Modification of mother-child instructional interaction was examined in relation to specific tasks and the age of the child. Thirty-two mothers taught their 6- or 8-year-old children one of two laboratory classification tasks resembling a home or a school activity. The home task involved putting grocery items on shelves in a mock kitchen, and the school task involved sorting photographs of common objects into a tray divided into brightly colored compartments. It was expected that mothers would compensate for the perceived difficulty of the school task for the younger children by providing greater instructional assistance to this group than to the other three groups (older children in either task and younger children in the home task). The instructional interaction was coded in terms of number of directives, open-ended questions, and nonverbal instruction provided by the mother; involvement of the child; and the extent to which mothers rehearsed their children in preparation for the learning test which followed the instruction (the children were tested on their learning of the organization of items in the tasks). Multivariate analysis showed that the younger children in the school task received more instruction than either group of older children or the younger children in the home task. Univariate analysis showed this pattern to be significant for almost all of the instructional variables. This modification of instruction was accompanied by a slightly better performance on the learning task by the younger children in the school task than by children in the other three groups. (Author/MP)