Understanding School Boards and Their Use of Different Models of Governance

In this article, the authors apply Mel Gill’s (2002) description of governance model types to a national sample of school board members in the United States. The authors find that the majority of school board members take a policy-driven approach to board governance, while a substantial percentage take a traditional approach that delegates clear authority to the organization executive. Multinomial-logistic regression analysis and a series of analysis of variance tests are used to identify the structural and group dynamic differences between difference governance model types. The authors find that governance models have an impact on the group dynamics of organizations and that board approaches to governance differ substantially by area, concluding that future studies of governance models should consider the differences in governance strategies across functional areas.

Attend any public or nonprofit management conference in which practitioners are involved, and you are likely to find certain skepticism regarding best practice governance models designed to improve the performance of boards.The authors have been asked firsthand on multiple occasions: does this apply to my case?Implicit in the question is the reality that board governance is a value-laden enterprise that is impacted by the individuals serving on the board, the structure of the organization, the customers the board is serving, the regulatory environment, the region, the type of service of services provided, etc.
Despite the complex nature of governance, the pursuit of governance models with potential to improve organizational performance remains of interest to scholars and practitioners.Why?An approach to board governance that transcends the situation-specific nature of public and nonprofit organizations could have a positive impact on public performance.Take, for example, the case of public education, the focus of this article.A large body of research, summarized nicely in Ravitch's (2010) sprawling history of the American public education system, demonstrated that school and school district performance is largely a function of the types of pupils enrolled in a school.Imagine, however, if the governance behaviors on a school board, something board members can control, could offset the impact of some of the many things that a school board cannot control.Indeed, previous research by the authors (Ford & Ihrke, 2015) and others (Grissom, 2014) demonstrated that school board governance can, under the right circumstances, have an impact on the academic performance of school districts.Other research by Svara (1990), Herman and Renz (2000;2004), and Brown (2007) similarly suggests and/or demonstrates a link between board governance and the performance of public and nonprofit organizations.However, research linking governance to performance, while explanatory in nature, does not prove the worth or existence of a comprehensive governance model.
In this article, we use Gill's (2002) descriptive language of four governance models (traditional, operational, policy model, and management) and national data from elected school board members in 49 states to answer three research questions: 3. Does the specific model type used by board members impact the dynamics and performance of a school board?
This study is exploratory in nature.While we hypothesize based on our previous work (Ford & Ihrke, 2015) and the work of Herman and Renz (2000;2004), Brown (2007) and Grissom (2014) that board members reporting the use of different models have variation in their board dynamics and performance, little research exists to determine exactly how we should expect those variables to vary across models.In addition, because we allow survey respondents to pick the language that best describes how their board goes about making decisions, the results should not be seen as an indictment or celebration of any specific governance model but rather an indicator of determinants and impacts of the perceived governance behaviors of school boards in our dataset.In the following sections, we first lay out the background of governance models and review the corresponding literature; then we provide an overview of our data and methodology, followed by answering our research questions one-by-one using various quantitative measures.Finally, we provide a discussion of the implications of our findings.

