Writing and publishing papers

Introduction

Writing scientific publications is one of the core activities of doctoral and postdoctoral research. Through scientific publications you document your research and make it accessible to the wider community. This is important and an essential part of doing research: scientific findings that nobody knows about are of no use. Publications are also a way to communicate with your peers, to become known in the wider field, and to find interesting collaborators – in short: to establish yourself in the scientific community. And whether we like it or not, where and what you publish, and how often your papers are cited remain the most important indicators of individual scientific excellence and success. This is particularly the case in early career stages (doctoral students, postdocs), where publications will be something that potential employers and funding bodies, inside and outside academia, will evaluate when you apply.

Developing your writing skills is, therefore, an important part of a PhD and a postdoc. Nobody writes well at the beginning of a scientific career and learning to write well takes time and a lot of energy. That said, it can be learned by everyone – and is much less of an art than you might think. There are many guidelines on scientific writing (see next section) and this wiki section will not be another one. Instead, we intend here to document some policies and procedures we try to follow in the our Lab.

Writing guides

There are many guides on how to write. Scientific writing is to a large extend about how to structure thoughts and text, how to write clear and unambiguously, how to craft key messages and bring them across, how to build paragraphs and sections, and how to avoid common mistakes. Read these guides! A few examples of such guides that some people in the lab have found useful are listed here, but there are many more:

Alley, M. (2005): The Craft of Scientific Writing. Springer.
Day, R.A. & Gastel, B. (2006): How to Write and Publish a Scientific Paper. Greenwood.
Silvia, P. J. (2007): How to Write a Lot: A Practical Guide to Productive Academic Writing. American Psychological Association (APA).
Fischer, J.: Writing a journal article (Online tutorial by Jörn Fischer)
Joshua Schimel (2012): Writing Science: How to Write Papers That Get Cited and Proposals That Get Funded, Oxford University Press, USA, 2012
Glasman, Hilary (2010): Science research writing: For non-native speakers of English. Imperial College Press, London, 2010.

Scientific writing classes are also regularly offered by Humboldt-University’s continued learning program and the Humboldt Graduate School. Taking one or more of these classes is highly recommended.

In addition to guides on how to structure and write, there are also guides on the correct usage of terminology and style (e.g., Ukraine or the Ukraine?, What are SI units?). We have a copy of The Chicago Manual of Style, which is often what journals refer to in terms of style (but there are others too).

Steps to a paper, and when to ask for advice

Before you start writing your own papers, there are at least four main things to do:

Read: Know the state-of-the-art of your field! Reading, reading, and reading more is absolutely essential to developping good ideas, to see where the research frontier is, to identify research gaps, to link your work into the wider field, and to learn how to write well.
Do the research: well, easier said than done, but we do not treat this here.
Use a reference manager: If you do not do it already, start using a reference manager such as Endnote (free license, also to use at home, by HU Berlin), Mendeley, Refmanager, Citavi (also via a HU Berlin license) or Zotero feature. This keeps your references organized and citations correctly formatted in papers. Using a reference manager will save you many hours and days of work.
Learn how to use MS Word properly: This means, learning how to use styles and headings, line spacing and paragraph spacing (no empty lines!), referencing in terms of tables and figure (captions), and how to use your reference manger for citing while you write (and automatic formatting of citations and reference list). You don't have to use MS Office (e.g. OpenOffice or LibreOffice are good alternatives).

When you are ready to write, the first step is to draft the story-line of your article. It is crucial to tell a story that stimulates readers (and reviewers) and makes them interested in your research. The key to such a story is a clear research question or hypothesis and one or a few (main) results and interpretations (=messages). Your research question has to be of interest to the wider field, something that your particular study object/region/case exemplifies. Thinking about the story-line is very well-invested time and will make it a lot easier to write the full paper later on. A good time to write your outline, based on the PhD proposal in case of PhD students, is when you have first results. It is important to realize that neither the story-line nor the full paper is a documentation of all the things you did – only those that matter to understand the key results you achieved and the messages you want to include. Very often, researchers will not include all methods tested to find the one method that was useful, or include all results (e.g., model runs). Getting a clear message across often involves leaving out things that could be interesting, but are not essential for the main message – even if it took a lot of time.

The outline can be seen as the skeleton of your paper, and it should contain the main headings (introduction, material and methods, results, discussion, for some journals conclusion, acknowledgements, and references. The first outline should contain bullet points, where each bullet point represents the one main message of that paragraph.

