The internal representation of simple geometric networks: Experiments with sequential part presentation

Abstract Three experiments are described on the development of internal representations (IRs) of networks resembling city road maps when these are seen only part by part. In the first study four ways of extracting information from a subject about the structure of his IRs were compared under two modes of sequential part presentation establishing an absolute and a relative frame of reference, respectively. No differences were found between response or reference modes. The second study employed yet another reference mode, the egocentric, and varied the speed of presentation. It was found that in the egocentric mode subjects developed an adequate IR via a purely sequential stage of processing, while a part-to-whole strategy prevailed in the other modes. In the egocentric mode it took longer to develop an adequate IR than in the other modes. No effect of presentation speed was found. In the third experiment the effects of two variables (number of corners, number of crossings) possibly related to a network's complexity were investigated. The results did not make a wholly clear picture and pointed to the necessity of further experimentation.