This study is an exploration of a set of variables associated with instructional leadership of secondary school principals in Virginia. Four variables--(a) clarity of instructional goals, (b) performance efficacy, (c) autonomy, and (d) instructional expertise--were used to predict the instructional leadership of the principals. Data for the predictor variables were gathered from practicing principals with a mailed questionnaire. A total of 200 questionnaires (64%) were returned; one hundred eighty-four (59%) were usable. The measure of instructional leadership was developed around seven dimensions of leadership commonly found in the literature on effective schools. Data on these dimensions were gathered from teachers in each participating principal's school. A total of 729 questionnaires (69%) were returned by teachers; seven hundred one (66%) were usable. A principal components analysis of the items on the teachers' questionnaire revealed seven components of instructional leadership. Multiple regression analyses were run for total leadership behavior and for each of seven scales derived from the principal components analysis. Three of the analyses produced significant, but small, R-squares: using praise, rewards, and feedback (R² =.05, p<.05); establishing an environment for change (R² =.05, p<.05); and maintaining an orderly environment (R² =.04, p<.05). The partial regression coefficients indicated that the principals' perceptions of autonomy in decision making were positively associated with their use of praise, rewards, and feedback (b=.04, R<.05) and with their establishment of an environment for change (b=.15, R<.05). This study lends some support to those who believe that the provision of increased authority to the local school principal will provide a climate for innovation and change. / Ed. D.