Limited Magic of Binary-to-Octal Recoding and Memory Span

Abstract

The purpose of this study was to verify the effect of binary-to-octal recoding on memory span. Four experimental conditions were used: (1) all-binary condition consisting of lists of triplets of binary digits where binary digits were used as input and output in a memory span task; (2) binary-octal-binary condition in which the subject encoded binary triplets into octal digits, retained octal digits and at the time of recall decoded octal digits back into binary triplets; (3) binary-octal-octal condition in which the subject encoded the presented binary triplets into octal digits and at the time of recall reproduced the material in the encoded octal form; and (4) all octal condition with octal digits used as input and output in the same memory span task. Memory span was determined for all conditions in terms of octal digits or the corresponding binary triplets. Three subjects were tested repeatedly in all of the four conditions in a block-randomized order after having learned the binary-to-octal code. Results yielded significant increase in memory span in all three subjects when decoding at recall was not required and significant increase in memory span in two subjects when decoding was required in recall. The increase in memory span due to recoding was moderate and far less dramatic than predicted by constant-chunk hypothesis.