How to write a paper

General suggestions to Ph. D. students

(-> an updated version 10 Aug 2010 is available for download as a pdf)

Title

Try to find a title that is short and appealing with an emphasis on the problem under investigation at the expense of taxonomic detail. Attract the attention of people outside your specialty, without being too provocative or inaccurate. The title is what makes people decide whether they want to see more of your paper or not.

Abstract

Never forget that the abstract is often the only thing readers will read from your paper. Pay attention to it! Make it very concise, clear-cut, easy to read. Avoid technical terms or complicated reasoning.

• First 1-2 sentences: introduction to the concept.
• Next 1-2 sentences: the system and the specific question.
• No methods, unless unavoidable or new.
• Next x sentences, a brief summary of the results.
• Final 1-2 sentences: conclusion.
• You can end with a sentence that hints at the major topic of the discussion. For instance: “We discuss the relevance of these results for the evolution of low-altitude flying in Dumbo’s and other anti-gravitational objects.”

Introduction

• Themes (to be addressed in this sequence, and each about half a manuscript page long!)
• The general framework. VERY GENERAL
• The (more) specific question
• The system under investigation (essentials) ß
• The working hypothesis
• How this hypothesis will be tested VERY SPECIFIC
• Do not
  • try to review all the literature (unless you are writing a review). Only mention the most relevant papers.
  • mention exceptions unless their relevance to the central scope of the paper is clear.
  • talk about problems that are not directly related to the data.
  • mention any of the results of this manuscript.
  • refer too much to unpublished material (minimise it).

Methods

• The following sequence is advised
• More details on the animal system
• Collection sites
• Culture conditions
• Experimental procedure(s)
• Statistical procedures (very brief, unless special)
• Be accurate! It must be possible for readers to reconstruct the complete experimental procedure. It is possible to refer to other papers for standard techniques. But at least summarize them briefly when you think they might not be commonly known.
• Make sure that you include only those data sets that are relevant to your main topic. Cluttering your paper with all the data sets you have must be avoided when there are many, heterogeneous sources. Showing a few convincing data sets is the only good solution.
• Explain why you collected different data sets.
• Descriptive data that do not contribute to the main topic of the manuscript, but that are needed to show what material was available for the “main” analysis, may be added to the methods.

Results

• Be extremely concise and stick to a subdivision that relates directly to what you “promised” your audience in the introduction.
• Only describe and interpret your data objectively. Do not speculate, that is part of the discussion.
• Do not describe in detail those data that are already presented in tables or graphs. Summaries tables and graphs instead and point out overall effects and trends.
• When mentioning statistics, always mention (i) the test used (ii) the statistic-value (e.g. F or t value) (iii) the degrees of freedom and (iv) the probability of the statistic for the df.
• Round numbers to 2-3 significant digits.
• Move statistics to the legends of tables and figures if possible. Limit stats in the text.
• Average ± SD is preferred over average ± SE.
• When P = 0.000, write P < 0.001.
• Avoid bibliographic references.

Discussion

• Start out by summarising your results in a few sentences.
• Then link your results to the “hypotheses” formulated in the introduction and point out any similarities or differences from what you had expected originally.
• You can further speculate about any of your results, but be careful not to “take off” in daydreaming.

Acknowledgements

Your choice. Usually all the people who commented on the ms are acknowledged, as well as the technical assistance etc. who cooperated. Do not forget to mention funding agencies (e.g. the DFG) when your project was supported by external sources.

References

• Limit yourself to the most relevant references.
• Make sure that all the references mentioned in the text are also here, and that this list does not contain references not mentioned in the text (double cross-check!).

Tables

• Always put one table on one separate page for scientific manuscripts.
• Are you sure you cannot represent your data graphically? A graph is much better in transferring a message than a table.
• Write a good, clear legend that can be understood without having to read anything else!
• Clear, open layout. Nowadays, a lot of journals only use horizontal lines in tables. No vertical lines are used to separate columns.
• Reduce the information in the table as much as possible. If it becomes too long, consider moving some of the statistics to the legend, or spilt one table in two.

Figures

• Write a good legend (see tables).
• Take into account that figures will be reduced. Hence, select large fonts for labels and titles or make the graph small. Do not use very thin hairlines: they may become invisible after reduction.
• Axis-titles have to be brief, but complete. Look for the briefest (meaningful) expression you can think of. Always mention units (between brackets).
• Select axis-ranges in such a way that not too much empty space is left.
• Select very distinct patterns or levels of grey (e.g. black an white) to distinguish between subgroups in the graph. Patterns (hatched or cross-hatched) are better than shades. The latter may become smeared after reduction and reproduction.
• Always use the same standard pattern and colour for the same subgroups throughout the
• Manuscript (e.g. white for females, black for males).

