Complexity of contour in the recognition of visual form.
James Deese · PsycEXTRA Dataset · 1956
The experiments reported here were aimed at the problem of finding out something about the influence of complexity of contour of visually presented forms upon accuracy of recognition of these forms. The results showed that, for forms made of abrupt, right-angled changes in contour, complex forms (those with many changes in the contour) are more accurately identified in an immediate recognition test than simple forms, if the observers are required to remember only one form at a time. If observers are required to remember 10 or 25 forms for recognition, simple forms are more accurately identified than complex forms. For forms with abrupt, varied-angled changes in contour, there is no difference in accuracy of recognition between simple and complex contours. This lack of a difference between simple and complex contours for these forms is associated with a high incidence of reported verbal labelling or naming of the forms. This suggests that if observers can verbally code the forms, the relationships between complexity and recognition disappear. These findings led the author to the conclusion that forms which present relatively large amounts of information are more discriminable (unique), and are therefore easier to identify, but because they contain moreinformation, they are harder to remember. If the observer is able to code forms with previously stored information, the information conveyed in the form may make little difference in a test of recognition. Thus, any attempt to reduce the information content of contours in order to facilitate memory for them may run into difficulty because of reduced discriminability among the forms. If the forms can be coded in terms of previously acquired information, however, both discrimination and retention may be facilitated.