REASONING WITH RAVEN - WITH AND WITHOUT HELP

Richardson (1991) showed that children's performance at items from Rav en's Standard Progressive Matrices could be boosted by converting the elements to real-life objects and setting them in a realistic context. From this, he concluded that people reason exclusively by using conte xt dependent schemas and hence that measures of abstract reasoning abi lity may give a misleading impression of children's skills and potenti al. However, each realistic context item also had a cover story that w as read to the child (ostensibly to activate the necessary schema), an d this study investigated the possibility that these cover stories wer e giving the children additional help, perhaps by focusing their atten tion or by inadvertently giving clues as to the rules. Three factors w ere investigated; whether the items were abstract or set in context, w hether each item was given with or without a commentary, and whether e ach commentary gave weak or strong guidance as to the rules. It was fo und that although items set in context were easier than abstract items their advantage was reduced when like was compared with like, and tha t the strong guidance commentaries were responsible for the greatest f acilitation. It is concluded that the existence of domain-free reasoni ng processes cannot be ruled out, and that it is unlikely that people reason exclusively by using knowledge.