Abstract
A study of Morss (1984) suggested that the characteristically poor performance of Down''s syndrome (DS) infants on object concept tasks could be enhanced by embedding the standard task within a short training sequence. This paper reports an attempt to replicate that study and to test for any carry-over of improved performance to subsequent testing sessions 2 and 6 weeks later. Ten DS subjects (mean age: 19 months) took part in the study. The search behavior of almost all subjects improved, both in qualitative terms and in terms of the frequency of successful retrievals of the hidden object. Many also, however, showed evidence of a highly-specific form of ''switching out'' of the testing sessions. The problems that this cognitive avoidance cuuses for establishing statistical criteria for attribution of success/failure on these tasks are discussed.