My Authors
Read all threads
Designing better figures for small studies: Why you shouldn’t use bar graphs for continuous data and what to do instead (A Thread in Q&A format, with figures from our new @CircAHA paper)

journals.plos.org/plosbiology/ar…
“Can I still use a bar graph if my data are normally distributed?”
jbc.org/content/292/50…
“What should I use instead of a bar graph?”

It depends…
ahajournals.org/doi/pdf/10.116…
“Do I need expensive software to do all this?”

No – there are many free tools that make it easier to create more informative graphics.
See table 3 for a list of tools: ahajournals.org/doi/pdf/10.116…
“I still prefer bar graphs because they convey a clear message. Sometimes it’s hard to see what’s going on in dot plots.”

Emphasize the data points and de-emphasize the dots to convey a clear message while allowing readers to critically evaluate the data
ahajournals.org/doi/pdf/10.116…
“Can I just add dots to my bar graph?”

This is better than a bar graph, but it won’t be as clear as a dot plot.
“What is a histogram with dots and why should I avoid this type of figure?”

Upload irregularly spaced data to determine whether your software or code is making a dot plot or a histogram with dots.
“When is it appropriate to show summary statistics like error bars, box plot and violin plots?”

Summary statistics are only meaningful when there is enough data to summarize
ahajournals.org/doi/pdf/10.116…
“Why shouldn’t I use box plots for bimodal data?”

Box plots conceal the two peaks. Bimodal distributions are easier to see with dot plots (smaller samples) or violin plots (large samples)
“How do I make effective dot plots with box plots or violin plots?”
ahajournals.org/doi/pdf/10.116…
“Doesn’t the y-axis always have to start at 0?”

Always consider the purpose of the visualization and properties of the data when designing your figures.
“Does it matter how I did my stats?”

Yes – The structure of the figure gives the reader information about your experimental design and analysis. Avoid sending mixed messages
ahajournals.org/doi/pdf/10.116…
“How do I design figures that match my study design & analysis?”

For simple analyses of small datasets, it’s often best to show one graph per analysis. The graph should include all groups, time points, or conditions from the analysis.
ahajournals.org/doi/pdf/10.116…
“What about statistical reporting…?”

See our @elife "science of science" paper: Why we need to report more than “Data were analyzed by t-tests and ANOVA” on how to avoid problems caused by standard reporting practices. Clear reporting benefits everyone.
elifesciences.org/articles/36163
“What can I do to encourage others to use more informative figures in papers, talks and posters?”
“What can I do to increase the likelihood that authors will improve their graphics when I request more informative figures as a reviewer or editor?”

Provide citations to references and resources that will help your colleagues to understand the problem and make better graphics
This thread is from our new online course on identifying and fixing common problems with visualization and stats in small sample size studies. If you liked this thread, please check out the rest of the course at bit.ly/SampleSizeMatt…
Missing some Tweet in this thread? You can try to force a refresh.

Enjoying this thread?

Keep Current with Tracey Weissgerber

Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!

Twitter may remove this content at anytime, convert it as a PDF, save and print for later use!

Try unrolling a thread yourself!

how to unroll video

1) Follow Thread Reader App on Twitter so you can easily mention us!

2) Go to a Twitter thread (series of Tweets by the same owner) and mention us with a keyword "unroll" @threadreaderapp unroll

You can practice here first or read more on our help page!

Follow Us on Twitter!

Did Thread Reader help you today?

Support us! We are indie developers!


This site is made by just three indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3.00/month or $30.00/year) and get exclusive features!

Become Premium

Too expensive? Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal Become our Patreon

Thank you for your support!