StattLabWiki

User Tools

Site Tools

Trace: • write_paper
literature_and_writing:write_paper

Table of Contents

Write a scientific paper

Gather all your results and sort them logically, which is not necessarily in the order you obtained them. Think about the main 2-4 messages you want to convey, write them down in simple/plain sentences. Go back to all your results and figure out what figures and data you need in support of your main claims/message. From here, try to get a “story line”, i.e. a logical pathway from the beginning to the main statements. Now you have main goals and a pathway to get there, you can write down the discussion/results section, which should include figures/data. Leave formatting the figures nicely for the end, just use the figures you have currently as placeholders. Once you have a rough outline (might be in bullet-point list style) of the results section, you can write the methods/model section. Go back and forth between the two, including methods needed to understand the results, but keep it concise. Tangential information can be move to supplemental information. Once the bulk of results, methods, and model sections are written, write the conclusions and introduction sections. Gather and add references as you go, do not leave them till the end. Again it might be useful to iterate between the two. Think of them as the “bookends” for your story. Leave the abstract for last. Once you have text everywhere, go over it and check that your main messages are there and your story line makes sense.

General tips:

  • Write what you mean, plain, short sentences
  • It's easier to edit bad text than to write good text right away
  • Often starting in the middle is easier than at the beginning
  • Show a first draft to someone else, get feedback on main messages, scope and flow early on
  • Give it a working title, decide on actual title last
  • Identify co-authors as early as possible. Consider all contributions and ask Antonia and senior authors for advice on composing the author list. Then, ask co-authors for permission to include them and iterate the paper draft with them when appropriate.
  • Your aaffiliation is your department or program; Antonia's main affiliation is the MSE department.
  • Acknowledgements in papers cover helpful collaborators and funding attributions. For the latter, list your fellowships and ask Antonia for her input.
  • Journals are selected by the target audience of a paper, the specific research fi eld, and the expected impact of the findings. Important aspects to take into account are journal impact factor, open-access publishing policies, and your co-authors' opinion and guidance. (Beware of predatory open-access publishers.)
  • Use official journal abbreviations for citations (as given on journal websites or on Wikipedia).

Also check out the additional resources below before starting to write.

LaTeX

The online editor Overleaf is great to get started with LaTeX, alternatively, there are a bunch of editors one can use as well, for example texstudio. The math department has a collection of resources for LaTeX, just ignore the Presentations/Beamer section.

A really great starting point is revtex as style guide or template, in overleaf this is as simple as changing the documentclass to revtex4-1.

Here is a pretty good document from Amherst College with advanced formatting tips and tricks: LaTeX for a Math thesis.

Overleaf has a really good help overall as well, for example:

The simplest latex document looks like this: 

\documentclass[letterpaper,10pt]{article}
 
\begin{document}
 
\section{Introduction}
Let's write something!
 
\end{document}

Equations in Latex

Equations are just a matter of knowing the right commands, which can take some getting used to. For example, the LJ equation looks like this: 

U_{LJ} = 4 \epsilon (({\frac{\sigma}{r}})^{12}-({\frac{\sigma}{r}})^{6})

which renders to: $$ U_{LJ} = 4 \epsilon (({\frac{\sigma}{r}})^{12}-({\frac{\sigma}{r}})^{6}) $$ LaTeX doesn't automatically know which brackets belong together and how large to make them, so they look a bit funny. Curly brackets are used by LaTeX to indicate what belongs together and are not rendered, e.g. \frac{top}{bottom} turns into $\frac{top}{bottom}$. If you want text to be “upright/normal” and not cursive, you have to indicate this with a \text{top}.

We can make it a bit more pretty by using more advanced formatting: 

U_\text{LJ} = 4 \epsilon \left[\left({\frac{\sigma }{r}}\right)^{12}-\left({\frac{\sigma }{r}}\right)^{6}\right] \quad \text{for } r< r_\text{cut} \quad .

which renders to: $$ U_\text{LJ} = 4 \epsilon \left[\left({\frac {\sigma }{r}}\right)^{12}-\left({\frac{\sigma }{r}}\right)^{6}\right] \quad \text{for } r< r_\text{cut} \quad . $$

Greek letters are usually available in math/equation mode and are simply their name with a \ in front, e.g. $\sigma$, turns into $\sigma$.

