23 February 2018

Scientific Paper

image description

Scientists love to refer to scientific papers, research results published in the scientific publications. Why you wonder?
Here is an answer from the scientist. Scientific papers have to follow a certain form and procedure before publication. And most of the scientists face the same procedure and follow the same form if they wished to publish their research results.
The form of the scientific paper follows the requirements of scientific research itself. And in a way, a scientific paper describes the path through research.


The first part of a paper is introduction where a scientist explains where her research results fit into the overall puzzle of the area. And that reflects the preparation time when she was going through existing literature, learning what is behind the chosen research, and how to correctly formulate a research question.
Surprisingly this part is quite important because correct research question advances understanding, while crappy one results in a waste of time and resources.


Next part of a paper is describing the methods used for research. And that’s basically next step in the research itself, deciding what methodology is the most appropriate to tackle the research question, deciding what kind of data one needs and how to collect them. Pre-processing of the collected data follows the collection of the data. In the data mining industry, this is called data mugging, a scientist calls it cleaning the data. Every data point that is removed has to be removed before analysis and reason for removal have to be stated. (Like in one of my research projects, I had to remove all oscillations detected at the time when my lab coworker decided to jump around to warm himself, due to the extreme sensitivity of detectors.)
Then, the analysis methods are chosen and described in details. And I really mean details. Detailed enough that anyone from that particular field of research can repeat the analysis, including the average grad student who is just learning how to do proper scientific research.


The third part is where results are presented in the paper. And result sections corresponds to the analysis part of the research project. This part is for me the most fun because that’s when actual results appear. As the analysis progresses, and results keep on popping out, a scientist can see was the research question answered. Sometimes wrinkles appear. Scientists love wrinkles. Each time results show that answer is not straightforward and more research is needed research is declared success. My last paper was one of those. I got results I expected, but everything was slightly too fast, indicating that what I thought is the main method of the energy transfer was wrong. The mystery behind the research topic was deeper than I thought. So more research was needed to fully understand what is going on.


The fourth part of the paper is a discussion. Here scientists put the results in the context, explaining how the result fit with previous research and what new questions arose. In this part, a good paper will also discuss what other possibilities could cause this particular results to appear. And elaborating what did researchers do to control for such possibilities and which of the mentioned possibilities are not controlled. Some papers do not mention the last part, which possibilities are not controlled simply because there might be too many of those. In such cases, it is up to the reader to interpret the results only within the boundaries explained in the method section and not assume that the result can be generalized.
During the research process, this reflects all the consultations and conversations with colleagues in the research group, musing about the results, double-checking analysis and oneself. I remember once freaking out because I realized that all my results might be wrong because I did not control for solar rotation. I re-run the analysis and got the same results because I forgot to write down that I did remove solar rotation during pre-processing of the data. Yeah…

The rest

Some papers finish with summary or conclusion answering their research question.
But all papers list all other scientific works used in their research. Actually, citations are what makes reading the text a bit hard, because a sentence is often interrupted by citation. But this part is done to show that results are part of the overall science. Also, a number of times scientist’s research is cited is an important measure. The more citation scientist has--distributed over several different papers--the better scientist that person is. (There are also tricks to accomplish high citation numbers, if one publishes a paper describing the novel method used in many other scientific works, then citations surge. Also sometimes, paper with wrong results tend to be cited a lot, because scientists are humans after all, and love to rub in a variant of that ‘I told you so’.
The number of citation is also used to determine the quality of the scientific journal. The journals who published papers that are often cited are considered the best. The measure of the quality of a journal is called impact factor, and means really that, how much impact the journal is made on scientific research, or deep down how often papers from scientific journal were used in other scientific research. Not surprisingly, Nature and Science are two journals with the highest impact factor, mostly because they cover the whole science and tend to publish exclusively just breakthrough research, i.e. something completely new.


Before the paper can be published in any decent journal it has to go through a peer-review process. This means that my carefully composed paper is given to one or two scientists with expertise in the same field of research. Those scientists carefully examine everything that is written, and ask questions, suggest improvements, poke holes and do grammar-Nazi duty.
I published more than 30 papers. Not one of them was accepted outright. The best ever response of referees I got was ‘Minor changes’. Meaning I had to fix how I presented the results, few sentences, or words. But usually, the judgment was ‘Major changes’, basically requesting from me to redo part of the analysis.
It might take from two months to a year before paper passes through this process and it is finally published. My average time was around half a year, allowing me to publish around 2 papers per year.

No comments:

Post a Comment