How do you read a scientific article?
A scientific article in a journal or scientific publication, if you have little research experience, can seem to be a difficult and complicated text. However, most scientific articles have a clear structure to make reading them just that much easier.
By reading a scientific article in a structured manner, you can better determine if it’s relevant and useful for your dissertation. In this (non-scientific) article, we explain how you should read a scientific article.
In this article we will use the following scientific article as an example:
Perrett, D. I., Burt, D. M., Penton-Voak, I. S., Lee, K. J., Rowland, D. A., & Edwards, R. (1999). Symmetry and Human Facial Attractiveness. Evolution and Human Behavior, 20, 295-307. Retrieved from http://www.grajfoner.com/Clanki/Perrett%201999%20Symetry%20Attractiveness.pdf
This article is about the possible link between facial symmetry and the attractiveness of a (human) face. We will concentrate on Experiment 1 in the article.
Before you start
The very first question that you should ask yourself is, could this article be relevant to my dissertation? You answer this by scanning the article. In other words, read only the title and the headings. If you notice right away that the article is not relevant to your subject, then you are better to look for another article.
Quality of the article
Now that you have determined that the article is interesting for your own study, it is worth considering whether the article is of high quality, because you can’t just assume that every scientific article is a high quality one.
When an article of poor quality is used as a source in your dissertation, you run the risk of drawing incorrect or unsubstantiated conclusions. Your supervisor will also always look at the quality of your sources to determine whether your conclusions are well founded.
There are a number of points by which you can determine whether the article is of a high quality.
Did you know?
Scribbr editors correct an average of 150 errors per 1,000 words? Prevent your thesis, essay or paper from being rejected based on language. Have it proofread by Scribbr’s language experts!
Now that you know that the article is relevant for your own research and the article is of high quality, you can get started reading the article in more detail.
Step 1: Read the introduction
Many students begin by reading the abstract, but you can better start by reading the introduction. The abstract is concise and often contains difficult language, and it is difficult to understand the abstract if you haven’t yet read anything of the rest of the article.
Step 2: Determine the big question within the research field
What is the “big question” that the researchers in the field of study want to answer?
When you know what the underlying big question is, you understand better why the research in the article was done. The article is, in fact, often just one small part of a much larger study about which more researchers write articles.
Look for the reasons for performing the research. Often, a study builds on a previous study. See which studies were done previously, which limitations these studies had and how this research adds to the prior research. You don’t always have to search for this information yourself, because it is often provided in the paper itself.
In the case of the example article, the big question in the field of research is: “Which factors determine attractiveness?” Possible predictors of attractiveness, such as facial symmetry, are researched. This study focuses on the attractiveness of the human face. Research is being done on the effect of facial symmetry on attractiveness, and a link has been found between symmetry and attractiveness.
Step 3: Determine the research questions
Which research questions are the authors trying to answer, exactly? There could be multiple questions, but there could also be just one. Write down the research questions for yourself.
Sometimes there aren’t any questions but rather hypotheses. with hypotheses instead of research questions, the research determines whether the expectation of the author (the hypothesis) is correct. In that case, write down the hypotheses.
In the article from the example, there are no clear research questions or hypotheses to be found, so you’ll have to determine them yourself from the text. Two experiments are done in the research, studying whether a certain expectation that the authors have is correct. This is, then, research that assasses hypotheses. It appears from the introduction that the expectation is that people find a symmetric face more attractive than a face that is not symmetric. This brings us to H1 and H0.
H1 and H0
H1 = People find a symmetric face more attractive than an asymmetric face.
H0 = People have no preference between a symmetric face and an asymmetric face.
Step 4: Look at the approach
What do the authors do to answer the specific questions? What is the plan of approach?
Surprisingly enough, in prior studies of the attractiveness of human faces, a preference for asymmetry was found. The researchers of this article think that this is due to the fact that the faces in the previous studies were made unnaturally symmetric. With this type of research, a photograph is taken of a face and this is then manipulated to make the face symmetrical. This resulted in unnatural properties and changes in the structure of the skin. It is, then, not strange that the participants had a preference for the naturally asymmetric faces. In this research, therefore, a new manipulation technique was used to make the faces symmetric. The form of the facial features is varied and skin structure is kept constant.
Step 5: Read the methods section
Write down exactly what the authors have done per experiment. Describe it, for example, in a clear outline but make sure that you record all the details so that you can understand the big picture from your outline. This goes more quickly by hand than on the computer, but for clarity we have made an example in Word.
