s unique in letting you customize the environment you interact with another person (or a group of people) in.

In meatspace, you might often talk with someone in the same place – a bedroom, a college dining hall – and that interaction takes on the flavor of that place.

Even if not, in meatspace, you have an experience in common, which is the surrounding environment. It sets that interaction apart from all of the other ones. Taking a walk or going to a coffee shop to talk to someone feels different from sitting down in your shared living room, or from meeting them at your office.

You also have a lot of specific qualia of interacting with a person – a deep comfort, a slight tension, the exact sense of how they respond to eye contact or listen to you – all of which are either lost or replaced with cruder variations in the low-bandwidth context of text channels.

And Messenger doesn’t do much, but it adds a little bit of flavor to your interaction with someone besides the literal string of unicode characters they send you. Like, we’re miles apart and I may not currently be able to hear your voice or appreciate you in person, but instead, we can share the color red and send each other a picture of a camel in three different sizes, which is a step in that direction.

(Other emoticons sometimes take on their own valences: The game master in an online RPG I played in had a habit of typing only “ : ) ” in response when you asked him a juicy question, which quickly filled players with a sense of excitement and foreboding. I’ve tried using it since then in other platforms, before realizing that doesn’t actually convey that to literally anyone else. Similarly, users of certain websites may have a strong reaction to the typographic smiley “uwu”.)

Reasoning from fictional examples

In something that could arguably be called a study, I grabbed three books and chose some arbitrary pages in them to look at how character’s emotions are represented, particularly around dialogue.

Lirael by Garth Nix:

133: Lirael “shivers” as she reads a book about a monster. She “stops reading, nervously swallows, and reads the last line again”, and “breaths a long sigh of relief.”

428: She “nods dumbly” in response to another character, and stares at an unfamiliar figure.

259: A character smiles when reading a letter from a friend.

624: Two characters “exchange glances of concern”, one “speaks quickly”.

Most of these are pretty reasonable. I think the first one feels overdone to me, but then again, she’s really agitated when she’s reading the book, so maybe that’s reasonable? Nonetheless, flipping through, I think that this is Garth Nix’s main strategy. The characters might speak “honestly” or “nervously” or “with deliberation” as well, but when Nix really wants you to know how someone’s feeling, he’ll show you how they act.

American Gods by Neil Gaiman:

First page I flipped to didn’t have any.

364: A character “smiles”, “makes a moue”, “smiles again”, “tips her head to one side”. Shadow (the main character) “feels himself beginning to blush.”

175: A character “scowls fleetingly.” A different character “sighs” and his tone changes.

The last page also didn’t have any.

Gaiman does more laying out a character’s thoughts: Shadow imagines how a moment came to happen, or it’s his interpretation that gives flavor – “[Another character] looked very old as he said this, and fragile.”

Earth by David Brin:

First two pages I flipped to didn’t have dialogue.

428: Characters “wave nonchalantly”, “pause”, “shrug”, “shrug” again, “fold his arms, looking quite relaxed”, speak with “an ingratiating smile”, and “continue with a smile”.

207: Characters “nod” and one ‘plants a hand on another’s shoulder”.

168: “Shivers coursed his back. Logan wondered if a microbe might feel this way, looking with sudden awe into a truly giant soul.” One’s “face grows ashen”, another “blinks.” Amusingly, “the engineer shrugged, an expressive gesture.” Expressive of what?

Brin spends a lot of time living in characters’ heads, describing their thoughts. This gives him time to build his detailed sci-fi world, and also gives you enough of a picture of characters that it’s easy to imagine their reactions later on.

How to use this

I don’t think this is necessarily a problem in need of a solution, but fiction is trying to represent the way real people might act. Even of the premise of your novel starts with “there’s magic”, it probably doesn’t segue into “there’s magic and also humans are 50% more physically expressive, and they are always blushing.” (…Maybe the blushing thing is just me.) There’s something appealing about being able to represent body language accurately.

The quick analysis in the section above suggests at least three ways writers express how a fictional character is feeling to a reader. I don’t mean to imply that any is objectively better than the other, although the third one is my favorite.

1) Just describe how they feel. “Alice was nervous”, “Bob said happily.”

This gives the reader information. How was Alice feeling? Clearly, Alice was nervous. It doesn’t convey nervousness, though. Saying the word “nervous” does not generally make someone nervous – it takes some mental effort to translate that into nervous actions or thoughts.

2) Describe their action. A character’s sighing, their chin stuck out, their unblinking eye contact, their gulping. Sheets like these exist to help.

I suspect these work by two ways:

  1. You can imagine yourself doing the action, and then what mental state might have caused it. Especially if it’s the main character, and you’re spending time in their head anyway. It might also be “Wow, Lirael is shivering in fear, and I have to be really scared before I shiver, so she must be very frightened,” though I imagine that making this inference is asking a lot of a reader.
  2. You can visualize a character doing it, in your mental map of the scene, and imagine what you’d think if you saw someone doing it.

Either way, the author is using visualization to get you to recreate being there yourself. This is where I’m claiming some weird things like fictional body language develop.

3) Use metaphor, or describe a character’s thoughts, in such a way that the reader generates the feeling in their own head.

Gaiman in particular does this quite skillfully in American Gods.

[Listening to another character talk on and on, and then pause:] Shadow hadn’t said anything, and hadn’t planned to say anything, but he felt it was required of him, so said, “Well, weren’t they?”

[While in various degrees of psychological turmoil:] He did not trust his voice not to betray him, so he simply shook his head.

[And:] He wished he could come back with something smart and sharp, but Town was already back at the Humvee, and climbing up into the car; and Shadow still couldn’t think of anything clever to say”

Also metaphors, or images:

Chicago happened slowly, like a migraine.

There must have been thirty, maybe even forty people in that hall, and now they were every one of them looking intently at their playing cards, or their feet, or their fingernails, and pretending as hard as they could not to be listening.

By doing the mental exercises written out in the text, by letting your mind run over them and provoke some images in your brain, the author can get your brain to conjure the feeling by using some unrelated description. How cool is that! It doesn’t actually matter whether, in the narrative, it’s occurred to Shadow that Chicago is happening like a migraine. Your brain is doing the important thing on its own.