Introduction: Main paragraph structure, key references, clear research questions and objectives at the end
Study region: Main paragraph structure
Datasets and methods: Main sections and analyses done, in the order in which they will be presented
Results: First drafts of key results (figures or tables) together with the main points you want to make based on them. Develop ideas on where you need a figure or a table to support your story. Do not have the same information in figures and tables, and do not duplicate information between figures/tables and text.
Significance: Statement on why your results may be relevant. New scientific insight/new method/conservation implications.

The first outline should also contain a first abstract (fully written, not bullet points). Writing an abstract requires you to condense every section to 1-2 sentences – the research gap you identified, how you addressed it, and what your main finding is should become very clear from it. Writing the abstract always helps to get clearer on what the main story-line is, and where inconsistencies or weaknesses remain. Remember to always state in the abstract why your results are relevant/significant. Always end your abstract with the answer to the question: Why should someone care?

Once we have iterated the outline once or twice, the skeleton of the paper should be clear. Now is a good time to decide on a target journal and to really start writing. Introduction and Methods can be written even while you carry out final analyses. Once you have the first sections written (everything up to the results), it is time to structure the Discussion section (again with bullet points) and to have another round of feedback from your advisor(s). Now is a good time to make a list of who should be asked for co-authoring the paper, and to send the next version to them for feedback. Ask them to be critical and see it as an opportunity to improve the paper. If your colleagues do not follow your story-line don’t try to convince them, but think about improving the story-line. You do not have the chance to discuss with reviewers or readers either.

When writing a manuscript, please use the manuscript template that you can find here (and which includes additional tips and tricks on the structure and content).

Selecting the journal

Reading helps to pick the right journal to publish your paper in. Good ways of finding out where to publish are to look for similar papers to the paper you are writing, the journals you cite when writing, and generally the journals that are often in your own literature database. Generally, it is good to be ambitious and to aim high (within reason), but you should also have a plan B (and C) in case the best option does not work out. Discuss your strategy with more experienced researchers and your advisor.

Once you picked a journal, carefully check the “Guidelines for Authors” of that journal and follow them closely in every aspect. It will save you a lot of time to format your paper, your tables and figures, and the references correctly from the start. Many journals have strict word limits – make sure you are aware of them and what they entail. Writing concise requires a lot of experience and your advisor(s) will help you with that. However, try to be close to the word limit when submitting a draft for internal revision to your advisers. As a general rule of thumb, a paper should not exceed 6000 words, with introduction, methods (incl. study region), results and discussion (incl. conclusions) accounting for about 1500 words each. However, this varies widely and some journal guidelines will provide a different structure or limit the number of words much more rigorously.

For many journals, it becomes increasingly hard to take the first hurdle – the editor who decides whether your paper is a fit for the journal and interesting enough to be published there. It is, therefore, important to draft a very convincing cover letter, highlighting what is exciting about your paper and why it fits the journal. Do not wait with this until you want to submit! The cover letter is the most important document in many journals now, so we should work on it for a while to make it shine. Journals might also ask you to provide highlights or answers to specific questions (e.g., Why is this research novel? Why is it a good fit for the journal?). Know about these issues and prepare answers early on (with the cover letter), so we can work on them. Here, you can find examples of cover letters from some of our previous papers.

Typically, journals will also ask you for the contact details of potential reviewers. Think of who could be potential reviewers and discuss it with your advisor(s). Good reviewers are people who are good fits thematically, are not too close to you or coauthors, and have at least a neutral view on the paper. Don’t list the five most important researchers in your field (they won’t have time to review) or people whose work you criticize (they will be asked by the editor anyways). Ideally, you know reviewers a bit and they remember you (e.g., from a conference).

Getting the paper to a final, ready-to-submission stage can require several rounds of revisions, normally between you and your supervisor. At some point, the manuscript will be ready in terms of its text and content. Before submitting a manuscript to a journal though, it should be spotless. This involves that you carefully crosscheck the entire manuscript in terms of language (asking a native English speaker for help is always the best option) and again check the style guidelines. It is not your advisors' task to format your paper.

The review process

After submitting a paper, the editor or subject editor will either reject your submission right away or pass it on to several reviewers. In case of an immediate desk rejection, the most important thing is to get the paper out again. Normally, other than reformatting and minimal rewriting nothing is required, but discuss this with your supervisor.

Once you get a paper rejected after review – don’t be devastated. It happens all the time to all of us. It is a normal process and does usually not say much about the quality of your research, let alone the quality of you as a researcher! If a paper is rejected it makes sense to carefully go over the reviewers’ comments and incorporate those that are easy to incorporate. More major comments, if they are justified or would represent substantial improvements of the paper, can be implemented, but discuss it before you start working.

