34 Writing a Report
Writing is a form of structured thinking. As you explain your methods and results, you will often discover gaps in your reasoning, questionable assumptions or flaws in your logic. However, this does not mean that you are writing only for yourself. You are writing for a reader whose perspective is very different from yours. They have not spent time with your data, and they are unlikely to care which R function you used. Readers are generally focused on the results.
Assignments sometimes condition students to prove that they know what they are doing by showing every step they took. That is not the purpose of a report. A report is a communication tool, not a diary of your work. Readers want to know what you found and have enough information to understand how you found it, but they do not need to see every step along the way. Most importantly, they want to know what the results mean and what, if anything, they should do about them.
34.1 The structure of a report
A common instinct is to structure reports in the same order as your work was done, starting with data acquisition, then cleaning, methods, results, and interpretation. That order records the analyst’s effort, but it does not match the reader’s priorities.
The structure of a report will vary depending on the audience and the purpose of the report. One reasonable structure for your report is:
- Introduction section: The introduction provides the question that your report will be addressing and overviews the answer. It should also provide a bit of the motivation for why this is an important or interesting topic.
- Background section: The background section provides context for the question and answer. It should provide a bit of the history of the topic, and any relevant information that will help the reader understand the question and answer. This is where you could provide some maps and yield trends for your RM.
- Analysis and results: This section would contain your main analysis and the results of this analysis. You should make it clear what method you are using (e.g., a linear regression with a certain response variable and particular explanatory variables). You should also provide the results of your analysis, including any tables or figures that support your answer. Importantly, you should also interpret these results for your reader – what do these estimates mean?
- Conclusion: The conclusion should summarize the main findings of your report, and provide any recommendation. It should also discuss any limitations of your analysis, and what you cannot say about the relationship between weather and yields.
34.2 Some general tips on writing your report
- When reporting numbers round to the appropriate level of precision, and include the units.
- Provide context for your numbers. Not all readers will know that 20 bushels per acre is a poor yield for wheat in Saskatchewan.
- Even better than context is comparison to a benchmark. “Yields in 2022 were 20 bushels per acre, compared to an average of 50 bushels per acre in the previous five years.”
- Report only the numbers that are needed to support your argument. A table with 40 rows is not a summary if you only discuss three values. Move supporting detail to an appendix or provide the data file separately.
- Use a table when the reader needs exact values. Use a chart when the comparison, pattern or change matters more than individual numbers. Every table and chart must be introduced and discussed in the text. Captions should identify the measure, groups, time period and units. They should not repeat the whole paragraph above the figure.
- The method section should let a knowledgeable reader understand and reproduce the analysis without becoming a diary of software commands. For example, write
We calculated the mean 2025 wheat yield for each soil zone. Four of the 150 fields had no reported yield and were excluded from the averages.
Instead of:
We clicked Data, made a PivotTable, dragged yield_bu_ac to Values and changed the calculation to Average.
Include the limitations of your data and methods. This will actually increase the credibility of your report and the reader’s confidence in your findings. Some limitations include:
- Coverage: data that cover only one province or a short time period.
- Missing data: observations or variables absent in a systematic way.
- Assumptions: a weighting rule, category mapping or price used in a calculation.
- Causation: are you reporting an observed relationship that was not created by a controlled experiment (Chapter 22).
State each limitation plainly and explain its consequence. “The data include insured acres only, so the provincial average may not represent uninsured production” is more useful than “results should be interpreted with caution.”
Aim for simplicity not complexity in both your writing and your description of results. Many students reach for “two dollar words” to impress the reader, but this often makes the report harder to read.
34.3 Formatting reports in Word
There are several tools for writing reports. The most common are Microsoft Word, Google Docs and LaTeX, each with its own strengths and weaknesses. In this course, we will focus on Microsoft Word because it is widely used in industry and in other classes you will take.
That does not necessarily make Word the best tool for every report. LaTeX and Quarto provide stronger support for reproducible reports, particularly when a document contains equations, citations, tables, figures and results generated by code.
The document below provides guidance on Word-specific formatting.
Use the Course 261 Word Formatting and Submission Guide for Word-specific instructions on numbered sections, paragraph styles, page layout, figures, tables, equations and final submission checks. Unless an assignment explicitly requests another file type, submit the finished report as a PDF so that its margins and formatting remain fixed.
34.4 A checklist before you submit
Run through this before you hand anything in. It is the list I use on my own papers, so some of it is aimed at longer research work – not every submission has equations or a bibliography. Skip what does not apply.