(Possible Facebook messenger equivalents: 1) “I’m sad” or “That’s funny!” 2) Emoticons / emotive stickers, *hug* or other actions 3) Gifs, more abstract stickers.)

You might be able to use this to derive some wisdom for writing fiction. I like metaphors, for one.

If you want to do body language more accurately, you can also pay attention to exactly how an emotion feels to you, where it sits in your body or your mind – meditation might be helpful – and try and describe that.

Either might be problematic because people experience emotions differently – the exact way you feel an emotion might be completely inscrutable to someone else. Maybe you don’t usually feel emotions in your body, or you don’t easily name them in your head. Maybe your body language isn’t standard. Emotions tend to derive from similar parts of the nervous system, though, so you probably won’t be totally off.

(It’d also be cool if the reader than learned about a new way to feel emotions from your fiction, but the failure mode I’m thinking of is ‘reader has no idea what you were trying to convey.’)

You could also try people-watching (or watching TV or a movie), and examining how you know someone is feeling a certain way. I bet some of these are subtle – slight shifts in posture and expression – but you might get some inspiration. (Unless you had to learn this by memorizing cues from fiction, in which case this exercise is less likely to be useful.)

Overall, given all the shades of nuance that go into emotional valence, and the different ways people feel or demonstrate emotions, I think it’s hardly surprising that we’ve come up with linguistic shorthands, even in places that are trying to be representational.

[Header image is images from the EmojiOne 5.0 update assembled by the honestly fantastic Emojipedia Blog.]

Social games for fun, bonding, and blackmail

[Salad bowl image from fir0002 / flagstaffotos.com.au, under a CC BY-NA 3.0 license.]

At a party, or hanging out with some friends or strangers, and not sure what to do or how to get to know each other? Try a social game! The ones here fall loosely into a couple categories: improv, communication, affinity, and inference.

Don’t get me started – improv

The simplest of improv games. Possibly, it will get you comfortable generating and discussing opinions, but even if it doesn’t or you’re already comfortable with that, it’s a bunch of fun.

The game goes in a circle. Person A comes up with a topic, and tells it to Person B. Someone starts a 3-minute timer. Person B energetically rants about the topic for 3 minutes. At the end of the 3 minutes, Person B writes a new topic for Person C, and the game proceeds.

The purpose of the game is to rant, not to necessarily say things you agree with or even think are factually correct – trying to come up with a coherent critique on the spot is fun, but something like Cecil Palmer’s thoughts on the existence of mountains is also a great outcome.

Some notes: People’s tolerance for ranting about things they actually care about, or are close to, vary in a party context, so let people veto suggestions. There is no “losing”, there’s just continuing to rant until the timer is up.

Salad Bowl – improv / communication

A slightly more complicated improv game.

Start by separating your group of 5-12 people into two teams. Everybody gets 6 pieces of paper (more or less depending on how long you want the game to be), writes a word or short phrase on it, folds it, and puts it into a bowl. The bowl is shuffled.

For each round, take 30 seconds per person. One person draws a sheet from the bowl, and tries to get others on their team to guess the word. If their team gets the word, the person puts the sheet aside and draws another. At the end of 30 seconds, hand the bowl to the next person on the opposing team.

With an odd number of players, one person doesn’t get assigned to a team – on their turns, everybody gets to guess. The sheet of paper goes to whichever team guesses the correct answer.

At the end of each round, tally and write down how many sheets of paper each team has won. Put the papers back in the bowl, and move on to the next rounds.

Remember, the rounds go in order!

Round 1: Taboo. You can say any words except for the one (or ones) written on the card, or versions of them. (E.G., if the card says “dank memes”, then “rare Pepes” or “cats from the internet with words on them” is fine, but “meme”, “memes”, “memetic”, or “memery” are not.)

Round 2: Charades. Act out the word.

Round 3: One word. You can say exactly one word (that’s not the word or a version of the word on the card) to get your teammates to guess what’s on the card.

Round 4: Pose. Say “close” when your turn starts. Everybody on your team closes their eyes. Strike a pose that represents your word or phrase. Say “open”. Hold the pose, and your teammates guess based on the pose.

Post-it Pictionary – communication

For n people (where n = 4-10), give everyone a pile of n post-it notes. Everybody writes a sentence or phrase on the bottom post-it note. Then they pass it to the right.

The next round of people look at the bottom note, then, on the post-it above it, draw a picture to represent the sentence. Then they pass it to the write.

The next round of people look at only the most recent note, then write the phrase they think is described by the image.

Continue passing stacks, alternating looking at the most recent note and drawing a picture or writing a sentence. Once the note reaches its original owner, go around and show off what happened to your note.

Hot Seat – affinity

Do you want to know a group of people way, way better? This game is the fine craft nitro porter to “Truth or Dare”’s 6-pack of Budweiser. I think it came from the Authentic Relating community.

Find a smallish group of people among whom there’s a decent amount of trust. Put everyone in a circle somewhere where other people won’t wander in (e.g., if you’re in a party, walk to a park or find a room and close the door.) Start a timer. (5 minutes is good, make it more or less depending on the size of the group and how long you want to spend playing.) Everybody asks any question they want to the person “in the hot seat”, who answers. This person is allowed to skip questions. At the end of the timer, go to the next person.


  • If the person in the hot seat doesn’t want to answer a question, they cede their turn to the next person.
  • At the start of their turn, the person says a number from 1-5 designating the amount of invasiveness of the questions they want. (In my experience, question-askers aren’t very good at translating a number into a nuanced level of invasiveness, but your group may be different.)
  • The version described under “Hot Seat” in this PDF.

Some notes: The people I play this with call it “intimacy hacking”. For the game to go successfully, I think the people asking questions do have to be ready to ask personal questions, but to try not to hurt the person in the hot seat. It actually gets easier to play around people you don’t know very well.

