How to read a Scientific Paper?
Almost all well-known universities have published their own little guidelines on how to read a scientific paper. From the University of Cambridge over Stanford University, to Harvard, they all have some advice on how to READ.
First of all, the Cambridge paper written by S. Keshav gives three reasons for reading a scientific paper. The reasons are to do a literature survey, to keep up to date or to review them for conferences or a class. For most of us, however, it probably is a necessity to obtain these resources for our thesis or dissertation. So many of the papers we come across will be new to us. How do we get started?
Well, reading a paper is not a linear process, from top to bottom (like a book), but depends on your prior interests. Compared to other media, paper reading often requires simultaneaous evaluation of the source at hand. Thus the researcher needs to answer the question whether it is worth investing the time in the first place. So Keshav suggests getting an idea of the paper first, and many scientists seem to follow this advice.
The Science magazine has questionned twelve scientists on the topic and most of them state that they first look at the title, abstract and the figures of the paper before digging deeper into the topic. These parts allow you to grasp whether the paper is relevant for you. Sometimes also the discussion or conclusion are considered in order to extract the main points of the paper. However, when trying to get a deeper understanding the scientists read the entrie article, sometimes taking notes or highlighting text along the way. If the paper seems overwhelming, these scientists suggest to break down the text into pieces, to work through a paper with collegues or to use more easily accessible resources, such as blog posts.
After reading the paper, it may be worth rereading it again. This time the scientist can try to be critical about it, asking questions such as:
-
Did the authors choose a good approach to answer the question?
-
Are their assumptions and interpretations reasonable?
Or the scientist may prefer to use a creative eye by asking:
-
Are there alternative ways of answering the problem the paper addresses?
-
What are good ideas in the paper?
If you really want to judge the quality of a paper, it can often might help to choose the same research question and try to come up with your own approach. By comparing your approach with the one used in the paper, you can criticice it or see its innovation. Additionally, you can present the paper to other people, which will further deepen your understanding.
An important aspect of paper reading is how you deal with the new information. Of course, you can now go ahead and use your favourite reference manager to imbed the paper's citation into your script. But often we are not yet ready to do this. We may be at the beginning of our research and not yet have a script. Thus this information needs to be stored long term. As described above, many researchers say that they take notes and/or highlight the text. But how do you organise these notes? One option would be to use Logseq to import highlights from Zotero and to add your notes there. You could then summarize the paper in one or two sentences and build a database with all the papers you have read so far. The biggest benefit of this is that it can help you to understand the informtion of your paper in context. You will be able to easily compare results and see relationships between different studies.