Advisory Board Performance: Managing Ambiguity and Limited Commitment in Public Television

Citizen advisory boards are important to non-profit and governmental organiza tions, yet these boards face fundamental problems of ambiguous responsibilities and limited board member commitment. In the present paper a model of these proposi tions is developed and tested. Board performance is operationalized as productivity and board impact, and is expected to be dependent on the development of opera tional objectives and a subcommittee structure, which in turn is facilitated by the financial support of management, and impeded by a large board membership. The model is tested using path analysis on a national sample of federally-mandated Community Advisory Boards to public television stations. We find that the estab lishment of operational objectives and subcommittees is significantly associated with productivity but only weakly with the impact of the advisory board. Board size is unassociated with board performance. The findings support the assertion that successful boards must clarify their roles and develop efficient operating structures, but suggest that the advisory board-management relationship is complex.

Citizen advisory boards are important to non-profit and governmental organizations, yet these boards face fundamental problems of ambiguous responsibilities and limited board member commitment. In the present paper a model of these propositions is developed and tested. Board performance is operationalized as productivity and board impact, and is expected to be dependent on the development of operational objectives and a subcommittee structure, which in turn is facilitated by the financial support of management, and impeded by a large board membership. The model is tested using path analysis on a national sample of federally-mandated Community Advisory Boards to public television stations. We find that the establishment of operational objectives and subcommittees is significantly associated with productivity but only weakly with the impact of the advisory board. Board size is unassociated with board performance. The findings support the assertion that successful boards must clarify their roles and develop efficient operating structures, but suggest that the advisory board-management relationship is complex.
Virtually all organizations have boards of outsiders who pass judgment on organizational policies. These judgments may be binding on the organization or they may be purely advisory. As pervasive as these structural features are we know surprisingly little about these boards. The present study is concerned with the performance of one kind of board-the advisory board.
Advisory boards are composed of non-organizational members who meet intermittently to advise the organization's managers, with members almost always appointed or approved by these full-time managers. Advisory boards are particularly important to non-profit organizations, since they are frequently imposed on non-profits by those (especially government) providing funding. Advisory boards are assumed to provide a &dquo;public&dquo; check in return for public monies. However, uncertainty about what non-profit advisory boards can and should do is a frequent problem. There is confusion about who these boards' constituencies should be and about their roles in the organization. These confusions are exacerbated by the often limited commitment advisory board volunteers give to overcoming the barriers such uncertainty creates. Hence, board members and managers are frequently dissatisfied with advisory board performance (Sewell, Phillips and Phillips, 1979).
The number of non-profit and government advisory boards grew explosively as a result of the federal citizen participation mandates of the 1960s. Fur-thermore, they have been identified as one of the citizen participation techniques most likely to be used during the 1980s (Creighton, 1980). Yet advisory boards are considered a controversial form of citizen oversight by some (Lewis, Houghton and Hannah, 1978) and an outright waste of resources by others (Arnstein, 1969).
Despite several decades of debate there is limited research and no empirical tests of such views. There have been numerous studies of advisory boards and committees (e.g., Public Administration Review May/June and October, 1972) but few of these focus on advisory board &dquo;effectiveness,&dquo; and virtually none are concerned with the relationships between those features board members can control (i.e., its own procedures) and board performance. Studies of the demographic and structural features of the communities from which members are drawn are of little direct use to board members seeking assistance.
The present study tests a model of the efficacy of board practices, such as subcommittee structure and the setting of clear objectives, in the hope of providing useful guidance to non-profit board members seeking tools for building successful boards. This problem is addressed by developing a causal model based on research on boards from the management and citizen participation literatures, and social psychological research on group decision making. This model is then tested on a sample of federally-mandated public television community advisory boards using path analytic techniques.