Background and Literature Review
Governance is a relatively new topic of study for scholars of organizations.Most scholars agree that we have a limited understanding of not only what governance is but also how it works (Bradshaw, Hayday, & Armstrong, 2007).Paradoxically, there is a lot of talk by academics and practitioners about governance "models," as if they are an available and accepted option for overseeing organizations.We are not sure why this is the case given how little research has been done on different governance models as to their effectiveness.
There is increasing pressure for organizations to adapt and change, particularly in a world heavily influenced by forces such as globalization and technological innovation.In response, we have seen the evolution and emergence of new organizational structures in all three sectors (Miles, Snow, Mathews, Miles, & Coleman, 1997).For example, we are seeing more and more examples of organizations from the public, private, and nonprofit sectors working together to solve problems and, in some cases, save money and other resources.While these organizational forms continue to evolve and emerge out of necessity, there is a corresponding need to better understand how these forms are to be governed.The options available for governing are often referred to as "models" of governance, suggesting there are different ways of going about the governing process and that finding the right fit is one of the challenges of leading these new emerging forms.Interestingly, we have observed these calls for using models to govern these new and emerging forms, but we remain uncertain as to what the research tells us about the effectiveness of governance models for single organizations.
Within the nonprofit sector, it is common to hear academics and practitioners discuss different models of governance and their experiences with them, both good and bad.Renz (2007), one of the foremost thinkers on governance in the nonprofit sector, suggests that "[g]overnance is the process of providing strategic leadership to a nonprofit organization.It entails the functions of setting direction, making policy and strategy decisions, overseeing and monitoring organizational performance, and ensuring overall accountability.Nonprofit governance is a political and organizational process involving multiple functions and engaging multiple stakeholders" (p. 1).How nonprofits go about governing can vary; ostensibly, we are told that this variation can be captured in different types of models.
Academics tend to be critical of any kind of claim that there is one model of governance that fits all circumstances, particularly in the nonprofit sector (Abzug, 1996;Brudney & Murray, 1998;Dornstein, 1988).Practitioners want help with dealing with their boards and the idea of a model that helps the board get its work done is attractive because of its simplicity.So what gives and how can we contribute to the many challenges of understanding how the governance of organization works and why it matters?
There are many examples of models of governance; for the purposes of this research, we will utilize only those that pertain to the public and nonprofit sectors.We will further limit our discussion to models of governance relevant to local governments, the category of governments that schools fall under.Here we will examine four models of governance common in the literature on both public and nonprofit organizations.
The authors readily admit that any discussion on the nature and extent of the use of governance models is potentially fraught with problems, particularly when any comparisons are done between organizational sectors, for a several reasons.With this research, we borrow from the literature on nonprofit organizations to inform our research on school boards, normally considered a type of local government, but today schools come in numerous alternative forms including the nonprofit form.We borrow from the nonprofit literature simply because there has been more written about governance models in this literature than in the public sector literatures.
We do not consider nonprofit boards to necessarily be the same as traditional public school boards, the members of which are elected rather than appointed as with nonprofit boards.This institutional feature of how members get on these respective boards has implications for what we expect are the dynamics on these boards, such as the amount of conflict board members experience while governing.Nonprofit boards, we surmise, tend to have too little conflict due to the volunteer nature of board service and the reputations and relationships that must be managed by board members in the communities where they serve.By managing reputation and relationships, nonprofit boards tend to avoid or suppress difficult discussions on controversial topics that could actually lead to better decision-making.We also surmise that traditional school boards, with elected board members, tend to have too much conflict.Many school board candidates run for office for the simple reason of getting rid of administrators or board members currently in place.When they get on the board, they are ready to make changes and are often unwilling to work with the other board members to come to solutions about difficult problems.
We have no doubt that the institutional context matters as to the dynamics that take place on governing boards in the public and nonprofit sectors.What we do not know, however, is whether the institutional context matters in the extent to which boards in these two sectors use different governance models, and it is also not the focus on this manuscript.Here, we explore the extent to which democratically elected school boards use different types of governance models and ultimately whether governance model usage has an impact on the dynamics and performance of school boards.
In the public sector, governance refers to processes of regulation, coordination, and control (Rhodes, 1997).A traditional distinction found in the literature on cities has involved a discussion as to whether their forms of government are "reformed" or "unreformed."Reformed governments are a product of progressive era reactionary reforms intended to take the graft and corruption out of governments that had been a part of the spoils early in the 19 th century.
Reformed governments tend to have smaller councils and feature at-large elections with nonpartisan ballots.Unreformed governments have larger councils and feature district elections with partisan ballots.Structural features are what distinguish these governments from one another and, as intended by reformers, one ends up getting a different type of governance depending upon which form of government is used by a given community.Svara (1990) has laid out how governance differs in these two types of communities.Frederickson, Johnson, and Wood (2004) have taken the government form distinction a step further and suggest that cities can range from highly "political" on one end of the continuum to highly "administrative" cities on the other end of the continuum.In between these extremes are "adapted political," "conciliated," and "adapted administrative" cities.Yet research on local governments tends to continue to use the traditional dichotomy of government form (Nelson & Nollenberger, 2011).
There are other structural features that can vary across boards, such as the extent to which they use committees and subcommittees to divide labor.But structural features are not the only aspects of governing boards that can vary.They also can vary in terms of who -board and/or staff -participates in governance decision-making.Further, they can vary in terms of their focus.Some boards will focus on writing policies, while others will focus on the day-to-day operations of the organization they govern.Out of all this variation have come numerous attempts to categorize boards and how they govern in governance models (Gill, 2002).
Gill identified four models that are commonly used in the public and nonprofit sectors.The operational model is the first of these models and, in the nonprofit sector, tends to be the model of choice for new organization that have no staff and that must rely largely on board members and other volunteers to achieve their aims.Operational boards also have management responsibilities but are distinguished from management boards by their lack of staff support.
With the operations model, the board has as its primary focus the operations of the organization.
The second model identified by Gill (2002) is the managerial model.The board manages operations, although it may have a staff coordinator.Board members actively manage finances, personnel, and service delivery directly or as committee chairs and report directly to the board.Staff reports to board member managers either directly or through a dual reporting line to a board member and a staff coordinator.With the managerial model, the board has as its primary focus the management of operations.
The third model is the traditional model.With this model, the board governs and oversees operations through committees but delegates management functions to the CEO.Committees, established along functional lines (e.g., finance, human resources, programs) that parallel management functions, are used to process information for the board and sometimes do the work of the board.The committee structure and ambiguity in roles may invite board interference in management functions (Gill, 2002).In most cases, the CEO has a primary reporting relationship to the board through the chair.Gill suggests that, with the traditional model, the board has as its primary focus the governance of the organization, which Houle (1997) supports in his classic work on nonprofit boards.
The final model is the policy model.With this model, the board governs through policies that establish organizational aims ("ends"), governance approaches or processes, management limitations, and that define the board/CEO relationship.The CEO has broad freedom to determine the "means" that will be implemented to achieve organizational aims.The CEO reports to the full board.In its purest form, the board does not use committees but may use task As with the traditional model, Gill (2002) suggests that the policy model has as its primary focus the governance of the organization.
There is a limited amount of literature involving the testing of some of these models as to their effectiveness in governing organizations, particularly when it comes to the traditional (Duca, 1997) and the policy (Brudney & Nobbie, 2002;Nobbie & Brudney, 2003) governance models.However, there does not appear to be much in the literature on public organizations and the effectiveness of these models, yet we know from experience that these models are commonly used in local governments.Our goal with this research is to explore the frequency of usage of these models by school boards and then try to assess their effectiveness.

