Guide to reading scientific literature

Here’s my recommended approach in a nutshell:
Before you jump into a paper, you need to know what information you’re looking for. First, identify your questions. Then, actively seek out answers to them by studying the tables and figures and consulting the text as needed.

Scientific papers are not textbook chapters

Here’s the #1 mistake that students make when they are reading scientific primary literature (original research articles): They try to read the paper like a news article or a textbook chapter. That is, they start at the beginning and read the words in order until they get to the end.

This is a great way to end up confused and discouraged. What commonly happens is that students spend hours slogging through a paper only to realize, at the end, that they still have no idea what the paper was about. Then they show up to class unprepared to discuss the reading, despite having spent hours trying to prepare!

Here’s the thing: Scientific papers are not designed to be read like textbook chapters. In many ways they are more like reference manuals. They have a standard format and organization, and they are designed to allow readers to skip directly to the parts that they are interested in. Yes, you can—and sometimes should—skip over certain parts of a paper. This is not a short-cut, it’s actually the way you are supposed to read this literature! (It’s what many experts do.)

So: When you have a scientific paper to read, don’t just jump in without a plan. Decide what information you’d like to find out from the paper, then go find that information.

If you’re used to reading papers more passively, this active approach—where you are not exactly reading, more like extracting information—may feel different. But with practice, you’ll see that for scientific papers, active information-seeking will produce better and faster comprehension than passively reading.

How to extract information from scientific papers

What kind of information should I be seeking?

That depends on your goal. In this guide I will assume that your goal is conceptual comprehension: you want to understand the basic findings of the study and how these were discovered. See the end of this document for how your approach should change if you have additional goals (e.g., if you are reading a paper for a graduate seminar or to learn about the study’s methods so that you can plan your own study).

If your goal is conceptual comprehension, you will need to find answers to the following questions:

What specific question did the researchers have?
What happened in the study? (Who were the participants, and what did they do?)
What was the dependent variable (the thing that the researchers measured)?
What were the independent variables (the things that the researchers changed or compared)?
What were the results? Specifically: How did the independent variable(s) change the measurements of the dependent variable(s)?
What was the answer to the researchers’ question?

(No idea what a dependent or independent variable is? Watch this 7-minute YouTube video for an explanation.)

How information is organized

Scientific papers have information in two places: (1) the text, and (2) the tables/figures. Contrary to what many students think, the tables/figures are most important. If you do not understand what is shown in the figures, you do not understand the paper. If you are short on time, skip the text entirely and just study the figures!

The text of a scientific paper is typically organized into the following sections:

Abstract: Brief summary of the paper, but not the whole story; watch out for authors’ spin!
Introduction: Gives context for the study, reviews related studies, and motivates the present study. Tip: The last paragraph of the Intro will usually tell you what the authors’ specific research question is.
Methods: Describes the design of the study (including the dependent and independent variables) and the specific procedures used
Results: Factual and concrete; tells you what the measurements were
Discussion/Conclusion (sometimes these are separate sections): Here the authors tell you what they think their results mean. You may or may not agree!

But again, the tables and figures are the first and most important places where you should look for information. The Methods section often has figures or diagrams showing how the experiment worked, while the Results section should include a figure or table (or several) showing the measurements of the dependent variable(s) and how these changed when the researchers manipulated other variables.

My recommended process

For achieving conceptual understanding, here is my recommended process:

Look at the title page: authors, journal name, date. The goal here is to be able to put the paper in context, so you can understand it better.

What is the authors’ discipline/subfield (e.g. psychology, linguistics, neuroscience)?
Who is the paper written for? You can often tell from the journal’s name! Shorter journal name = less specialized audience. This will affect the types of issues that the authors focus on, and how much detail they include.

Very general audience (Science, Nature, Proceedings of the National Academy of Sciences) – these journals are also very prestigious, so papers published here are usually very rigorous and well-controlled

Slightly specialized audience (short journal name like Cognition or Language)

Specialized audience (longer journal name like Language Learning and Development, Infant Behavior & Development)
When was this written? Is this a famous, classic paper or a newer (and perhaps more controversial) one?

Read the title and abstract.
Read the Introduction/Background. This is to give you context for the study. Remember that the last paragraph of the Introduction will usually tell you the authors’ research question.
Study each table and figure and read the captions. DO NOT SKIP THIS STEP! It is the most important step! Again, this is what expert scientists do.
Reference the Methods and Results sections as necessary to answer your questions and to clarify information in the tables and figures. For example, text in this section should help you identify the dependent and independent variable(s) and learn how measurements were taken.
Read the first paragraph of the Discussion section, which usually summarizes the results, to check your understanding. If you want to see how the authors interpret their findings, go ahead and read the rest. (Students in Language Acquisition or Psycholinguistics should do this.)
Re-read the abstract.

For steps 4 and 5 especially, remember that your goal is to extract information. Take each question, one at a time, and scan the paper looking for an answer. When you find a paragraph or figure that looks like it has relevant information, stop and read or study it. Keep scanning and reading until you’ve found all the information you need. Once you know the answers to your questions, you’re done—even if there are parts of the paper you haven’t read!
Optional: Write a 3-sentence summary to check to see if you have successfully extracted the information that you needed. Yes, this is an extra step, but I really strongly recommend getting into the habit of doing this. It is a fantastic way to train yourself to read papers in “information extraction” mode—you’ll quickly recognize when you’ve failed to answer one of your questions. It will also help you encode and remember the main points of the paper.

Adjust your process depending on your goals

The above process is what I recommend if your goal is conceptual comprehension. This is usually sufficient if you are reading a paper for an undergraduate class or introductory graduate-level course (like Language Acquisition or Psycholinguistics at Georgetown), for your job as a research assistant in a lab, or for some journal clubs.

But if you are reading a paper for a graduate seminar or advanced reading group, conceptual comprehension might not be sufficient; you might also need what I call theoretical comprehension. That is, in addition to understanding the basic findings and how they were discovered, you’ll also need to understand relevant theories of this phenomenon and how these findings fit (or don’t fit) with those theories. To do this, you will definitely need to read the Discussion section carefully (after making sure you understand the results). Don’t skip the footnotes! In addition to the questions above, you should be looking for answers to these:

What explanations did the researchers offer for their findings?
What problems are there with these explanations?
Are there alternative explanations?

Another goal might be methodological research. If you’re planning a research project of your own, you may want to learn about other research studies on the same topic in order to understand other scientists’ procedures. You may be less interested in the actual findings and the theoretical implications. Your focus here will be on the text of the methods section. You might have questions like these:

Where did the researchers get their participants?
What was the experiment like for participants? (Exactly how long did it take? Were there any breaks? Did they have any special tricks to get children to pay attention?)
How many children needed to be excluded due to not understanding the procedure?