Attribution and Social Influence1


  • 1

    The author expresses deep gratitude to Professor Bernard Weiner for numerous suggestions, criticisms, and continuous support in all phases of this research. He is also grateful to two anonymous reviewers for their constructive criticism. Remaining imperfections are the author's sole responsibility. Thanks are also due to Beatrice Knaapen, Karen Lloyd, Erica Stansell, and Alecia Wagner for efficiently serving as research assistants.

Correspondence concerning this article should be addressed to Arolodo Rodrigues, Department of Psychology, California State University, 5310 North Campus Drive, Fresno, CA 93740–0011.


This article utilizes an attributional approach to explore genotypic similarities among Raven's (1965) 6 bases of power. Two scenarios describing a successful influence attempt leading to a good or a bad outcome were created. Following a randomized blocks design, 60 subjects in each condition read 6 explanations given by the target of influence that reflected the bases of power of Raven's taxonomy (reward, coercion, legitimate, referent, expert, and informational). Rating scales following each power basis explanation assessed the perceived causal dimensions of locus and controllability attributed to the target of influence's behavior and how much the behavior led, in the good and bad outcome conditions, respectively, to pride (guilt), self-esteem enhancement (decrement), responsibility (as perceived by self and by others), and gratitude (anger). The results revealed that reward, referent, and informational influence are genotypically similar and perceived as more internal and more controllable than expert, legitimate, and coercive influence. These perceptions, in turn, led to higher degrees of affective reactions, confirming the predictions of Weiner's theory (1986). The results are related to Milgram's (1963) obedience studies and Kelman and Hamilton's (1989) crimes of obedience. Comments on why reward and coercive influence were found to be genotypically distinct in this study are also presented. Implications of the actor/observer bias in studies of this nature are indicated.