The ability to draw visual representations such as diagrams and graphs is considered fundamental to science learning. Science exams therefore often include questions which require students to draw a visual representation, or to augment a partially provided one. The design features of such questions (e.g., layout of diagrams, amount of answer space) could, however, influence students' ability to respond to the questions and present potential accessibility issues, which in turn could influence the validity of score inferences. This article reports on a small-scale study examining the accessibility of GCSE science questions involving the creation and augmentation of visuals (e.g., adding an element to a partially provided diagram) by analysing the patterns of question omit rates. Omit rates for questions involving creating or augmenting visuals were compared to those for questions without, and these comparisons were conducted across tiers, subjects, question position, maximum marks and facility values, as well as by gender and attainment group.