Chapter 3 Composing tweets
The most common question I get at social media workshops that I’ve run is, “What should I tweet?”. There is no easy answer to this, because every researcher and their subfield is different. But regardless of your subfield, the best way to engage your followers is to either entertain or educate. In other words, you should aim to help people either pass time or save time. Being entertaining doesn’t come naturally to most people, so don’t worry if this isn’t you. But as a scientist, you are very well-placed to educate, no matter your level of training.
Thanks to Chapter 1, you’re up to speed with the general mechanics of Twitter. You also just learned about the “entertain or educate” principle, with any scientist ready to educate. But even when arriving at this point, many scientists hit a wall, because they believe they can only tweet the finished product of their work, which is typically papers. But as soon as you realise that the process of your work is just as interesting as the output, then things become much easier. You’re going to see a few examples of this below.
When it comes to tweet frequency, it’s hard to give any firm recommendations. Whether you’re tweeting too much depends on how many people your followers are following, and how much they tweet. So, this means that the “too much” threshold is different for everyone. In my opinion, I think it’s fairly difficult to tweet too much. In the early days of Twitter (or any social network), where users didn’t follow that many people, it was pretty easy to flood someone’s timeline. But now, people tend to follow hundreds (sometimes thousands) of accounts, so this is harder to do. I think the upside of more tweets outweighs the downside a few people unfollowing you because they think you tweet too much. You never know which tweets other people will find interesting.
Some example tweets
3.0.1 Tweeting about your own research papers
This is one of the most common tweets you’ll see from scientists who aren’t very active on Twitter. When a new paper is published, they’ll log on, post the title of their paper with a link to the paper, and then log off until their next paper is published. There are so many more types of tweets that you can compose, but I’m going to walk through how you can lift your game with these types of tweets, which are one of the main reasons many scientists are on Twitter.
1. Add an image from the paper to go with the tweet. Have a look through the paper to see if there’s a striking image that can be used. If there are no figures in your paper, you can just take a screenshot of the abstract or a particularly interesting part of the paper. Skitch is a handy app for screenshots, as you can easily annotate and highlight images. If you’re on your smartphone, you can also just take a screenshot and crop the image to your needs. You can add up to four images per tweet, so if you have more than one striking image you can take advantage of this
2. Add a quote from the paper. Find a striking quote from the paper to include in your tweet. You can either write this as text, or take a screenshot of it and then highlight the quote, using Skitch or other software.
3. Tag your co-authors and the journal. Co-authored papers are a team effort, so you should acknowledge your team. In addition, this gives your followers the chance to find more people to follow. If you have a lot of co-authors, remember that you can also tag authors in images as well. This means that your co-authors will get notified of the tweet and others can see who else is involved in the resesarch.
4. Add your own commentary of the paper. If you’re not including a quote, you should share why you think the paper is interesting. Maybe it’s your first paper or it’s part of a project you’ve been working on for a long time.
5. Add a link to a non-paywalled version of the paper. People might see the tweet, but not bother clicking on the link if they do not have institutional access to the journal. However, if you include a link to a preprint or a postprint1, then people know they can easily access the paper.
In the following example of a paper I co-authored, I mention a summary of the results, highlight the sample size as this is a notable feature of the paper, tag my co-authors who are on Twitter, include a link to a postprint of the article, and attach a figure from the article.
New paper! Intranasal oxytocin abolishes the self-referential bias in terms of reaction time (n = 170), replicating prior results https://t.co/3m9TxBIYis Great work from Weihua Zhao, @bn_becker + Chengdu team and @ChrisMontag77. Non-paywalled version: https://t.co/lzQ1wknUQ1 pic.twitter.com/sKd1p3FAyU
— Dan Quintana (@dsquintana) August 13, 2019
Here’s another good example, with @GuyProchilo sharing his first first-author paper. Notice that Guy tags his co-authors, the journal, adds a hashtag (#IOpsych) relevant to his research field in this tweet (Industrial-organisational psychology), and includes a striking image of the paper.
I’m excited to announce my first first-author paper - published in @Meta_Psy!
— Guy Prochilo 🏳️🌈 (@GuyProchilo) November 20, 2019
“An Extended Commentary on Post-Publication Peer Review In Organizational Neuroscience”: https://t.co/VmbZvgcxE8
Coauthored by @WLouisUQ, @DLabMelbourne, @HannesZacher, & @Social_Neuro_M#IOpsych pic.twitter.com/g8tiY9xp6V
While single tweets announcing new papers can be an effective way to share your new work, I think a thread, which are a series of connected tweets, does a better job—see Chapter 4 for how to make an effective thread to introduce your new work.
3.0.2 Tweeting about other people’s research papers
There are only so many articles that you can co-author. So in addition, you should also share papers that you find interesting. If you haven’t already, you should set up new publication alerts using keywords that are relevant to your work. Pubmed and Google Scholar are popular options for setting up paper alerts. The advantage of Google Scholar is that it will also capture new preprints and theses.
Below is an example of tweeting a link to a new paper, in which I added an image and quote from the paper, and mentioned the lead author and the journal that it was published in.
Effects of route of administration on oxytocin-induced changes in rCBF in humans https://t.co/Ibfv97VZ3T
— Dan Quintana (@dsquintana) March 3, 2020
“....we also provide robust evidence confirming the validity of the intranasal route to target specific brain regions” Great work from @DanielArturAM + team in @NatureComms pic.twitter.com/J8H97nkgBT
3.0.5 Asking a question
Twitter can be a great way to get advice. If you don’t have many followers, you should also consider tagging experts that might know the answer (but don’t spam people, so use this sparingly). Here’s an example that I used, which was also used for researching this book.
