Document Design for Technical Job Tasks: An Evaluation

Abstract

Document design strategies have typically focused on either procedural text or instructional text. This study evaluates the design of yet another class of text, the nonprocedural job text. A nine-page segment of a technical text was redesigned by three separate companies, each specializing in document design. The companies were given complete freedom in redesigning the text. Evaluation was carried out by means of paper-and-pencil tests of fact comprehension and of inferential comprehension for fault isolation tasks. Both speed and accuracy were measured in these "open book" tests. None of the redesigns resulted in more accurate or faster comprehension performance. In fact, one of the redesigns was less comprehensible than the original, as measured on each test. The results are discussed in terms of "reading to do" versus "reading to learn," and in terms of the potential moderating effect of the readers' "format schema" for the traditional design.