During my researching and reading for my thesis project proposal (which I should finish soon!), I’ve come to realize that my ability to discern whether a given paper will be useful has improved a lot since I began graduate school. Before, I would have to read a paper in its entirety to figure out if I need it or not, which puts quite a few hours of valuable time down the drain.
My first semester in the program, I had a lot of papers to read each week for two reading-heavy courses, and I found myself struggling to keep up some weeks more than others. I spoke with professors and classmates and did some Googling to get some ideas of how to read long and/or dense papers more quickly without losing understanding of the material, but the biggest teacher has really been simple: just read more papers. This was one of the first pieces of advice I received, but at first, I wasn’t convinced it was very helpful. Why would reading more papers help if the sheer number and length of papers I had to read was what I was struggling with in the first place?
As it turns out, reading more papers was exactly what I had to do in order to experiment with different approaches and figure out what worked for me. I’m still no expert, but reducing the average amount of time spent on one paper from 4 or 5 hours to 2 hours is a big improvement!
In this post, I’ll share my process (thus far) for reading scientific journal articles (i.e. the actual research behind sensationalized science reported in the media). My process has two parts that can be summarized in two questions: 1) do you know enough about the topic to judge the relevance of the paper?, and 2) if the paper is relevant, what are the important bits that you need?
Part 1: Is the paper relevant?
This is the million dollar question, isn’t it? Figuring out if a 10+ page paper is relevant to what you’re working on will save you time in the long run, especially when you have found 10+ potentially-relevant papers on PubMed or Google Scholar.
My method for answering this question involves the following 3 “sub-questions”:
1. Do you have a working background knowledge of the main topic(s) in the paper?
If yes: Read the abstract to get a general overview of the paper. Look for the main findings, the significance of the research, any conclusions the authors have reached, and any key terms or phrases you are unfamiliar with. (Write these all down.)
If no: Skip the abstract (for now) and read the introduction. Take brief notes to gain some working background knowledge, including any key terms or phrases you are unfamiliar with. I often go to Wikipedia and/or YouTube for more information if I come across something I haven’t seen before.
2. After reading the abstract or introduction, do you understand the main topic(s) in the title of the paper?
If yes: Great! Go to the next question.
If no: Write down key terms from the title, including those that you are unfamiliar with. Look these terms up and write down brief definitions of them if they are not explained in the introduction. (This is where Google comes in handy!)
3. Is the information in the title, introduction, and/or abstract a good indication that the paper is relevant?
If yes: Great! Go to Part 2.
If no: Don’t read the whole paper. Instead, just read the discussion/conclusions section(s). (Some papers have one or the other, or both.) Is the paper still irrelevant? If not, go to Part 2!
Part 2: What bits of this paper do I need?
You’ve got a relevant paper, great! The next step is to figure out what parts of it are most relevant.
For papers with several figures that may have several parts – I had to read a paper once that contained 7 figures, each with panels (a) through (j), which was honestly torture to get through – I have found the following approach to be particularly useful.
1. Briefly look at the figures in the paper. Consider the following:
- How many are there (not including supplementary figures, if any)?
- How are they organized? Split into multiple panels? Large enough to read? In color or in black and white?
- What types of data are presented? Graphs? Tables? Images? Diagrams?
- Are there any forms of data that you are unfamiliar with?
This first step is meant to give you a general idea of what the figures look like so that you can find what you’re looking for without flipping back and forth between figures.
2. Read the figure legends to get the gist of the main findings of the paper. Take brief notes on a separate piece of paper. If there are any methods or techniques that you’re unfamiliar with, write down what they’re called to look them up in the methods section later.
This step helps you zoom in on what findings you are interested in most so that you can be more focused on what to read in the results section.
3. Look at the headings of the results section. How many are there?
This step gives you an idea of how many “main findings” there are in total, even ones that may not be mentioned in the abstract. Read the headings and try to summarize them in your own words based on your notes about the figures from the previous two steps.
4. Read the results section. Whether you read the material under every single heading depends on whether you understand that finding: how the authors found it, what they conclude, what the figure(s) mean, etc.
While you’re reading, reference the notes you have so far and go through the figures one at a time. I like to take notes directly on the figures as I go along, labeling them (e.g. what a particular bar in a graph means) and jotting down in a few words what the main finding is from a given panel or part of a figure. I also reference the methods section as needed to understand what was done to obtain a given result.
Most papers present figures and findings in order; that is, the authors ask a series of questions that build upon one another, and you need to read the whole results section to understand this progression. However, if you are more experienced and already have enough background knowledge, you may be able to jump around the results section to only pull out the “nuggets” that you need. (I’m not at this point yet, myself.)
5. Take a critical look at the data in the figures.
- Are the data convincing?
- Is the data statistically significant?
- Do you think the data has important implications for other areas of research?
- Are there any additional experiments that the authors didn’t do that you think are important for the questions the authors are trying to answer?
This step is one I still struggle with sometimes, as I’m still learning how to evaluate data and I’m not familiar with the process of interpretation for every kind of experiment. From what I’ve gathered thus far, this step is the most difficult, and takes the most practice.
Last but not least, step 6!
6. Read the discussion and conclusion section(s).
- Are your questions about the data from #5 answered in these sections?
- What are the limitations of the paper? Do the authors explicitly state any?
- Do the authors reference contradictory findings from other papers? (If so, look these up – you might find something interesting!)
These last two sections of any paper are, in my opinion, the most helpful to read. If you are bogged down in details and can’t see the bigger picture of what the authors found and why it matters, these sections are just the place to “zoom out” a little bit.
I hope this post has been helpful!
I understand how difficult it is to read dense scientific journal articles, and how equally challenging it is to overcome nagging feelings of ignorance or self-doubt at how little you think you know about a given topic. It’s easy to get caught up in how much you don’t know, only to forget how much you have learned already.
For more suggestions and advice on reading scientific journal articles, check out these resources: