Tag: communication

Meaning in Learning and in Communicating Science to the Public

lydiaandreyevnakrasilnikovaLeave a comment

I joined the Harvard Museum of Natural History’s Science Education Partners/Virtual Scientist in a Classroom (book me to speak with your classroom, if you want!) this winter, which means I’m spending three hours every Friday for a month learning how to communicate about science to 6th through 9th graders and therefore to everyone—in part because I want to feel like a Harvard student and do more things at Harvard before I graduate and in part because I want to improve my out-loud science communication and I miss teaching.

One of the exercises we did on the first day was to list out activities from our childhood that qualified as “a memorable learning experience that took place in an informal environment (meaning outside of school).” Here are mine:

  • Coding (in HTML) fan sites for my favorite band at the time, t.A.T.u. These were very embarrassing and you will never ever see them, and maybe you don’t consider HTML to be coding, but these fan sites were my first ever exposure to any sort of code. This was at El Valor in Chicago.
  • Coding my theme for MySpace, which would have been a few years later. I think this was a first coding experience for a lot of people in my generation. It must have been HTML with a little bit of JavaScript and some CSS. This is where I learned how to google anything and everything I wanted to learn to code—probably the most important skill.
  • Setting up a shop and pricing products in NeoPets—my first exposure to economics of any sort. (I’m not particularly good at economics.) Apparently there was also a stock market (a few of my friends learned about the stock market through NeoPets) but I missed that; I was more interested in the very small moneys of buying up snowballs at one neopoint and reselling them for three neopoints.
  • Reading fiction, which is both escapism and (I think) the best way to learn how to write. I’m bilingual so I learned to read a little late but I read a lot as a kid, a novel a day for a long while.
  • Reading popular science magazines, which for me was first Popular Science and then MIT Technology Review (it is not lost on me that it is absolutely nuts that I now occasionally write for TR; younger me would be beyond the moon). Since both of my parents are biologists my first exposures to science were of course through them; after that it was Popular Science, and Popular Science was also an introduction to science more broadly.
  • DeviantArt. I was obsessed with photography, especially macro photography, and I learned a lot about what makes a photo meaningful to me by looking at other people’s photos. Some people also posted about their process. And of course I learned a lot from posting my own art and getting feedback.

These experiences were meaningful to me because while they were fun they were also some of the times I learned the most. I was learning because I wanted to do something with the information, which is when I learn best.

It surprised me, listing these out, that all of them were in some way social. Fan sites for bands, MySpace, DeviantArt, and NeoPets are all directly social. Reading fiction and popular science magazines is also social, in a different way—absorbing other people’s voices and ideas and thinking about what I would say if I were part of the conversation. I like group work now—specifically I like groupwork at work, which is most of my job—but in school I most enjoyed work where I holed up on my own and did practice problems or wrote essays. I did not like lecture and I really did not like groupwork, especially in-class groupwork. So it’s surprising to me that my best and most memorable learning all involved other people, even if at a distance.

Next we talked with an actual 8th-grade science class. We went around the room introducing ourselves and our work and then we answered questions from the class. I was appallingly terrible at this part. I said I study how vaccines affect infection by and transmission of COVID-19, which is true but resulted in a lot of questions about vaccines—how they work, how they’re designed—which is interesting but not what I study. What I should have said was that and something about the process: that most of what I do is code at my computer and talk with other people about my ideas and their ideas, that we use the mutations in the SARS-CoV-2 genome to trace transmission events, and that we work with the Department of Public Health to connect what we learn from the genomes to real-world data about the actual people infected and learn more than we could from either data source alone.

I want to mention that I was very impressed by the 8th graders. They asked excellent questions and already spoke enough of the language of biology to have a conversation in it. In retrospect I think I was there too by 8th grade; it shouldn’t have surprised me.

After meeting with the 8th graders we talked about the differences between how we communicate about science with the public vs. with other scientists. The most important difference is not the language, which is what I had been focusing on, but that the order of things is inverted: in science we communicate

  1. the necessary background,
  2. then supporting details,
  3. then our results and conclusions;

when we communicate with the public we should be presenting

  1. the bottom line,
  2. then the so-what/why it matters,
  3. and only then the supporting details.

