We have urged that the maxim for good medical writing is the shorter word in the shorter construction. The same applies to illustrations: they should be simple; they should show what you want to show and no more; they should be honest and not distort your results; and they should be clear at first reading. This chapter outlines how to produce the most common type of illustration in medical scientific papers, the diagrammatic representation of numerical data: the graph. There is instruction on other types of illustration in some of the books in the bibliography; and we will not be giving more than the very simplest of statistical advice.

Tufte (see quote above) uses the idea of data-ink and non-data ink: ink that is needed to indicate a number, and ink that is embellishment. He writes, ‘Just as a good editor of prose ruthlessly prunes out unnecessary words, so a designer of statistical graphics should prune out ink that fails to present fresh data-information.’ The main reason for non-data ink is the now ubiquitous use of computer programs, which produce all types of fancy diagrams at the tap of a key. A magazine review of an upgraded graphic package began: ‘Presentation is about creating an impression, so it is of prime importance to get the material just right.’ For many users of these packages – business people in selling and advertising – style is as important, or even more important, than substance. For scientists, the substance is all that matters. If they use graphic packages inappropriately, they do nothing to improve the already poor presentation of much scientific data. Bernard Dixon was writing of slides presented at meetings when he wrote, ‘[in] a typical, glitzy business presentation . . . the medium truly is the message, as tawdry and vacuous sentiments are given a patina of momentous significance.