抄録
Woodrow performed experiments in weight discrimination with varying standards. And he worked out the theory of ‘set’ to explain the phenomena of time-order errors.
Our primary interest in the present study is in the tracing of his experiments.
Materials:-25 cylinders made of nickel were used, the weight of which ranges from 72 to 120 gr. by 2 gr. difference. They are all 25 mm. high and measure 25 mm. in diameter.
Method:-The weights were lifted by being grasped by the thumb and the forefinger. It took approximately 2 sec. to raise and 2 sec. to lower a weight, and 13 sec. for one comparison judgement, including 5 sec. between R1 and R2. The interval between pairs was 10 sec. Subjects were required to judge ‘heavier’, ‘lighter’ and ‘equal’ in terms of the second weight. Throughout the all experiments there were five variables; one of these was equal to the standard weight of the series, the other four were disposed at equal intervals of 2 gr. above and below the standard.
Experiment I Two experiments were made, one with the standard of 76 gr. and the other with the standard of 116 gr. With both standards, the. time-order errors were negative and approximately of equal magnitude. (See Japanese text p. 823) Experinnent II. was performed with varying standards. In a series 76, 86, 96, 106, and 116 gr. were used as stalldardg, 25 pairs of stimuli were mixed and presented to the Os irregularly in two time orders, and go 50 judgements were required.
The average time-order error was negative and nearly equal to the magnitude obtained with a single standard. With lighter standards it was positive and with the heavier negative. This confirms the results of Woodrow, in our results, however, the magnitude of time-order error seems to decrease with the heaviest 76 and the lightest standard 116.
In order to investigate the second attribute of Woodrow's ‘set’, i. e. the central tendency, we performed following experiments with two or three standards. The caombinatioh os standards is as follows:-76, 116; 96, 116; 76, 96; 76, 86; 106, 116. 76, 96, 116: 86, 96, 116; 96, 106, 116.
The results indicate that the effect of central tendency becomes more evident asthe number of standards increases in the range of weights from 76 to 116 gr. (See Japanese text p. 827)
With two or three standards, the cantral tendency seemed to be more effective when the difference between them was 20 gr. than when it was 10 or 40 gr.