BOARDS IN FOR-PROFIT ORGANIZATIONS
Governing boards in for-profit organizations have received substantially more research attention than have nonprofit advisory boards. Despite the greater legal power of for-profit governing boards the processes of these two kinds of boards appear to be quite similar. They both face the dual problems of ambiguous tasks and limited member commitment.
Ambiguous responsibilities. In his excellent review of research on governing boards, Mintzberg (1983) argued that there are three important ambiguities in the role of governing boards. Mintzberg's first ambiguity concerns the question of who has the right to membership on the board. He cites a 1977 survey of the largest 500 U.S. corporations which reports that only 1.6 percent of their directors represented major outside shareholders. His second ambiguity is the uncertainty over whose interests the board or board members should represent. Representation is frequently confused; are board members supposed to represent a particular constituency, such as shareholders, or the corporation as a whole? This ambiguity in the representational roles of board members is reflected in the controversy surrounding interlocking directorates which has dominated governing board research. Interlocks, or common board memberships among organizations, are a pervasive practice (Dooley, 1969). The most common interlocks are between companies with head offices in the same commercial centers and between financial and non-financial corporations (Dooley, 1969;Levine, 1972). Do such interlocks suggest the control of corporations by financial institutions, wealthy families, or some other external constituency or do they, as Mace (1948) argued, result from the managerial use of boards for information ? The debate over interlocking directorates is fueled by the ambiguity in the role of boards. Mintzberg's (1983) third ambiguity concerns the uncertainty about how governing boards actually govern or control organizations. Building on the research of Zald (1965) and Mace (1948), he concluded that part-time board members simply lack the information they need to make decisions. Mace (1948) and Zald (1969) found that boards only controlled their organizations in special cases of concentrated ownership or when the organization was dependent on the board for critical financial support such as fund raising or loans. In the absence of clear dependence on an external group using the board to control the organization, the board's responsibilities become ambiguous.
Member commitment. Not only are the responsibilities of governing boards often unclear, this problem is exacerbated by board members' part-time involvement. Those who have studied governing boards note that part-time board members have only limited information upon which to base their decisions. Mace (1948) suggested that board members usually do not ask discerning questions, because they know so little they fear looking foolish. In addition to the part-time involvement of board members their commitment to the organization is usually less than full-time executives because they are not as dependent on the organization as are full-time employees. Board members' limited involvement is reflected in their compensation; for corporate directors it is often nominal, and non-profit organizations are required by law to have an unpaid or voluntary governing board.
Given the ambiguous nature of board responsibilities and the limited commitment of most board members, in practice governing board members often look to full-time managers for guidance and direction. Mace (1948) suggested that more often than not managers select, direct, and judge board members, rather than the reverse.

ADVISORY BOARDS AND CITIZEN PARTICIPATION
The problems of responsibility, ambiguity, and management domination are even more severe for advisory boards than they are for governing boards. Advisory boards are supposed to represent organizations' clients or local community, but their members are usually selected and appointed by the organization's management rather than a specific constituency. Advisory boards exist to advise management, but what does this mean in practice? Does advise mean to make concrete recommendations, or to act as a &dquo;sounding board?&dquo; Should the board initiate proposals or react to management's proposals? In what areas should the board offer advice? Suppose the board members themselves disagree over their role?
Since the Economic Opportunity Act of 1964 mandated the use of advisory boards in local antipoverty programs, federal legislation has required citizen participation in transportation, urban planning, energy, education, and social services. Yet, as Langton (1978) notes, advisory boards are a particularly controversial form of citizen participation. Advisory board members are expected to represent the community's interests in the development of agency policies, but board members are usually selected by the agency, not by the community (see Hannah and Lewis, 1982, for a review of the political power of advisory boards). Activists who feel the citizen's role is to monitor agency performance have criticized ad-visory boards for assisting rather than monitoring, and have questioned their ability to represent the community (Langton, 1978;Arnstein, 1969). Further evidence of confusion about the role of citizen advisory board members is provided by Sewell, et al. (1979).