I think we’re pretty good at the supporting details, maybe even the bottom line; the so-what is more difficult. We listed out answers to the question, “Why is my work important?” Here’s what I came up with (I have very high opinions of the importance of my work):

  • It can show how a disease is transmitted (airborne, blood, food, sex, etc.).
  • It can show who is most vulnerable to a disease.
  • It can guide people’s decision-making on how to protect themselves and their loved ones.
  • It can guide public health recommendations.
  • It can guide how we allocate our resources.

We then thought about the follow-up question: “Why is that important?” Here’s what I wrote (again, I have a very high opinion of my work (which is why I do it, in addition to it being interesting)):

  • It can save lives (if done quickly).
  • It can give people agency (if communicated well and if the government helps).

Then we asked ourselves the next follow-up question, which you can probably guess: “Why is that important?”

To which I answered: because life and health and agency are important.

So when I talk about my work, that last one should go first (or maybe in this case be skipped), then the second one, and only then the third—and only then the information I am used to presenting first, the specifics of what we do and how it works.

Speaking of meaning—

The first Friday of Science Education Partners was on (Friday) the 13th, on Old New Year’s Eve on the Julian calendar. That night the tradition is to get together with friends and toast to the old year and then the new year, same as on New Year’s but less of a big affair/more cozy; I was back in Cambridge/Salem at that point and we went to a cocktail bar for karaoke in a small group. Then the next day was Old New Year, the last of my five holidays of the winter and the definitive end to 2022.

This winter we spent a month in Key Largo in a house along a channel, with kayaks and the ocean. The owners had a few plants outside that looked dead; over the month my grandmother and I watered them and nursed them back to health. By the end of the month they looked happy. Here they are:

What I’ve Learned About Making Scientific Posters

lydiaandreyevnakrasilnikova2 Comments

I was lucky to get to present our now published recent work, learning from last summer’s SARS-CoV-2 Delta outbreak in Ptown, at last winter’s Broad retreat. My goal was entirely to do justice to the work and the massive number of people who did that work, but to my surprise and delight we won a poster prize in the Computational Biology and Data Science category, which as you might imagine is one of the largest and most interesting categories at the Broad. I think our poster won primarily because of the project, not the poster itself, but I think the poster itself had to have also been pretty okay, at the very least clear and effective. I learned a lot putting it together, and I also learned a lot putting together other, not as astoundingly successful posters in past years and collecting my colleagues’ good advice.

Here is how the poster turned out:

The first thing I did was add the title, stretched in very large (size 85) font to cover the entire top of the poster, so that people could easily read it from far away. Our poster title was the title of our paper: Transmission from vaccinated individuals in a large SARS-CoV-2 Delta variant outbreak.

I divided the poster itself into two columns, expecting that after the title the reader would read the left column from top to bottom and then the right column from top to bottom. The left column starts with the author list and the graphical abstract side by side. The author list, which is huge, is in small-ish (size 18) font, with affiliations in even smaller font (size 12) below it. The graphical abstract is gorgeous and effective and was made by the Broad Pattern team in a process that was both extremely impressive and also very collaborative and iterative, a lot like the process of putting together the eventual journal cover with Thought Café. This project was my first time working with professional scientific illustrators, more scientific and detailed in the case of the graphical abstract and more abstract (a pun! yay!) in the case of the cover.

The rest of the poster dives into the science. Our abstract does a great job summarizing the work and our findings (the purpose of the abstract), so I organized our the bulk of the poster around the abstract, in its entirety, divided into five parts and shown in large (size 37) text.

The first sentence of our abstract covers background on the outbreak and gives context for the questions we ask in the paper. I put this sentence first after the author list and graphical abstract, on its own in the middle of the left column:

An outbreak of over one thousand COVID-19 cases in Provincetown, Massachusetts, in July 2021—the first large outbreak mostly in vaccinated individuals in the US—prompted a comprehensive public health response, motivating changes to national masking recommendations and raising questions about infection and transmission among vaccinated individuals.

The rest of the abstract presents the four takeaways of the paper, each in one sentence:

  1. To address these questions, we combined genomic and epidemiological data from 467 individuals, including 40% of known outbreak-associated cases.
  2. The Delta variant accounted for 99% of outbreak-associated cases in this dataset; it was introduced from at least 40 sources, but 83% of cases derived from a single source, likely through transmission across multiple settings over a short time rather than a single event.
  3. Genomic and epidemiological data supported multiple transmissions of Delta from and between fully vaccinated individuals.
  4. Despite its magnitude, the outbreak had limited onward impact in MA and the US, likely due to high vaccination rates and a robust public health response.

