Is an out‐of‐role act credible to biased observers and does it affect the credibility of neutral acts? 1

As a rule, out-of-role behavior is thought to have greater credibility than ln-roIe behavior in the sense that the former IS more informative about real behefs than the latter Thus even the simple-heart, seeking a reliable gmde to another's true feehngs and being aware of the social obligations under which this person labors, will discount behavior consistent with these obhgations and pay special attention to behavior inconsistent with the same The theoretical explanation is straightforward (Heider, 1958, Jones & Davis, 1965, Kelley, 1967) In order to evaluate the credibihty of an action the observer first must decide whether the action is internally determined ("he said X because he beheves X is true") or extemally determined ("he said X IS true because his position obhges him to say X is true") When both classes of determinants provide a reasonable explanation for the action (in-role behavior) each is given less weight than when only one offers a reasonable accountmg (out-of-role behavior) Thus, if the act is m hne with one's role obligations then both extemal and intemal causation are plausible, if, however, the act IS out of line, only the latter is plausible It follows then that out-of-role behavior is more credible than m-role behavior because the former is more likely than the latter to be attnbuted solely to intemal causes From a different theoretical stance Skinner and others make a similar distinction between tacts or behavior under the control of discriminative