This report describes efforts to test a model of leadership effectiveness that centers on "macro variables" and "discretionary leadership." Macro variables were represented by the complexity of the environment, context, and structure of a unit. Discretionary leadership was defined as influence over and above that typically vested in a managerial or command role. Empirical testing used a mixture of mail questionnaires and secondary data concerning 68 telecommunications units of the Army Seventh Signal Command. Using correlation and regression analysis, six major propositions and two exploratory aspects of the model were investigated. The results were as follows: (1) greater complexity was associated with more discretionary leadership; (2) structural complexity wa directly related to unit performance; (3) discretionary leadership was more closely related than traditional leader behavior to both performance and employee maintenance; (4) as complexity increased, more discretionary leadership was needed; (5) selected group characteristics did not alter the relationship between leadership and criteria; and (6) the expertise of the unit made no difference in the effectiveness of discretionary leadership. Theoretical extensions and specific applications are discussed in addition to supplementary supporting data. (Author/TE)