FOR several years researchers have asked managers and sub-ordinates to make a variety of judgments about their work rela-tionships, involVing issues such as control, supportiveness, satisfaction, and effectiveness, among others. Obviously, social interaction plays a significant role in these relational judgments, because without interaction relationships could not exist (Cappella, 1984). Although there is a substantial body of research directed at the interrelationship of these relational assessments (for example, consideration's relationship to job satisfaction), there is considerably less work on the actual sequences and structures of manager-subordinate interactions that underpin these relational judgments. The implications of this neglect are two. First, there is an

overreliance on perceptual data to measure and understand relationships, and its use has been much criticized (for example, see Eden & Leviatan, 1975; Ericsson & Simon, 1980; Nisbett & Wilson, 1977; Rush, Thomas, & Lord, 1977; Schriesheim, Kinicki, & Schriesheim, 1979). Such influences as selective memory, social desirability, implicit theories, or capricious answering in response to unexpected interpretive probing can confound the measurement of relationships.