Data
The data for this study comes from two sources.The first is an original survey of school board members conducted by the authors in late 2013 and early 2014.Individual board members were surveyed; thus the board member is the unit of analysis.However, we use the perceptions of board members to learn about the boards themselves.The survey instrument was informed by a national survey of school board members conducted by Hess and Meeks (2011) in cooperation with the National School Boards Association, municipal governance surveys conducted by Johnson and Ihrke (2004) and Ihrke and Niederjohn (2005), and original questions developed by the authors.The 89-item survey was sent, via the Qualtrics online survey tool, to the over 28,000 democratically elected school board members with a publicly listed e-mail address in all U.S. states excluding Hawaii (which has only one appointed state-wide school board).All e-mail addresses were mined from school district websites by the authors.Overall, the authors obtained a 17.7% response rate, a rate slightly lower than the previously mentioned Hess and Meeks (2011) and Ihrke and Niederjohn (2005) surveys.As a check against the possibility of response bias, we compared the characteristics of school districts represented in our sample with the population characteristics of all American school boards.The results, displayed in table 1, indicate that the graduation rates, racial demographics, and instructional spending of districts included in our sample are fairly similar to the population.While this comparison does not rule out the possibility of response bias, the similarities give us a degree of confidence in the representativeness of our sample.
Once data collection was completed, each respondent was matched with data from our second data source, the National Center for Education Statistics (NCES).The NCES annually collects and releases a wide variety of demographic and performance data on each of the almost 14,000 school districts overseeing the delivery of public education in the United States.The pairing of these two data sources enables us to combine soft governance measures collected via survey with hard measures of demographics and performance.Table 2 lists the summary statistics and their source for the variables used in difference aspects of the forthcoming analysis.The survey measures explain difference aspects of the group board dynamic as perceived by school board members.The NCES measures are all variables beyond the control of school boards, shown in previous research by Hanushek (1997), Ravitch (2010), and Ford and Ihrke (2015), to have an impact on the performance and behaviors of public school districts and the boards that oversee them.Tables 3 and 4 summarize the survey variables included in two additive indexes used to measure board conflict (see Ihrke & Niederjohn, 2005) and the presence of regenerative relations (see Golembiewski, 1995) on school boards.As can be seen in tables 3 and 4, both additive indexes hold together well with Chronbach Alphas of 0.84 and 0.87, respectively.
Table 5. Board Member Agreement with Model Descriptions Percent N Traditional: The board governs and oversees operations through committees established along functional lines (finance, human resources, programs) but delegates the management functions to the superintendent 30.66 979 Operational: The board manages, governs and performs the work of the organization.

