Research Report Architecture

How should I structure a research report?

Start from the decision it informs and how the audience will read, not from a template. Then answer four questions in order: do the objectives map to chapters or cut across each other, is this one argument or several parallel findings, what ordering principle follows from that, and what job does each chapter do. The conventional structure (summary, context, objectives, method, findings, implications, recommendations, appendix) is a reasonable first guess and a poor final answer. If your last four reports had the same sections, at least three of them were the wrong shape.

Should the report follow the order of the questionnaire?

Almost never by choice. Instrument order is the default only because it requires no decision, and it reliably produces chapters that end without conclusions and a report that finishes on demographics. Order from the argument the reader has to travel. The legitimate exceptions are a prescribed tender or regulatory format, a tracking template read side by side across waves, a data book or appendix, and the case where the instrument already followed a natural sequence.

Should each objective get its own chapter?

Test it rather than assume it. Lay your top findings against the objective list and see which objective each answers. If each objective is answered by a distinct body of evidence, the objectives can be the chapters. If the same evidence answers two or three objectives, a chapter per objective will present that evidence three times, and the chapters should instead be built around the evidence bodies with a coverage table showing where each objective is answered.

Where should the methodology section go?

It depends on the audience and there are three defensible answers. Before the findings, in full, where the reader must judge the evidence before accepting the conclusions: academic, regulatory, technical or contested work. A short statement early with the detail appended, for most commercial reporting. After the findings, for a senior audience reading for a decision. Whatever you choose, material limitations still travel with the findings they qualify: a base of 44 is disclosed where the number appears, not only in the method section.

How long should each section of a report be?

Set the budget by how much the section bears on the decision, then check it against the evidence volume. Left alone, sections grow to fit the analysis behind them, which is exactly how the largest chapter in a report ends up being the least important one. A study that ran forty questions on usage and eight on the thing being decided will produce a usage chapter five times too big unless someone sizes deliberately. Surplus evidence goes to the appendix and the budget holds.

What should go in the appendix of a research report?

Decide by a rule, before you are tired. A workable one: the main document carries what the reader needs to follow and check the argument, and the appendix carries everything else that was analysed. So full tables, the instrument as fielded, the complete method, and secondary findings go back. Three things never go to the appendix for convenience: contradicting evidence, material caveats, and anything moved there because it complicates the story.

Is a research report one argument or a set of findings?

Try stating it in three sentences. If the sentences want "because" and "therefore", it is one argument and the structure should be linear, with chapters that depend on each other. If they want "and", the findings are parallel and the structure should be modular, ordered by importance, with chapters that stand alone. If you write "and therefore", you have the common mixed case: build the linear argument as a spine and place the independent findings after it as visibly separate sections rather than interleaving them.

When should I decide the structure of a report?

After you know roughly what the study found, and before any prose exists. Too early and you are structuring the questionnaire. Too late and the report has already been written into the wrong shape, and changing it costs days and loses source references. The practical test is whether you can state the top five findings. If you cannot, structure the analysis first.

How do I write one report for two different audiences?

Build for the audience holding the decision, and make the other audience's chapters extractable: self-contained, with their own bases and context, so a chapter lifted into their meeting still stands alone. A structure that tries to serve two audiences equally usually serves neither. Where the two genuinely need different conclusions emphasised, the honest answer is one report plus a separate short executive artefact, not one compromised structure.

Should a tracking report keep the same structure every wave?

Separate two things. Conventions (base definitions, segment names, terminology, comparison rules) should be carried forward deliberately, because they are what makes waves comparable. Structure should be re-examined every wave, because the story changes even when the questionnaire does not. A tracking report with the same chapter order for six waves is usually reporting the instrument rather than the market.