In their review of twenty-two case studies of public participation, Sewell, et al., (1979) concluded that agency managers and advisory board members usually differed in their participation goals. Agency representatives saw boards as a way to develop public acceptance for programs, increase agency performance, and improve the agency's image. In contrast, citizens saw participation as a way to diffuse the power of agencies and insure that those affected by agency decisions would be able to influence policies. Therefore, it isn't surprising that Sewell, et al., (1979) also found that the citizens and administrators used different criteria to assess participation success: agency representatives measured success by the extent to which the final plans were accepted by the community and whether or not the public image of the agency had improved; alternatively, citizens measured success by the extent to which they modified agency plans. Lewis et al. (1978) suggested that the ambiguous mandate of advisory boards means that the administrators' expectations for their advisory boards is the most critical factor influencing their effectiveness. They argued that the &dquo;empty agenda&dquo; of these groups can lead to time wasted setting goals for the advisory board if there is no clear direction from administrators. Furthermore, membership on citizen advisory boards is a voluntary, unpaid, civic activity. Many advisory board members are likely to see their membership as one more voluntary contribution and so be reluctant to commit substantial time to the board.
Despite these inherent problems of ambiguity and limited commitment, advisory boards can perform important functions. They provide a forum for the exchange of information between managers and interested outsiders. Potentially, they can assist outsiders, by allowing them to influence organizational actions, and managers, by increasing community acceptance and understanding of their actions. Yet advisory boards cannot achieve these functions unless they overcome the inherent ambiguity of their role by developing clear operational objectives, and by finding an efficient way to use members with only a limited commitment so that members feel that they are contributing.

A CAUSAL MODEL OF ADVISORY BOARD PERFORMANCE
Given the ambiguous mandates of advisory boards and the differing expectations of important groups, it isn't surprising to discover controversy concerning the appropriate operational definition of board &dquo;effectiveness.&dquo; Lewis et al. (1978) provide a review of various effectiveness standards and found that they ranged from measures of &dquo;output&dquo; (number of new programs) to characteristics of the board itself (representativeness) and members' attitudes (reduction in board members' feelings of alienation); these authors adopted &dquo;having a significant impact on program&dquo; as perceived by the administrators as their preferred measure. Similarly, Kamienecki and Clarke (1982) defined effectiveness as the &dquo;value of the board's advice as perceived by managers.&dquo; Since the present paper focuses on the internal operations and management of advisory boards, a mea- Model of Advisory Board Performance with Unstandardized and (Standardized) Path Coefficients and Goodness of Fit Statistics sure of board effectiveness called &dquo;performance&dquo; composed of measures of both productivity and organizational impact is adopted.
It is proposed that high performing advisory boards must have both clear operational objectives and a subcommittee structure. Further, the establishment of operational objectives is hypothesized to be impeded by a large board membership and facilitated by a strong management commitment to a high performing board. This model of advisory board performance is diagrammed in the Figure. Objectives. The presence of concrete objectives allows board members to organize and direct their efforts, thus overcoming the inherent ambiguity of their responsibilities. The very process of establishing objectives forces board members to make their various expectations explicit and gives the board a focus. Lewis, et al. (1978) argued that advisory boards have vague mandates and spend much time deciding what they are supposed to do; we suggest that the more time they spend clarifying their objectives the more likely they are to be productive and have a significant organizational impact. This is consistent with the social psychological research on group decision making performance. Hackman and Morris (1979) proposed that group decision making performance is improved when groups spend relatively more time discussing task strategy. Hackman, Brousseau and Weiss (1976) found support for this proposal using a laboratory task that required coordination and sharing among subjects.
Structure. High performing advisory boards are expected to be more likely to adopt some form of internal division of labor, that is, subcommittees. As noted above, part-time advisory board members face a broad array of issues and information. A subcommittee structure facilitates the focusing of attention in a manner that helps compensate for the limited commitment that part-time volunteers give to board membership.