I gave each of these points its own results section on the poster, in order, in whatever way I was able to get them to fit in the two columns. The first point ended up at the bottom of the left column and the other three divided the right column more or less equally.

Each results section starts with its sentence from the abstract, in large font as a header for the section, padded by at least some whitespace above and below. The hope is that a person with not a lot of time to dedicate to the poster can grasp the main takeaways simply by reading everything that is in the largest font. In all cases but one the section’s header sentence is at the top of the section; in one case the content of the section was tall so the large-font sentence is in the top left, which is still easy to find and I think adds some pleasant variety to the layout.

Next, I pulled out the figure panels that contributed to the message of each results section. The figure panels in a section were not always from the same figures and were not only from the main figures. Here are the figure panels that ended up included in each section of the poster:

  • Section 1: Figure 1 in its entirety and Figure S1 in its entirety
  • Section 2: Figure 2A and Figure S4A
  • Section 3: Figure 3 in its entirety and Figure S5C
  • Section 4: Figure 2C and Figure S4B

I added the figure legends around the figures in very small (size 14) text, so that someone who really wants to can read them and someone who does not want to read them can easily ignore them. In a few cases I was able to fit the entire figure legend for a panel under that panel, but in most cases I had to clump the figure legends together with labels like Left, Right, Middle, Top, Left Top, and Left Inset. Like a game of Twister but for the eyes.

I did not include any figure and panel numbers or references to other figure panels, because I wanted the figures to be at home in the poster and not just visiting from the paper.

Finally, I added three QR codes, because I had seen QR codes on other posters and found them absolutely delightful and very novel. They are probably neither delightful nor novel at this point (and probably weren’t delightful or novel at that point either) but I think they were useful.

One QR code leads to our interactive tree, where you can examine the temporal and genetic relationships between our SARS-CoV-2 sequences in the context of other Delta sequences. This tree is most relevant to the second results section of the poster so I placed it there, with “Interactive tree” and a little arrow to the QR code.

The other two QR codes are relevant to the paper as a whole. One points to the preprint (the paper was not yet published in a peer-reviewed journal) and the other points to the tweetorial I wrote summarizing the paper both for people who are in our relatively small subfield of biology and hopefully also for people who are not. I put these two QR codes at the very bottom left of the poster, because that is the corner that had space. In addition to squishing them into a corner I also put a little box around them to signify that they are separate from the results section they are next to. I also labeled them: “Preprint” and “Tweetorial” with little arrows to the QR codes. (Really the arrows should point from the QR codes to the things the QR codes point to, and not the other way around, but that’s too complicated and not a hill I was willing to kill this poster on.)

And that brings us to the poster’s final form:

Here are the five lessons I learned:

  1. Information should be presented in order along the path you expect the eye to follow.
  2. The more QR codes the better. This is half-joking, but three was clearly not too many.
  3. The most important takeaways should be in the biggest font, with font size corresponding to importance.
  4. The poster should be skimmable to the reader’s preference of skimmability. I used to think there should be minimal text, and I still do, but now I think it is also important to give an interested reader something to look at more closely. But any text that is not an important takeaway should be unintimidating and easy to avoid reading if a reader does not have time or spoons to read it.
  5. I’ve seen a lot of posters that show the background, methods, results, and conclusions, each in its own section, and in most cases I personally much prefer when a poster focuses on and is organized around the big takeaways of the work. I don’t remember a lot long-term; it helps to be directed to the important parts. (Of course, it is entirely possible that the next project I make a poster for will be perfect for the background-methods-results-conclusions format and I will take this all back.)

Here is the poster as a zoomable pdf:

PTown retreat posterDownload

Here is a video of the poster printing (!). It’s silly, but it was exciting for me:

I couldn’t get my poster to send itself to the poster printer (the morning of the poster session, because as hard as I try that is the timeline I apparently work on) and a fellow Broadie helped me because that is just what the Broad culture is like, a cozy and collaborative (and absurdly smart) home.

