Studying from scientific papers
In many fields of study you’re going to encounter scientific papers at one point in your academic career. The following tips are taken from my six years of reading psychology papers.
Empirical articles are reports of studies that involved some type of data to draw conclusions. Examples of this are (quasi-)experiments, epidemiological studies, intervention studies, observation studies. These papers typically consists of four sections.
Introduction: background, literature overview, leading up to research questions and hypotheses. Sometimes, the authors give a 2-3 sentence summary of how they addressed these questions in the current study as well.
Methods: detailed report of what the authors did to address their research questions
Results: detailed report of the findings, with no interpretation. Typically filled with numbers and statistical analyses
Conclusions: after a brief summary (usually), the authors draw conclusions based on results, answer the research questions, link it to existing literature, comment on the studies’ strengths and limitations, give pointers for future research, and mention implications
In a well-written paper, the introduction and discussion can be read without reading the middle part and give you a good grasp of what the study is about. This is not to say the middle part is not important, but in some cases you just don’t have the time to go through the analyses before an exam.
Some empirical studies consists of several sub-studies (usually in purely experimental research. In my field of psychology this is usually by social or cognitive psychologists). Apart from a general introduction and a general discussion, the set-up of each individual experiment is the same as mentioned above.
Before you start, ask yourself “what do I want to get out of this article?”
Are you studying for an exam?
Know what kind of questions to expect. Ask for practice exam question if possible. This will be help you focus on what is important
Generally, understand why the study is important/what it added to our knowlegde, how the questions were addressed, type of measurements used, conclusions, limitations
Think about the paper, too, rather than just trying to memorise it. What did you think about it? What would you do differently? What did the authors improve? These kinds of questions might show up on the exam
Don’t study papers in isolution. See if you can explain certain findings in one paper with a model posited in another paper, for example. Or try to think of what one author would say about someone else’s methods
If you’re running out of time or energy, read only the introduction and the discussion. Although the methods and results are also important to get an idea of how reliable and good the study was, the main ideas and findings will be in the intro and discussion
Are you writing an argumentative essay?
Keep the topic of your essay in mind and focus on that
You’ll be tempted to only look for papers that provide evidence for your thesis, but try not to discard counter-evidence. How can you explain contradictory findings? There should be nuance
Are you writing a (bigger) research paper (e.g., your thesis)
When you’re working on a bigger project it’s useful to summarise all relevant papers in a table (I like to use Microsoft Excel). List the articles in rows and their properties in the columns (e.g. author names, year, journal, hypotheses, basic theories, N, sample characteristics, measurements used, findings)
If you prefer, you can also do it on paper and combine it in a folder. However, I prefer the flexibility of a digital file, because I often go back and change or add things once I get a better understanding of a certain paper.
I created two examples (one, two). These are taken/based on paper overviews I made for an essay and my master’s thesis. (I only left a couple of papers, but you’ll typically end up with way more articles of course!)
In the beginning phase, you want to focus on the introduction (to get an idea of what has been done in this field) and the discussions (where are the gaps), but later on you’ll move on to methods (what did they do and how it can be repeated/done better). Mark the papers that you like the best! They will guide your process to coming up with your own research questions, but also help you in a later phase when you don’t know how to structure a certain section
Reviews are papers that combine the knowledge we gained from previous studies that have been published. The great thing about these papers is that they kind of read like a text book! They are structured into sections and you are spared the trouble of reading and summarising all these papers yourself! However, there is no way to know how the authors worked on the paper, as it’s not systematic. Sometimes authors do wish to make a certain point and only discuss certain points. I would say review articles are an especially great starting point if you’re interested in a topic that you’re fairly new to and want to quickly learn about. Systematic reviews
If you’re looking for a more thorough type of review, there’s the systematic review. It’s relatively rare, but I love these babies. Someone (a team, probably) has gone through the effort of systematically going through all the papers published on a certain topic that meet certain criteria, read them, and organised all these findings into a paper that reads a bit like a mix between a text book and an empirical paper.
If you want to take the rigour even further, you’ve got yourself a meta-analysis. These authors are the real MVPs. In psychology, there is a huge replication crisis (with many “groundbreaking” studies failing to deliver the same results in replication studies) as well as a publication bias (it’s far easier to get published when you have interesting or unusual results). Meta-analyses help these issues a bit (does the field need a reform? Yes. Will I discuss that here? No). In meta-analyses, the authors systematically go through databases looking for papers that meet certain criteria (like in systematic reviews, but instead of just listing and organising previous results, and basing their conclusions of the authors’ conclusions (basically), they re-analyse the research question using the statistics from many papers studying roughly the same thing. The advantage is that you have a much larger sample, different measures may be used (reducing the risk of instrument-specific findings), and there is an estimate of what the effect size would be if there were no publication bias. Not always easy to read, definitely not easy to write, but they are helpful for the field.
Case studies describe a very low number of observations, common mostly in medicine. These are usually exploratory bases for more experimental research and should be treated as such. Studying these papers can be useful to get a good idea of a certain condition, especially if it is a very uncommon condition. Case studies are often used in introductory courses, because they tell a gripping tale of a real person; they are intriguing. However, it’s risky to draw conclusions from them.
