A sample of colleges and universities that advertised for a provost or vice president for academic affairs in 1979 was studied to determine participants' perceptions of search committee operations. For 45 participating schools, useable questionnaires were returned by 253 search committee members, 31 successful job candidates, and 25 department heads. Among the topics covered in the questionnaire are the following: the mechanics of the search process, details of committee composition, perceptions of the participation of various campus groups, reasons for candidate selection, and background data about members and successful candidates. Institutional size and type of institutional control were found to be very useful for distinguishing differences among career patterns of successful candidates. To qualify for a position as a chief academic officer, experience as a dean and other full-time administrative experience was usually more important than experience as a department chair in large, public institutions. The reverse was true in small, private institutions. The findings suggest that committee members do not select candidates who represent themselves "writ large" in professional terms or in demographic terms. In the five years preceding the search, more successful candidates than members had published books, chapters, or monographs, and more had at least one article in a national journal. Other differences are noted. Additionally, distinct differences emerged between committee members and the pool of candidates they select for consideration on several demographic variables (sex, race, age). The three study groups had quite different perceptions about what criteria were most important in the selection process. (SW)