I'm writing a scientific paper for a Journal

[Bloke of the forest]

Hello sciencefourms.net, it's nice to meet you (collective sense). I'm just some Bloke from the forest who's interested in science - feel free to derive profundities from my name, if you want.

 

 

That's the introduction out of the way...

 

So,

 

I'm not actually writing a scientific paper for a journal; not yet. I am going to though. A few weeks ago a Eureka moment hit me and now I think I have a scientific theory! It is about physics. And no, (cliché) I'm not going to "reveal" what it is because I would be worried that someone might...steal it. And so, I want to get it down into a paper and (hopefully) get it published. The only thing is (cliché) I'm not very good with articulating ideas in writing so I'm stumped when it comes to writing a scientific paper.

 

This is where you come in (cliché), I was hoping that some bright sparks here could give me advice on how to go about constructing a paper.

 

As for my experience with science: I have a decent A-level at Physics and Maths (from...let me think) and a undergraduate degree (Open university).

[timo]

Preparations:

A) Figure the message you want to convey. Sadly, that's "we did the same thing as the other guys, but with slightly different parameters/systems" in many cases. Possibly even more sadly: that's enough to get it published.

B) look through the results of your calculations/simulations/experiments. How do the data support the claim, and which of them do you want to show. Sadly, cherry-picking data (i.e. showing only the data that fit best and ignoring that some of the data not shown wouldn't even support the hypothesis at all) is very common there. Luckily, you might not need to because one can get away with rather dubious data anyways (a colleague recently got a paper accepted in which the referees said that the data are not entirely convincing, but probably the best that can be done with today's technology).

 

Certainly, a lot can be said about actual writing, but (cliche!) the process can often be cut down to realizing that B) is void. No offense meant, btw. I'd gladly say a few words about the actual writing once you figured out B). It's just that before that it is a waste of time.

Edited February 2, 2011 by timo

[Bloke of the forest]

Preparations:

A) Figure the message you want to convey. Sadly, that's "we did the same thing as the other guys, but with slightly different parameters/systems" in many cases. Possibly even more sadly: that's enough to get it published.

B) look through the results of your calculations/simulations/experiments. How do the data support the claim, and which of them do you want to show. Sadly, cherry-picking data (i.e. showing only the data that fit best and ignoring that some of the data not shown wouldn't even support the hypothesis at all) is very common there. Luckily, you might not need to because one can get away with rather dubious data anyways (a colleague recently got a paper accepted in which the referees said that the data are not entirely convincing, but probably the best that can be done with today's technology).

 

Certainly, a lot can be said about actual writing, but (cliche!) the process can often be cut down to realizing that B) is void. No offense meant, btw. I'd gladly say a few words about the actual writing once you figured out B). It's just that before that it is a waste of time.

 

Cheers.

 

So as for structuring, does it go: Abstract, Introduction, Middle bit*, Conclusion?

 

[*Contains the important stuff.]

 

Normally, would the references that are used go at a bibliography at the end or would they be in the margins?

[ajb]

I suck at writing papers and it takes me a long time. I never seem to spot all the typos and editing artefacts. I have mathematics and theoretical physics papers in mind with my advice, but I am sure it will cover most sciences.

 

My rough and ready advice would be:

 

1. A clear title Make sure you inform us quickly what the paper is about.
    
   
2. A clear and consise abstract So you got us this far, now make sure you tell us about what you have done. We may not get any further that this!
    
   
3. A good introduction Tell us in sufficient detail what you have done and very importantly why. Make sure you tell us the main point of your work. Sell it to us. Include a brief review of the literature and place any preliminaries here, like some conventions or notations. Include anything you would like us to know, but is not your own work, anything you would feel embarrassed to put later on in the main chapters. Remember most of your readers will not be experts in your field. How much you write here may depend on the journal requirements.
    
   
4. Main body Now let us have the beefy details. Exactly how you present all this will depend on the journal and your personal style.
    
   
5. Concluding remarks Not all papers need a conclusion. I tend to say something, just rounding up and suggesting further lines of investigation. My thoughts are that people may well read the introduction and concluding sections and not much more in between. Feel free to set the context of your work here and/or in the introduction. Pointing out short comings and restrictions on your work is ok. Better than a referee doing it!
    