If the person in the seat clearly stands out in some way from the other people playing (gender, background, appearance, whatever), you might still ask about that, but tread carefully and don’t only ask questions about that. Try not to use the game to hit on people or ask a lot of prurient questions only to people you’re into. Having a facilitator who can police questions if needed is good if you’re not all very comfortable with each other. Be sure that everybody knows what they’re getting into, and with whom, before you start and it becomes harder to duck out.

Aside from that, ask questions you’re curious about, questions that’ll help you know them better, or questions that are fun to answer. This game is easier to play than it sounds, and kind of magical when it goes well.

Chill Seat – affinity

Less replay value than Hot Seat, but still a lovely time.

Everybody goes around the circle, and gives a compliment to the person in the Chill Seat. Then go on to the next person.

Variations: We played a version at a going-away party, where everyone said nice things about the people who were leaving. It was adorable.

Ring of Fire – affinity

Conceptually similar to Hot Seat.

Go around the circle. The first person asks a question, and in turn, everyone else in  the circle answers – ending in the person who asked the question. Then the next person goes.

Some notes: This game tends to be easier to play than Hot Seat, but can still be intense. People have different tolerances of getting into long personal stories during the game – I find it kind of frustrating, some people think it adds a lot of value and enjoyment. If your group decide to stop playing, make sure to wait until everyone’s answered the current question.

“Why these and not those” games – inference

Good for trying some group problem solving. Described better by my friend here.

Flying Circus – inference

Like a chump, I’m writing this without having tried it myself. That said, I imagine an interesting group game is getting a hold of one of the Flying Circus of Physics (With Answers) books, or questions from it online, and trying to answer one of the questions in it as a group.

Remember some strategies for group problem-solving: make sure you understand the problem before proposing solutions, try coming up with several hypotheses, try coming up with experiments or observations that would disprove your hypotheses. Don’t look up information, but think of related phenomena you’re familiar with, and see if your theory works with them.

Probably works best for groups who are interested in physical phenomena, but for which no member is already especially knowledgeable.

Other games

Improv: List of improv games

Communication: Mad Libs, Telephone

Affinity: Truth or Dare, Never Have I Ever

Inference: 20 Questions, lateral thinking puzzles, Who Am I

Other classes of social games: Storytelling games, strategy games

How to design surveys that represent asexuals

[CW: mentions of sex.]

Most surveys that discuss the matter almost certainly misrepresent the asexual population in one way or another. Fortunately, if you’re creating a survey, or interpreting results from a previously-conducted survey, there are ways to make your results more accurate!

This post is based on my previous post about asexuality, which contains more detailed sources and reasons why I think this topic is important. The idea of this post can probably also be applied to representing diverse sexual preferences or even gender identities (e.g. allow varied responses, don’t make assumptions), but the specific suggestions are targeted towards asexuality. Feel free to share this with people who are designing surveys.

Remember that asexuality and aromanticism exist

If your survey touches in any way on romance, sexuality, relationships, or related behaviors, the most important thing is to know and account for the fact that asexuality exists at all.

The basics: Asexuality is an umbrella term for people who don’t experience sexual attraction. 1-8% of people are or could be called asexual (more info here). Asexual people aren’t an easily-dismissed minority, and they are in your sample demographic. (Probably.) Aromanticism, similarly, is not having romantic interest. We don’t know how many aromantics there are, but they’re certainly out there. People may be aromantic and asexual, or either one, or neither. Some people consider asexuality and aromanticism to fall under the LGBTQ demographic, some people don’t. (The extended LGBTQIA+ acronym does include asexuals – that’s what the ‘a’ is supposed to stand for.) More information can be found here.

In representing asexual people in your results, the first question is what you’re using your data for.

My survey is about general identity/demographic information

We might expect 2x-4x as many romantic asexuals as aromantic asexuals (where do these numbers come from?). This is important because people on the asexual or aromantic spectrum have multiple identities – they might be biromantic and gray-asexual, or aromantic and homosexual, or heteroromantic and demisexual. This means that a question like the following is likely to lead to inaccurate answers:

What’s your orientation?

  • Heterosexual
  • Homosexual
  • Bi/pansexual
  • Asexual

One community survey found that the number of asexuals doubled when asexuality was asked about separately. You could do the same thing:

What’s your orientation?

  • Heterosexual
  • Homosexual
  • Bi/pansexual
  • Other

Are you asexual?

  • Yes
  • No

A solution that might be less confusing for people who don’t know what asexuality is, is to allow respondents to check multiple boxes, e.g.:

Check which of the following best describe your sexual/romantic orientation:

( ) Heterosexual
( ) Homosexual
( ) Bi/pansexual
( ) Asexual

It would also be nice (and more accurate) to include some other options:

( ) Gray-asexual
( ) Demisexual
( ) Other

You could also ask about romantic and sexual orientation separately:

What is your sexual orientation?

  • Heterosexual
  • Homosexual
  • Bi/pansexual
  • Asexual
  • Other

What is your romantic orientation?

  • Heteroromantic(attracted to another gender)
  • Homoromantic (attracted to your same gender)
  • Bi/panromantic (attracted to all genders)
  • Aromantic (do not experience romantic attraction)
  • Other

(Edit, 3/4/17: Siggy points out in the comments that it’s important to include an “other” or write-in response on romantic orientation questions, as well as sexual orientation.)

You could also just have a write-in response:

What is your sexual/romantic orientation?  ___________________

You can then bin responses like “straight” and “heterosexual” as meaning the same thing, or, say, “aro-ace” and “gray-asexual lesbian” as both being on the asexual spectrum.

There is a downside in that people don’t necessarily know what “heteroromantic” means right away, even if they are that. (So if you’re going with options with less-familiar words, include definitions.)

Weed out troll answers with a lizardman question

The problem with more questions or write-ins is that those open up options to troll or confusing responses, perhaps from people who disagree with the basis of the question, or don’t understand.

Since people who troll on a gender or orientation question are likely to troll on other parts of the quiz, you could throw in a lizardman question – an absurd question designed to weed out troll respondents (or at least calibrate the honesty of participants).

