Reading and Understanding a Scientific Paper

March 11, 2019 4:49 PM

Science journalism isn't that great. The Non-Scientist’s Guide to Reading and Understanding a Scientific Paper from Elysium Health is here to help! "To illustrate the process, we’ve chosen a popular scientific paper published in 2015 in Proceedings of the National Academy of Sciences, “Evening use of light-emitting eReaders negatively affects sleep, circadian timing, and next-morning alertness.” You can access the paper for free here."

And a lighter take from Science including the steps optimism, fear, and regret.
(Previously)

posted by Anonymous (12 comments total)

But what if you use special blue filter glasses to read the kindle? Need more research!
posted by sammyo at 4:56 PM on March 11, 2019

The key problem is that most scientific papers are just genuinely badly written, even taking into account the constraints of the genre (dense technical language, authors usually not native English speakers, and so on).
posted by vogon_poet at 5:04 PM on March 11, 2019 [2 favorites]

This is pretty good advice! I sort of cobbled it together through years and years of reviewing literature (and it varies depending on discipline, e.g. the social sciences where the literature review is an important part of the research since it synthesizes theories together and tells you what the paper is speaking to), but it's cool to see it all together in a single layman's article.
posted by codacorolla at 5:14 PM on March 11, 2019

I was disappointed to see that they didn't mention the part about removing large chunks of explanatory text to fit a page limit. Just typing that is giving me flashbacks to trying to reconstruct someone's math in a journal where the authors apparently had to remove four of every five equations. I'm looking at you, Physical Review Letters. As bad as Fermat's margin quote.
posted by kovacs at 5:21 PM on March 11, 2019 [14 favorites]

This is a pretty good description of how to read an individual scientific paper, and it’s helpful in the problem the article describes of following up on a news article that references the paper.

However, I wish the post included some guidance on how to put that paper into context. Any single result may not turn out to be true, a working scientist reading a paper that says something controversial will go looking for related studies and try to figure out how it fits in. If you follow up on a clickbait headline to find a slightly-less-flashy paper that claims something weird, you should also take the time to figure out if the rest of the field agrees (and if not, why not).
posted by a device for making your enemy change his mind at 7:08 PM on March 11, 2019 [6 favorites]

I work at a university writing center. When I talk to anyone (grad student or faculty) who is about to submit a paper, I remind them what matters isn't how smart they sound but rather what their smart ideas are. (And just because everyone else writes incoherently doesn't mean they should too.)
posted by astapasta24 at 7:59 PM on March 11, 2019 [6 favorites]

If you follow up on a clickbait headline to find a slightly-less-flashy paper that claims something weird, you should also take the time to figure out if the rest of the field agrees (and if not, why not).

I agree, but this is also something that's really hard for a non-scientist to do, let alone to explain the skills and techniques behind it, when you don't have institutional access to papers, don't know the terminology in the field to use as search terms, don't have many tools to view citations (Google Scholar helps here), don't have ready access to research librarians, and only have so much time to devote to the task without doing a full-on literature review. This is really a place where the publicly-accessible tools need to get a lot better and where science journalism can play a role in figuring out how to make this process more accessible.
posted by zachlipton at 8:12 PM on March 11, 2019 [7 favorites]

The duller the writing style is, the easier it is to defend.
posted by ovvl at 10:05 PM on March 11, 2019 [1 favorite]

I was disappointed to see that they didn't mention the part about removing large chunks of explanatory text to fit a page limit.

Ahahaha, I just edited a paper for publication that had to be cut by 1,000 words, or 1/5 of its length. It was still far over the limit after working some editing magic, so I sat down with the author (a friend of mine, luckily) for a few hours (that I didn't charge her for) and we cut down the whole paper together. In the end, it worked well and the paper is coherent, but I can see how an author who doesn't have a relationship with their editor or even (*gasp*) no editor at all would be tempted to just cut stuff willy-nilly. Result: chaos.
posted by lollymccatburglar at 12:33 AM on March 12, 2019 [3 favorites]

Ugh, yes, length restrictions… To say it in the words of a coauthor, “We did it! The paper is now empty.”
posted by wachhundfisch at 3:17 AM on March 12, 2019 [7 favorites]

Supplementary Material is where you put everything you had to cut from the main paper because of length restrictions. It also has the advantage that there are rarely any formatting requirements for Supplementary Material so can just cut and paste into it any old how.
posted by drnick at 4:18 AM on March 12, 2019 [2 favorites]

Indeed. We almost look at the paper now as an extended abstract for the real meat in the Supplemental. Though some idiot journals are over-limiting those now too. Why? pdf electrons are (almost) free.
posted by bonehead at 5:06 AM on March 12, 2019 [3 favorites]