ABSTRACT
Imagine a situation in which an individual's reputation for leadership was an accurate index of his actual exercise of leadership. Imagine, further, that finding the reputation of individuals for leadership, together with a little interviewing, would enable a researcher to determine the distribution of leadership in a community. Granted these assumptions, the problem of power would become amazingly simple. For if reputation pointed to the under-
. lying reality, there would be no need to painstakingly study many decisions in order to observe leadership in operation. One would have the far simpler option of asking who are reputed to be leaders, charting their policy preferences, and inferring from this data the realities of the structure of power in the community.
