Statistics ReportFirst, learn how to write a report about original research. A statistics project report is generally similar to other original research reports. Statistical analysis is different from other methods, but before you learn about those differences you should put it all in perspective by reading our article about how to write a great research proposal. After you review the way to write a great research proposal, you can use the tips in this article to prepare for the specific challenges unique to statistics reports.
The introduction should include the conclusion. This is true of most articles and reports. The introduction is written last, after everything else is complete. That way, you can introduce the main idea you're sharing with the reader. That means you must use the introduction to explain what you did (the statistical method) and what you concluded (the results).
Do not explain the basic principles of statistical analysis! Write the report in a way that assumes the reader is familiar with statistical methods. If a restaurant menu had a detailed explanation of the process by which you would order the food, wait for it to be cooked, receive it from a server who brings it to your table, pay the bill – it would be funny, but kind of ridiculous. So in your statistics project report, only explain any unusual/uncommon steps you took during the project, and do not waste time explaining basic principles of statistical methods.
Do not try to say too much in one paragraph. Do not allow yourself to express multiple ideas within a paragraph. Statistics projects are complicated, so keep the paragraphs simple and focused. The first sentence of the paragraph gives the idea shared in that paragraph, and the rest of the sentences explain it. Use short sentences. If you limit each paragraph to one idea, the paragraph becomes a tool for expressing that single idea with excellent clarity.
Test whether your report enables readers to reproduce the procedure you used. Do this by showing the report to one of your peers and asking them to tell you whether they understand it and if they'd need any additional information they would still need in order to reproduce the procedure. You can, in turn, do a check like this for your peer's report. At the same time as you help each other to ensure the projects are reproducible, you can also proofread for one another.
Explain what you used the software to do, NOT how to use the software. This is a common error! Statistical analysis projects can be managed using software for data management, coding, analysis, editing and reporting, but when writing your report DON'T explain procedures in terms of the software you used. You must simply and clearly explain what you used the software to do. Rather than writing, “The researcher used the SPSS function IsElementInVersions,” you would explain, “The researcher used statistical software to determine if the element exists in all versions of the variable.”
The difference is a matter of professionalism. An inexperienced student works through the writing of the report as though it is written only for the professor to see, but a professional stays mindful of the purpose of the report, which is to communicate the project to other people.
Miscellaneous Advice on Statistics Reports
Data management
- Project structure is created (data, code, figures).
- Raw data should not be modified but taken into statistical software and then modified for cleaning or transformation (proc import or proc export).
Coding
- Source code should be organized in logical units.
- Use [R] custom functions in a file so they can be retrieved when needed in a new environment in order to avoid overpopulating.
Analysis
- Set/record the seed when calling RNG or stochastic algorithms (e.g. k-means).
- Use multivariate (Trellis) display instead of one plot per variable and use interactive visualization tools. (example: GGobi).