In middle school, we got drug use surveys that asked us to check if we had ever done marijuana, heroin, hallucinogens, amphetamines, perscription drugs, inhalants, or derbisol (also known as DB, dirt, wagon wheels, or hope.) We asked the health teacher what “derbisol” was after the test, and she looked it up, and derbisol isn’t real – it’s a lizardman answer. (Apparently, 18.2% of high-schoolers in some groups have claimed to use derbisol. Remember: if you don’t talk to your kids about wagon wheels, bloggers will.)

The point is that you can adapt a lizardman question to a variety of contexts.

My survey is about sexual/romantic/relationship behavior

The keys here are A) remember that asexuality and aromanticism exist, and B) ask about behavior or preferences rather than making assumptions.

  • Many asexual people date people.
  • Some asexuals sometimes have sex.
  • Some people who don’t identify as asexual still don’t want to have sex for whatever reason.
  • Someone who’s gray-asexual may normally round themselves off as “asexual” on surveys, but have experienced sexual attraction before.
  • Some people are asexual but don’t know it.
  • Asexual people may or may not identify as queer.
  • Etc.

So if your question is about, say, attitudes from people who have or want to have sex with women, don’t ask if they’re heterosexual/bisexual men or homosexual/bisexual women. Instead, ask if your respondent has or wants to have sex with women.

Same goes for relationships.

The Asexual Identification Scale is 12 questions about behavior and preferences that capture 90% of asexual people, and can also identify asexual people who don’t realize they’re asexual. If you’re curious specifically about asexual-type behaviors, this may be your answer.

My survey is gathering data for both demographics and behaviors

State what you’re using the data for. For instance, if you have one question to ask college students about their orientation and who they’re likely to date, state that your study is  about dating preferences.

You won’t get a complete picture of people’s orientations, but you weren’t going to anyways with one multiple-choice question. And people with complicated identities (like “biromantic asexual”) are more likely to write in the part that represents who they’re planning to date, not have sex with. If you’re using the response to gather information about STD risk, make it clear that your question is about sexual activity. (And then clarify what “sexual activities” you’re talking about, since people define that differently too and it’s probably relevant to STD risk. Specificity counts!)

2. Avoid over-generalizing from your results. If you’re using data from a question like the first one (“pick one: homosexual, heterosexual, bi/pansexual, or asexual”), realize that your answers for who dates or has sex with whom are necessarily fuzzy, because your results are representing asexuals and aromantics poorly.

Throw a prediction party with your EA/rationality group

TL;DR: Prediction & calibration parties are an exciting way for your EA/rationality group to practice rationality skills and celebrate the new year.

On December 30th, Seattle Rationality had a prediction party. Around 15 people showed up, brought snacks, brewed coffee, and spent several hours making predictions for 2017, and generating confidence levels for those predictions.

This was heavily inspired by Scott Alexander’s yearly predictions. (2014 results, 2015 results, 2016 predictions.) Our move was to turn this into a communal activity, with a few alterations to meet our needs and make it work better in a group.


  • Each person individually writes a bunch of predictions for the upcoming year. They can be about global events, people’s personal lives, etc.
    • If you use Scott Alexander’s system, create 5+ predictions for fixed confidence levels (50%, 60%, 70%, 80%, 90%, 95%, etc.)
    • If you want to generate Brier scores or logarithmic scores, just do 30+ predictions at whatever confidence levels you believe.
  • Write down confidence levels for each prediction.
  • Save your predictions and put it aside for 12 months.
  • Open up your predictions and see how everyone did.

To make this work in a group, we recommend the following:

  • Don’t share your confidence intervals. Avoid anchoring by just not naming how likely or unlikely you think any prediction is.
  • Do share predictions. Generating 30+ predictions is difficult, and sharing ideas (without confidence levels) makes it way easier to come up with a bunch. We made a shared google doc, and everyone pasted some of their predictions into it.
  • Make predictions that, in a year, will verifiably have happened or not. (IE, not “the academic year will go well”, which is debatable, but “I will finish the year with a 3.5 GPA or above”.)
  • It’s convenient to assume that unless stated otherwise predictions that end by the next year (IE, “I will go to the Bay Area” means “I will go to the Bay Area at least once in 2017.”) It’s also fine to make predictions that have other end dates (“I will go to EA Global this summer.”)
  • Make a bunch of predictions first without thinking too hard about how likely they are, then assign confidence levels. This post details why. You could also generate a group list of predictions, and everyone individually lists their own confidence levels.

This makes a good activity for rationality/EA groups for the following reasons:

  • Practicing rationality skills:
    • Making accurate predictions
    • Using confidence intervals
  • Accessibility
    • It’s open to many different knowledge levels. Even if you don’t know a thing about geopolitics, you can still give predictions and confidence intervals about media, sports, or your own life.
    • More free-form and less intimidating than using a prediction market. You do not have to know about the details of forecasting to try this.
  • Natural time and recurring activity
    • You could do this at any point during the year, but doing it at the start of the year seems appropriate for ringing in the new year.
    • In twelve months, you have an automatic new activity, which is coming back together and checking everybody’s predictions from last year. Then you make a new set of predictions for next year. (If this falls through for some reason, everyone can, of course, still check their predictions on their own.)
  • Fostering a friendly sense of competitiveness
    • Everyone wants to have the best calibration, or the lowest Brier score. Everyone wants to have the most accurate predictions!

Some examples of the predictions people used:

  • Any open challenges from the Good Judgment Project.
  • I will switch jobs.
  • I will make more than $1000 money in a way that is different from my primary job or stock.
  • I will exercise 3 or more times per week in October, November, December.
  • I’ll get another tattoo.
  • Gay marriage will continue to be legal in Washington state.
  • Gay marriage will continue to be legal in all 50 states.
  • I will try Focusing at least once.
  • I will go to another continent.
  • CRISPR clinical trials will happen on humans in the US.
  • A country that didn’t previously have nuclear weapons will acquire them.
  • I will read Thinking Fast and Slow.
  • I will go on at least 3 dates.

