Problem solving is recognized as a valuable educational experience in science. Thus genetics, essentially a problem-solving science included in almost all high school biology courses, offers a fruitful area for studying student problem-solving performance. The research reported in this article describes the performance of 30 high school students solving 119 problems generated by the computer program GENETICS CONSTRUCTION KIT (Jungck & Calley, 1985). Solving GCK problems requires students to plan experiments, generate and interpret data, and reason from effects (phenotypic data) to causes (genotypic data). Research data consisted of transcribed audiotapes of students thinking aloud as they solved problems and computer printouts of initial data and sequence of crosses. Transcripts were analyzed for common actions and comments made during the problem-solving process in terms of initial data redescription and interpretation, hypothesis generation, cross data redescription and interpretation, solution synthesis, and solution confirmation. This study was done in an effort to add to the understanding of student problem-solving strategies and to develop a model of student performance—a model that when combined with a model of expert performance may serve as a basis for improving genetics instruction.