Both operational objectives and subcommittee structure are expected to lead directly to better advisory board performance. But what is the causal relationship between objectives and subcommittees? The development of operational objectives should lead a board to establish subcommittees. The formation of subcommittees implies implicit, if not explicit, objectives, since each subcommittee would be expected to investigate whether or not it is prepared to offer advice on the topics covered by its subcommittee. Therefore, as the Figure indicates, operational objectives are assumed to antedate the formation of subcommittees.
Size. Large boards are hypothesized to find it more difficult to agree on operational objectives than are small boards. The greater the number of members the greater potential variety in their expectations and goals for the advisory board, and so size is expected to be inversely associated with the establishment of operational objectives. With greater diversity more time is required to discuss and clarify operational objectives, time that part-time volunteers may be unwilling or unable to commit. This idea has found support in laboratory research on decision making groups. Large groups seem less able to fully use all of the available member resources (see Hare, 1962, or Steiner, 1972, for reviews). The larger the group the more limited the opportunities for individual participation (Indik, 1965), and the greater the conformity pressures (Gerard, Wilhelmy and Conolley, 1968).
Managerial Support. Although advisory boards may be imposed on organizations, full-time managers still retain a great deal of control over them. Above it was noted that managers frequently control the governing boards which are their legal superiors. With an advisory board, managers retain all of the advantages of full-time commitment and information, as well as the power to decide who shall serve. Further, advisory board members (more than governing board members) are dependent on management for guidance, since they are assumed to want to give advice that will be used by managers and are dependent on these managers for resources. Lewis, et al. (1978) found that the more effective advisory boards had the assistance (but not dominance) of professional staff. Thus, advisory board members are expected to be receptive to managers' suggestions on operational objectives, and a positive relationship between managerial support and the establishment of operational objectives is proposed.

Sample and Procedure
The Telecommunications Financing Act of 1978 required all public broadcasting stations, as a condition of receiving federal funds, to establish Community Advisory Boards by May of 1979 to provide taxpayer oversight. Each station was required to make &dquo;good faith efforts to assure that the composition of its advisory board reasonably reflects the diverse needs and interests of the communities served&dquo; and that &dquo;the board shall be permitted to review programming goals established by the station, service provided by the station, and the significant policy decisions rendered by the station (Public Law 95-567, Section 307).&dquo; Survey instruments were mailed to all 167 public television stations listed in the 1980 Directory of Information Sources for Public Television in April 1981, two years after the stations were required to have their advisory boards in place. Each station received a questionnaire to be completed by the station manager. A followup letter to non-responding station managers was mailed by May. Feedback summarizing the findings was mailed to responding stations in October, 1981. We received 59 usable manager responses, a 35% response rate. Despite this disappointing response rate, analyses indicated that the responses were representative of geographical regions, and large and small stations in the population, but there was a relative oversampling of community licensed stations (39% of responding managers) and undersampling of television stations operated by states, cities, and educational institutions.

Measures
Two indicators of advisory boad performance are available. Board &dquo;Productiv-ity&dquo; is indicated by a dichotomous variable from the questionnaire item &dquo;Has your CAB produced any written documents?&dquo; with a &dquo;yes&dquo; response scored as &dquo;1&dquo; and a &dquo;no&dquo; response as &dquo;O.&dquo; The variable representing &dquo;Board Impact&dquo; is the number of board-initiated changes using responses to the questionnaire item &dquo;Have there been any changes in the station as a result of the CAB?&dquo; with choices including &dquo;No changes, policy changes, programming changes, public access and outreach changes, financial management changes, operations or organizational changes&dquo; summed to get an estimate of the range of the advisory boards' impacts.
