Scientific writing - theses and paper

Gregor Weihs

About this document: This is a more or less random collection of tips that I gathered by seeing the same mistakes time and again in thesis and paper drafts. Much of it is geared towards German native speakers and their typical mistakes when writing in English. If you are a student in my group, use this as a checklist to save me and yourself a lot of time.


  • While it appears that there is no general consensus on the use of the grammatical tenses, I prefer the following rules, which are also recommended by several articles on scientific writing.
    • You are almost always reporting something you did in the past, therefore use the past tense throughout, never present perfect. For writing a thesis the majority of guides clearly recommends to use the past tense for reporting about work you did before writing the thesis.
    • Present tense is only used for general truths, e.g. "The sun rises every day", or atemporal facts like "This paper presents ..."
  • Voice
    • Try to use the active voice, whenever possible and especially if a person did something. "We measured ..." instead of "... measurement was performed." If something happens and the subject is completely irrelevant the passive voice may be more suitable.
  • Check here for the use of "because", "since", and "as".


  • Defining ("that") vs. non-defining clauses (", which").
  • Prepositions are difficult for non-native speakers of English because there are many false friends. Here is a list of the ones I see mixed up most frequently:
    • depend on
    • independent of
    • associated with
    • related to
    • look at
  • Number correspondence is stronger in English than in German. It is "the signal and idler wave intensities" not the singular.


  • "in contrast to" – "on the contrary"
  • use "in order to" sparingly, often a simple "to" says the same
  • There is a subtle difference between "allow" and "allow for". While the misuse of the latter seems to be so rampant that I have seen wrong uses even in prestigious outlets, it is still wrong to use "allow for" when you mean "allow", see also the dictionary definition. Also think about other words that can express similar or even better fitting meaning: enable, permit,  support, foster, etc.


Compound modifiers need a hyphen: while it is "... information encoded in time bins.", it is "... emits time-bin entangled photons". In the latter fragment "time-bin" modifies "entanglement"


Be aware of the differences between the hyphen and the (en/em)-dash.

In LaTeX use `` (two accent characters) and '' (two single apostrophes) to get proper English curly quotes. 

Technical things

  • Most physics texts are written in LaTeX, which makes it easy to stick to established mathematical typesetting rules. Nevertheless one still has to take care that the various elements appear correctly.
    • Essential packages: amsmath, graphicx, hyperref, siunitx
    • New operators: LaTeX knows many operators and function names out of the box, but some have to be defined, otherwise they will look like a string of variables. amsmath provides a command for easy definitions: \DeclareMathOperator{\sinc}{sinc}.
    • Quantities, units, and uncertainties: the siunitx package is a heavensent for typesetting numbers, quantities, units, and uncertainties. Don't try to do it manually, it almost never comes out right.
    • Multiletter subscripts should be typeset in roman shape, i.e. P_\mathrm{Input} to distinguish them from products of variables.
  • When you embed graphics, use vector graphics for line plots and drawings, never bitmaps. 
  • In cross-references it is good style to capitalize the name of the element you are referring to, i.e. "... is shown in Figure 5". Also we never vary the name of the element. LaTeX automatically labels them as "Figure", so don't call it "Image 5" or anything else. 
  • Use bibtex (or biblatex) for organizing and including your references and always include DOIs. This makes your document so much more useful, because the reader can go to the reference with a single click.
  • The bracket format for typesetting uncertainties, e.g. 52.7(34), is preferred over the form 52.7 ± 3.4, mostly because it is much simpler in the case where the quantity has a unit and easier to read in an equation.


Nach oben scrollen