The evaluation of two computerised instruction programs for arithmetic word-problem solving by educable mentally retarded children

Abstract

To investigate the impact of instruction aimed at improving text-analysis strategies for arithmetic word problems compared to instruction aimed at improving modelling the specific semantic structure of the word problem with concrete material, eighty-four educable mentally retarded children (aged 8.6–16.8) received computer-assisted instruction in a 2 (Text-Analysis) × 2 (External-Modelling) pretest-posttest control group design. The results showed that in test situations with modelling facilities the children who received External-Modelling instruction outperformed the other children, whereas on a test without those facilities the children who received Text-Analysis instruction performed better than the other children. It is argued that children who received Text-Analysis training learned to construct problem representations mentally, which they may also use in the paper-and-pencil situations, whereas children who received External-Modelling training were handicapped in these situations because they were still dependent on the presence of materials with which they could construct a representation of the problem externally.