Presentation on theme: "Will G Hopkins Auckland University of Technology Auckland NZ Reviewed/enhanced by Steve Olivier University of Northumbria, UK Writing Up: Thesis, Journal."— Presentation transcript:
Will G Hopkins Auckland University of Technology Auckland NZ Reviewed/enhanced by Steve Olivier University of Northumbria, UK Writing Up: Thesis, Journal Article, Lay Report, Poster Thesis: format, feedback. Article: content, authorship, impact factor, dealing with editors/reviewers. Lay report: content, language, statistics. Poster: See separate presentation for content, language, statistics. Research Skills Workshop Minor updates July 2008, Sept 2012
Thesis Write it as paper(s) for a journal, if your institution allows it. Put extra material into appendices. If necessary, expand the paper(s) into traditional thesis chapters. Supervisors: give feedback by marking a double-spaced manuscript with a pencil+eraser. Do it in the presence of the student as often as possible. Identify problems. Avoid rewriting more than the occasional phrase or sentence. Its OK to have major input into the results chapters. But you should have only minor input to the preface and conclusion chapters. Give feedback to remote students over the phone, via Skype, and/or via a single-spaced tracked-changes Word doc.
Journal Article General Choose a journal with the right focus and impact. Format headings, tables, figures, and references in the style of the journal. Bounce the tracked-changes Word doc between authors. Get feedback from a colleague or expert before submitting it. The following guidelines are mainly for articles based on quantitative methods.
Title Write it first. Try to state the finding in the title. Don't make it too specific, or people might not read or cite it. Authors Agree about authorship issues early on in the project. Here is the usual arrangement… First author is the person who wrote the first draft. Should be the student if paper arises from the thesis. Corresponding author is the person who can best address queries of reviewers and readers. Anyone who has made a substantial contribution to the study should be an author and should have an opportunity to contribute to the manuscript. Have you left anyone off?
Summary/Abstract It's by far the most important part of the paper. Write it before the rest of the paper, not after. Rewrite it after you have finished the rest of the paper. Start with a rationale for the study: state why you did it, not what you did. Include as much detail of methods as possible. Include probabilistic magnitudes of effects in the results. End with the main conclusion: state why or how it's useful, not a rehash of what you found. It should be within a few words of the prescribed length. Be as economical with words as possible, but do not compromise grammar. Avoid abbreviations here and throughout. Dont make up any. Do not include references, figures, or tables.
Introduction Pose the problem in one or two sentences. Review the literature briefly in a manner that leads to the importance/justification of your study. Do not state or test hypotheses. Do say you estimated effects. Methods Use subheadings for Subjects, Design, Measures, Statistics. Use a figure for a complex design. Cite references for published methods. Describe others fully. Statistics Justify sample size in terms of precision of estimation, available resources/subjects/time, and/or prior studies. Describe the modeling of dependent and predictor variables. Mention confidence limits, not p values. Mention clinical or mechanistic inference, not statistical significance. Mention standard deviations, not standard errors of the mean.
Results Show subject characteristics as means and standard deviations. In an intervention… show pre-test means and SDs as subject characteristics; show change-score means and SDs to give an impression of any individual responses; show differences in mean changes, with confidence limits; calculate any individual responses as a standard deviation. Show scattergrams only for a good reason (outliers, non-uniformity of error, non-linear trend). Summarize multiple outcomes in a figure or table. Never show test statistics (t, F, 2 ). State chances of magnitudes for the most important effects. Avoid repetition of outcomes in figures, tables, or text. Do not discuss the findings or interpret them qualitatively.
Discussion State the main finding, then discuss how technicalities (reliability, validity, specific problems with equipment, ) might have impacted it. Interpret the magnitude of the main and any other findings qualitatively. Reconcile the finding with those in other articles: a qualitative mini meta-analysis and Bayesian interpretation of your finding. Explain possible mechanisms and confounders. Include variables you have and haven't measured. Devote space to discussion of a finding in proportion to the certainty of its magnitude. Introduce no new results. Explain any major limitations in generalizing to the real world.
Conclusions/Recommendations/Applications If the journal has such a section… State the main findings and/or applications in plain language, without being too repetitious. It must stand alone; therefore… cite no references refer to no tables or figures. Make no substantial new points of discussion. Avoid generalizations and "should"s that go beyond your findings. Indicate useful future research, if you haven't already in the Discussion.
Acknowledgements Remember the funding body. It may help to have consulted and acknowledged an expert. Name the people who have provided help with equipment or testing. References Use Endnote to avoid omissions, extras, and bad formatting. But the references have to be entered correctly in Endnote! Figures See Preparing Graphics for Publication in Sportscience. Create with GraphPad Prism or SigmaPlot, then copy into PowerPoint or other vector drawing editor, ungroup, and clean up. Excel+ PowerPoint 2007/2010 produces permanent ugly thick lines when you ungroup. Excel+Powerpoint 2003 work OK.
Submitting the Article Take impact factors into account when you choose a journal. Take the copyright policy into account, too. Follow the instructions for authors! Check you have the most recent version. Until you are a big shot, opt for anonymous review, if possible. The peer-review process is flawed. If your paper gets rejected outright for bad reasons, ask the editor to send the manuscript out for further reviewing. Don't accept unscientific criticisms. Justify your stand. But some reviewers spend a lot of time on manuscripts, so try to be courteous and grateful. When you become a reviewer or editor, be a Mother Theresa and nurture the authors, regardless of their attitude or ability.
Lay or Popular Report Get one or two images into it somehow, Take digital pics during the project. Ask subjects if it's OK to put their pictures in the report. If you use images like this off the Web, be wary of copyright infringement. Posing a useful question early on holds attention. Get the main finding across as soon as possible. Then report the details. Put boring stuff in appendices, if possible. US Olympian Dominique Dawes
Use plain language What's more, not moreover. So, not therefore. We, you, not the passive voice or one should… No technical terms (jargon) without a simple explanation. No double negatives like the previous point. Explain magnitudes and uncertainty/precision qualitatively. Don't use difficult statistics like correlation coefficients, Cohen's effect size, p values, or confidence limits. Do say it is unlikely, likely, very likely that whatever has a trivial, useful, major or whatever effect on whatever. Do say that the true effect of whatever might be as small as… or as large as…
This presentation is available from: See Sportscience 7, 2003