169
Policy Model: The board governs through policies that establish organizational aims (ends), governance approaches, and management limitations.These policies also should define the relationship of the board with the superintendent.The superintendent has broad freedom to determine the means that will be used to achieve organizational aims.

1,958
Management: The board manages operations through functional committees that may or may not have a staff coordinator.

Results
In this section, we use the data described in the preceding section to explore the previously stated research questions.Descriptive statistics are used to answer the first research question: what governance model are school board members using?To answer this question, we first asked school board members which of the statements, as listed in table 5, best describes the way in which their board goes about making governance decisions.The model descriptions were adopted word-for-word from Gill (2002).As can be seen, the majority of respondents (61.32%) indicated that the description of the policy model best described the way in which their board makes decisions.Almost one-third of respondents (30.66%) indicate that the traditional model description in which the board oversees operations through committees and gives the executive management authority best describes their board governance behaviors.A small percentage of board members (5.29%) believe that the operational model in which the board is highly involved in the day-to-day operations of the school district best describes their governance behaviors.Last, a very small percentage of respondents (2.72%) agree that their board manages organizational operations through committees.
At first glance, these results appear to indicate that American school board members generally take a policy-driven approach that defines overall organizational goals, and then gives the executive broad authority to meet those goals.However, in addition to asking board members directly which model description best describes their board, the authors attempted to operationalize Gill's (2002) governance models by asking which of the statements below best describes the way in which their board goes about making decisions in each of five key functional areas (Gemberling, Smith, & Villani, 2000) 1 (as follows):  Operational: "The board as a whole deliberates and makes decisions." Management: "The board makes decisions based on committee recommendations." Traditional: "The board delegates decisions making authority to the superintendent." Policy: "The board follows its established policies when making decisions."We expected to find consistency in the approach to decision-making in the key areas of financial management, personnel management, student academics, public perception of the district, and dealing with state government.However, as can be seen in figure 1, the approach to decisionmaking varied widely depending on the functional area in which that decision was being made.
In the areas of financial management and managing public perception of the district, board members favored a hands-on operational approach to governance.In the areas of personnel management, student academics, and relations with state government, board members favored a traditional model approach where authority is delegated to the superintendent.Notably, the policy model, where the "The board follows its established policies when making decisions" was chosen as the best descriptor of board decision-making by about 20% or less of board members in each of the key functional areas.Given that over 60% of board members chose the policy model description as the best descriptor of their overall governance behavior, there is clearly a large disconnect between the ways in which boards view their overall governance behavior, and their governance behaviors in regards to specific areas.More discussion of the possible meaning of this disconnect is included in the conclusion section of this article.
We answer our second research question (do school district characteristics predict the type of model used by board members?) using a multinomial logistic regression model predicting board member responses to the statements listed in table 5. Multinomial logistic regression is appropriate given the categorical nature of the dependent variable (Long & Freese, 2006).We note that all independent variables included in the model were obtained from the NCES, while the dependent variable is from the survey instrument, thus mitigating any potential problem of common source bias (Favero & Bullock, 2015).
The results of the model, as displayed in table 6, compare the impact of each independent variable on the likelihood that a school board member will identify his or her board governance  Included in the table are relative risk ratios, which, when "above one" indicates that an increase in the independent variable increases the likelihood of a respondent selecting that particular model description compared with the policy model description, and when "below one" indicates that an increase in the independent variable decreases the likelihood of a respondent selecting that particular model description compared to the policy model description.
