UPDATE 9/29/2023:

Now published in JAMA

See also this blog

and summary thread

Here’s a protocol for a study of visual abstracts. I submitted to, who (correctly) declined to post it as it doesn’t meet their definition of human participant research. The study is already under way; overall getting it registered hasn’t been a top priority as this is a relatively small study on how VAs may affect journal site traffic and social media metrics. The main goal in posting this here is to show we have not cherry-picked outcomes or done statistical data torture after collecting our results.