College and Research Libraries PAUL METZ Administrative Succession in the Academic Library Data from a national sample of academic library directors indicate that di- rectors hold their positions longer than had previously been supposed. Li- brary size does not appear to contribute to rapid turnover at the top . A multivariate analysis suggests that female library directors are much more likely than male directors to have been hired to their positions from within the library and that this relationship cannot be accounted for in terms of institutional characteristics. The attitudes of hiring committees or the con- strained geographic mobility of many female librarians may account for this finding, though only future research could determine this with authority. wHETHER IN A UNIVERSITY, an army, ali- brary, or any other form of bureaucracy, the inauguration of a new head administrator frequently marks the culmination of a pe- riod of deep concern and the beginning of a time of reevaluation and redirection. The task of appointing a new leader consumes significant organizational energies, an ex- penditure of resources well justified by the power that leaders hold to shape the futures of their institutions. For these reasons , the issue of turnover in top administrative posi- tions has excited keen research interest, both among social scientists interested in formal organizations in general and among academic librarians concerned with the effects of succession on their particular organizations. The interest that administrative succes- sion holds for social scientists can be traced to a series of seminal articles by Oscar Grusky, who noted that the twin potential of succession for constructive change or for destructive internal jealousies made succes- sion a rewarding research topic. 1 Studies by Grusky and Kreisberg. focused on the frequency of succession and indicated that frequency of succession may be positively associated with organizational size. 2 Paul Metz is special assistant in the Congres- sional Research Service, Library of Congress . 358 I Somewhat less research attention was paid to determining the conditions that af- fect the choice of internal or external suc- cessors, but when this secondary issue was addressed, studies by Newcomer and by Helmich and Brown suggested that larger organizations also tend to hire their top administrators from within more than do smaller organizations. 3 Within the library field, the most influen- tial contribution has been McAnally and Downs' important essay, "The Changing Role of Directors of Academic Libraries." Al- though McAnally and Downs' findings rep- licated the previously found positive rela- tionship between organizational size and frequency of succession, their principal interest was in demonstrating that, at the time of their study and within the domain of the large academic libraries they consid- ered, succession was occurring with shock- ing frequency. They found in 1973 that the median ten- ure of the directors of Association of Re- search Libraries (ARL) institutions was only three years. Sixty percent of the directors of the larger libraries in their sample and 45 percent of the directors of smaller libraries had succeeded to their positions within the previous three years. McAnally and Downs interpreted frequent succession as a conse- quence and sign of heavy pressures borne by the directors of research libraries, such as the information explosion, the pincers of growing enrollments and financial hardship, increased administrative distance from the leadership of the parent colleges and uni- versities , and changes in management theories and in the attitudes towards admin- istrators of library staff. 4 Other library studies have considered var- ious aspects of succession; for example, Louis Kaplan has recently pointed out that various individual factors may play a role in the retirement decision and has provided historical data on the frequency of early retirement among ARL directors . 5 But McAnally and Downs' essay remains the only nearly comprehensive piece of research devoted to the topic. While the research currently available on administrative succession in academic librar- ies increases our knowledge of the phenom- enon, it has left some questions unanswered and others unasked . We do not, for exam- ple , know where new library directors come from or what factors shape the choice of a successor. The purpose of the present study is to illuminate further the most significant aspects of succession by addressing the questions of frequency of succession and the determinants of what kinds of successors are chosen, The study particularly deals with the factors that influence the critical choices made between insiders and external candi- dates to library directorships. METHODOLOGY The data reported here come from the author's doctoral study in sociology at the University of Michigan , "The Academic Li- brary and Its Director in Their Institutional Environments. " 6 Most of the quantitative data come from a questionnaire that was mailed to a random sample of the directors of academic libraries in ·the United States in 1976. The original sample, taken from The American Library Directory, included 311 academic libraries. Of these, forty-five were eliminated from the sample, principally be- cause their libraries served more than one institution (six cases), or because other ref- erence sources did not list the parent in- stitutions as bona fide grantors of the bache- lor's degree (sixteen cases), or because the Administrative Succession I 359 potential respondents indicated that they were simply acting directors (thirteen cases) . The original mailing and one follow-up yielded 215 responses from the 266 remaining institutions for a rate of re- sponse of 80.8 percent. The data from these questionnaries were collated with reference data describing the libraries and their par- ent institutions. 7 Because the reference data had been coded for all 266 eligible institutions, it was possible to compare responding and non- responding institutions in order to test for response bias . The tests revealed no sig- nificant, or nearly significant, differences between the groups in terms of size (de- fined as number of professional librarians on the staH) , public or private sponsorship, or level (graduate or undergraduate only) of the institutions . It therefore seems reason- able to conclude that the data represent all American academic libraries meeting the sample criteria: those libraries that serve only one institution, whose directors hold permanent appointments , and whose parent institutions are fully accredited four-year colleges or universities. · As a supplementary source of insight, in- formal interviews were conducted with ten practicing library administrators and library educators, representing broad varieties of experience and perspective . Because these respondents were not selected randomly, it would be unwise to consider their com- ments strictly as data . Rather, the inter- views should be seen as informed sources for illuminating the quantitative findings. THE SAMPLE Data from the various reference sources and the questionnaire itself provide a profile of the responding institutions and of their directors. The libraries that fell into the final study population employed an average of 9. 