This paper reviews five studies in which developmental changes in visual selective attention were assessed with a component selection measure. In the first study, performance on the component selection task (in which two components of the stimuli were redundant and could both serve as functional cues) was compared with performance on two incidental learning tasks (in which one stimulus component was task-relevant and the other was incidental). Three sets of integrated stimulus components were used. Attention to the nondominant component increased with age when this feature was redundant with the dominant component but not when it was incidental. These results suggest a developmental improvement in the flexibility of attention deployment; as children grow older they tend increasingly to distinguish between situations in which it is useful to attend to several stimulus features and situations in which it is more advantageous to attend selectively. Studies 2, 3 and 4 examined the flexibility of children's attention deployment in response to instructions. With colored shapes as stimuli (Studies 2 and 3), 5- and 8- to 9-year-old children were able to vary attention to the nondominant color component according to instructions. In addition, 8- to 9-year-olds decreased attention to the dominant shape component when told to attend to color. This result suggests that the older children exercised "attentional trading," withdrawing attention from the dominant component in exchange for increased attention to the other. Instructions to attend to both stimulus components proved relatively unsuccessful with 5- and 9-year-olds when the components were integrated shape and color (Study 3). However, 9-year-olds could accommodate these instructions when the components were separate pictorial elements (Study 4), showing that stimulus integration is an important factor in assessing development of attentional abilities. In Study 5, substitution of nonsense figures for standard geometric forms in color-shape stimuli caused a shift in attention from shape to color as the primary functional cue for learning (measured across subjects). This shift was more pronounced at age 9 than age 5. In general, it appears that children grow better able to alter the basis for identifying stimuli. With age, children show a more complete withdrawal of attention from a component that normally serves to define the stimuli, when it becomes advantageous to redirect their attention to another feature.