05 - Thoughts about research and publication

I had planned to write a post about 'bridge collapses in 2016' since last October, but somehow I never found the time to put together the data I had collected. Anyway, I hope that in the near future you can also read about that topic in this blog, but today there is something else I want to talk about. I have some questions in my mind that I've been trying to answer... So basically in this post I'm just going to think loudly. Meaning that there will be more questions than answers :)

Question 1. Why would anyone want to be a researcher?

Basic question, isn't it? In my opinion if you want to be a researcher you should:

• be willing to do something for the good of the society or for the good of science;
• believe that you can make at least a small difference in the world with your work;
• be ready to learn continuously, seek for new ideas, be creative and innovative and
• be willing to share your work and be open for collaborations, co-operations.

Like in fairy tales, right? You can say now that I'm a bit naive but I still believe in the above points and I still find them crucial. (OK, let's say that the last one is 'optional', especially if you are a real genius but without the ability to listen to others. I mean, in this case you won't be able to be a pleasant company to others, but still you can manage to be a good scientist, I don't doubt that….)

Question 2. What is the connection between research and publication?

Tougher question, I know… I tell you first what I would have answered to this question few years ago.
'It's obvious. There is a topic that you are an excellent expert of. You are working on it for a long time, you are seeking for new ideas, you are trying new concepts, new algorithms etc. Some of your ideas work, some don't. Some results are worth attention, some aren't. After some time some of your work make you realize that you chose the wrong direction and it led you nowhere, while some of them remain promising and open many new doors for you. So basically you have done a huge amount of work in a particular field and you have obtained many kinds of results and conclusions. Then you define a threshold for yourself in terms of 'usefulness', 'importance', 'novelty' and whenever your actual results pass this threshold you write it up and submit it to a journal.‘ 
OK, OK, perhaps I was naive... :)

To be honest, I still can't see how the whole system of scientific publication really works but even what I can see is quite different from my ideas I quoted just above. The main points that became obvious for me very fast are these:

• There is a huge amount of publications in any field of science;
• Many of these publications are not available to common man without paying and
• Lot of publications have a long list of authors.

So the next questions are ready...

Question 3. How can there be such a huge amount of publications? Are we really developing so fast in every field?

Have a look at the graphs below (just because I love to create colourful but somehow still minimalist graphs), where you can see data obtained from sciencedirect.com searching for different keywords. First let's see these data related to civil engineering keywords.

Number of articles found by sciencedirect.com's search engine for some civil (and/or mechanical) engineering keywords

Only in 2016 there were more than 25000 papers published related to 'residual stresses'; and almost 20 000 related to 'push over analysis'. (I am aware that these results may be related to other fields and the numbers can't be trusted 100 %, but they are still good indicators...) And these numbers are basically 4 times more than the numbers of 1999.

But let's see what about other fields of science!

Number of articles found by sciencedirect.com's search engine using some 'random' keywords of different fields of science

The situation is not too much different. But in general there is a higher level of variability. While the yearly amount of publications related to 'string theory' didn't even become double in 2016 compared to the level of 1999, publications related to 'mild cognitive impairment' became more than 6 times more. But are these differences showing the real difference between the development or the level of innovation of these fields? Or do these differences show us something else? I really don't know...

At last, we can have a look at the 'relative' increase in publications considering that the annual number of 1999 is the 100 %.

The relative increase in the number of annual publications compared to the state of 1999

Looking at these graphs, the naive me would say, 'So great! Our research is well funded; we are producing more and more important results! No one can stop us; with such a progress rate there is no problem in this world we can't cope with!' But somehow I'm doubtful now. Well, I even stopped believing in unicorns a few years ago…

From my point of view the main problem is not the huge (and still increasing) amount of publications (they are mostly distributed online, so at least they don't 'kill' too many trees...), the problem is that in my opinion it is not possible to produce so many really significant and reliable results. And at this point it becomes personal. I want to do research, I want to have an idea what is out there, I want to use the results and conclusions of others but I don't know whom I can trust. In most of the cases the results are not reproducible, the methodologies are not clear, and it even happens that the abstract and the conclusion parts are opposing each other. And since I cannot trust the results of others (to be clear, I'm not saying that most of the results are not reliable, I'm saying that some of them are surely not and it can be hard to distinguish one from the other), I have to do everything from zero by myself and it shouldn't be like this.

However this post is already far too long so let's finish with a last short question (skipping for now one of the most interesting questions related to the 'economy of science' and to the 'business of publishing'... maybe I can talk about these next time ;)).

Question 4. How can a paper have 3,4,5,6 etc. authors?

I used to believe that being an author means that you actually wrote something based on the definition of the word. But it seems that I was wrong (?). An author of a scientific paper is often much more a person who has something to do with the finished work, either helped with some ideas, with funding, wrote a paragraph, provided data or anything. But is this just a matter of definition? Who decides what 'author' means in the scientific world? Is there a general concept? Is there any general agreement in this? It's confusing... And there is also another issue that confuses me: the order of authors. It is often not based on the amount of work they invested in the paper (which would be logical I suppose) but on lot of other things. It is sometimes just question of prestige to be the first author. In my mind, the first (or only?) author should be the most relevant person; the one who really knows where the results come from, who understands every decision, every reason behind choosing this or that method or algorithm. But I tell you more. Let's turn this question upside down, who wants to have his name first on a paper without being 100 % clear about the content, about the codes behind, about every single decision? Or a better question would be: if someone does, what is the main motivation behind this?

Let me finish this post with a question I found on researchgate.com: 'Is modern research becoming more and more “publication oriented”?' It was asked by Dr Naveen Sathyan in December 2013 and it has received its last answer in December 2016. More than 500 answers in total… It seems that I'm not alone wondering about such topics :) Anyway, my short answer to this question (after all the above 'loud thinking') would be simply: 'I believe that it shouldn't…’