Each of these indicators of advisory board performance is limited; yet they appear to be quite distinct, measuring independent dimensions of board performance (see the Table). Productivity does not indicate what was written, but is used because written reports take more effort and deliberation than verbal reports. They are also more likely to command management attention because they are available to outside groups. For example, the first written report of the KCET (Los Angeles) advisory board received coverage in Variety and The Los Angeles Times Oanuary 25,1980), and the management of KCET subsequently published a public response to each of the advisory board's recommendations and implemented many of them. Board Impact ignores the relative importance of each change, but a simple count of the number of areas in which change has been initiated should have a rough correspondence to overall influence.
The variable &dquo;Board Size&dquo; was taken from responses to the question &dquo;How many board members were originally appointed? &dquo;Managerial Support&dquo; is the managers' estimation of the dollar cost (including staff time) of their advisory boards during the prior twelve months. &dquo;Operational Objectives&dquo; is represented by the dichotomous responses to the question &dquo;Does your CAB have specific written responsibilities other than those identified in the Telecommunications Act?&dquo; scored &dquo;yes&dquo; -&dquo;1&dquo; and &dquo;no&dquo; = &dquo;O.&dquo; &dquo;Subcommittee Structure&dquo; is indicated by the number of subcommittees used by the advisory board; the questionnaire item is &dquo;We have subcommittees with the following responsibilities:&dquo; summed for number listed. This variable is surprisingly independent of Board Size (r = .10, n.s.). The means, standard deviations, and intercorrelations among the variables appear in the Table. RESULTS The proposed model is tested using path analytic techniques (c.f., Duncan, 1975;Pedhazur, 1982).~ These techniques allow the testing of a set of hypothesized relationships among variables. In path analysis a set of relationships (paths) is proposed (like those appearing in the Figure). These hypothesized relationships (the over-identified model) are compared with the &dquo;null&dquo; or just-identified model (all variables are significantly related to all other variables). Support for the proposed model is ascertained in two steps: first, are all of the paths that are hypothesized to be significant, in fact significantly greater than zero? And second, does the proposed model &dquo;fit the data&dquo; better than the &dquo;null model&dquo;? The unstandardized path coefficients appearing in the Figure are the &dquo;beta&dquo; weights from the regression equations, the standardized coefficients can, for convenience, be interpreted like correlation coefficients (i.e., O=no relationship, 1=perfect relationship). The standardized and unstandardized path coefficients are presented in the Figure and the original and reproduced correlation coefficients appear in the Table. Managerial Support The proposed model receives only limited support. Board Size and managerial support are independent of the Operational Objectives. However, Operational Objectives are significantly associated with Subcommittee Structure, and both have a positive association with one performance indicator-Productivity. The other performance indicator, Board Impact, is weakly associated with both Operational Objectives and Subcommittee Structure, but their separate path coefficients did not reach statistical significance. The proposed, over-identified model is compared with the just-identified model (i.e., all paths among all variables hypothesized to be significant) with a null hypothesis that the proposed model fits the data. The relevant statistics are reported in the Figure. We are not able to reject the null hypothesis.2 Since a Q5 of .84 and a Q6 of .89 are only moderately close to 1, the proposed causal model appearing in the Figure is a moderately good fit.

DISCUSSION
The proposed model obtained limited support. Attempts to reduce the negative effects of ambiguity and limited commitment through the establishment of operational objectives and subcommittee structures were strongly associated with advisory board productivity, but unassociated with the impact of the board. The exogenous variables of board size and managerial support for the board were, contrary to expectations, unrelated to the establishment of objectives. Although, board size was unrelated to any of the other variables, the Table suggests that managerial support has a direct association (significant correlation coefficient) with the performance indicators, rather than the indirect path hypothesized. Therefore, the present test suggests that the establishment of op-erational objectives and a subcommittee structure leads to greater board productivity, and managerial support is directly associated with both board productivity and impact. The limitations of the present study, the differential impact of objectives and structure on performance, and the direct relationship between managerial support and performance merit discussion.