Here is the poster on its poster board, at poster spot 2, poster spot 2 itself being a great honor:

Because of its layout, the poster was very easy to present. The poster session was in person, which was also very fun. I was able to simply gesture to and summarize the main points and describe the takeaways from the figures that led to the main points. When a new person joined I was able to quickly interrupt myself to summarize the project overall, I think along with a quick list of the main points while gesturing to them, and then jump back into the point I was presenting. I cycled through the main points as people came and went.

I was lucky that I was very much on that day; it was absolutely one of my better presentation days, and the project itself had filled my entire brain for months. I really enjoyed presenting to the smaller groups that visited our poster; it felt more like a conversation than a presentation. I liked that people were comfortable interrupting me to ask questions and I liked that I was able to think about the questions for longer than I usually feel comfortable stopping during a talk. I have noticed over the past year that I really like being able to make eye contact with people and feel like we are in conversation. The more conversational and less formal the situation feels, the better I am at presenting the work, a little correlated (R2 being maybe about 0.65) with the number of people I am talking with. I have not recently given a talk in person, but the pattern seems to hold at least over zoom and in this one in-person poster session. I’d like to learn how to bring the same energy to all my presentations, including in contexts that are more challenging or high stakes. Something to strive for and remember very fondly. This poster session was amazing. I want all my communication about my work to feel like it felt.

The Story of Our Cell Cover

lydiaandreyevnakrasilnikova4 Comments

Possibly the coolest thing that’s happened to me, definitely the coolest thing that’s happened to me professionally, is getting to be a part of a project that not only got to be in Cell, but also got to be on the cover of Cell. In this blog post, I want to tell you about the process that created that cover, because it’s very different from anything else that has happened to me and I think it went extraordinarily well.

We worked with Thought Café, whom you might recognize because they also illustrate CrashCourse videos. We worked with Julia Nadeau, Head of Production, and with amazing illustrator Eric Diotte—check out Eric’s instagram and web site:

Initial Brainstorming

A few of us (Katie S., Gage, Bronwyn, Danny, Steve S., and I) from deep in the science of the project on the Broad side of the collaboration met and brainstormed as a group before engaging the illustration team. This was a very fun meeting—we all pitched ideas for the cover and riffed off each other. After our brainstorming session, I sent an initial email to the illustration team. In this initial email, I included:

I also included a summary of our group’s initial brainstorming—

Looking at existing Cell covers, we think we’re going to want to go for something more abstract, with mostly metaphorical connection to the work and with some elements added in literally.

(Can you tell I love Escher?)

The one thing I did not include but should have included is examples of existing Cell covers that we liked. (Julia asked for links to existing Cell covers and we provided them in a follow-up email.)

Initial Sketches

Because we had brainstormed ahead of time, Eric already had sketches of the four most promising of our ideas ready for our initial meeting.

We talked through each of the four:

  • In A, a dandelion flower has SARS-CoV-2 virions in place of the seeds. Its seeds, the SARS-CoV-2 virions, blow off the flower and into the air; in the background are rolling hills of many, many dandelion flowers.
  • In B, a tree has many dense branches (like a phylogenetic tree of an outbreak) and roots made of genetic material, with SARS-CoV-2 virions growing like flowers on its branches and appearing in the air around it.
  • In C, a wave represents the different “waves” of COVID, SARS-CoV-2 virions appearing in the foam.
  • In D, colorful wildflowers shaped like SARS-CoV-2 virions represents the many introductions of Delta into the outbreak; in the background are fields of identical flowers showing that only one plant had taken over, one Delta introduction accounting for most of the outbreak.

Eric modified the dandelions live during our meeting, confining them to a bubble to convey that the Ptown outbreak was successfully contained.

At this stage the dandelions were my personal favorite because I love dandelions as a metaphor for pretty much anything, but the dandelions were getting complicated in their bubble and they didn’t neatly fit the actual messages we wanted to convey. We couldn’t decide on a favorite as a group, so we polled the covid side of the lab over slack to choose between the two top contenders, Hokusai and the dandelions. Hokusai was of course a very clear winner.

