A great deal of interest in the field of education has been generated in recent years over leadership studies. Researchers and practitioners have evidenced a growing awareness not only about leadership but also with leadership behavior. Associated with every position or status in a social system, such as a public school system, there is a set of socially-defined expectations concerning what is appropriate behavior for a person occupying a leadership position. These expectations constitute a "behavioral model" for the incumbent of a position, providing him with a pattern to which he may adjust his own behavior. In addition, to the extent that the incumbent conforms to these expectations, he permits other persons with whom he interacts to anticipate his behavior in prescribed situations and thus enables the interacting individuals to function collectively as an integrated unit. Within this framework, the adult education administrator in the Public School System of the District of Columbia, as the officially designated leader in charge of the adult education program in his specific community, was confronted by major sets of responsibilities. He was responsible to his superiors, his students, his community, but just as importantly, he was also responsive to his own staff. Thus, the purpose of the study was to obtain information to aid in a better understanding of the nature of the leader behavior of the adult education administrator and to provide empirical evidence to serve as a basis for future studies in adult education administration. To fulfill the above purpose, the study was designed to ascertain the degree of divergence or congruence between the actual and ideal leader behavior of the adult education administrator as perceived by the adult education administrator himself and his instructional staff. The Leader Behavior Description Questionnaire, Form XII, was used to measure certain acts of leadership of the adult education administrators. The adult education administrators' self-opinions, as well as, the opinions of their instructional staffs were obtained on 200 items contained in the two parts (actual and ideal) of the instrument. Twelve variables of leadership were measured regarding the leader behavior of ten adult education administrators of regular adult education centers. The data were gathered in ten separate sessions, following initial interviews of the adult education administrators at their respective centers. This information was requested on the basis of how "does" the adult education administrator behave (real behavior) and how he "should" behave (ideal behavior). A total of ninety-one instructors and ten adult education administrators participated in the ten group interviews. Two inter-group hypotheses and two intra-group hypotheses in null form were tested for significance through three analyses of variance procedures: a one-way Multivariate Analysis of Variance, a one-way Univatiant Analysis of Variance and Multiple Correlated t-Tests. The results of the study were such that conclusions could be supported to the effect that both reference groups were essentially in agreement about the adult education administrators' real and ideal behavior; differences of opinions were most apparent when real behavior was compared to ideal behaviors from instructors' descriptions of leader behavior; both the expectations of the adult education administrators in describing their own leader behavior and the expectations of the instructors in describing their administrators' leader behavior were scaled slightly higher than descriptions by both groups on real leader behavior; finally, administrators' expectations were considerably higher than those of instructors on ideal behavior. The research findings also implied a need for a multiple criteria approach to the study of leader behavior effectiveness and an examination of leader behavior from the standpoint of role differentiation in future leadership studies. Ed. D.