The Non-Scientist's Guide to Reading and Understanding a Scientific Paper

Key Takeaways:

• More than 3 million new English-language scientific papers are published each year in over 30,000 peer-reviewed journals.

• While many are paywalled, there are also prestigious open-access journals where you can read articles for free.

• Reading articles will help you make more informed decisions in the areas of life that concern you, and better understand and participate in the public debate about important scientific issues.

• Here are the basic steps: focus on the big picture the scientists are addressing; read the Abstract, Introduction, and Discussion, in that order; think critically about the conclusions the scientists make; and conduct follow-up research.

• For practice, we provide a link to a popular scientific paper on light-emitting e-readers.

We live in a golden age of scientific research. The top five countries in scientific research and development—the U.S., China, Japan, Germany, and South Korea, respectively—spend over $1 trillion on it each year. But where do all the resulting discoveries and eureka! moments go? Eventually, they may find their way into textbooks or form the foundation of a life-saving therapeutic, but first, most of them go onto the page, in a scientific article.

According to the International Association of Scientific, Technical, and Medical Publishers, over 3 million new scientific papers are published each year in more than 30,000 verified journals that use the stringent "peer review" system, whereby multiple scientists who are specialists in the relevant field of study provide a critical and in-depth review of a new paper. The process takes months and is overseen by a journal editor and several reviewers who read the study; only once the editor is convinced the author has addressed notes from peer reviewers in such categories as originality, importance, manner of presentation, and critical flaws, is it accepted into a journal.

The most common form of a scientific article is a primary research article, an original report of research that chronicles an experiment in such a way that it can be replicated and the results reproduced by other scientists—core tenets of the scientific method. (Another type is a review article, where several primary research articles are discussed and their findings placed in greater context.)

The internet has made disseminating that sea of information easier than ever, one result of which is a deluge of "A Groundbreaking New Study Finds…" headlines on websites and in magazines. While overly reductive reporting on scientific breakthroughs is not new to the media —the general public has, at times, felt whipsawed by science and health reporting for decades—the total quantity of scientific research and media sources is only increasing. More often than not, those articles grab the biggest "finding" and speculate on what it might mean a decade into the future, which often falls short of explaining the study in context and helping readers form an educated understanding of what the scientific research actually showed.

So how do you distinguish hyperbole from scientific evidence? By reading the papers yourself.

That's not easy, even for scientists, who readily admit that reading these papers can be akin to torture. (An article in Science, enumerating the steps of reading a paper included "fear," "regret," "bafflement," "distraction," and "rage.") Rather than charging headfirst into several thousand words of science-speak, follow this plan of attack for primary research articles. With a little practice, you can do more than just understand them: you can replace conventional wisdom with knowledge, make more informed decisions in the areas of life that concern you most—health, fitness, and diet, for example—and better understand and participate in the public debate about important scientific issues.

Editor's Note: 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.

1. Locate the Article

You'll likely begin your search one of two ways: either by tracking down a paper cited in a news story (in which case: Google), or by searching for papers on a topic that interests you. The highest-profile sources of peer-reviewed articles are Nature, a British journal, and Science, its American competitor. There are credible journals, however, for every branch and sub-branch of science. The best place to find those is on PubMed.gov, a database holding more than 36 million citations to credible journals.

Note that many scientific papers are not available for free. As of 2022, only 35% of articles are “open access” (free, immediate, online availability of the entire article with no access fees). Paywalls are a source of frustration for a large swath of the general public (including academia, industry, and media), who argue that open access to research hastens innovation—and indeed that the public has a right to access the research it funds with tax dollars. This friction between publishers and everyone else has given rise to a variety of responses, including open-access journals, the search engine Sci-Hub, and policies from funding sources, like the Bill & Melinda Gates Foundation, that require research to be disseminated to the public free of charge.

