Stefan Slak

Human ability to randomize sequences as a function of information per item

Sixteen Ss were asked to produce a random sequence of binary and ternary numbers in a counterbalanced within-S design. For each S and each condition, relative redundancy up to the fifth order was determined. Relative redundancy increased as a function of order of redundancy estimate significantly faster in the ternary-digit condition. It was concluded that it is more difficult to randomize items containing greater amounts of information.

INFORMATION, SIMILARITY AND VOCALIZATION IN SERIAL LEARNING OF DIGITS

32 Ss were given a serial learning task involving either quinary or decimal numbers. The reading of items was silent for one half and aloud for the other half of the Ss in a 2 × 2 factorial design. As evidenced by trials-to-criterion, quinary numbers were significantly more difficult to learn. Vocalization resulted in slower learning, but the difference was not significant. There was no interaction between the two factors. Additional analysis revealed a strong difference in the serial position curve between the two vocalization conditions, the aloud condition resulting in a stronger recency and weaker primacy effect. In learning numerically coded information, the decimal code was interpreted as superior to the quinary code because of higher information per item and lower intralist similarity.

Alprazolam Withdrawal Insomnia

A 49-yr.-old man volunteered to be interviewed about withdrawal effects of alprazolam after having taken alprazolam for 3 mo. During the alprazolam-intake period the subject reported improved sleep though he had not suffered previously from insomnia. Upon withdrawal, he experienced bad mood for a few days and severe insomnia for at least 3 wk. The conclusion was made that more attention should be paid to withdrawal-induced insomnia in case of intermediate-acting benzodiazepines such as alprazolam.

Sequence redundancy under conditions of randomization and spontaneous activity

The purpose of the study was to compare the sequence redundancy under conditions of randomization with the sequence redundancy under conditions of spontaneous activity. One group was instructed to randomize three alternative push-button responses, and the other group was asked to produce a spontaneous sequence of such responses. Both tasks were self-paced and performed twice. Redundancy up to the fourth order was computed. Completion times and information production rates were determined. Redundancy was significantly higher in the spontaneous activity condition. Information production rate analysis confirmed that there was a greater degree of nonrandomness in the spontaneous activity condition. The effect of test order on redundancy was explained in terms of the difference in completion times between randomization and spontaneous activity conditions. Redundancy was characterized as a possible measure of behavioral stereotypy.

Long-Term Retention of Random Sequential Digital Information with the Aid of Phonemic Recoding: A Case Report

A single S learned a random sequence of 1,152 digits in 34½ hr. with the aid of a prelearned phonemic recoding system, while he was able to learn only one-third of that length without phonemic recoding in the same period of time. After a 3-mo. retention interval, S remembered two-thirds of the learning sequence, while in the control condition there was no evidence of recall. Utility of multiple recoding of binary sequential information was demonstrated. Results were compared with previous findings on phonemic recoding. Practical applicability of phonemic recoding of sequential digital information was suggested.

Simple reaction time as a function of preparation for a subsequent choice reaction time task

The purpose of this study was to test the hypothesis that an increase in the number of alternatives in the choice reaction time task would lengthen the reaction time in the inunediately preceding simple reaction time task because of higher attentional preparation demand. The simple reaction time task involved an auditory click and was performed between the warning light and the visual choice reaction time task. The hypothesis was confirmed. The increase in simple reaction time as a function of information load of the subsequent choice reaction time task was explained in terms of an increased amount of conunitment of limited-capacity attentional system and was interpreted as signifying that there is at least partial parallel processing of the two tasks. The relation of these results to some recent data on the effect of expectancy and limited-capacity short-term memory was discussed.

Phonemic recoding of figural information and memory span

The purpose of the present study was to verify the effect of phonemic recoding of figurai items on memory span in two individual subjects who, prior to the experiment proper, learned a phonemic code allowing them to recode bivariate figurai items into pronounceable two-phoneme syllables. Memory span was determined with the use of a phonemic code and with the use of the natural language where subjects in a memory span task read the items aloud in terms of the phonemic code or the natural language, respectively, and reproduced the items in figurai form in both cases. The hypothesis that reduction in the length of verbal representation with the use of phonemic recoding would increase the memory span was supported by the results. Phonemic recoding was compared with the traditional phonetic mnemonic system.

Information, similarity and vocalization in free learning of numbers

Thirty-two Ss were given a free learning task involving either high-similarity-low-information quinary numbers or low-similarity-high-information decimal numbers. Reading of the items was either silent or vocalized. The design was a 2 by 2 factorial. Quinary numbers were more difficult to learn. Vocalization resulted in slower learning. There was no interaction. Decimal code was interpreted as superior to the quinary code.

Isolation effect based on digit composition of numbers and its stability

Forty Ss were given a serial-learning task involving 10 three-place numbers. The sixth item either had similar digit composition or was isolated by having a different digit composition. A strong and stable isolation effect was observed. The isolation effect was interpreted as being due to a lesser degree of formal similarity, and its stability to the absence of spread of isolation effect.

Short-term memory for figural items as a function of the number of variable dimensions

A total of 108 undergraduates were given a memory span task for figural items varying in one, two, or three dimensions. Memory span in item units significantly decreased, and memory span in bits of information significantly increased as functions of the increase in the number of variable dimensions and bits of information per item. Results are interpreted as additional evidence that the constant chunk hypothesis is wrong and as further extension of previous data on univariate and bivariate information storage.