The "confrontation technique" has been used successfully in training designs in complex organizations, and this study at once reinforces and extends existing experience. Specifically, what follows is a report on an application of the technique that induced favorable attitudinal changes in the marketing area of a large firm. The similarities of this study with existing ones encompass both design and results. Basically, the design required the exchange of "images" between individuals or organization units. The public sharing of these data apparently reduced the amount of "unfinished business," freed up organization members, and led to favorable attitudinal changes on a variety of before/after comparisons. Thus the study is consistent with a growing major emphasis in the contemporary literature. Numerous observers have cited the necessity and positive consequences of "confronting" in organizations (Davis,:1967; Greiner, 1967, pp. 66-70; Lorsch, 1965, pp. 95-110). Several change agents have successfully varied the basic design of confronting persons in complex organizations with data about how they are seen by some significant others. As in this case, to note one variety of design, the participants may do the actual image building and communicating (Beckhard, 1967; Blake, Mouton, & Sloma, 1965). More or less oppositely, the consultant can play a more active role by interviewing and then synthesizing an image to be communicated to the client system (Argyris, 1962, 1967). A range of variations lies somewhere between these two types (Benedict, Calder, Callahan, Hornstein, & Miles, 1966; Miles, Calder, Hornstein, Callahan, & Schiavo, 1966). Also the design and results of this study are distinguished from other reports in the literature, particularly in four senses. First, the present training experience deals with several functionally related departments. Second, the design encompasses several hierarchical levels down to the first level of field supervision. (Other efforts in the literature tend to deal with two major groups, e.g., labor and management, and apparently with only a few levels of organization toward or at the top.) Third, an unusually wide variety of data was gathered in the present training design. Only gross attitudinal changes in the total population are reported here; i.e., the gross focus here is on changes in how everyone saw everyone else at time, and time2. But the cooperation of the participants permits a range of analysis broader than that of existing reports in the literature. Fourth, the confrontation experience was part of a large and longrange program of organization change. In contrast, existing reports tend to deal with ad hoc efforts.