Also relevant:

  • 16 types of useful predictions
  • Brier values and graphs of ‘perfect’ vs. actual scores will give you different information. Alexander writes about the differences between these. Several of us did predictions last year using the Scott Alexander method (bins at fixed probabilities), although this year, everybody seems to have used continuous probabilities. The exact method by which we’ll determine how well-calibrated we were will be left to Seattle Rationality of 2018, but will probably include Brier values AND something to determine calibration.

(Crossposted from LessWrong.)

Science for Non-Scientists: How to read a journal article

Scientific journal writing has a problem:

  1. It’s the main way scientists communicate their findings to the world, in some ways making it the carrier of humanity’s entire accumulated knowledge and understanding of the universe.
  2. It’s terrible.

It’s terrible for two reasons: accessibility and approachability. This first post in this series discussed accessibility: how to find papers that will answer a particular question, or help you explore a particular subject.

This post discusses approachability: how to read a standard scientific journal article.

Scientific papers are written for scientists in whatever field the journal they’re published in caters to. Fortunately, most journal articles are also written in such a way that you can figure out what they’re saying even if you’re a layperson.

(Except for maybe math or organic chemistry synthesis. But if you’re reading about math or organic chemistry as a layperson, you’re in God’s hands now and I can’t help you.)

Okay, so you’ve got your 22-page stack of paper on moose feeding habits, or the effects of bacteriophage on ocean acidification, or gravitational waves, or whatever. What now? There are two cardinal rules of journal articles:

  1. You usually don’t have to read all of it.
  2. Don’t read it page by page.

Journal articles are conveniently broken into sections. (They often use the names given, or close synonyms.) I almost always read them in the following order:


1. Abstract

The abstract is the TL;DR of the article, the summary of what the studies found. Conveniently, it’s first. The abstract is very useful for determining if you actually want to read the rest of the article or not. Abstracts often have very dense, technical language, so if you don’t understand what’s going on in the abstract, don’t sweat it.

2. Introduction

As a layperson, the introduction is your best friend. It’s designed to bring the reader from only a loose understanding of the field, to “zoom in” to the actual study. It’s supposed to build the context you need to understand the experiment itself. It gives a background to the field, what we already know about the topic at hand, historical context, why the researchers did what they did, and why it’s important. It’ll define terms and acronyms that will be crucial to the rest of the paper.

It may not actually be easy language. At this point, if you encounter a term or concept that’s unfamiliar (and that the researchers don’t describe in the introduction), start looking it up. Just type it into Wikipedia or Google, and if what you get seems to be relevant, that’s probably it.

3. Conclusions

In a novel, skipping to the end to see how the suspense plays out is considered “bad form” and “not the point.” When reading papers, it’s a sanity-saving measure. In this part of the paper, the researchers write about what conclusions they’re drawing from their studies,and its implications. This is also done in fairly broad strokes that put it in context of the rest of scientific understanding.

4. Figures

Next, go to the figures that are strewn around the results section, just before the conclusions. (Some papers don’t have figures – in that case, just read the results.) Figures will give you a good sense of the actual results of the experiments. Also read the captions – captions on figures are designed to be somewhat stand-alone, as in that you don’t have to read everything else in the paper to tell what’s going on in the figures.

Depending on your paper, you might also get actual pictures of the subject that illustrate some result. Definitely look at these. Figure out what you’re looking at and what the pictures are supposed to be telling you. Google anything you don’t understand, including how the images were obtained if it’s relevant.

In trying to interpret figures, look at the labels and axes – what’s being compared, and what they’re being measured by. Lots of graphs include measurements taken over time, but not all. Some figures include error measurements – each data point on a graph might have been the average of several different data points in individual experiments, and error measures how different those data points were from each other. A large percent error (or error bar, or number of standard deviations, etc) means the original data points were far apart from each other, small error means that they were all close to the average value. If you see a type of graph that you’re not sure how to read, Google it.

5. Results

The section that contains figures also contains written information about the researchers actually observed in the experiments they ran. They also usually include statistics, IE, how statistically significant a given result is in the context of the study. The results are what the conclusions were interpreting. They may also describe results or observations that didn’t show up in figures.

Maybe read:


Methods are the machinery of the paper – the nuts-and-bolts, nitty-gritty of how the experiments were done, what was combine, where the samples came from, how it was quantified. It’s critical to science because it’s the instructions for how other researchers can check what you did and see if they can replicate the results – but I’d also rather read Youtube comments on political debates than read methods all day. I’ll read the methods section under the following circumstances:

  • I’m curious about how the study was done. (You do sometimes get good stuff, like in this study where they anesthetized snakes and slid them down ramps, then compared them to snakes who slid down ramps while wearing little snake socks to compare scale friction.)
  • I think the methodology might have been flawed.
  • I’m trying to do a similar experiment myself.
snakes on a plane.gif
Snakes on a plane! || Gif from this video.

Works cited

Papers cite their sources throughout the paper, especially in the introduction. If I want to know where a particular fact came from, I’ll look at the citation in the works cited section, and look up that paper.

Acknowledgement/Conflicts of Interest

Science is objective, but humans aren’t. If your paper on “how dairy cows are super happy on farms” was sponsored by the American Dairy Association and Dairy Council, consider that the researchers would be very biased to come to a particular conclusion and keep receiving funding. If the researchers were employed by the American Dairy Association and Dairy Council, I’d be very tempted to just throw out the study.

Science for Non-Scientists: How to find scientific literature

Scientific journal writing has a problem:

  1. It’s the major way scientists communicate their findings to the world, in some ways making it the carrier of humanity’s entire accumulated knowledge and understanding of the universe.
  2. It’s terrible.

This has two factors: Accessibility and approachability. Scientific literature isn’t easy to find, and much of it is locked behind paywalls. Also, most scientific writing is dense, dull, and nigh-incomprehensible if you’re not already an expert. It’s like those authors who write beautiful works of literature and poetry, and then keep it under their bed until they die – only the poetry could literally be used to save lives.  There are systematic issues with the way we deal with scientific literature, but in the mean time, there are also some techniques that make it easier to deal with.

