How to read papers?
Before we delve into reading a paper, we start by discussing the general structure of papers. While papers are generally organized as having a title, abstract, introduction, etcetera, I’m talking about the higher level structure of a paper which is spread across different sections of a paper. Following that we will walk through a paper as an example, and show you how to parse the high level structure of the paper.
Structure of a paper:
Section 1: The Quest for Knowledge - Research Gap and Questions
Every paper tries to answer a question that has not been answered. Otherwise, what is the point of the paper? The question stems from a gap in research that exists somewhere in the field. For example, suppose one day you wake up in the middle of the night and ask yourself “Is there clostridioides difficile in pleural fluid?”. You end up using your favorite search engine, Yahoo search, and find nothing online related to it. You only find a case study, where C. diff was reported present in the pleural fluid of a patient in Italy aged 60 years.
The absence of research papers would mean that you might’ve come across a research gap! The presence of the case study means that there could be C.diff in pleural fluid that has not been studied yet widely, because the presence of C.diff for one patient only is not super conclusive! Having found a gap in literature/research, authors would generally try to ask a question that is specific. Such a question can be “Is there C.diff in the pleural fluid of patients in Italy in the age group of 60-80 yrs?”. This is how research gaps and questions differ, a gap can be quite wide but a question would start off being very specific, and slowly broaden in scope. For example, once the aforementioned question is answered through a project studying 20 patients from Italy in the age group of 60-80, the authors might decide to study adult patients of any age, to answer a research question with broader scope.
All papers generally try to provide a finding in domains where research gaps exist, therefore every paper has a specific research gap/challenge as motivation. The authors therefore try to answer a research question in this gap and try to answer the question via experiments.
Section 2: Backstory Chronicles - Related Works
Once a potential gap is found, authors look for research articles surrounding the gap, necessarily to define the perimeter of the gap and scope out the research question(s) they are interested in answering. Therefore, in every paper, authors also provide related papers to tell the readers what has been done before, and what their new contributions are.
Section 3: Setting Up Camp - Experimental Setup and Results
Research papers generally have an experimental setup through which to collect evidence that helps answer the research question(s). For example, to answer whether there is C. diff in pleural fluid or not, the authors would:
i) have to decide on which patient group to study
ii) obtain samples of pleural fluid
iii) perform microbial culture or other controlled experiments to quantify C. diff prevalence and
iv) analyze the results across multiple samples to find statistical significance of C. diff prevalence. One important thing to note is that there is not always one way to find the answer to a research question. The findings are also subject to how the experiments are designed and performed. Therefore, authors spend a good amount of time-space in research papers, describing the experimental setup, trying to justify why results from their experiments are conclusive and sufficient to provide answers for the research questions.
Section 4: Grand Finale - The Conclusion
Given the results, the authors try to draw a conclusion on what the experiments entail. Recall that each paper starts with a research question and a gap. Based on the experimental results, the authors try to draw conclusions answering the research questions and making conclusive statements regarding the research gap. Authors might also mention limitations of their study and future directions for exploration.
Now that we have a higher level idea of the components of a research paper, we are going to go through a particular paper titled “Multi-country surveillance of Clostridioides Difficile deomstrates high prevalence of spores in non-helthcare environmental settings", published in the Journal: Anaerobe. Through this paper I will demonstrate how to read a paper. Once again, there is no right or wrong in this, this is just a method that I use (which might evolve and change in the future as well) to read papers.
Paper Expedition:
“Multi-Country Surveillance of Clostridioides difficile Demonstrates High Prevalence of Spores in Non-Healthcare Environmental Settings”
Jinhee Jo, Anne J Gonzales-Luna, Chris K. Lancaster, Jacob K. McPherson, Khurshida Begum, M Jahangir Alam, and Kevin W Garey
This is an open access paper, meaning the full version of the paper is accessible to anyone here:
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9197859/
First, let's start off by reading the title and abstract. Notice that for this paper the abstract is broken into multiple sections which will help us find answers to the following questions quite easily. Not all papers might be structured in such a way, but similar information should still be present in the title and abstract.
Okay first question:
What is the research gap this paper is trying to address?
