Some Aspects of Problem Solving: Sequential Analysis of the Detection of Embedded Patterns
Report Number: AMRL TDR 62-148
Author(s): Kolers, Paul A., Zink, Donald L.
Corporate Author(s): Antioch College
Laboratory: Behavioral Sciences Laboratory
Date of Publication: 1962-12
Pages: 16
Contract: Laboratory Research - No Contract
DoD Project: 7183
DoD Task: 718303
PB Number: PB114930
Identifier: AD0400310
Abstract:
Performance on a forced-choice target detection task was examined in terms of the sequences of responses subjects made. In about 50 percent of all cases resulting in a correct detection, subjects made a specific error immediately before identifying the correct target. In about the same percentage of cases, that same error was given instead of the correct response. Practice on the task resulted in a general imporvement in speed and number of correct detections, and also in a change in the proportional occurrence of certain errors: the most likely but wrong response (X) was given more often, and the least likely and wrong response (Z) was given less often, with practice. In addition, while the best 25 percent of the subjects detected more targets to criterion and detected them more rapidly than the poorest 25 percent, the proportional occurrence of specific response patterns was approximately the same in the two groups for the targets that were correctly detected. Scaled similarity of the alternatives was an important determiner of erros, while perceived orientation had no relation either to the perceived similarity or to the response sequences. A mechanism is suggested for these sequential acts involving a decrease in probability of detecting sepcific kinds of forms following perception of another kind.
Provenance: RAF Centre of Aviation Medicine
Author(s): Kolers, Paul A., Zink, Donald L.
Corporate Author(s): Antioch College
Laboratory: Behavioral Sciences Laboratory
Date of Publication: 1962-12
Pages: 16
Contract: Laboratory Research - No Contract
DoD Project: 7183
DoD Task: 718303
PB Number: PB114930
Identifier: AD0400310
Abstract:
Performance on a forced-choice target detection task was examined in terms of the sequences of responses subjects made. In about 50 percent of all cases resulting in a correct detection, subjects made a specific error immediately before identifying the correct target. In about the same percentage of cases, that same error was given instead of the correct response. Practice on the task resulted in a general imporvement in speed and number of correct detections, and also in a change in the proportional occurrence of certain errors: the most likely but wrong response (X) was given more often, and the least likely and wrong response (Z) was given less often, with practice. In addition, while the best 25 percent of the subjects detected more targets to criterion and detected them more rapidly than the poorest 25 percent, the proportional occurrence of specific response patterns was approximately the same in the two groups for the targets that were correctly detected. Scaled similarity of the alternatives was an important determiner of erros, while perceived orientation had no relation either to the perceived similarity or to the response sequences. A mechanism is suggested for these sequential acts involving a decrease in probability of detecting sepcific kinds of forms following perception of another kind.
Provenance: RAF Centre of Aviation Medicine