| Abstract: | Sentence‐writing behaviour, including use of words self‐selected and self‐located independently of teacher modelling, was established with eight mildly retarded children in a special class. The procedure comprised three components: 1) provision of word lists; 2) praise for self‐selecting and self‐locating words from supplied word lists; and 3) response cost for responses dependent on teacher modelling. The latter were termed Words Shown and Words Told, while self‐selected and self‐located responses were termed Words Found. In a sequential component design it was found that after baseline, supplying instructions and a word list (Phase 2) and continuing the use of word‐lists together with praise for self‐managed behaviour (Phase 3) decreased the frequency of Words Told, but Words Found and Words Shown were not appreciably affected. Words Shown decreased with the introduction of response cost for responses dependent on teacher modelling (Phase 4). Also, in Phase 4 Words Found increased substantially above levels in all previous phases and Words Shown decreased to zero frequency. These changes within sentence‐writing behaviour are discussed in terms of the need for accurate discrimination of and selective attention to positive and negative instances in the acquisition of self‐regulation. |
