1996 年 11 巻 p. 46-60,145
This study investigates the process model of public opinion formation mediated by perceived media impact.
Davison (1983) coined the the term “third-person effect”: i. e., individuals tend to perceive a stronger impact of media message persuasiveness on others rather than on themselves. Several studies, the reafter, found the evidence of this tendency to overestimate the media's impact on others as compared with the self (Perloff, 1993, for a review). Davison also proposed that individulals are inclined to cope with perceived others' attitude change as a consequence of the media's impact, i. e., people change their own attitude or behavior in response to the perceived others' change, which means that they themselves are influenced by media messages in question (the third-person effect hypothesis). This hypothesis suggests that perception of media's impact mediates the actual impact.
The third-person effect is related to several social psychological phenomena. First, the notion of self-other distinction (perceived discrepancy between self and others) is relevant to “fundamental attribution error” (e. g., Ross, 1977). Second, “pluralistic ignorance”, which means misperception of social distribution of opinion, is related to the perceived discrepancy. Third, the idea that people's expectations are the key to their actual behaviors is substantially paralleled to the argument of “spiral of silence” hypothesis (Noelle-Neumann, 1984). The hypothesis suggests that those who perceive themselves as minority hold their tongues in fear of expected isolation.
Relating to these phenomena, the present author proposed the process model of public opinion formation through the third-person effect as follows; The greater the perceived third-person effect is, the larger the discrepancy between one's own opinion and expected public opinion will be (Hypothesis 1). Also, as the discrepancy increases, the perceivers will change their attitudes or behaviors all the more (Hypothesis 2).
These hypotheses were confirmed by the author's two studies. n study 1, the third-person effect was correlated with the expectation of discrepancy between one's own opinion and public opinion. Study 2 showed that the third-person effect facilitated the intention to speak out, which was not predicted by the spiral of silence hypothesis.