Thread by @danicat83

Have you ever been overwhelmed when researching scientific papers on a new subject? If your answer is yes, this thread is for you.

I present you "Reading Scientific Publications 101"

Just as a small preamble, I was taught this on a course about Evidence Based Medicine back on my medical student days and adapted it to my current circumstances.

I've just used it recently on a technology research subject and the results are consistent with my medical research.

First: know your keywords and databases. Seems silly right? But this might save you a lot of time. After all, "garbage in garbage out" is a thing.

I usually research big publishers like Springer and O'Reilly, as they have a good amount of papers, but also Google Scholar and sometimes specific platforms like Pubmed (for healthcare). ACM Digital Library is also a good place to start.

Once you found a good article, here is how to read it.

1. Read the title
2. Read the abstract
3. Read the last paragraph of the conclusion
4. Read the last paragraph of the intro
5. Read the intro
6. Read the conclusion
7. Read the results
8. Read the methods and discussion

Sounds a bit mental right? But here is the logic: in every step of the chain you will be seeking to confirm if the information in the article is relevant or not to your current research. Stop at any stage if this condition is not satisfied anymore.

The abstract should be self explanatory. The last paragraph of the conclusion tells about what they achieved with the paper: is this what you are looking for?

It will also tell if further research is expected and what are the limitations of the current research.

The last paragraph of the intro sets the tone for the whole paper and it will help you figure out if the information you are looking for (usually the "how") will be explained through the article.

The results session will go into details of what they achieved. You may also apply the technique of reading the last paragraph first before the whole section if you want to save some time.

Finally, once you are convinced the paper has the information you need, you can read the rest of it including materials and methods, discussion, and whatever other sessions they might have.

If you need to expand your research, looking for citations on the intro of any article on the subject will help you get to the source of a given technology / technique, with the most referenced articles being the ones you should focus.

It is also very useful to look for the most recent review article on the subject, as they tend to compile the most relevant articles in a structured way. You might use it just as an index to deepen your research.

This technique allows me to get up-to-date in almost any subject in 2-4 hours of research, assuming I have the right keywords to look for.