This study is limited in several ways. First the data are limited to survey responses from the station managers alone. Their views may be different from those of board members, and there is always the potential problem of methods bias (all responses taken from the same self-report instrument tend to be positively correlated). Several of the variables are only limited representations of their concepts (for example, the financial costs of the board would be expected to be only imperfectly associated with managerial support). Finally, the sample size of 59 was disappointingly small. Although this study is a very limited one indeed, it is important to note that it is the first to systematically test hypotheses about the effects of board procedures on board performance. As such it has lead to some interesting findings. The fact that operational objectives and subcommittee structures lead to board productivity but not to board impact could indicate that other, nonmodeled variables may be more important in influencing board impact. It may be that task clarification and an efficient internal structure are more strongly associated with &dquo;quantity/output&dquo; variables like productivity than quality indicators such as impact. After all, for a board to have a large impact it must not only establish clear priorities and efficiently organize its activities (be productive) but it must pick important and meaningful priorities. That is, clear goals and structure are probably necessary but not sufficient for board impact; whereas productivity probably has a stronger and more exclusive dependence on clear objectives and structure.
Management support of advisory boards was not associated with the establishment of operational objectives. However, its significant direct correlation with both productivity and board impact is suggestive. In developing the model it was assumed that managers would take the lead in initiating and directing boards toward their own (managerial) operational objectives. It may be, however, as Lewis, et al. (1978) found, that as representatives of specific constituencies or the local community, the board felt uncomfortable allowing managers to establish their priorities. After all, advisory boards, unlike governing boards, are designed to institutionalize oversight or a kind of &dquo;opposition.&dquo; Our informal observations indicated that board members were quite sensitive to management domination. It could be that managers influence, not through control of the boards' operational objectives (&dquo;setting the agenda&dquo;), but through their support of board activities after the objectives have been set. If managers see that their boards are becoming helpful they can provide staff support and documents, if not, they will probably find that they are too busy to support their advisory boards.
These results also suggest the importance of empirical verification. Both the management literature on governing boards and the citizen participation literature provide useful insights drawn from case studies. However, we could locate no empirical tests of the causal statements in either body of writings. In the present test, we found support for the expected relationship between operational objectives and subcommittee structures and one measure of performance-productivity-but not for the other-board impact. The results of tests such as these help to remind us that empirical research is necessary precisely because it can lead to a more complex and complete understanding of the role of advisory boards in non-profit organizations.
In conclusion, this study suggests that, to be productive, advisory boards need to overcome the ambiguity of their responsibilities and the limitations of member part-time commitment through developing clear objectives and an efficient subcommittee structure. What we do not know is why some boards are unwilling or unable to establish objectives and subcommittee structures. Are they unable to agree on objectives because their individual goals are divergent? Do the participants actually want an active and effective board, or are they content to merely have their names associated with a worthy cause? The present study does indicate that advisory boards wishing to be productive are more likely to be successful if they devote time to resolving the ambiguous nature of their role and the limited commitment of their members through the development of operational objectives and a subcommittee structure.

NOTES
Research supported in part by the Faculty Research Fund of the Graduate School of Management. Special thanks to Bill Stevenson; and to Jane Kaupphan, Betsy Youd Amador, Diane De Moulin and Georgi Wright for research assistance. 1. Productivity is a binary dependent variable, and several writers have suggested that this creates special problems with non-normal error terms, non-constant error variance, and constraints on response function. However, the present analysis was conducted based on Neter and Wasserman (1974, p. 323) who suggest that "the method of least squares still provides estimators which, under quite general conditions, are asymptotically normal." 2. A cautious approach prevents a conclusion that the proposed causal model does fit the data. Joreskog (1974) suggests that the X 2 's should be interpreted with caution, since they are influenced by sample size and in cases such as the present, in which the sample size is relatively small it could lead to retention of the null hypothesis even when the model is not a good fit. The more useful statistic is the nonsample dependent Q, consisting of the ratio of unexplained variance in the over-identified model to the unexplained variance in the just-identified model. Values of 1 would indicate perfect fits (Pedhazur, 1982).