A method was devised for examining developmental changes in children's component selection--i.e., the disposition to discriminate among stimuli on the basis of a selected component. In an experiment to develop an optimal measure, problems of three levels of difficulty were administered to children of ages 4, 8, 10, and 12. According to data from an initial phase of the task, the problem of intermediate difficulty level, involving five stimuli, proved to be the most appropriate for developmental assessment of component selection. Results for this problem indicated a general tendency for children to select the shape component of stimuli differing in color and shape, with little developmental change in this tendency. In a second experiment, using similar materials, component selection was found to decrease from ages 4 to 8 and to remain constant from ages 8 to 12. These results were interpreted to be inconsistent with the notion, implied by certain theories, that children's tendency to exercise component selection should increase with age. Further information regarding this process was provided by (a) a three-dimensional measure in the first experiment and (b) two variations of the task in the second experiment that were designed to ensure attention to shape.