Investigated the effects of learner-generated drawing on learning outcome. 196 9th-grade students with a mean age of 14 years were asked to read a science text about the chemistry of doing laundry for comprehension with either no additional support (control group) or under 1 of 4 experimental conditions (drawing groups). In the drawing groups, students were instructed to generate drawings representing the content of the text using a tool bar showing all relevant elements for drawing a picture. In 3 of the drawing groups, students were additionally instructed to either highlight information in the text relevant to the drawing, use a mental imagery strategy, or both. Subjects then completed 3 posttests on transfer, retention, and drawing. Results showed that subjects in the drawing groups displayed higher comprehension of the scientific text than subjects in the control condition. This is attributed to a generative effect of drawing, in which additional generative processing is fostered. Furthermore, students who produced high-quality drawings performed better on the posttests than students who produced low-quality drawings. This result is attributed to a prognostic effect of drawing, in which quality of drawing reflects quality of generative processing. It is concluded that drawing can serve as a generative and prognostic activity.