Scientific writing

This workshop is an introduction to the very basics of science writing for those with little or no experience. Work through the tasks below and do add to the tips and list of resources.

TASK 1: Gather a list of the research papers that interest you most. Extract a list of the journals these authors have published in. Remember that journals have impact factor ratings and you get more points for higher quality journals. Check ISI Web of Knowledge (Sussex Web of Knowledge) and the 'additional resources' link on that page to find out which journals (Sussex electronic journals link) rank the highest - aim realistically high.

Each of these journals will have different slots for original research papers, reviews and other types of paper. For example, the journal Biomedical Imaging and Intervention describes the different kinds of papers as:

• Original article
• Case report
• Technical note
• Pictorial essay
• Review
• Commentary
• Editorial
• Letter to the editor
• Others
• Non-scientific material

Follow the link to get a description of each of these. Which ones do you think are the most time-dependent (This will also depend on the journal)? Compare these descriptions to those used by the journal Nature.

Task 2: Look up your list of journals online.

If you are affiliated with a university, the library will have an 'electronic journals' page (e.g. (Sussex electronic journals). If not, you may find that the electronic journals offer early views of Online First articles for free. Download a few papers from different parts of the journal and compare them.

Take a look at this image. It is a description of a research paper by Jorge Cham, the creator of PhD Comics. For incremental science, it is certainly a case of 'many a true word spoken in jest', as you can see from this handout from MIT, which emphasises the importance of establishing a research territory, a niche, and showing the importance of the research area. Whether this is to be described in scientific or socio-economic terms has been a political hot potato for a long time. (See George Monbiot's article in the Guardian this week).

TASK 3: Write two or three sentences describing why scientists think your research field matters. If you can, add another few sentences relating this to important social concern. How might you use this to explain to a policymaker what your research is about, and why it is relevant to policy and society?

The title of your paper will strongly hint at the major research question or hypothesis in your work. If you pick the title well, it will also refer to other work and show how your work relates to what has gone before, or what is new or distinctive about your approach.

TASK 4: Look at your list of journals and think about the research questions/hypotheses explored in your work. Which topic looks most appropriate for which journal? You will know this by being familiar with the papers from the journals - read, read, read! If a good match jumps out at you, you have a place to start.

TASK 5: Write down a working title and specify the journal, and article type.

The best overall web resource for writing a research paper is this one from Columbia University.

This MIT presentation gives good tips on how to pick a title, and the structure of a paper. Here is another good article, this time looking at how to structure a paper in maths.

It goes without saying that you should do a thorough review of literature. Dont just rely on PubMed, go for a research database with a wider disciplinary reach, such as Web of Knowledge. If you are locked out of academia, try searching Google scholar, or doing an advanced Google search limited to academic domains. Similarly, if you want to look beyond the academic literature, don't be afraid to gain insight from trade press, other grey literature and news media (just be aware of the advantages and limitations of each source type). Work your networks to get access to data sources as well as literature. Write it up as you go along, so you can capitalise on the energy of discovery.

NOTE: It is well worth getting in touch with some of the best authors in the field, perhaps asking them for a copy of a specific paper if you cannot get it, or asking them if they are about to publish anything new tackling a specific issue. Researchers are surprisingly willing to send copies of papers 'in Press' to those who demonstrate a real engagement with the topic.

TASK 6: Create a structure for your research paper. Who are the key authors that you will need to cite in the introduction? Acknowledge the historical basis of your work. This is especially important because some of these people may well be your peer reviewers.

TASK 7: Check out the 'Guide for authors' for the journal you wish to write for. What system of referencing does it use? How many words should your paper be? Which section are you going to write for? Who is the editor of that particular section? Do you know anyone who has written for this journal, or who knows the editor of this journal? Ask around, or do a search for your department or institution to find out.

You can also see the parallels between Jorge's illustration of the importance of the graph, and this presentation, 'How to Write a Research Paper' about the importance of graphics, from MIT. The MIT presentation will give you lots of tips about how to improve your graphs and illustrations.

TASK 8: Create a list (just one or two to start with - maybe that is all you need) of potential figures, tables and graphs.

TASK 9: After getting advice from those who have published in the journal before, begin writing (if you havent already - it is easier to write as you are going along). Start with the section you find easiest to get over writing inertia. This can be the research questions or methods section. Dont start with the abstract! This is the last thing you should write.

TASK 10: Finish the paper! You end by writing the abstract and maybe tweak the title.

Use either the Numeric (Vancouver) or Harvard system of referencing. Sussex has a good page describing these styles too.

The content of a research paper is the product of a long effort and communications process. Frank Stockdale of Stanford explains the purpose and process of scientific writing in this video is available from ITunesU. A companion lecture on writing grant proposals is also available.

Why not join the Nature Network? Here there is a group called 'Ask the Nature editor' for scientists who want to learn more from the journal editors about writing papers and other aspects of getting their work published in the Nature journals. You can post questions in the forum. Another Nature Network group is the Good Paper Journal Club, in which members post and discuss examples of well-written papers.

Keep them short! Yes, one page will do. Tips on how to write policy briefings: here is a great set of instructions from Citizens for Public Justice. Also here is another one from an AIDS charity.

STEMPRA is the organisation for PR in Science in the UK. You can find examples of Press Releases at AlphaGalileo and other sources such as Science Daily.

See an Introduction to Science Journalism, and the activities of the Association of British Science Writers.

To find more opportunities for communicating science, look at the Introduction to Science Communication in the UK and the British Council's Talking Science.