220 INTERRELATIONSHIPS AMONG TYPE OF REINFORCEMENT, ANXIETY, GSR, AND VERBAL CONDITIONING Koji Tamase Department of Psychology, Nara University of Education, Nara, Japan This investigation examined the effects of type of reinforcement and anxiety level on vevbal conbitioning(vc). The Cattell Anxiety Scale revised in Japan was conducted to 237 undergraduate students to evaluate their anxiety level. From the highest end and the lowest end of the CAS score distribution, 60 Ss were ranbomly assigned to one of the three experimental conditions, with 10 Ss in each cell. The Taffel type procedure was used. The <Ss were tested individually with 60 white index cards, each contained a three-syllable present tense verb selected from Kashu's "T" association list (1962) and the four pronouns: I, YOU, HE, and SHE. Each card was presented at the S's own rate after 10 sec. interval of his response to the preceding card. The first 15 trials were not reinforced in order to establish each S's operant level. For the following 30 trials,e said "Good"in a flat manner (Gr. 1), or "Mm-hmm"accompanying head nod (Gr. 2), whenever an S made a sentence with HE or SHE. The last 15 trials were not reinforced to assess the extinction effect. During the 60 trials the control group (Gr. 3) was not reinforced at all. GSR was recorded throughout the experimental session to see the js"s physiological states. After the last trial all 5s were administered the awareness questionnaire. The following results were obtained: (1) "Good" was likely to be effective as a reinforcer for low-anxiety Ss, and "Mm-hmm" plus nodding seemed to be effective for high-anxiety Ss, but the reinforcement effects were slight as a whole (Table 3); (2) GSR deflections seemed to be greater for low-anxiety Ss than for high-anxiety Ss in Gr. 2, but no such differences were found in Gr. 1 and Gr. 3 (Table 5); (3) "Mm-hmm" plus nodding appeared to influence GSR deflections compaired with "Good"(Table 6); (4) When the data were dichotomized by orienting reflex (OR), measured by GSR deflections for the first reinforcement stimulus, high-or groups tended to be more easily conditioned than low-or groups, though statistical level of confidence was not obtained (Table 7); (5) The frequencies of the four pronouns were significantly different from each other in the operant level (Table ll); (6) Reanalysis of the data revealed that the Ss who showed high-preference for HE and SHE indicated no increasing trends through reinforcemet, but that the Ss who showed low-preference for HE and SHE were fairly condotioned through reinforcement (Table 12).

221 Discussion was made on several possible bases of the finding that the verbal conditioning was not established sufficiently as a whole. The reinforcement by anxiety interaction, relationship between awareness and VC, and relationship between S's preference and VC were also discussed.