This first post in this series will discuss accessibility: how to find papers that will answer a particular question or help you explore a subject.

The second post in this series discusses approachability: how to read a standard scientific journal article.

How to Find Articles

Most scientific papers come from a small group of researchers who do a series of experiments on a common theme or premise, then write about what they learned. If your goal is to learn more about a broad subject, ask yourself if a paper is actually what you want. Lots of quality, scientifically rigorous information can be obtained in other ways – textbooks, classes, summaries, Wikipedia, science journalism.


blog science stack
The great food web of “where does scientific knowledge come from anyways?”

When might you want to turn to the primary literature? If you’re looking at very new research, if you’re looking at a contentious topic, if you’re trying to find a specific number or fact that just isn’t coming up anywhere else, if you’re trying to fact-check some science journalism, or if you’re already familiar enough with the field that you know what’s on Wikipedia already.

You can look at the citations of a journal article you already like. Or, find who the experts in a field are (maybe by looking at leaders of professional organizations or Wikipedia) and read what they’ve written. Most science journalism is also reporting on a single new study, which should be linked in the article’s text.

If you have access to a university library, ask them about tools to search databases of journal articles. Universities subscribe to many reliable journals and get their articles for free. Your public library may also have some.

Google Scholar is a search engine for academic writing. It has both recent and very old papers, and a variety of search tools. It pulls both reliable and less reliable sources, and both full-text and abstract-only articles (IE, articles where the rest is behind a paywall.) Clicking “All # Versions” at the bottom of each result will often lead you to a PDF of the full text.

If you’ve found the perfect paper but it’s behind a paywall- well, welcome to academia. Don’t give up. First up, put the full name of the article, in quotes, into Google. Click on the results, especially on PDFs. It’ll often just be floating around, in full, on a different site.

If that doesn’t work, and you don’t have access through a library, well… Most journals will ask you to pay them a one-time fee to read a single article without subscribing. It’s often ridiculous, like forty dollars. (Show of hands, has anyone reading this ever actually paid this?)

But this is the modern age, and there are other options. “Isn’t that illegal?” you may ask. Well, yes. Don’t do illegal things. However, journals follow two models:

  1. Open content access, researchers pay to submit articles
  2. Content behind paywalls, researchers can submit articles for free

As you can see, fees associated with journals don’t actually go to researchers in either model. There are probably some reasonable ethical objections to downloading paywalled-articles for free, but there are also very reasonable ethical objections to putting research behind paywalls in general.

How good is my source?

Surprise! There’s good science and bad science. This is a thorny issue that might be beyond my scope to cover in a single blog post, and certainly beyond my capacity to speak to every field on. I can’t just leave you here without a road map, so here are some guidelines. You’ll probably have two goals: avoiding complete bullshit and finding significant results.

Tips for avoiding complete bullshit

  • Some journals are more reliable than others. Science and Nature are the behemoths of science and biology (respectively), and have extremely high standards for content submission. There are also other well-known journals in each field.
  • Well-known journals are unlikely to publish complete bullshit. (Unless they’re well known for being pseudoscience journals.)
  • You can check a journal’s impact score, or how well-cited their work tends to be, which is sort of a metric for how robust and interesting the papers they publish are. This is a weird ouroboros: researchers want to submit to journals with high impact scores, and journals want to attract articles that are likely to be cited more often – so it’s not a perfect metric. If a journal has no impact score at all, proceed with extreme caution.
  • Watch out for predatory journals and publishers. Avoid these like the plague, since they will publish anything that gets sent to them. (What is a predatory journal?)
  • Make sure the journal hasn’t issued a retraction for the study you’re reading.

Once you’ve distinguished “complete bullshit” from “actual data”, you have to distinguish “significant data” from “misleading data” or “fluke data”. Finding significant results is much tougher than ruling out total bullshit – scientists themselves aren’t always great at it – and varies depending on the field.

Tips for finding significant results

  • Large sample sizes are better than small sample sizes. (IE, a lot of data was gathered.)
  • If the result appears in a top-level journal, or other scientists are praising it, it’s more likely to be a real finding.
  • Or if it’s been replicated by other researchers. Theoretically, all research is expected to replicate. In practice, it sometimes doesn’t, and I have no idea how to check if a study has been replicated.
  • If a result runs counter to common understanding, is extremely surprising, and is very new, proceed with caution before accepting the study’s conclusions as truth.
  • Apply some common sense. Can you think of some other factor that would explain the results, that the authors didn’t mention? Did the experiment run for a long enough amount of time? Could the causation implied in the paper run other ways (EG, if a paper claims that anxiety causes low grades: could it also be that low grades cause anxiety, or that the same thing causes both anxiety and low grades?), and did the paper make any attempt to distinguish this? Is anything missing?
  • Learn statistics.

If you’re examining an article on a controversial topic, familiarize yourself with the current scientific consensus and why scientists think that, then go in with a skeptical eye and an open mind. If your paper gets an opposite result from what most similar studies say, try to find what they did differently.

Scott Alexander writes some fantastic articles on how scientists misuse statistics. Here are two: The Control Group is Out of Control, and Two Dark Side Statistical Papers. These are recommended reading, especially if your subject is contentious, and uses lots of statistics to make its point.

Review articles and why they’re great

The review article (including literature reviews, meta-analyses, and more) is the summary of a bunch of papers around a single subject. They’re written by scientists, for scientists, and published in scientific journals, but they’ll cover a subject in broader strokes. If you want to read about something in more detail than Wikipedia, but broader than a journal article – like known links between mental illness and gut bacteria – review articles are a goldmine. Authors sometimes also use review articles to link together their own ideas or concepts, and these are often quite interesting.

If an article looks like a normal paper, and it came from a journal, but it doesn’t follow the normal abstract-introduction-methods-discussion-conclusion format, and subject headings are descriptive rather than outlining parts of an experiment, it might be a review article. (Sometimes they’re clearly labelled, sometimes not.) You can read these the same way you’d read a book chapter – front to back – or search anywhere in it for whatever you need.

What if you can’t find review articles about what you want, or you need more specificity? In that case, buckle up. It’s time to learn how to read an article.

