Abstract

The purpose of the present study was to test experimentally the Fest ingot's (1957) dissonance theory predictions about information selectivity. Subjects were 721 senior high school students. The experiment was conducted under the pretense of reading interest survey. Scholastic aptitude test which was important to high school students was administered for the preparation of dissonance manipulation. The experimental groups were received the test score which was randomly manipulated, while the control groups were not. Three forms of inters! survey questionnair which consist of 14 magazine article titles were administered for measuring the seletivc exposure to informations. The experimental conditions were divided by consonant-, dissonant-information presentation condition and the simultaneous presentation conditions of both information according to the nature of critical article titles of interest survey questionnaires. Interst in relevant information was measured by checking on the 10.5cm， 16-point graphic rating scales below the article titles. Opinion survey questionnaire was administered for measuring the magnitude of dissonance and evaluation of source of dissonance. The main results of present study were as following. 1. Generally the lower the score, the greater the magnitude of dissonance in all conditions. The magnitude of dissonance(pleasant-unpleasant) varied in accordance with the conditions of information presentation. Several dissonance groups showed inversed dissonance to the direction of experimental manipulation. 2. The curves of separate presentation conditions markedly differed from the curves of simultaneous presentation condition of descrepant informations. 3. The active seeking of consonant information from both contitions was not appeared. 4. There was active avoidance o£ dissonant information in the separate dissonant information presentation condition, but the seeking of dissonant information in the extreme dissonance group was not shown in this condition. The consonant information was also avoided as the dissonant information in the moderate dissonance groups of simultaneous presentation condition. 5. The curve of consonant information in simultaneous presentation condition was similar to the curve of Festinger's (1957) gambling study. 6. The degree of seeking consonant information was weaker than the degree of avoiding dissonant information. 7. Interest m relevant informations of the control groups was greater in the separate presentation conditions than in the simultaneous presentation condition (p<.04. one tailed). 8. The Rho coefficient between the magnitude of dissonance and the evaluation of source of dissonance was . 92.