Before You Start: The Anatomy of a Primary Research Article

2. Wrap Your Head Around the Big Picture

Read the Abstract for the bottom line of the study and then dig into the Introduction. While the Abstract can be dense prose, the Introduction is where the authors provide big-picture

context for the study and explain previous research in the field, often in a narrative style that’s easy to follow. It's helpful to imagine each scientific paper as a single chapter in a long novel; this section is where you’ll gain an understanding of what chapters came before. It’s also where the authors broadly explain their hypothesis—or, where they're guessing their own chapter will lead.

You want to come away from the Introduction with an understanding of the central problem the particular field of science is dealing with, as well as a subset of questions (related to the problem) that the authors of the paper plan to address. They will offer a hypothesis about the answer to those questions and suggest a plan of attack—the experiment(s)—for investigating whether it's true.

In the e-reader article, the central problem is, How significantly are human sleep cycles, and therefore health, affected by technology? The authors explain that the “previous chapters” in this field have discovered that a circadian timing system directly affects when and how we sleep and that the circadian cycle is directly affected by exposure to light in the early evening and nighttime, which suppresses the release of melatonin, a hormone that tells our bodies to sleep; the resulting lack of melatonin acutely increases alertness. The specific questions posed in the paper are: How does the light of an e-reader's screen affect the circadian rhythm of test subjects; and how does that affect the quality of their sleep and their alertness the next day?

3. Skip Ahead to the Conclusion

The Methods and Results sections come next, but skip ahead to the Discussion, alternatively called the Conclusion or Interpretations, which will summarize the findings—at least what the authors think they have shown—in a digestible way. This section includes how the authors interpret the data from their experiment and what it means for their original hypothesis. At the very end, they will speak to how their findings change our understanding of the bigger picture, those surrounding chapters in the novel chat make up the scientific field's entire progress. Keep in mind that scientific articles have limitations, often acknowledged in the paper, which require further research.

The scientists behind the e-reader study address both the larger problem and their unique questions in their Conclusion. "These results indicate that reading an I.F. (light-emitting) eBook in the hours before bedtime likely has unintended biological consequences that may adversely impact performance, health, and safety," they write. Specifically, based on their data, they suggest that beyond just disrupting single instances of sleep, the light from e-readers may lead to a pattern of sleep deficiency by delaying the circadian timing system, which in turn reduces the REM sleep that would otherwise happen closer to waking.

They describe a harrowing big picture given these findings: teenagers are spending upwards of 7.5 hours a day consuming media on readers, phones, and computers, much of which happens in the evening and nighttime. The authors are especially concerned because studies of night shift workers have found a relationship between chronically suppressed nocturnal melatonin release and colorectal, breast, and prostate cancer.

4. Understand the Results

The Results section and its attending figures and tables present the data without interpretation from the paper's authors. As a result, this section will usually be the most difficult for the non-scientist.

A few statistics terms will help you navigate the data: "significant” and “non-significant.” This basic statistics terminology is used by scientists to describe events that could be the result of random chance (“non-significant”) versus data that could represent meaningful discoveries (“significant”). If you're not familiar with statistics, consider reading a primer like this one from Harvard Business Review. Be aware that the Results section (and Methods) is sometimes challenging even for other scientists who are reading outside their field of expertise.

Pay attention to the figures and charts, which pack a lot of clear information into visuals. Read the captions closely, since they tend to explain the results with simple, clear language. And circle back to the simplest questions: What is being measured? Given what you read in the Introduction and Discussion, what does the data illustrate?

The e-reader study includes several results from its experiment that support the interpretation in the Discussion section, all fairly easily understood using the figures and captions. For example, after five consecutive days of reading using electronic light, the e-reader group's melatonin onset levels were significantly suppressed in the evening when compared to the group reading traditional books. Those using the e-reader also took an average of nearly 10 minutes longer to fall asleep than those in reading books. And those using e-readers had significantly less REM sleep on average.

***

Assessing the Subject: Cells, Model Organisms, and Humans

A small but important detail: What are an experiment's test subjects? There's a big difference between the implications of results found studying model organisms (non-human species) or in human cells ex vivo (outside the body) versus results discovered in clinical studies (humans)—one that articles in the news usually don't usually mention in the first few paragraphs.