Revision and Final Version

Having chosen our overall vision of Hokusai’s Delta wave, we then went through a series of (very) fast revisions with Eric and Julia. Eric revised, Julia sent us the revisions, we discussed internally and suggested, Eric revised, and so on. This was a very collaborative and iterative process: seeing the image converge in the direction of its final form made that direction much clearer, and in almost all cases, we didn’t know what we wanted until we saw something almost but not quite there. The final, polished version looks like this:

I’m very proud of this image. Eric did an extraordinary job, and every detail is something special to us. Here is the write-up I sent Cell describing each of the details:

This cover is a reference to Hokusai’s The Great Wave off Kanagawa—in this case illustrating instead the SARS-CoV-2 Delta wave in Provincetown, Massachusetts. In “Transmission from vaccinated individuals in a large SARS-CoV-2 Delta variant outbreak,” we use genomic sequencing to learn from the first wave of the Delta variant in Massachusetts, which resulted in a large outbreak in the coastal tourist town of Provincetown. The proposed cover image shows waves containing SARS-CoV-2 virions crashing into a lighthouse. The lighthouse shines a light at the waves, illuminating the single-stranded RNA genomes in the SARS-CoV-2 virions. In this illustration, the waves represent waves of the Delta variant, while the lighthouse represents Provincetown, the site of the outbreak we studied. The light emanating from the lighthouse represents genomic sequencing. The wave that is illuminated, or sequenced, is the first wave approaching Provincetown; it represents the Provincetown outbreak, which was the first wave of Delta in Massachusetts. The Provincetown outbreak was not the largest wave of Delta in Massachusetts, and so the illuminated wave in the illustration is similarly closely followed by another, larger wave. The water to the right of the waves is calm, to signify that Provincetown had very few COVID-19 cases before the outbreak. The lighthouse itself recreates the distinct and consistent coloration and silhouette of the real lighthouses in Provincetown.

If you’ve enjoyed the cover and the process behind it, you might also enjoy the paper it’s about, “Transmission from vaccinated individuals in a large SARS-CoV-2 Delta variant outbreak,” out in Cell.

Tweeting About Your Science: My Guide/What I’ve Learned

lydiaandreyevnakrasilnikova2 Comments

All of the notable things I have ever done—being born, my weird space cow art, my long-form blogging about my mental health, my long-form blogging about underwear, probably other things—were on October 21st rapidly surpassed or at least matched in measurable viewership by 35 words, one emoji, two images, and a link.

The characters in question live in my twitter thread summarizing our study on the July COVID-19 outbreak in Provincetown, Massachusetts—then a preprint, now out in print on the cover of Cell. As I write this sentence, the first tweet of the thread has been viewed over 100,000 times, most of those views having happened within days of the tweet—which means that more than twice the population of my hometown saw at least a summary of our main findings at the moment they had the greatest potential for public health impact. I am quite delighted that the research we all poured our minds and hearts and summer and autumn into was seen by so many people so quickly. I am very impressed by how fast it spread and to whom.

Tweeting about science isn’t something I thought I was good at (and this was my first time doing it), but since it went so well I want to share with you what I’ve learned—for a long time as a consumer and now apparently as a creator. Here is my guide to writing a good twitter thread (tweetorial or tweetstorm, though I’m not a fan of either word) about your research.

Step 0: Before You Tweet

Before you start drafting your tweets, think about who your audience is and what you want your audience to take away from your work:

  1. What are your key scientific findings?
  2. How do you hope your key scientific findings will contribute to people’s everyday lives, immediately or someday?
  3. What action do you want to inspire? Are you hoping that people will alter their behavior? Are you hoping that people will use your tool? Are you hoping that people will build on something you’ve started?
  4. As you wrote the above, whom did you think of? Who are the people who are positioned to use your work immediately? Who are the other people who might also be interested in knowing about it? Who else might stumble on it?
  5. Do you have strong emotions about your work? Do you feel delighted? surprised? grateful? horrified? Is there a particular emotion you want your audience to feel? Are you hoping to delight? to surprise? to horrify? How strong is your own emotion? Does it contribute to your message or does it detract from your message?
  6. In what ways can your work be misinterpreted, misunderstood, or misused, unintentionally or maliciously? Can your work be used to fuel misinformation? Can your work be used to fuel hatred? What content did you personally find helpful in understanding how your work fits into a compassionate, good world? What content might help correct possible misinterpretions of your work?

Step 1: The First Tweet

The first tweet in your thread is the most important tweet of your thread. Most people will not click your link to read your paper and most people will not click your first tweet to read the full thread. For many people, this first tweet is all they will see of your work. It should communicate enough of your main findings and your call to action that not seeing the rest is more or less okay.