I’m putting something together on twitter profile/bio pages for an upcoming talk.
— Dan Quintana (@dsquintana) March 3, 2020
What do you think makes a ‘good’ twitter profile, when it comes to images, text, web URL etc...? What makes you think, “Yeah, I’m going to follow them...”?
This book is written for both HTML (online) and PDF (print) output. This is a book about Twitter, so I use a lot of tweet examples. One benefit of an online version is that you can embed tweets, which means that readers can visit the tweet’s original source and read any replies to the tweet. It’s also easy to see at a glance how many people have engaged with the tweet, as these numbers are attached to the bottom of embedded tweets. However, converting a HTML element into a image for a PDF is not as straightforward as you would think, as you essentially have to take a screenshot for each the tweet. To avoid this extra work, I put a call out on Twitter:
{bookdown} folks: I'm trying to knit a PDF version of a HTML book that contains HTML elements (embedded tweets).
— Dan Quintana (@dsquintana) June 24, 2020
Is there a way to automatically take a screenshot of embedded tweets for PDF output?
Using the {webshot} package + PhantomJS didn't work...#Rstats
Fortunately, Garrick Aden-Buie (@grrrck) found this tweet and updated his ‘tweetmd’ R package, which solved my problem.
Thanks to prodding from @dsquintana, I added `include_tweet()` to {tweetrmd}. Automatically embed the HTML tweet in HTML #rstats R Markdown docs or a screenshot in other formats, like PDF.https://t.co/DIyuQgGmvz pic.twitter.com/v4JNipxXBr
— Garrick Aden-Buie (@grrrck) June 25, 2020
3.0.6 Answering questions
A great thing to tweet when you have have nothing to tweet is to help other people by answering their questions. There are several hashtags that are often used by scientists to ask questions, such as #AcademicChat and #PhDchat. What might seem like an easy solution to you may save someone else hours of work.
3.0.7 Replying to tweets
One of the fastest ways you can build connections on Twitter and to establish yourself as an expert in your area is to reply to other people’s tweets. This seems counterintuitive, as one would think that conventional tweets are a better way to achieve these goals, but replies are almost guaranteed to be seen by the person you’re replying to, whereas conventional tweets may not be seen at all (or quickly scrolled past), especially if you’re new to Twitter. If others are following both yourself and the person you’re replying to, they’ll see the original tweet and your reply in their feed.
Document your experiences
I hardly ever sit down and think, “what will I tweet?”. Most of my tweets are simply offcuts of work I’m already doing, or things I’m already discussing in real life with colleagues.
For example, I was recently having a chat with some colleagues about grant applications. I told them a principle I used to write my applications, which they found useful. So I turned this into a tweet, which also seemed to resonate on Twitter.
Every grant application introduction I write indirectly answers these four ‘why’ questions:
— Dan Quintana (@dsquintana) February 29, 2020
1. Why this?
2. Why me?
3. Why here?
4. Why now?
This thread is another example of “showing my work”. I didn’t have to think of something new to tweet, as I essentially copy and pasted my reply email.
Got an email from a new grad student asking for recommendations for resources to better understand research design and statistical inference. Here’s what I’m going to tell them...
— Dan Quintana (@dsquintana) July 14, 2019
Peer-review is a task that most of us do, so you can take this opportunity to make a wider commentary about your research topic. For instance, I recently shared a point about power analysis using a peer-review I was conducting.
Reviewing a paper with no power analysis, so I did one for them!
— Dan Quintana (@dsquintana) March 4, 2020
The main outcome was tested using a paired-samples t-test, with 25 participants in total. So I plugged these numbers into the @jamovistats 'jpower' module and got these results ⤵️
1/2 pic.twitter.com/ImOmhhTMDI
Let’s break down this tweet:
- I used a line break to clearly separate the two parts of this tweet.
- The screenshot image is eye catching (especially the plot) and conveys a lot of information.
- The ‘down arrow’ emoji adds a tiny bit of flair.
- I tagged a tool that I mentioned @Jamovistats, so that followers could easily find out more.
- I numbered this tweet “1/2”, to signal that this is a two-part tweet. My followers will see these two tweets back-to-back, but people who see this tweet as a retweet will now know there is more than one tweet on this topic, and can then check the full conversation.
Of course, if you’re going to tweet about peer review you will need to keep things as anonymous as possible! Also consider that the author of the paper may follow you on Twitter or come accross your tweet if it gets retweeted.
You can also tweet about being at the other end of the peer review process.
Reviewer 1: Expand the third paragraph of your discussion as it’s critical for your paper
— Dan Quintana (@dsquintana) October 16, 2019
Reviewer 2: Delete the third paragraph of your discussion as it weakens your paper pic.twitter.com/t6OPE5MWUx
Remember, the key is sharing the process of your work, not just the final product.
A preprint is a version of an article that is posted to a preprint repository before peer-review. Most journals allow you to submit papers that have been posted as preprints and do not consider this dual-publication. Check your journal’s policies. A postprint is a version of paper that has undergone peer-review but has not been typeset. These are typically posted to the author’s personal or institutional website. Postprints are usually just a PDF version of the final Word document that was sent to the journal. Unlike preprints, journal policies vary regarding postprints, so it’s best to check before uploading one.↩︎