反復試行に伴う同異反応時間の変化と情報処理特性

松田 隆夫

doi:10.4992/jjpsy.45.313

EFFECTS OF PRACTICE ON SERIAL AND PARALLEL PROPERTIES OF HUMAN INFORMATION PROCESSING IN AN ARRAY MATCHING TASK

TAKAO MATSUDA

Author information

JOURNAL FREE ACCESS

1975 Volume 45 Issue 6 Pages 313-323

DOI https://doi.org/10.4992/jjpsy.45.313

Details

Published: February 10, 1975 Received: July 08, 1974 Available on J-STAGE: July 16, 2010 Accepted: - Advance online publication: - Revised: -
Correction information
Date of correction: July 16, 2010 Reason for correction: - Correction: ABSTRACT Details: Wrong : The present study investigated the modes of human information processing for an array of colored alpha-numeric stimuli arranged in 2 rows of 1, 3, 5, and 7 items each, to which each S made the “same” response by pressing one of two keys if both rows had exactly the same spatial arrangement with respect to a given stimulus attribute or information source of either class (letters and digits) or color (red and blue), and the “different” response was to be made if otherwise. In the information processing approach to an array matching task as described above, it will be necessary to give careful consideration to the stimulus attributes yielding the information, and for different stimulus attributes, different processing modes and processing rates must be observed in the “same-different” reaction times (RT). Since the previous study of Matsuda (1971b) indicated that the spatial informations from the class attribute of arrays were processed in the serial self-terminating mode, and that those from the color attribute were processed very rapidly, though the processing mode could not be determined to be either serial or parallel, the purpose of the present study was to clarify how the extended practice might affect the processing modes and processing rates as well as to reexamine the previous findings. Number of stimulus arrays used were 160 in total, 10 for each condition of Attribute (2)×Item-pairs (4)×“Same”-“Different” Responses (2), and the experimental procedure was analogous to the previous one with a main exception that the same 160 trials in a daily session were repeated 8 times (1 to 8 sessions) about 3 days apart. Data from 3 undergraduates out of 4 were analyzed because one S could not complete all trials over 8 sessions. Main results are summarized as follows:
In general, previous findings were all supported by both averaged and individual data. Stimulus attributes had significant effect on both “same” and “different” RTs under almost all conditions of item-pairs by sessions. RTs obtained under the condition of class attribute were significantly affected by the number of item-pairs (N), and the functional relation between N and RT was linear, slopes for “same” responses being 275msec/pair in 1-2 sessions and 219msec/pair in 7-8 sessions, and those for “different” responses 165msec/pair in 1-2 and 112msec/pair in 7-8, on the average. Difference between “same” and “different” RTs under the condition of color attribute was not found, and although the slope of N-RT function in early sessions was very gentle but significant, RTs and N under 7-8 sessions were almost independent with each other, slope being near to zero (4msec/pair). These results were considered to be evidences that spatial informations from class attribute, which resulted from the identification and categorization of items, were processed in serial self-terminating mode in spite of the extended practice, though the latter accelerated the processing rates to some extent, and that those from color attribute, on the other hand, were processed in parallel mode, the practice eliminating some concomitant operations which were not essential but accompanied the parallel processing of colar attribute. Temporal properties of some component processes involved in an array matching task were also estimated tentatively from the simple RT, two-choice RT, and “same-different” RT; that is, central processing time for naming and categorizing each item-pair was approximately less than 290msec, and other internal processes including response selection (about 120-130msec) required approximately 320msec.
Right : “Same-different” responses to arrays cosisting of 2 rows of 1, 3, 5, and 7 colored alpha-numerics were observed over 8 sessions of a four-week period with respect to whether every pair of the upper and lower items had the same attribute of either class (letter and digit) or color (red and blue). In the class attribute, the relation between the number of item-pairs (N) and reaction time (RT) was linear in every condition, and slopes for “same” were 275 msec in 1-2 sessions and 219 msec in 7-8 sessions, while those for “different” were 165 and 112 msec, respectively. These findings supported the hypothesis of the serial self-terminating processing of spatial information from class attribute, though the practice accelerated the processing rates to some extent. In the color attribute, difference between “same” and “different” RTs was not found, and RTs and N were almost independent with each other in 7-8 sessions, though the slope of N-RT function was still significant in early sessions. These suggested that spatial information from color attribute was essentially processed in parallel mode, and the practice had an effect to eliminate some nonessential operations accompanied in early sessions.

Abstract

“Same-different” responses to arrays cosisting of 2 rows of 1, 3, 5, and 7 colored alpha-numerics were observed over 8 sessions of a four-week period with respect to whether every pair of the upper and lower items had the same attribute of either class (letter and digit) or color (red and blue). In the class attribute, the relation between the number of item-pairs (N) and reaction time (RT) was linear in every condition, and slopes for “same” were 275 msec in 1-2 sessions and 219 msec in 7-8 sessions, while those for “different” were 165 and 112 msec, respectively. These findings supported the hypothesis of the serial self-terminating processing of spatial information from class attribute, though the practice accelerated the processing rates to some extent. In the color attribute, difference between “same” and “different” RTs was not found, and RTs and N were almost independent with each other in 7-8 sessions, though the slope of N-RT function was still significant in early sessions. These suggested that spatial information from color attribute was essentially processed in parallel mode, and the practice had an effect to eliminate some nonessential operations accompanied in early sessions.

Corresponding author

Correction information

Register with J-STAGE for free!