Typing style

• In general: periods, spaces, commas, semicolons, lower or upper case, italic, underlined or bold, roman or Arabic numerals, abbreviations … - you should think about each of these, and use them consistently throughout the manuscript. Look at a recent paper in the target journal! This is one of the most tedious parts of writing a manuscript – but a very important one. If your writing style is inconsistent – the editor will tell you to make it so and your referees might consider this “sloppy writing” which may affect their general judgement.
• Do not forget to add page numbers. A manuscript sometimes may have to be reassembled (e.g. when your supervisor drops it when falling asleep in the train …).
• Standard style is: never a space before, but always exactly one space after ; , . : ? ) } ]
• In the US, people often use two spaces after a period (.). Check the instructions of the journal! Some specifically say that they do not want two spaces.
• Avoid writing text in brackets as much as possible.
• Titles: Do not number major (sub)titles. Number paragraphs only if really necessary. It is often a good idea to use similar subtitles in the introduction, methods, results and discussion. This makes it easier to see the link between the 4 main parts of the manuscript.
• It used to be commonplace to underline text that is supposed to be printed in italics (e.g. species names). At the moment, more and more journals request the manuscript as an electronic file. If so, it is best to write in italics and not underline (see instructions to authors).
• Do not use boldface or italics for references in the text.
• Write references in the text consistently in the same format. Separate authors by semicolon. Check with the target journal for the format to use. For instance:
  • (Baur & Chen 1990, 1992; Leonard 1994)
  • (Baur and Chen, 1990, 1992; Leonard, 1994)
  • (BAUR & CHEN 1990, 1992; LEONARD 1994)
• “et al.“ is used for references with more than 2 authors and replaces all authors but the first. Some journals require you to write all authors when there are 3 and when they are mentioned for the first time.
• The normal sequence is: oldest first. For more references of exactly the same authors in the same order: write authors once and add successive years of publication
• Many journals “indent” the first line of a paragraph.
• Be concise. This means: eliminate ALL superfluous words and expressions. Shorten sentences when possible. Cut long sentences in shorter ones. Be aware of possible ambiguities.
• Be fluent. Connect sentences in a “logical, intuitive” manner by using e.g. “therefore”, “on the one hand … on the other hand ...”, “however”, “(al)though”, “yet”, “despite”, “whereas”, “consequently”, “as a result ...” … but use them correctly!
• Use a spell-checker, but realise that spell-checkers are stupid because they do not recognise a typing error when the result is an existing word (e.g. women and woman). Grammar checkers can give some useful advice too, so use them – but think, too.
• Avoid repetitions of the same word. Use a thesaurus to find proper alternatives.
• Spelling is learned best if you use a dictionary frequently. But be careful! For non-native English speaking people, dictionaries are often a source of “wrong” words resulting in funny expressions.
• Do not use “julian date” (days since 1 January) in a manuscript, use the real date instead. Only in data files it makes sense to use the julian date because of calculations and graphical representations.
• ·Always write years in full: 1994, not 94.
• Think about the date format: December 6, 1996 is US style. 6 December 1996 is British style. Months are often abbreviated to the first 3 characters (Jan Feb Mar …).
• Other commonly used abbreviations are: days = d, hours = h, liters = l, molar = M, …
• Use line spacing 1.5 or 2 if your ms will corrected by a colleague or referee.

Note

• It is worthwhile to checkout the very good, extended instructions to authors on “how to write a paper” in journals like e.g. Animal Behaviour.
• A personal view on authorship (scientific manuscripts)
• Authorship is a delicate matter as usually more than one person is involved in any scientific study. Therefore, it is essential to make clear who is going to author a paper, and in what order, before data are analysed or, even better, before the start of the experiment. It is often necessary to combine data or logistical effort or ideas from several people for a coherent story. This inevitably results in multiple co-authorship. See it as a benefit, not a disadvantage. You will be author on more papers when you are cooperative.
• First authorship is a matter of common sense. The first author is the person responsible for the opinions and analyses presented in the paper. Ideally, the same person develops the experiment, collects the data and writes them up as the first author. Sometimes, however, this is not possible. In such cases, you should talk openly about authorship and find a compromise. The person who has done the analyses, who has been responsible for the conceptual background and who has written the ms is usually the first author.
• As a rule, Ph.D. students should be first author on most if not all papers they have in their thesis. In fact, having some published before you finish your Ph.D. should be the normal situation. Your supervisor(s) will co-author these manuscripts. Exceptions may apply and should be discussed.
• As a rule, “Diplom-“ and “Staatsexamen” Students write their project in German. Their supervisor may then write up the study for a scientific publication based on the student’s work with the student(s) who did the work as co-author(s). In exceptional cases, it can be agreed that undergraduates write up their project as the first author