This study asked participants to rate models of computer-generated text on the perceived ease with which it could be read and studied. These ratings were submitted to a Q-mode factor analysis to identify the underlying criteria used when the reader/perceiver formed a judgment related to the "study-ability" of the text model. Subjects were undergraduate student volunteers from the University of Nebraska-Lincoln Teachers College. Stimuli were 64 models of computer text designed to use variables that have been studied in both print and CRT legibility research and that are frequently used in text design. The stimuli reflected combinations of six frequently used format variables: leading, directive cues, paragraph indication, hypertext, position of heading, and line length. Subjects performed a Q-sort procedure, sorting the stimuli into seven piles according to the "study-ability factor." They were then interviewed about the criteria used during the task and asked why they rated stimuli as highest and lowest. While conscious of such stimulus characteristics as the presence or absence of directive cues, the length of lines, and double or single spacing, the participants were guided in their judgments by the overall structure, simplicity, and spaciousness of the documents. Twenty-seven references are listed. The experimental stimuli and instructions to subjects are included. (LMM)