The background subsection of the abstract answers this question. The research gap is the absence of worldwide analysis of C.Diff spore contamination, even though C.Diff is frequently found in both hospital and non-healthcare settings.
What are the specific research questions?
The research question being asked in this paper is quite wide in scope. No wonder this is a multi-country survey! The question can be posed as something like:
“Is there C.Diff spores present across different countries in:
Hospitals
Non-healthcare buildings
Outdoor Environments
Shoes?”
What is the main finding/contribution/conclusion?
While the abstract already has a ‘finding’ subsection, the answer to this question is not clearly stated there. Why you ask? Because we are looking for the one main finding of this paper. Or you could say we are looking for a statement that is an answer to the research question above. That can be found in the interpretation section of the abstract. It states that there is high prevalence of C.Diff in non-healthcare settings, especially floors and shoe soles!
What methodology did the authors use to look for evidence?
Generally in papers, this is mentioned in short in the abstract. If not, it comes up in the introduction at the very least so no need to squint your eyes for this! For this paper we can see this mentioned in the methods subsection as well as the findings subsection of the abstract. The authors collected swabs and typed them using PCR-fluorescent ribotyping. Do these terms sound alien to you? Now is the time for a short google search!
What are the keywords for the paper?
We are done parsing the abstract! But before we move forward lets do some pondering. Are there any words that sound important but you don’t understand? Is it making the research question hard to understand or even detect? Do a google search so that you understand all the important keywords!
In this paper the authors do mention the keywords separately. Based on the abstract, could you guess what else could be considered a keyword that is not mentioned in this list? Maybe ‘multi-country’?
Moving on: Introduction/Background and Conclusion.
Based on the Title and Abstract, you already have an idea of what the main components of this paper is, i.e., the research gap, questions, experimental methods, results and conclusion. Now as you keep on reading the paper, your target would be to get more information on each of the structural components we mentioned above and furnish your idea of the paper. For example the next step would be to read the conclusion section (there is no separate conclusion section for this paper and the conclusion can be considered the last few paragraphs of the discussion section) and the background section. In many papers there might be an introduction that comes before the background. The two paragraphs of background try to provide some related past works and motivate the research gap. The second paragraph also elaborates on the objective of the research. Note that there might be a lot of jargon or technical terms coming up already. Some terms may not give you a lot of benefit in understanding the four key points of the paper. Therefore, you may skip digging into these terms for now, you can Google them later when trying to understand all the details of the paper.
After reading the intro/background and conclusion, it should be quite clear what the authors are trying to convey, at least at a higher level. As a researcher, one might have to read a large number of papers everyday to stay updated with what is happening in the field. In that case, reading the title, abstract, reading lightly into the conclusion and introduction/background should be good enough to understand the 4 key structural components of the paper. Therefore, you can pat your back now, you have completed the first milestone in reading the paper, something that you’ll end up doing very fast and frequently in your research journey.
Experimental Highlights: Figures
Next up, instead of jumping right into the methods section and reading sequentially, we are going to take a look at the figures and try to understand what visual results the authors are trying to provide. This might be different based on the field of study, but in papers where statistical analysis of data is involved, there should be some visual representation of the results which you can consider the experimental highlights! Start looking at Fig. 1, Fig. 2. and Fig. 3. Note that there are some recurrent plot types and annotation styles that are being maintained. At first, try to figure out what each axis of the plots denote and what type of a plot it is. For example, all of these plots are bar plots. Fig 3. specifically is a stacked bar plot, where there seem to be segments in the bar denoting group specific values. Note that the plot types (e.g., scatter, bar, box) only tell you how variables are represented in these plots. You need to figure out what the variables are as well, for example, in Fig. 1A the variable is the environment from which swabs were collected. For Fig 1B each environment is split into three based on the location of swabbing, e.g., floor, non-floor or bathroom. Notice how in 1B, we can see that for different environments, the contamination rate (what percentage of swabs collected had C.Diff) is very similar when we collect samples from the floor! Floor in a healthcare setting and floors in outdoor settings have similar amounts of C.Diff! Fascinating!
