A popularly held assumption that young children learn more effectively when they can manipulate " real " objects than when the same concepts are presented pictiorially was tested in an experiment with 36 four-year-old children in the Los Angeles Children’s Centers. A supplementary question was whether these young children could learn concepts and vocabulary concerned with the home environment in a replicable instructional program.

All the children were pretested and assigned on a stratified random basis to one of three treatments. One group received training with three-dimensional manipulanda, a second group received exactly the same program except that the instructional materials were line drawings, in color, of the three-dimensional objects, and a third group served as a control. On the posttest, both experimental groups were significantly superior (p<.01) to the control group. Contrary to expectation, however, no significant difference attributable to the types of materials used for instruction was found. The only difference between the two instructional procedures was that the program with manipulanda tended to take more time.