So You’re Not Ready To Go Vegetarian

[Content warning: Moralizing about what food you should eat, descriptions of bad things happening to animals, eating bugs. Also, lots of people can’t go vegetarian or significantly alter their diet at all due to health, cost, time, sensory issues, strong preferences, lack of options, inability to pick your own diet, etc. Most of the ‘alternatives’ posed here take money, time, or majorly changing your habits. If reading this post is likely to make you feel guilty or bad in an unproductive way, feel free to skip it.]

This is a rather utilitarian list of approaches to improving the lives of animals even if you still eat meat. I’ll start with some general strategies, ranked roughly in order from “least  to most weird”. See what works with your diet, resources, and preferences.

 Basic ideas:

  • Eat less meat in general.
  • Eat less chicken, eggs, beef, and farmed fish.
  • For other animal products, eat Animal Welfare Approved, Certified Humane, or 100% Grass-Fed meat, or buy from a source where you know how the animals are treated.
  • Eat species that suffer less, either in farms or at all.
  • Pay other people to go vegan for you.
  • Support animal welfare by donating money effectively.

I suspect that some people will object to the notion that it’s ever alright to kill or use an animal, and that encouraging people to do this in a “less bad” way is just making compromises with the devil. (As opposed to veganism, which is merely selling your soul to Seitan.) If you’re one of these people, you’re probably already a vegan and this essay isn’t for you.

Not that I entirely disagree- many more people should be vegetarian. That’s not the point, though. Many people are Vegetarian Sympathizers, as I once was. As a young person, for instance, I knew that I had moral issues with the idea of eating animals- that a cow’s brain wasn’t very different from a cat’s, which also wasn’t very different from a human’s. I also knew that meat had unfortunate impacts on the environment and that global warming was a serious problem. But my developmental environment had lots of meat. And also, I had a very strong objection- cheeseburgers.

Pictured: The Seattle restaurant that was the source of my conflict. The mind is willing, but the flesh is weak. | By Jmabel (CC BY-SA)

This wasn’t a rational objection. But we’re not rational creatures, and the Cheeseburger Objection was the actual thing standing in between me and vegetarianism. And if I’m going to eat cheeseburgers anyways, why not eat steak, chicken, fish, etc.?

Honestly, the Cheeseburger Objection is a pretty good one. One cow makes a lot of cheeseburgers. One cheeseburger might make you very happy. Acknowledging that isn’t a reason to stop caring about animal welfare entirely. And Cheeseburger Objectionists can still make extremely meaningful contributions to animal welfare without depriving themselves of that cheesey goodness.

1. Only go vegetarian sometimes.

Meatless Mondays are a thing- don’t eat meat just one day a week. That’s 1/7 fewer animals you’re eating, and gaining valuable practice in cooking and eating vegetarian. If that’s too easy, up it to two days a week. Repeat.

Some other strategies that have worked for people: eat vegan before 5 o’clock (IE, meals before dinner), only eat meat outside the house, only eat meat inside the house.

Or, if you’re inclined towards vegetarianism- except for cheeseburgers- (or orange chicken, shrimp, your uncle’s venison, baseball stadium hotdogs, etc.-) consider just being a Cheeseburger Vegetarian. I think there’s this tendency to think that if you’re not doing something 100% all the way and identify as that, any tendency you have towards it doesn’t count at all. But that’s completely untrue. Given that we live in a world where most people do eat meat, conspicuously eating less meat both saves animals, and is a talking point that puts vegetarianism on people’s radars.

(Of course, if you’re being a Cheeseburger Vegetarian and hoping to talk to other people about it, people might take you less seriously. This might be a problem. You could either keep your cheeseburger habit private and secretive, hoarding McDonald’s in the dark like the world’s most gluttonous dragon – or you could acknowledge that if someone’s going think that plant-based diets are a joke and not important, they can already find whatever reason they want to do that.)

If you don’t know how to cook food or eat meals without meat, maybe the problem is educational. Look for recipes that contain tofu, beans, lentils, TVP, or vegetables. If you only know one kind of cuisine, broaden your horizons- Indian, Ethiopian, Mexican, Chinese, etcetera, all have lots of opportunities for low meat dishes.

We live in a golden age of easily available recipes. PETA, Vegetarian Times, and Leanne Brown’s free cookbooks are a few good resources. Google it. Also, if you want to make a favorite Food X vegan or vegetarian, look up “Vegan Food X” and you will instantly get 4,000 hits including step-by-step photographs and people’s life stories as told through salad dressing recipes. The internet is a magical place.

2. Eat humanely sourced meat.

This is way harder than it sounds. The good news is that meat is given labels which reflect how it was raised. The bad news is that some of these labels are regulated, and some aren’t, and it’s difficult to determine which labels actually correspond to good living environments and which are symbolic or easily falsified.

Look for the following words on packages:

Certified Organic animals may still be subject to a variety of inhumane conditions. The label means that hormones, antibiotics, and some other treatments are not allowed, and that the animal must be allowed to “exhibit natural behaviors.” I suspect that organic animals are somewhat harder to mistreat, because farmers are incentivized to raise animals in low-disease environments, so organic may be better than conventional if those are your only two options. *

Animal Welfare Approved is an independently-verified certification that has very high welfare standards, including for slaughter. Certified Humane is a less strong but similar certification. There are probably other good ones- look for what they require and how they’ve verified.

Hoofed animals: Look for 100% Grass-Fed, a legally-defined term in which all animals must be raised entirely on pasture (grass, etc) and not fed harvested grain. It seems much harder to mistreat a cow raised this way, since it can’t be confined. This is different from grass-finished, pastured, or normal grass-fed, since all cows eat some grass before they arrive at feedlots.

3. Be careful with chicken.

Chickens are extremely common and live extremely bad lives in factory farms, probably moreso than any other animal.

Cage-free or free-range eggs are better than alternatives, but I don’t think they’re humane. A cage-free chicken may have a somewhat better and more natural life than a non-caged chicken, though they’re newly at risk of fighting with other chickens, which caged chickens aren’t. They may still be subject to having their beaks cut off, slaughter of male chicks (half of all egg-laying chickens are killed shortly after hatching), bird flu, crowded environments, being raised in darkness, starvation-based forced molting, etc.