To maximize the chance that someone scrolling will read and understand it, your first tweet, more than any of the other tweets in the thread, needs to be short, be easy to understand, and not look like a block of text. Where possible, I recommend using bulleted lists and breaking up your text with plenty of vertical whitespace.

Your first tweet should include a link to the paper (though some people save it for late in the thread, as a sort of reward). If you don’t like the preview that appears when you paste in the URL, you can attach an image. I attached two images: one with the title and one with the abstract. Maybe you have a very pretty image, or a graphic that nicely summarizes your main findings—those would be nice to attach instead. We did not have a graphical abstract yet when I posted this thread, but if I were posting it today I would attach our graphical abstract instead of the images I did attach.

Your first tweet should have some indication that there is a thread attached to it. That indication can be the thread emoji (🧵) or the word thread, perhaps with an arrow (⬇), or it can be a a 1/, or it can be something else.

The first tweet also needs to do the important work of appearing in search results when people look up keywords. Try searching terms related to your work. Look at what results appear on twitter, how many results appear, what kinds of discussions are happening around those terms, and if those discussions involve the audience you are hoping will see your work. In my case, I knew that my twitter thread needed to contain the following keywords:

  • COVID-19
  • SARS-CoV-2
  • Provincetown
  • Ptown
  • outbreak
  • Delta
  • vaccinated
  • vaxxed
  • vaccine
  • public health
  • symptoms

(I don’t see any reason to use hashtags for these keywords—the terms appear in search results regardless of whether or not they have a hashtag in front of them, and having that many hashtags just seems a bit alarming. But I could be wrong.)

These 11 words account for almost a third of the 35 total words of the tweet. I made sacrifices to ensure that they all appeared. (Vaccinated, for example, is much longer than vaxxed, and Provincetown is much longer than Ptown—but they each produce their own search results and I wanted our paper to appear in all of them.)

If you would like and if it comes naturally, you can also (or alternatively) include in your first tweet some kind of a hook—a funny joke, a cliffhanger, an unanswered question, or content that elicits an emotional response. With a hook people will be more likely to click to read the details, but they are also less likely to walk away understanding your work if they don’t click. If the thing that is important to you is that as many people as possible walk away with some understanding of your work, you probably want to summarize your work in the first tweet. If you would like to connect more deeply with a smaller number of people, rather than shallowly with a large number of people, it might be better to use a hook. (I of course did not do that.)

Finally, I think it is important to include yourself in the work. The science is done by people and people like other people. I start the first tweet with Our and many of my other tweets in the thread start with We. I am a person who likes exclamation marks so I include exclamation marks—because exclamation marks best capture how I feel about this paper.

Things to Know

Here are seven general principles to keep in mind as you draft your twitter thread.

  1. I recommend you draft your tweets in Google Docs or similar, and that you get feedback from your co-authors and revise at least once before posting. This thread took me three days with feedback from multiple co-authors: Katie, Bronwyn, Gage, and Pardis all helped me revise this thread and made it much better. (In its original form, it didn’t even have capital letters.)
  2. Each tweet should be its own complete thought. If a person retweets one tweet from your thread and not any of the others, its message should be clear. It should not be possible to take any individual tweet out of context to mean something you did not intend.
  3. Where possible, and especially when making important points, make your text easier to digest by breaking it into bullet points or breaking it up with vertical whitespace.
  4. If you include a url in your tweet, your tweet will have a nice link preview. If you attach any images to your tweet, the link preview will not appear. The link preview in the final, posted tweet will look the same as it looks in the draft tweet before you post.
  5. If you end your tweet in a url (if the url is the very, very last thing in your tweet) and you have not attached any images, the url itself will not appear in the tweet when it is posted (but the link preview will).
  6. Most people do not click on images. You should size your images so that the preview that people see on twitter looks the same as the image itself. If you are attaching one image, it should be 1500 pixels wide and 850 pixels tall (or scaled with that same aspect ratio). If you are attaching more than one image, all of your images should be square. The final cropping of the images does not always match the preview of your tweet draft, nor is it consistent between devices. If it is important that people see the entire image, including the edges, add some white space all around your image so that the content of the image will still be visible even the edges are slightly cropped. (If you do not have image editing software, you can download and use GIMP for free.) I do not recommend attaching more than two images—the previews will be too small to see without clicking, and most people will not click.
  7. Create a secret fake twitter account to test out your tweets before posting them for real. Delete your test tweets immediately. Do test post the actual images you intend to post to make sure they appear as you expect them to appear and are cropped as you expect them to be cropped. Do not link to your work or post any of the actual text you intend to post—not even for a moment. Other than images, replace your actual text with dummy text and your actual urls with dummy urls when you test post. You can also use this account to draft your actual tweets (without posting them) to make sure they fit within the character limit—don’t rely on other applications to count characters; they all seem to count in their own ways.