   
6. References and acknowledgements A healthy number of references is essential. It shows you know your subject and have an awareness of what other people have done. If you struggle to find references then you will have to ask yourself if the community will be interested in the work. You should also mention the people who have helped you, but don't qualify as coauthors.
   

 

Make sure you have the right journal in mind for the paper and follow any rules the journal suggest. They usually have quite clear guidelines on the presentation of the paper.

 

I tend to get a rough draft ready pretty quickly. I find this will help with making my work clearer. Having it all in one document, rather than in a folder and in note books allows me to find any parts of the work that need clarification. This means not just the actual presentation of the results, but also the results themselves. I can spot the weaknesses in my work this way and it helps me to find other things to investigate. The first rough draft is nothing like the paper I submit! That is if it develops into something of near publishable quality.

 

Whatever happens you will find your own style that works. It may just take a while!

 

Good luck

Edited February 2, 2011 by ajb

[timo]

I think there is no formal need for structure (except for the existence of an abstract which you usually write last, anyways). If your structure is as non-saying as "intro, middle part, conclusion" you can as well write without a structure. "Problem statement -> very brief summary of previous attempts -> general idea -> introduction of the terminology -> calculation or description of experiments/simulations -> results -> explaining the relevance of the results, how they relate to the idea or problem, mention possible remaining issues/questions" might be a good rough guideline. The abstract is best written last.

 

For format, just have a look at physics papers, e.g. any random ones on www.arxiv.org - better even some from the category your work is supposed to be in. Format is not a real problem, anyways. You can still re-type the whole thing with a professional program to make it look cooler, once you've finished writing.

 

I'd like to emphasize the absolute need for what I named points A) and B) in my previous post. "I think maybe the big bang wasn't a singularity but a quantum fluctuation from a previous universe. It makes sense!" is not a piece of scientific work.

Edited February 2, 2011 by timo

[Bloke of the forest]

Generally speaking, should there be a contents and "numerical headings" e.g. 1.2...4.1....e.t.c?

[Cap'n Refsmmat]

That's really down to the journal's style guides, but most papers I've seen do not use numerical headings.

 

How long is this paper? It shouldn't be long enough to require a table of contents or numbered headings, if you want it published in a mainstream journal. It's best if you can get your ideas across in 5-15 pages.

[Bloke of the forest]

That's really down to the journal's style guides, but most papers I've seen do not use numerical headings.

 

How long is this paper? It shouldn't be long enough to require a table of contents or numbered headings, if you want it published in a mainstream journal. It's best if you can get your ideas across in 5-15 pages.

 

I was just about to get onto that. I don't think my idea will cover many pages and five probably will be the maximum(?) It isn't that my idea is actually too simple to be scientific or not thought through enough, it just doesn't cover much ground and should be fairly concise. I was going to ask whether there is a "standard/minimum" general length or does the length not matter. (Quality over quantity)

[Cap'n Refsmmat]

Some journals only accept very short papers (a few pages or less); others will accept papers ten or fifteen pages long. You should identify some scientific journals in your field and check their guidelines.

[Bloke of the forest]

Some journals only accept very short papers (a few pages or less); others will accept papers ten or fifteen pages long. You should identify some scientific journals in your field and check their guidelines.

 

I will do.

[ajb]

I was going to ask whether there is a "standard/minimum" general length or does the length not matter. (Quality over quantity)

 

It will depend on the journal. Some will accept longer papers than other. Something like 5 to 25 pages is "normal", but longer papers are common in some journals.

 

You will find once you have your main ideas down the paper will grow. You will find more things to say, find examples, clarify certain points, add more references etc. So don't worry too much about the length of the first draft.

 

The best advice here is to have a look at the guidelines of the journals you want to submit to. They maybe a little flexible on the length. If you are having trouble with the length then ask the managing editor for advice.

[hypervalent_iodine]

If you're stuck on how to write a paper the best thing that I find to do is to go out and read a bunch of papers in a similar vein to your research and work from there. I did that a lot when I was in undergrad and I was asked to write journal style articles with no clue as to how to set it out. As has been mentioned, most journals have set guidelines about article presentation and exact formatting, so you will most likely have to modify it somewhat. Still, it's good to get an idea of how people who have been published go about writing their reports, particularly if you've never done it before.

Edited February 3, 2011 by hypervalent_iodine

[ajb]

If you're stuck on how to write a paper the best thing that I find to do is to go out and read a bunch of papers in a similar vein to your research and work from there.

