The present information processing research provided a comparison of visual search and memory search processes with respect to search rates and search strategies. An innovative paradigm was evaluated which varied search type (i.e., visual or memory) randomly between trials of a single choice reaction time (RT) task. The following sequence of events for each of the 280 trials occurred. A ''memory" list of letters appeared, which subjects (N = 12) turned off by pulling simultaneously a left and right-hand RT trigger. After 3 sec, a warning buzzer sounded followed by a variable interval ranging from .5 to 1 sec and the "recognition" list. The subject pulled the left or right-hadn trigger as quickly as possible to indicate whether the memory and recognition lists included the same letter. A single target letter appeared as the memory list on the 140 visual search trials and as the recognition list for the 140 memory search trials. The search set (presented either as a memory list or recognition list) contained three, five, or seven letters.

Memory Search was faster and less accurate than was Visual Search in the present task, The RT slope and serial position data indicated that Visual Search involved a serial, self-terminating comparison process, whereas the Memory Search data confirmed Sternberg's serial, exhaustive model. It was noted that physical characteristics of the task could have biased Visual Search toward a self-terminating strategy. Suggestions for future research in this area were discussed, as well as the author's plans to apply the paradigm to the study of information processing deficits in learning disabled children.