Step 2: Context, Background, Collaborations

This part is easy. I initially had this tweet closer to the end, but my co-authors encouraged me to move it to the front and I think it is much better here. We start off by acknowledging the organizations involved in this study, because the list was huge. If it were a much smaller study, I might start off by acknowledging the individual authors. It is also helpful to start off with some context—has this work been peer reviewed? Does this work build on previous work? Does this work build on anything familiar to the audience?

If someone has clicked your first tweet to read the thread, they are already in and willing to skim. Your second tweet does not have to be very exciting.

You might notice that while this tweet includes an attractive preview of the link, it does not include the actual url. That is because the url is pasted after the 2/. Because the url is the very last thing in the tweet and there is no image attached to the tweet, the actual url does not appear.

Step 3: Paper Summary

The paper summary will likely be the bulk of your thread. As much as possible, each tweet should include one complete and clear idea and be retweetable on its own. Where relevant, you should include figures from the paper. Try to add some emojis. You can also include links to previous work or work that you use or build on. I start many of these tweets with We.

The order in which ideas appear and the ways you present those ideas might not match their order or presentation in the paper. Your audience on twitter is almost definitely broader than the audience you wrote the paper for, and your language and narrative should adapt. Your focus should be on communicating the findings that are most relevant to this broader audience in a way that is understandable (and not condescending), with minimal jargon. In some cases you might want to delve into more detail, and that’s okay—tweets including jargon should be easy to skim and easy to grasp at least the purpose of without extensive background in your field and should not be necessary for understanding the overall story.

Step 4: Findings Summary

Your paper summary should end in a tweet summarizing the findings that are most relevant to your audience, or most actionable.

Step 5: Your Findings Out in the World

This is the hardest part.

This section is about the role you hope your work will play in the world, and it is a part of your efforts to shape that role. Here, you need to anticipate the ways in which your work could be used to misinform or hurt. You should address possible misinterpretations explicitly and head on. You should provide links to informative content people can engage with to address possible misinterpretations before they arise. Wherever possible, that content should be readable by people outside your field.

If you are less worried about misinterpretation or misuse of your work, you can instead or additionally use this space to explore the life you hope your work will live in the world—how you hope people will use or build on your work, or your own emotional response (if you have one you’d like to share) to your findings.

In either case, it is nice to end with a positive or hopeful feeling before your call to action, because hope is empowering.

Step 6: Key Takeaways/Calls to Action

Finally, at the very end, I have the final takeaways (the third time I am adding takeaways—but that is what this thread is). These takeaways do not relate to the content of the paper itself; instead, they are focused on how I hope the paper’s findings will be useful in the real world. I end by making explicit the calls to action I tried to imply in the first tweet and have been building toward throughout the thread.

Step 7: The End/Thank Yous

This part is fun (and easy). Thank your co-authors—the first and last authors first, if the full author list is very long, and the full author list listed out or in summary. If anyone is on twitter, tag them.

When you are ready to post, use twitter in a web browser on your computer (not the app on your phone) to prepare the thread in advance and then post the entire thread at the same time.

The time you post your thread is important. Don’t tweet when your target audience is having dinner or asleep. I think it’s probably best to time your thread for when your target audience is browsing twitter at work, but there are tons of more informed articles on timing of tweets that you can read instead of using my guesswork.

Finally, the number of people twitter shows your tweet to is determined not only by the number of likes and retweets your tweet gets, but also by the speed at which it gets them. Tell your colleagues and co-authors about your tweet immediately after you tweet it. Hopefully they will engage with it quickly and help it spread.

That’s all! I hope this guide helps your work reach its audience.

If you would like, you can view (and retweet!) this thread in its home on twitter and read the paper the thread is about. You can also read an article about the power of twitter to disseminate scientific research, which I found on twitter.