If you get a paper back with minor revisions, major revisions, or rejection with the opportunity to resubmit, it is time to celebrate (cautiously) because you made the second big hurdle (the first was the editor) towards getting your paper published. Revising your paper and getting it out again should now be the #1 priority, regardless of what you were doing when the news arrived.

Before working on the revision, read the reviewer comments carefully! Try to take criticism constructively (even if it might be formulated harshly) and seriously. Usually, it is best to follow the advice of the reviewers, responding to each comment individually, even the smallest comments. Keep in mind that the purpose of the response is not to convince the reviewer, but to convince the reviewer and editors that you took the comment up and changed something in response (to avoid that other readers have the same issue or misunderstanding). Your response can be done either in text form or, easier in my view, in the form of a table. In the response letter, do not leave out anything the editors or reviewers said – specially not the good things (i.e., do not only respond to the criticism). It is important for potentially new reviewers you might get in round 2 to see the value that made the previous reviewers suggest you revise (rather than suggest to reject) your paper!

Be open-minded, reply respectfully, and – even if you don’t feel like it – value the reviewers' comments. It is hard work to review a paper carefully and it is a voluntary service by a reviewer to the community. Most reviewers try to be constructive – and even if it sometimes doesn’t feel like it, the review process improves research papers substantially in almost all cases. Also, keep your replies short and to the point. Simply think of what you would want to read as a response if you made a review. Include all necessary information, including page numbers of the revised manuscript version, in your response – make it as easy as possible for reviewers and editors to go over this document and to see that you took their remarks seriously.

While you do not have to follow every single remark or suggestion made, only rarely you should disagree with the reviewers and rebut a comment. However, if there are obvious flaws in the review, i.e. if following their suggestions would weaken your paper, you should explain in detail why you did not follow the suggestions. When in doubt, discuss this with your supervisor.

In practical terms, start out by filling the table and incorporating the easy comments. The more substantial comments, for instance those that require re-analyses, should be discussed with your advisor before you start working on the comments. Once you are at the point of re-submitting your revised paper, the response document is submitted with the revised manuscript and a letter to the editor explaining the main changes. You can find a template for a response document here. We are also providing some examples of how we have responded to past reviews.

Being a reviewer

It will not take long and you will be asked to review other researcher's work - and you will get more requests to review than you can answer positively pretty quickly as you advance in your scientific career. When you get an invitation, the first question to ask is whether you should accept. Keep in mind that doing a review thoroughly requires time - typically several days. a good rule of thumb is that you should only review papers that are in your core expertise (so that you can actually evaluate the work and say something constructive about it) and in journals where you can envision to publish yourself. While it is important to carry out reviews, you should also not accept too many invitation (a good rule of thumb is that you should review 1-3 papers per year) as an early or mid-career scientist.

When you do a review, write comments that you would like to see yourself as an author - so: be friendly, be constructive, be fair and respectfully, and ALWAYS be focused on content in your criticism (not the authors). Every paper has been a lot of work to prepare and if a paper has flaws, this can be said in way that does not devastate the authors. Also note that increasingly many journals tend to publish a paper's review history - including the reviewer comments and the responses to these comments. So make sure to address comments thoroughly, and provide constructive criticism when you write your own reviews.

Authorship and co-authorship

Because publications have become such an important criterion for assessing researchers’ performance, who is a co-author, and where they stand has become very important. Decisions on who to include and the order of co-authors are, therefore, sometimes tricky! There are guidelines on co-authorship in dissertations by the Geography Department – know them!

In our field, the most important person is the first author. This should normally be you. The order of authors refers to the contribution these authors made – the person with the highest contribution would be second-author and so on. The last author (also: senior author) can be, but does not have to be, your advisor or the principal investigator of a project. There are some guidelines on who to include, based on five key contributions that can be made:

Generating the research idea. Normally this would be the doctoral or postdoctoral researcher, typically together with the advisor (e.g., because he/she developed the original research proposal). Ideas are often developed collaboratively and credits should be given to all who have contributed in it substantially.
Contributing important data sets or methods. Some collaborators will share data and/or methods they have developed. A good rule of thumb is to ask yourself ‘Could this paper have been done in the same way without these data or tools?’. If the answer is no, then co-authorship may be warranted (otherwise --> Acknowledgments).
Performing the research. Apart from you, this might include other researchers or collaborators and if their input was substantial, they should get credits for their support. However, if you pay colleagues for providing, for example, lab analyses or data, they would normally do not become co-authors.
Contributing to the writing of the paper. Usually one researcher writes the bulk of paper and will be the first author. Co-authors are expected to contribute to the writing, in an iterative process, led by the principal author.
Providing the resources. Carrying out research means someone had to write a successful proposal and make sure you have the infrastructure to carry out research.

