Research Report Design

How do I make a research report look professional?

Make a small number of decisions and hold them without exception: line length, leading, a type scale of about four clearly distinct sizes, one primary idea per page, a headline that states the finding, a fixed position for bases and sources, and one meaning per colour throughout. Consistency is worth more than sophistication. A plain document applied consistently reads as competent; a sophisticated one applied inconsistently reads as careless, and the carelessness transfers to the analysis in the reader's mind.

What should a chart headline say?

The finding, not the topic. "Reasons for Lapsing" is a label. "Lapsed members left over scheduling, not price" is a headline: it makes a claim, it can be disagreed with, and it survives being quoted in an email with no chart attached, which is what happens to headlines. Test that it is true of that page and nothing more, and watch for causal words the design does not license. "Support contact destroys satisfaction" claims causation that a cross-sectional survey cannot support.

What type size and line length should a report use?

Line length matters more than size and is more often wrong: aim for roughly 55 to 75 characters per line for continuous prose. Set leading at about 1.4 to 1.5 times the type size for body text. Size depends on the reading mode, which has to be decided first; screen reading generally needs larger type than research documents use, and every element including bases and axis labels sits above a stated accessibility floor. Use about four clearly distinct sizes. Sizes that are close together read as inconsistency, not hierarchy.

How much information should go on one page?

One primary idea. Test it by covering the headline and asking a colleague what the page says. If they name two things, one of them is being ignored by every reader, and the page should split. If they say "I don't know", the page may have no finding on it at all, which is common and is usually the discovery that lets a document lose a third of its pages without losing anything.

How do I make a research report accessible?

Measure contrast rather than judging it: at least 4.5 to 1 for body and small text, 3 to 1 for large text and meaningful non-text elements. Set a type size floor and hold it for bases and captions too. Never carry meaning in colour alone, which means direct labelling, position, pattern or shape in charts and a word alongside any colour-coded status. Use real heading levels rather than large bold text, mark table header rows as headers, give meaningful images alternative text, and check every chart in greyscale, which takes two minutes and catches most failures.

Why does my deck feel like a wall of charts?

Because every page carries equally weighted visuals with topic headlines, so no page has a point and the reader has to build the argument themselves. The fix is three things: one primary idea per page, a headline that states the finding, and the deletion of the charts that exist because a distribution existed rather than because a finding did. Where a page carries several charts, one should be visibly dominant and the rest subordinate. Three charts of equal size is not a hierarchy.

Where should the base size and caveats go on a page?

In a fixed position on every page, at a size that can actually be read. A base line in six-point grey satisfies the letter of a disclosure while defeating its purpose. And a material caveat is not a footnote at all: if a reader would act differently knowing it, it belongs in body text next to the finding it qualifies, not at the foot of the page or in an appendix.

How do I make charts made by different people look consistent?

Fix five things once and apply them everywhere: what each colour means (and keep each segment's colour constant through the whole document), whether axes start at zero and how any exception is signalled, the order of categories, whether labels are direct or in a legend, and the weight of gridlines and axes. Rebuild inherited charts rather than patching them. The diagnostic that this has failed is that a reader can tell which analyst built which chart.

Should a research report be dense or readable?

Choose, and write the choice down. A narrative document carries one idea per page and reads well but answers specific questions badly. A reference document is dense, consistently placed and searchable but persuades badly. Both are legitimate. The failure is not choosing, which produces a document too dense to read and too incomplete to search. The usual right answer is a narrative main document plus a genuinely dense appendix, designed to different rules and visibly different.

Can design change what a report says?

It must not, and this is the main risk of the stage. A headline shortened to fit a column is a claim altered by a layout constraint. A base dropped to clean up a chart makes a finding uncheckable. A number rounded to fit a label loses its source. And a page with room for a fourth bullet will get one, which is how content that was never in the evidence enters a document that arrived clean. Design moves and sets things. It does not edit evidence.