 

Which is really a prerequisite of doing research.

[hypervalent_iodine]

Which is really a prerequisite of doing research.

 

Not so much in first year undergrad, which was where I was referencing my analogy from. Anyway, I know that, but there's a difference between researching content and analysing the way a paper has been written.

Edited February 3, 2011 by hypervalent_iodine

[ajb]

Not so much in first year undergrad, which was where I was referencing my analogy from. Anyway, I know that, but there's a difference between researching content and analysing the way a paper has been written.

 

Yes indeed. The problem is the OP states that he has some "theory" or whatever, but has no idea at all how to present it. I do find this a little strange as one would have looked at published papers when working. So, he should have some idea. Details and subtleties maybe be missed of course. Working in science but never seen a published paper or a preprint should set alarm bells ringing. However, I personally will be very open minded at this stage.

 

Seeking advice on writing papers is good. I admit I suck at writing papers, I welcome any advice and am happy to pass on what I have discovered from my personal experience.

[Bloke of the forest]

One more question (hopefully): Would it be a good idea to include some history on the subject? My idea for example draws on some other ones, so would it be advisable to explain the context?

 

I do find this a little strange as one would have looked at published papers when working. So, he should have some idea...Working in science but never seen a published paper or a preprint should set alarm bells ringing.

 

What do you mean "working in science"? And what do you mean by "alarm bells ringing" - as to what? I have seen published papers before, I just wasn't making rigorous notes on their structure.

 

[ajb]

What do you mean "working in science"? And what do you mean by "alarm bells ringing" - as to what? I have seen published papers before, I just wasn't making rigorous notes on their structure.

 

Let me say I want to be realistic and not offensive in anyway.

 

If you are writing a paper in physics then you will have discovered something new, maybe theoretically a new way to view something known, or a new way to calculate something, or maybe something novel and really unexpected. To get the background knowledge to start such research you will have read existing literature. Part of this may well be books, but for the cutting edge of science you will most likely need to examine published papers or their online preprint counterparts. This is also necessary in placing your work in a larger context, which is important. Your work should be correct, but it should also be of some interest to other people.

 

So, if one has not looked at journal papers and similar I would say that one is unlikely to have anything of much worth to say.

 

Now, as you have looked at papers maybe you have a few on your desk? Or as pdfs on your computer? If so you do have some idea of what is required and the general layout, which of course will vary in details.

 

You don't seem to be asking for much general advice on writing papers, but concentrating on the layout.

 

As for your "history question", it is important to set the context of the work in the introduction. This is the place to say something about other approaches, what motivated the work and point to related works. However, I would avoid a detailed history lesson.

 

Have a look at the arXiv. It will help you.

[Bloke of the forest]

Let me say I want to be realistic and not offensive in anyway.

 

If you are writing a paper in physics then you will have discovered something new, maybe theoretically a new way to view something known, or a new way to calculate something, or maybe something novel and really unexpected. To get the background knowledge to start such research you will have read existing literature. Part of this may well be books, but for the cutting edge of science you will most likely need to examine published papers or their online preprint counterparts. This is also necessary in placing your work in a larger context, which is important. Your work should be correct, but it should also be of some interest to other people.

 

So, if one has not looked at journal papers and similar I would say that one is unlikely to have anything of much worth to say.

 

Now, as you have looked at papers maybe you have a few on your desk? Or as pdfs on your computer? If so you do have some idea of what is required and the general layout, which of course will vary in details.

 

You don't seem to be asking for much general advice on writing papers, but concentrating on the layout.

 

As for your "history question", it is important to set the context of the work in the introduction. This is the place to say something about other approaches, what motivated the work and point to related works. However, I would avoid a detailed history lesson.

 

Have a look at the arXiv. It will help you.

 

Okay, cheers. Another question occurred to me: when using references, what kinds of sources should you use, other papers? How is their accuracy determined?

[ajb]

Another question occurred to me: when using references, what kinds of sources should you use, other papers?

 

The best is papers published in a peer review journal. In mathematics and physics preprints on the arXiv are often referenced, if they have a published version then this should be cited. There are also conference proceedings that are not peer reviewed, and on the arXiv write ups of talks also exist.

 

Books can be used, though the advice is to cite the page(s) also.

 

I have seen some references to websites and blogs. For example nLab I have seen in a paper. I think this will become more and more common. Wikipedia is not cited in general in papers of any worth. Crackpots like it though!