The results in table 6 show several statistically significant relationships between school district characteristics and model type.First, we note that we control for the number of respondents per-board to prevent serial correlation.We find that board members overseeing districts with higher enrollments are less likely to identify the operational model as the best descriptor of their board's governance behavior.Board members overseeing districts with higher percentages of minority pupils are less likely to identify the traditional model descriptor as representing the way in which their board makes decisions.In addition, board members overseeing districts with higher percentages of special needs students are more likely to identify the traditional and management models as the best descriptor of their board's governance behavior.Board members serving in suburban districts are more likely to identify the traditional model.Finally, board members overseeing districts with larger per-pupil revenues are more likely to identify with the management or traditional model description; however, the size of the effect is extremely small and substantively meaningless.Overall, the results support the hypothesis that certain district characteristics, including the percentage of minority and special needs students, as well as suburban location, does have an impact on governance models reported to be used by American school board members.We explore our third research question (does the specific model type identified by the board members impact the dynamics and performance of the school board?) through a series of Kruskal-Wallis rank tests and one ANOVA test.As discussed in the introduction, part of the allure of a comprehensive governance model is that model governance behaviors can be adopted by boards and, in turn, improve the dynamics of the board, i.e., reduce conflict, improve productivity, enhance leadership, in ways what ultimately improve the performance of a public or nonprofit organization.The Kruskal-Wallis and ANOVA methodologies, though limited by their inability to show causation, does allow for the identification of differences in the group dynamics and performance of boards using different governance models.This is a crucial first step, as it is necessary to determine if there are differences by model type if scholars hope to explain why those differences exists.
Table 7 displays the rank sums of multiple variables collected from the authors' survey instrument and the NCES, along with a corresponding χ 2 statistic showing whether or not the group differences are statistically significant using the nonparametric Kruskal-Wallis test.After each test, Dunn's pairwise comparison post-hoc statistics are calculated to see which models are significantly different that one another.The first variable, the conflict index, is the previously explained additive index of negative conflict types (Ihrke & Niederjohn, 2005;Jehn, 1995).There are significant group differences in the level of conflict reported by school board members across models, with post hoc tests showing differences between all groups except operational and management.The second variable is the regenerative relations index consisting of variables measuring the level of trust, and owning of decisions, perceived by board members (see table 4).Board members giving a higher score have better perceived group dynamics, which, in theory, will improve overall organizational performance (Gabris & Nelson, 2013;Golembiewski, 1995).As shown in table 7, there is significant variation in the regenerative relations index, with post hoc tests showing differences between all models expect operational and management.
The next variable was obtained from a survey item measured on a 5-point Likert scale where board members were asked to state the extent to which they agree their board is highly productive.The higher the score, the more strongly board members indicated agreement.Here again there is significant variation, though differences only exist between operational and traditional and operational and policy.The next two variables, also measured on a 5-point Likert scale, are measures of relationship and entrenched conflict included in the conflict index (Jehn, Northcraft, & Neale, 1999).Both of these questions also showed significant variation across models, with post-hoc tests showing differences between all groups expect operational and management for both variables, and operational and traditional for the entrenched conflict variable only.
The next variable deals with board member perceptions of leadership.We asked board members to state their level of agreement that there is a clear leader on the board.The measures were  (Gabris, Golembiewski, & Ihrke, 2001).Significant variation existed across groups, though post-hoc tests revealed group-to-group differences are limited to operational and management, operational and policy, and management and policy.
Last, we compare the extent to which high school graduation rates vary across identified governance models using and ANOVA test.While high school graduation rates are an imperfect measure of school district performance, in particular as many school districts in the United States do not serve high school students, it is nonetheless one intuitive comparable indicator of the extent to which school districts are satisfying their mission.We find significant variation across groups, as indicated by a significant F-test of 7.01.However, differences are limited, as can be seen in figure 2, to the policy and traditional model, with the traditional model showing significantly higher graduation rates.
Overall, we find that measures of group dynamics, and one measure of school district performance, vary significantly across identified governance model types.In general, the board members who identify traditional and policy model descriptions also perceive lower levels of board conflict, higher levels of trust and owning, and higher levels of productivity than board members who identify the operational and management descriptors as their governance model.In addition, the traditional model boards (as perceived by board members) have higher graduation rates than all other identified board types, and the management model shows substantially lower levels of perceived board leadership than all other identified board types.In the concluding section, we reflect on the meaning of our results and propose a new approach to understanding board governance models.