A couple examples:

  • Free-range – the amount of time or space required for “outdoor access” isn’t legally defined, and varies from facility to facility.
  • A cage-free chicken is still raised in barns or warehouses. They may have no outdoor access, or have their beaks cut or burned off without anesthesia.
  • Organic eggs still aren’t treated with antibiotics but can still be raised in factory farms.
  • More info on labels.
Putting a picture of happy chickens here seemed disingenuous, so here’s some eggs, I guess. | The Home Front In Britain, 1935-1945.

Any given egg source may well not do some or all of these- for instance, I’ve heard that there are some egg producers that don’t slaughter male chicks, and the cost of raising them is passed to the consumer as a higher price. The key here is to do your research. If you buy based on label X or Y without further investigation, even at a “nice” natural foods store or co-op, your chicken will probably have been raised in painful, inhumane conditions.

I think your best chance at getting humanely raised chickens or eggs is to buy from a home farmer or very small permaculture farm, ideally where you can see the chickens. These are likely to be significantly more expensive than other options. Farms may still slaughter male chicks.

4. Eat species that suffer less.

Quantification of animal suffering is a new field, and practices for calculating it are general estimates. That said, its numbers come from easily understandable ideas- that it’s worse to be a factory-farmed chicken than a feedlot cow, for instance. Some other ideas include that being killed is painful, so an animal that produces more food over a long period means less suffering per food unit (assuming said animal’s day-to-day existence isn’t terrible.) Also, that having a more complex brain probably means you can suffer more. It’s not an exact science, but it’s what we’ve got.

Brian Tomasik, who has studied animal suffering extensively, suggests using this metric that by eliminating chicken, chicken products, and farmed fish from your diet, you reduce the suffering you inflict on animals by an enormous amount.

Clams and mussels have very simple nervous systems and probably do not feel much pain, while full of nutrients comparable to other animal foods. Ozymandias at Thing of Things suggests that eating bivalves and dairy, and otherwise being vegan, can be a good trade-off between health, enjoyment, and helping animals. Also, you still get to eat clam chowder (if it doesn’t have bacon.)

The jury is still out on whether insects experience suffering. On one hand, insects are pretty simple critters; on the other hand, to produce any significant amount of food, you need a lot of insects, so however much moral weight they do have gets multiplied by a lot. On the third hand, about a quintillion die every year, so your own contribution is pretty marginal. (That number is extrapolation- I suspect most insects live less than a year, so the number is probably higher.)

Chingrit thot by Takeaway (CC BY-SA)

What is known is that insects are nutritious and environmentally friendly. Sourcing insects is difficult and pricey, so try raising your own.

Exotic meats. I suspect that exotic meats (deer/venison, buffalo, ostrich, etc.) are more likely to be raised in more ethical environments, because as species they’re less domesticated, and therefore harder to mistreat as in a factory farm. However, I have no evidence for this.

5. Eat environmentally sound meat.

Most of this list comes from a moral argument, but the negative environmental impacts of standard meat is so well-established that it’s worth discussing. 30% of the world’s non-frozen dry land is currently devoted to feeding or raising animals, and 18% of human-produced greenhouse gases came from agriculture. Lamb and beef have disproportionately high greenhouse gas emissions. You’ll note that chicken is rather low on this ranking, but as in the above section, there are other reasons to avoid it.

“Don’t non-animal-product foods also have carbon emissions?” Not that much.

Source and more info.

Fish is extremely nutritious, but many species are overfished. Eat conscientiously to avoid making the problem worse- the Monterey Bay Aquarium Seafood Watch has detailed recommendations for the consumer based on your location, sorted into handy “okay to eat” and “avoid this” categories. Bycatch ratios are another thing to beware: shrimp fisheries are the worst, trawling up an average 6 times more non-shrimp than shrimp.

6. Convince someone else to go vegan.

A review (again by Tomasik) of organizations that run ads promoting vegetarianism suggest that the cost of converting a someone to be vegan for a year is, conservatively, about $100. Do you have the money to spare, and think there should be more vegans, but eating meat is worth more than a hundred dollars to you?

Utilitarianism: it works.

cool skeleton
Utilitarianism: It’s this cool. And the ends justify the memes.

This approach won’t work forever, of course – if everybody decided that they individually would eat meat but convince others not to, the cost of getting anyone to go vegan would skyrocket. But not everybody is, and for the time, it’s still low-hanging fruit.

7. Donate to effective charities.

Can we do even better? The average vegetarian saves ~25 land animals per year (and perhaps 371-582 animals per year including fish and shellfish) according to the blog Counting Animals.

The Effective Altruism movement, which is near and dear to my heart, has produced several lovely projects, including Animal Charity Evaluators– a highly evidence-based group that researches which animal welfare organizations have the most bang for your buck. (Sort of the Givewell of the greater biosphere.) An $100 donation to any of their top three charities is estimated to indirectly save or spare the lives of 7,597 animals. (Via outreach, undercover video filming, corporate outreach, and more.)


A final note: People sometimes get annoyed at vegetarians or vegans because they think they’re being smug or morally uppity. This always seemed to me like a strange criticism – the problem is that they’re doing something good? – but if you think it has merit, imagine how smug you can feel in the knowledge that every year, you donate $100 to a certain charity, and that has the same effects as going vegetarian for thirteen years, every year.**

Updated 4/14/2017.

Further reading:

* Michael Pollen says in his book The Omnivore’s Dilemna that it’s difficult to get Organic certification, which has many requirements and regulatory steps, so some small and comparatively extremely humane farms may not (despite meeting many or all criteria for the certificate.)

**Note that you’re not allowed to use this to smugly dismiss vegetarianism unless you have actually made a substantial donation to ACE charities. If you don’t, and proceed to use the fact that that someone could make such a donation to be a dick to vegans, you’re doing negative good and the Utilitarianism Skeleton will get you.