 

You can also make references to private communications with other researchers.

 

 

How is their accuracy determined?

 

Peer review is not infallible. To some extent some trust is required.

 

However, I would be careful of using papers you really have no clue about.

[mississippichem]

Okay, cheers. Another question occurred to me: when using references, what kinds of sources should you use, other papers? How is their accuracy determined?

 

There are organizations that rate the "impact factor" of various peer reviewed journals. I'm not sure if there is some universal method for determining impact factor, but I think it always involves the number of citations to that journal. Journals that are listed as having high impact factor are more likely to be sound resources generally speaking.

 

I'm sure there are great journals that aren't often cited but when you're first beginning to look for references the highly cited ones are a good place to start.

[hypervalent_iodine]

Yes indeed. The problem is the OP states that he has some "theory" or whatever, but has no idea at all how to present it. I do find this a little strange as one would have looked at published papers when working. So, he should have some idea. Details and subtleties maybe be missed of course. Working in science but never seen a published paper or a preprint should set alarm bells ringing. However, I personally will be very open minded at this stage.

 

Seeking advice on writing papers is good. I admit I suck at writing papers, I welcome any advice and am happy to pass on what I have discovered from my personal experience.

 

The tone of his post made me suspicious that he hadn't read enough papers to become particularly familiar with how to present his topic, hence my suggestion. Certainly, he should have read extensively during the course of his research - as to whether he paid attention to the layout, etc. is another question altogether.

 

What do you mean "working in science"? And what do you mean by "alarm bells ringing" - as to what? I have seen published papers before, I just wasn't making rigorous notes on their structure.

I have to question how rigorous your research has been? No harm intended, but as ajb mentioned you should have done some pretty thorough background reading if you intend on publishing a new theory (or even an extension to an existing theory). Potentially, what you have discovered has in fact already been addressed by other people or groups who are doing (or have done) research in the area.

Seeking advice on writing papers is good. I admit I suck at writing papers, I welcome any advice and am happy to pass on what I have discovered from my personal experience.

As do I. Or more, I can never work out how to start things. It usually takes me until I've written the first paragraph before things start to flow in any sort of direction. One of the best things I've found to do is once you've written the first draft, leave it for a day or two and then read back over it. I am always surprised at the number of things I end up changing or fixing when I come back to it.

[ajb]

There are organizations that rate the "impact factor" of various peer reviewed journals.

 

 

You have to be careful when comparing impact factor across subjects. Typically, journals in mathematical physics score low, Communication in Mathematical Physics has IF of about 2. This is probably the top journal for mathematical physics. Then the IF of Nature is 34! The Lancet (medical journal) has an IF of about 30.

 

You can send papers in mathematical and theoretical physics to Nature, but they need to be very topical and ground-breaking. Probably not the journal to send your first paper to, unless it really is that good!

 

Also journals that publish a lot of review papers can bump up their IF a these tend to get cited more than original papers.

[mississippichem]

You have to be careful when comparing impact factor across subjects. Typically, journals in mathematical physics score low, Communication in Mathematical Physics has IF of about 2. This is probably the top journal for mathematical physics. Then the IF of Nature is 34! The Lancet (medical journal) has an IF of about 30.

 

You can send papers in mathematical and theoretical physics to Nature, but they need to be very topical and ground-breaking. Probably not the journal to send your first paper to, unless it really is that good!

 

Also journals that publish a lot of review papers can bump up their IF a these tend to get cited more than original papers.

 

Thanks, I should have pointed that out. The best way to compare impact factors is within a given field for sure. By the way, I've never really been a fan of "Nature". I've always felt that "Nature" wanted to be an interdisciplinary journal but secretly is a biology journal. Notice that every now and then they tag on some cool physics or physical chem just to stay relevant to us super nerds.

[CharonY]

Nature has its ups and downs. I would consider it mainly a bio journal, though. At least much in it is not rigorous enough for pure bio, but have some sexy stuff (often interdisciplinary) to sell it. I assume due to its format, about any discipline has something to complain about Nature (or, for that matter, Science).

 

 

Formats of papers (whether it includes a long or short intro, discussion or not, numbering, etc.) are outlined by each journal, so that is usually the first thing to check. However, it is impossible to write a decent paper without knowing the background of the topic (and thus knowing what is the novel aspect of ones own work).