You can copy most equations from wikipedia pages and get the LaTeX code, there is just a bit of additional junk surrounding them (directly copy pasted from here: https://en.wikipedia.org/wiki/Radial_distribution_function): 

{\displaystyle g^{(n)}(\mathbf {r} _{1}\,\ldots ,\,\mathbf {r} _{n})
={\frac {V^{n}N!}{N^{n}(N-n)!}}\cdot 
{\frac {1}{Z_{N}}}\,\int \cdots \int \mathrm {e} ^{-\beta U_{N}}\,
\mathrm {d} \mathbf {r} _{n+1}\cdots 
\mathrm {d} \mathbf {r} _{N}\,}g^{(n)}(\mathbf {r} _{1}\,
\ldots ,\,\mathbf {r} _{n})={\frac {V^{n}N!}
{N^{n}(N-n)!}}\cdot {\frac {1}{Z_{N}}}\,\int \cdots 
\int \mathrm {e} ^{-\beta U_{N}}\,\mathrm {d} \mathbf
{r} _{n+1}\cdots \mathrm {d} \mathbf {r} _{N}\,.

Is actually simply the same equation twice: 

g^{(n)}(\mathbf {r} _{1}\,\ldots ,\,\mathbf {r} _{n})=
{\frac {V^{n}N!}{N^{n}(N-n)!}}\cdot 
{\frac {1}{Z_{N}}}\,
\int \cdots 
\int \mathrm {e} ^{-\beta U_{N}}\,
\mathrm {d} \mathbf {r} _{n+1}\cdots 
\mathrm {d} \mathbf {r} _{N}\,.

which renders to: $$ g^{(n)}(\mathbf {r} _{1}\,\ldots ,\,\mathbf {r} _{n})={\frac {V^{n}N!}{N^{n}(N-n)!}}\cdot {\frac {1}{Z_{N}}}\,\int \cdots \int \mathrm {e} ^{-\beta U_{N}}\,\mathrm {d} \mathbf {r} _{n+1}\cdots \mathrm {d} \mathbf {r} _{N} $$

Figures in Latex

Figures with figure captions can be created like this: 

Aggregation rates for different shapes are shown in fig.~\ref{fig:label}. As visible, the cube aggregates...
 
\begin{figure}
    \centering
    \includegraphics{figure.pdf}
    \caption{Figure caption for figure}
    \label{fig:label}
\end{figure}

Use pdf/eps figures whenever possible, since they scale and one doesn't need to be worried about resolution. When writing a paper, do not be worried about figure placement in the final manuscript, worry about that at the very end, since it will shift a lot during writing of the text.

If you want to reference a figure in the text, do NOT write fig.1, instead give the figure a label \label{somelabel} and refer it, eg. \ref{somelabel}. The advantage is that as things shift around and are changing, the numbering stays correct automatically and the label, if chosen well, is more descriptive than a number. This principle works not only with figures, but also with tables, sections, pages,.. anything that has a number essentially.

Refrences in Latex

Refrences, ie. paper citations, are confusing. There are multiple different ways of doing it. Overleaf has a good help: https://www.overleaf.com/learn/latex/Bibliography_management_with_biblatex Also see https://tex.stackexchange.com/questions/25701/bibtex-vs-biber-and-biblatex-vs-natbib for a good comparison between different systems.

Here is a simple document to get started. 

\documentclass[letterpaper,10pt]{article}
\usepackage{biblatex} %Imports biblatex package
\addbibresource{sample.bib} %Import the bibliography file
 
\begin{document}
Let's cite! Einstein's journal paper \cite{einstein} and Dirac's
book \cite{dirac} are physics-related items. 
 
\printbibliography %Prints bibliography
 
\end{document}

sample.bib needs to contain the references: 

@article{einstein,
    author = "Albert Einstein",
    title = "{Zur Elektrodynamik bewegter K{\"o}rper}. ({German})
    [{On} the electrodynamics of moving bodies]",
    journal = "Annalen der Physik",
    volume = "322",
    number = "10",
    pages = "891--921",
    year = "1905",
    DOI = "http://dx.doi.org/10.1002/andp.19053221004",
    keywords = "physics"
}
 
@book{dirac,
    title = {The Principles of Quantum Mechanics},
    author = {Paul Adrien Maurice Dirac},
    isbn = {9780198520115},
    series = {International series of monographs on physics},
    year = {1981},
    publisher = {Clarendon Press},
    keywords = {physics}
}

You can download these references from the publishers websites or from google scholar. Mendeley also has an export function for creating the biblatex references. Generally, the convention is to use FirstauthorYear to label the references.

Formatting

Proofreading/Polishing

  • Leave serious proofreading for a final draft
  • Finding spelling mistakes: read your manuscript backwards, word for word.
  • Check consistency of used acronyms, greek letters, definitions (include figure captions, labels, headers etc)
  • Avoid too many acronyms

Resources

Where to publish?

Papers in our group often are published in those “standard” journals:

If it's an experimental collaboration or very “shiny”, the following can be considered:

If it is intended for a specific audience/research field, there are journals that reach the intended audience:

Other journals to consider if something needs to be published:

literature_and_writing/write_paper.txt · Last modified: 2022/11/21 20:35 by statt
Donate Powered by PHP Valid HTML5 Valid CSS Driven by DokuWiki