Whoever contributes considerably to at least two out of the five points explained above, can be eligible for becoming a co-author.

Important: Again, this is a tricky issue and, therefore, always discuss who to include or not with me, before sending your first draft out – it will be much harder to “drop” someone from the list than to ask someone to join as a co-authors. Also, ask if you are unsure about the order of co-authors. Finally, always be explicit about co-authorships up-front. Research collaborations can end in conflicts if co-authoring expectations are not met. In order to allow people to contribute to the writing, you also have to send potential co-authors a draft at a stage where they still can do so – that is, not too early (so it is clear where the paper is heading) and not too late (when the manuscript is more or less done and inputs can only be limited). Discuss with me when the right time to involve co-authors is. It is a good idea to invite people individually (not in mass emails) and to send people a first draft without a list of co-authors (to see if they will actually contribute). Those people who have contributed in minor, but still important ways should be mentioned in the Acknowledgments.

Affiliations to use

The affiliation in your papers should be the one where the research was primarily conducted and where the lab in which you worked is based (and in case of multiple affiliations, in the order of importance regarding this criterion). In other words, if you join the Lab and publish a paper from prior work, we should not be your first affiliation (you can still mention your current address or, if part of the work was done with us, use a second affiliation). If your paper is published after you have left the lab, the Lab should be your first affiliation and you might want to add your current address as a second affiliation.

Please use the following affiliation:

Firstname Lastname 1,*

1 Geography Department, Humboldt-Universität zu Berlin Unter den Linden 6, 10099 Berlin, Germany

*Corresponding author: firstname.lastname@geo.hu-berlin.de,

phone: +49 30 2093 xxxx, fax: +49 30 2093 6835

After your paper is published

Communicating the paper well

Once a paper is published, it is important to communicate about it so it will find the interested audience you are hoping for. We have a paper release procedure, including several steps of how to spread the news about your paper. Please read the Communication and Dissemination section of the wiki and particular the paper release procedure.

Making sure the data are safe

Archiving your data properly is extremely important. You have created a valuable dataset and spent hours of processing and coding - and this work should never be lost. People might want to continue your work, wonder about how you solved a particular issue, or simply want to show your work in presentations or publications and need high-resolution figures for that. Finally, we often get requests from researchers outside our lab, sometimes many years after people have left the lab, for data and/or methods or code (i.e., your R-scripts or py-Scripts). Lastly. Altogether, this raises the need for a consistent data archiving protocol following the publication of your paper. This document serves as a basic guidance to better archive your data. In case you have specific questions, however, feel free to either consult Benjamin Bleyhl or Matthias Baumann.

Where to create a folder and how to name it? - You need to map our Biogeo SAN drive and navigate to the folder _DATA_ACCEPTED_PAPERS. Within the folder, you should create a folder that contains the year of publication, your surname, and the abbreviation of the journal – each separated by an “_” (e.g., /2018_Baumann_RSE/).

Which data to store? - This, to some extent, depends on the type of work you have been doing. Whatever you store, it should be possible to reproduce your results and graphs from it. If you feel uncertain on what to include, please ask Benjamin Bleyhl or Matthias Baumann (or have a look at existing folders in the archive). In any case, the folder should contain AT LEAST the following parts:

A folder with all figures from the publications in the highest resolution you have (i.e., at least 300 dpi). Name them in accordance to the order in the paper (i.e., Figure01.png, Figure02.png, etc.)
The main resulting datasets produced during your analysis. In case of a remote sensing classification, only the final map should be preserved. Any earlier version should be deleted.
A folder containing all predictors that are not easily reproducible. For example, an Excel file including your response variable (e.g., species richness or presence/absence of leopards) with the predictor variables in columns). The individual predictor maps (e.g., distance to forest edge) can be deleted if they are easily reproducible. If you have performed a remote sensing analysis, please copy your training data and the validation data into separate folders (have a look at 2017_Baumann_GCB for an example).
RAW-Data that are not reproducible and need to be archived! This is particularly important for data collected during field trips.
A folder containing all scripts you have produced during the analysis. These scripts should contain substantial comments, so that at any stage in the script, the reader can follow what has been done.
The PDF of the final paper.
A README-File with the most important explanations to the files and datasets related to this publication. If possible, please also provide an email address through which we can reach you after leaving the lab and Humboldt-University Berlin (not mandatory).