Submissions to Bayesian Analysis: General Stylistic Considerations
- An over-riding concern is to make effective use of the electronic medium. Papers should be split into two distinct parts: a main manuscript and supplementary material. The main manuscript should be as succinct and readable as possible; the supplementary material can be any length (and in some cases may be unnecessary). The supplementary material may contain additional figures and tables, details about data, mathematical derivations, further explanations, or additional simulations or other supporting studies. The supplementary material should be placed after the main manuscript and before the references. Keeping the pieces distinct will allow someone to download the main manuscript without the supplementary material - but a serious student of the subject would want both. Main manuscripts should contain those parts that authors believe to be essential for all readers. Authors should keep in mind that readership is likely to be increased when the main manuscript is kept short.
- Figures should be used to help with exposition. Where possible, one or more figures should provide illustrations of main results. Figures should be integrated with the text rather than appearing at the end. These figures may be split into multiple parts. The legends on these figures should have enough detail that someone can get an impression of what is being displayed just by looking at the figures. (Explanations of how to read figures should appear in the caption rather than in the main text.) In most cases the captions should articulate the actual findings rather than simply identifying which figure is which. All relevant units should be identified, preferably on the axes (and color or gray scales). Extraneous axis labels and headings usually should be avoided unless they are large enough to be read easily and clear enough to be understood without effort. An often-preferable alternative is to explain the plots more fully in the caption.
- Abstracts should include a brief summary of the main research findings. Generic statements such as "an example is presented" should be avoided. Instead, minimal details about examples should be included, as in "we illustrate the benefit of our approach by analyzing data from a rocket science experiment to provide improved estimates of escape velocity".
- Case studies should highlight both the main results for the scientific, technological, or policy application and also the main points of interest to a statistician working in a different field. The latter are often appropriate for a concluding section.
- Concluding summaries should be provided only in the context of giving additional perspective on the work.