Conclusion and Discussion
In this article, we used data collected as part of the largest study ever conducted on American school board members to conduct an exploratory analysis on the school board governance models identified by school board members, the extent to which district characteristics predict the governance models used by board members, the ways in which governance approaches differ by functional area, and the variation of group dynamics and one performance variable across identified governance models.Broadly, we find evidence that governance models identified by school board members matter.They are linked to district characteristics as well as perceptions of positive group dynamics.Specifically, the traditional and policy model descriptors are chosen by board members overseeing higher-functioning boards.
Our study does, however, have numerous limitations.First, a number of board respondents did not answer our questions regarding model types.While the specific nonrespondents did not have different personal demographics than respondents, and did not oversee districts with demographics dissimilar to respondents, there is still the possibility that a population of school board members is unable to identify a governance model used by their boards.In addition, the descriptors created by Gill (2002) may be imperfect descriptions of specific model-types.Hence we caution the reader to remember that board members were agreeing to the description of their board governance model-not the actual name of the model.Finally, as we described in the results section, board members often differed from their general identified model of governance in their approach to governance of functional areas.Despite these limitations, our findings move the study of governance models forward.The clear group dynamic advantage on policy and traditional boards (as identified by board members) bears more exploration.Why are identified policy boards higher functioning, and why are these same boards not obtaining higher graduation rates than the other board types identified by board members?We speculate board members indicating they serve on policy boards, as well as traditional boards, set up clear lines between day-to-day operations and governance, enabling the boards to stay on task and leave the professional tasks to the professionals.However, future studies on specific boards adopting policy and traditional models could help answer the why question.
Last, and most important, are the ways in which board members differ in their governance approach in specific functional areas.This finding suggests that scholars are being too simplistic in their search for a comprehensive model of board governance.In may in fact be that that there are functions where an operational model approach works best, others where a policy model works best, etc.We suggest that studies of governance models move toward a hybrid approach where a governance model for a single board incorporates different governance models broken down by functional area.In addition, we believe that the hybrid model must be dependent on the organizational needs and structural characteristics.Though the multiple dimensions of such a model invite complexity, it would provide a better roadmap to improving organizational performance through governance.
Both the nonprofit and public administration literatures continue to advance scholarly understanding of the connection between the governance and performance of public and nonprofit organizations.Despite the complexities of governance, generalizable knowledge that can be used to improve public and nonprofit performance through governance is attainable through research approaches that embrace these complexities.This exploratory study shows, using the example of school boards and their members, that the complexities of governance can be measured and used to develop the next generation of evidence-based board governance models.accountability, and school choice.He is a member of the Thomas B. Fordham and American Enterprise Institutes' Emerging Education Policy Scholars class of 2015-2016 and a 2016 ASPA Founders Fellow.Prior to joining academia Michael worked for many years on education policy in Wisconsin.Douglas M. Ihrke is a professor and chair of the Department of Public and Nonprofit Administration at the University of Wisconsin-Milwaukee (UWM) where he also serves as executive director of the Helen Bader Institute for Nonprofit Management.He also holds appointments in the Department of Political Science and the Urban Studies programs at UWM.He has done extensive research on issues relating to the management and governance of nonprofit and public organizations, and he teaches courses in those same areas at UWM.He has published numerous articles in journals in the fields of public administration, nonprofit management, political science, and urban studies.Further, he consults regularly with public and nonprofit organizations and enjoys hunting and fishing in his free time.

Figure 1 .
Figure 1.Percentage of Respondents Using Model Descriptions in Decision-Making by Area

Figure 2 .
Figure 2. Mean High School Graduation Rates by Governance Model

Table 1 .
Comparison of Sample and Population School District Characteristics

Table 6 .
Multinomial Logistic Regression Results Predicting Model Type (Base Outcomes = Policy)

Table 7 .
Kruskal-Wallis Rank Test Results (Numbers indicate rank sum by group)