Bipolarity said:
How do you publish a scientific paper?
Once you've done a piece of research, you write it up as a manuscript and then submit it to a journal that's appropriate for publishing that work. The journal editor will then select referees to review the work and will ultimately decide whether or not they will publish the paper.
How do you know that your idea has not already been published somewhere? How would you go about making sure your idea is new?
By reading and networking. In order to do research, you need to be familiar with all the journals that publish research in that field and read them regularly. You should also attend conferences where you can interact with others in the field, view posters, and attend presentations and discussions so you know what others are working on. Just keeping current in a field is an enormous effort.
What happens if you wrote a paper in a field that someone already researched on without knowing?
The chances of this happening are actually quite small if you've done your homework. However, what can happen is that something you're working on comes into publication while you're working on it. I've had this happen to me before. In a worst case scenario it means that your work is no longer publishable. However, in most cases you end up with some overlap and some unique work. It will depend on the referees as to how much overlap is permissable. There is a certain scientific value in independent groups working on the same problem though, so just because someone else has published a result, doesn't necessarly mean your own work will be rejected.
Who do you speak with to start a paper?
Because you're asking these questions I assume that you're a student. In that case, you would start on a project of some sort with a mentor and once you've done something novel and of value to your field, you can start writing up a paper.
What are requirements for writing a paper?
Generally speaking you need to have done a piece of scientific work that is (a) novel, (b) of some value to the scientific community, and (c) of interest to the readership of the journal you're submitting to. Specific requirements are usually listed on the submission page of the journal's website.
Do they have to have experiments if they are in physics/chemistry? What if it's an economics paper? Can papers in these subjects be written using only analytical arguments with no resort to experiments?
You can write a paper without having done an experiment. Many papers are based on analytical calculations, or simulations. But in order to do this, you need to be able to base the work within the body of data from experiments that have been done (if any).