9 professional librarians. The mean number of volumes in their collections was 229,000. Thirty-four percent of the parent institutions were publicly sponsored col- leges and universities, and 59 percent of them offered graduate programs of some sort. The mean enrollment of the parent in- stitutions was 3, 990 students. It should be noted that, because the data represent an 360 I College & Research Libraries • September 1978 unweighted national sample, the population of responding libraries includes very few large research libraries. As a consequence, inferences from this population to the body of ARL libraries more frequently considered in the literature should be made with great caution. A number of variables from the question- naire provide a summary introduction to the population of responding library directors. Their average age was 47.5 years. Seventy percent were men. Twenty-one percent were in their thirties, 32 percent were in their forties, and 30 percent were between fifty and sixty years old. Not surprisingly, this general sample of academic library di- rectors is somewhat younger, less ·broadly educated, and more apt to be female than the selective group of ARL directors studied by Jerry Parsons. 8 Yet as a group, their educational preparation for their positions was impressive, as table 1 shows. TABLE 1 EDUCATIONAL ATTAINMENTS OF ACADEMIC LIBRARY DIRECTORS DEGREE PERCENT HAVING DEGREE Fifth-year bachelor's in library science 16% Master's in library science 84% Master's, not in library science 33% Ph . D. in library science or D.L.S . 8% Ed . D . or Ph.D. not in library science 8% Professional degree, not in library science 5% n = 215 FINDINGS The first conclusion supported by the data is that McAnally and Downs may have overstated the problem of high turnover in the administration of academic libraries or may have identified a problem that has abated since the troubled times in which they wrote. Whereas McAnally and Downs found that the median tenure of directors in their 1973 sample of ARL directors was only three years and suggested that the average was no more than five or six years, the me- dian tenure of directors in the present sam- ple was six years. The mean survival in the head administrative post was 8.1 years. Of course, McAnally and Downs' sample was restricted to large research libraries; but even when the present analysis was re- stricted to the 24 libraries in the sample with 500,000 or more volumes in their col- lections, the average tenure of directors was 10.9 years, somewhat more than the overall mean. While the differences between the samples must make any conclusion tenta- tive, it appears thaWibrary directorships do not change hands with their earlier fre- quency.9 The data did not provide useful expla- nations for variations in the frequency of succession, except to suggest that earlier sociological findings about administrative succession in formal organizations in general may not apply to academic libraries. When the tenure of head administrators in their positions was regressed on library size, public or private sponsorship, and degree- granting levels of the parent institutions, not one partial correlation exceeded .10. None was statistically significant at the .05 level. This negative finding suggests that. at least within academic libraries, it may be more profitable to explain succession on a case-by-case basis than to seek an overall explanation in any set of organizational or environmental characteristics. The issue of whether successors should be hired from within the library or from the outside is a sensitive one and evoked re- sponses from the interview respondents that indicated its delicacy. Most noted that hir- ing from within raises a number of organiza- tional problems. The mere consideration of an internal candidate may evoke intralibrary jealousies, and the selection of an insider may indicate an excessive resistance to change. Yet the prospect of advancement to the top position may promote motivation within the library, and the appointment of an insider may foster a useful continuity of policy. 10 While selection committees may not formulate explicit policies for preferring insiders or external candidates, these con- siderations must inevitably help to shape their attitudes toward specific candidates. The data show that the dilemma is typi- cally resolved in favor of external candi- dates. As table 2 demonstrates, internal administrative succession is relatively un- common in academic libraries. This finding, TABLE 2 PROFESSIONAL EMPLOYMENT OF LIBRARY DIRECTORS PREVIOUS TO PRESENT POSITION STATUS Employed by present library Employed by another library In nonlibrary position Attending school Unemployed n = 215 NUMBER 60 121 14 16 4 PERCENT 28% 56 7 7 2 which is consistent with Parsons' inferences from his sample of ARL directors, probably reflects a mixture of a conscious avoidance of internal candidates (who may be too well known, as several interview respoQdents noted) and a desire to seek new directions. 11 DETERMINANTS OF SUCCESSOR ORIGIN While it is important to know how com- mon internal succession in academic librar- ies may be, it is equally important to dis- cover the circumstances that account for variation in hiring patterns. Such a causal analysis is essential to an understanding of the phenomenon. The analysis aimed at de- termining the circumstances that promote internal succession was first performed along traditional lines of sociological analysis, which would suggest that charac- teristics of the employing organization are most significant. Contrary to earlier findings in the sociological literature, internal succession was not found to be a function of organiza- tional size. The data revealed that, if any- thing, larger libraries are more likely to hire from the outside. However, as table 3 shows, this relationship was not strong and did not achieve a satisfactory level of statis- tical significance. Stronger results were obtained when the professional origins of library directors were TABLE 3 ORIGIN OF SUCCESSORS IN SMALL AND LARGE ACADEMIC LIBRARIES Smaller libraries Larger libraries (six or more professionals) Chi-square = 2.5 Internal External Percent Successors Successors Internal 39 21 82 73 32.2 22.3 .10 .50, one degree of fr eedom % Int . 44.4 55.6 Public Private TABLE 8 Men Internal External 7 19 53 68 Chi-square = 2.5 % Int. 11.7 21.8 .10