Employee - centered leader focused on the people, their personal success, and the quality of the social system that formed within the work unit. Such leaders had high-performance goals for their work units and communicated their performance expectations to their subordinates. Employee-centered leadership combined a strong concern for the social aspects of the work unit with high-performance expectations. The Michigan researchers felt that their research showed employee-centered leadership more likely led to higher work unit performance than production centered leadership. They also felt that production centered leadership could to high productivity but had several latent dysfunction. The dysfunctions were poor employee attitudes with resulting higher turnover or absenteeism, little group loyalty, and high levels of distrust between subordinates and their leaders. The University of Michigan had an effective and productive program in leadership behaviour and this was carried out by people such as Likert, Katz, Maccoby, Kahn and Seashore at the survey Research Centre. the initial study was conducted by Katz, Maccoby and Morse (1950). The study was done in the home office of the Prudential Insurance Company and was matched with reference to the kind of people, the number of people, and the type of work performed. Each pair was so formed that one section in a pair was a high productivity section and the other section in the same pair was a low productivity section. The level of productivity of each section, whether high or low, was determined from prior work records. The heads of each section were then compared. The comparison was made on a number of variables to see if there was any significant difference between that of a high productivity section supervisor and a low productivity section supervisor. The Michigan studies were different from the Hawthorne studies. In Michigan studies a systematic measurement was made of the perceptions and attitudes of supervisors and workers whereas in Hawthorne studies the researches failed to develop quantitative measures for variables affecting supervisors and workers. In Michigan studies, factors such as type of work, working conditions, and work methods were controlled. When a comparison was made between the productivity of the two section it was found that with regard to demographic variables like age, sex and marital status there was no difference between the high and low supervisors. But when the attitudes and orientation were studied it was found that supervisors in charge of high producing section were found to be employee centered in term of their attitudes, i.e. they had concern for employees as uppermost in their minds. On the other hand, supervisors in charge of low-producing section were found to be production centered in their orientation, i.e. concern for production, was uppermost in their minds, and this was often at the cost of concern for employees.