For Experiment 1, two photos were made of each face. A photo was taken of a test subject and this photo was then manipulated using the new technique to make a symmetric face. Thus, there was an original photo and a more symmetrically formed version of the photograph. The photographs were presented in pairs to 49 evaluators. For each pair, they had to choose the most attractive photo.
Step 6: Read the results section
Write one or more sections to summarize the results of each experiment, each figure and each table. Don’t even think about what the results mean; just write them down as they are. Often, the results are summarized in the figures and tables, so look at these carefully!
Also pay particular attention to the words “significant” and “not significant”. These specific words have an important statistical meaning.
A result is significant if the probability is smaller then 5% that the difference found or the link found is coincidental. If the probability that the observed result is coincidental is equal to or greater than 5%, then the result is not significant. The probability that the result found is coincidental is also indicated with ‘p = …’. This means that a result is significant when the number after the ‘p’ is smaller than 0.05 (p < 0.05). Some studies speak of significance at only 1%. In these studies, the ‘p’ must be smaller than 0.01.
Suppose you are researching the influence of studying on the grade of an exam, and you do this research on 100 test subjects. It appears from your research that the average grade increases with more hours of study. Now, your finding doesn’t mean that you can immediately conclude that this result is always the case. It is possible that the results of your research are purely coincidental. That’s why you test for significance. Only when your result is significant may you conclude that more hours of study contribute to a higher average grade on the exam.
Example table 1: Extent of asymmetry in the photographs
Of the 30 faces, the average asymmetry of all 13 facial features is not more than 1 pixel. Only at the height of the outer corners of the eyes was a significant asymmetry observed.
Example piece about preferences
At the end of the experiment, the number of symmetric faces chosen was calculated per evaluator. In 57.8% of the cases, the evaluators preferred a symmetrical face. The t-test shows that the average deviates significantly from 50% (or, no preference). If the result was 50% then this would mean that the evaluator had no preference for symmetry or asymmetry. It was also tested whether the preference for symmetry differed with photos of men or women, and whether there was a difference in the preferences of male and female evaluators. Finally, it was investigated whether the gender of both the evaluator and the person in the photograph had an interactive effect on the number of preferred symmetrical faces. For all of these tests, the ANOVA test was performed and no significant results were observed. The analysis has even been done in reverse as well. Now it was examined per photo whether there was a preference for the symmetrical face more often than for the asymmetrical face. This examination gave the same results. Of all the evaluators, 75% did not realize that the faces were manipulated and did not think that this had influenced their judgment. It turns out that, also with only this group of evaluators, the preference for symmetrical faces is significant and is 56%.
Step 7: Determine if the results answer the specific questions
Form your own interpretations before you read those of the authors (in the discussion). Ask yourself at this step: what do the results mean? If you are a beginner in reading scientific articles, then this will be more difficult than when you are more experienced.
In the beginning, you will often need to adjust your opinion to that of the authors themselves. Later, you will probably be more critical.
The results show that the evaluators did find that symmetrical faces are more attractive than asymmetrical faces. The gender of the evaluator and of the faces could have perhaps influenced the preference, but this was not evident. Thus, H1 is confirmed.
Step 8: Read the conclusion and discussion
Now read what the authors think that the results mean. Do you agree with their interpretations? Also pay attention to what the authors identify as shortcomings of the research and what they propose for follow-up research. Don’t assume that they have done everything correctly – be critical.
Did you see any shortcomings that they didn’t mention? Do you agree with their proposal for follow-up research?
I agree with the interpretations of the authors. They identify a number of shortcomings whereby they immediately propose follow-up research to improve the completed research. I noticed that the study sample is small and that the ratio between the number of men and women is not very equal. This they don’t identify as a shortcoming, but perhaps the research could be improved with a larger and more equal sample. In addition, they have used only white respondents for the research, but it is naturally interesting to study whether the results also hold for the other races, such as Asians.
Step 9: Go back to the abstract
Now you can read the abstract. Does this reflect what the authors say in the article? Does the abstract match your interpretation of the article?
The abstract fits well with the rest of the article. I have interpreted the article as it was described in the abstract.
Step 10: Save the article and always reference the source
Now that you’ve read the article intensively, is it still relevant and useful for your research? If so, take the following steps:
- Save the document. As of recently, you can save your found articles in Google Scholar via ‘My library’. You can activate this by clicking on ‘My library’ to the upper left of the search bar. Note: you do need a Gmail account for this.
- If you don’t have a Gmail account, then save the document preferably in the Cloud (for example, Microsoft OneDrive or Dropbox). For a document name, you can use the authors and the title of the article.
- Immediately note the article in a reference list. Often this must be in the APA Style. You can use the Scribbr APA Generator, Mendeley or References in Word for this purpose.