Abstract
60 children aged three, four and five years were given simple addition and subtraction problems in a variety of task forms. Overall, children showed a considerable degree of competence on these tasks: however, performance was significantly affected by age, social class, size of number involved and form of task presentation. Most children were successful when the tasks were embedded in concrete situations ‐‐ either real or hypothetical ‐‐ and when the numbers involved were small. A significant number of children also succeeded on large‐number versions of these embedded tasks. Very few children, however, succeeded when the task was phrased in the formal code of arithmetic ('what does one and two make?') irrespective of the size of number involved. The educational implications of these findings are discussed.
Keywords