Effects of label distinctiveness and label testing on recognition of complex pictures

Two experiments examined the effects of label retrieval upon subsequent recognition of complex, scenic pictures. Contrary to past research, i.e., Bahrick and Boucher's 1968 study, retrieval of labels which accompanied pictures at input was associated with high recognition of "same-photo" (copy) cues but not false recognition of similar "same-scene" cues on a subsequent "photo" recognition test. The label-retrieval effect can be attributed to actual rehearsal of pictorial information and not to item selection. In addition, the beneficial effects of label retrieval did not vary with the informational content of the label itself, i.e., its power to distinguish between same-photo and same-scene items. However, the effects of label-type as well as label retrieval varied with recognition test instructions. In a "scene" recognition test, requiring recognition of both same-photo and same-scene items, effects of label-type were stronger than in the "photo" test, while effects of label retrieval were not significant. The results highlight the complexity of the relationships between label retrieval and picture recognition and suggest several factors which might determine such relationships.