Structuring and writing academic papers

These are notes of a presentation by Patrick Dunleavy (at University of Canberra on 22 March) on structuring academic papers. More details (and resources for writing) are on his blog @Write4Research. The presentation proposes two major models for structuring an academic paper:

  1. The conventional model
  2. The design model

Note that whatever method you use, it is important to set up, plan your paper before hand. Check tips in Storyboarding researchHow to do smarter, proactive planning-and-realization of projects, reports and articles, from the outset (Dunleavy, 2014 @Write4Research)

Conventional papers

The advice draws on Thomas Bassboll’s blogs at

A paper should have 40 paragraphs in eight sections

  • 5 paragraphs for introduction and conclusion
  • 5 paragraphs for the general literature review. Do not make it too long!
  • 5 paragraphs for theory that informs research
  • 5 paragraphs for the methods.  This section tends to be ultra-conventional in many papers. Think about what details readers need to know: do not put more, nor less.
  • 15 paragraphs for results, including at least 5 paragraphs discussing the results.
  • 5 paragraphs setting out implications. These can be theoretical and/or practical – one per paragraph?

The beginning and the end are the hardest to write. Ideas at the beginning and the end frame what has to be said. The beginning has to establish a commonality between you and the reader. Do not underestimate readers’ resistance to new work: establish a commonality in the first paragraph.  To write the overall conclusion: mirror the last paragraph of your introduction (have the last paragraph of the introduction reappear at the start of your conclusion – in slightly different wording).

The advice provided is based on an 8,000 words article. If your article has to be shorter, reduce the number of paragraphs on a pro-rata basis.

Designed papers

Such papers are developed to focus on the reader, to present a paper that the reader will notice and that will be cited.

The designed sequence contrasts with papers adopting ‘given structures’ such as

  • record of work done

followed by

  • pure conventional structure used in technical disciplines

Three basic designs are feasible:

  1. The focus-down model.  Perhaps 90% of work done uses the focus down model.
  2. The opening out model (useful in physical sciences)
  3. The compromise model

The core in all these models refers to the value added. You have to decide what the core is from a reader’s point of view.  These would be the value-added materials, these where you talk about new facts or show independent critical power.

The focus-down model


This model has a long literature review and a quick finish. The focus here seems to be on the literature.

The opening-out model

2016 CANBERAA SESSION Structuring and writing journal articles_057

This model proposes a quick set up to move straight into the core. This can be hard to do. Some disciplines do not like it.

The compromise model

Compromise model Start by engaging attention, by motivating the reader. Then continue maintaining the attention of the reader through signposting.

Starting a paper

A high impact start engages readers’ attention (see example below) – to do so it includes a lot of text that frames or sets up the content. This includes

  • Motivation material designed to move readers to engage with the analysis to come. You need to show why your work is important, interesting…
  • Signposts briefly point forward to the sequence of topics in the main sections to come. Title and subtitle are important, they need to fit really closely with what you say – the title tells the story of what follows and as such frames it.

Finishing a paper


Lead-out materials look back and draw out the conclusions of the analysis and their implications, revisiting each signpost  .

Links outward and forward connect the findings or conclusions here to the rest of the relevant literature

Coping with criticisms and ‘working up’ a paper that is not quite ‘there’ yet

Changing your text in response  to reviews and comments

  • Don’t ignore what reviewers say, however much you disagree
  • Even wrong-headed comments have value – they show you where misunderstandings arise
  • Generate a grid box of comments, your reactions, and the changes you’ve implemented
  • Always make some change for every comment

Four structural things to try

  • Do one thing well
    cut out digressions, and secondary materials
    Apply the BBC test – does every element “build, blur, or corrode” the paper? Every piece in your writing does one of these three. Writing that builds moves things forward. Waffle or unclear or un-necessary writing needs to be corrected or deleted. Corroded writing needs to be deleted.
  • Flatten the structure. If you have too many headings without text in between, your structure is probably to complex. You need to flatten it, have fewer sections/building blocs (8 in conventional structure, maybe 4 in designed sequences)
  • Say it once, say it right. Do not say argument in drips and drabs. Group similar points in one place
  • Try paragraph re-planning

Four explanatory things to try

  • Make the motivation clearer
    never under-estimate the ‘strangeness’ of new work
    and resistance to (recognizing) originality
  • Strengthen the argument tokens
  • Improve the data and exhibits. Make it easy to understand. When possible, visualise
  • Refresh the literature review . How ‘old’ is your current one? How new are the latest references?

Further resources

For more guidance see:

Bastow, S, Dunleavy, P and Tinkler, J. (2014). The impact of social sciences. Sage: London

Dunleavy, P. (2012). Authoring a PhD. How to plan, draft, write and finish a doctoral thesis or dissertation. Palgrave study guides.

University of South Australia Learning and Teaching Unit, Structuring an academic ‘argument’ within a journal paper

Writing for research. Twitter: @Write4Research

The impact blog: The impact of social sciences