ࡱ> @ :jbjbqq J lBBB4v(&&&$P" ,N 6 6 "X X X X n $ DFFFFFF,! #r& X X r&&X X 6  &X &X DJDDDD&& DZDh&& 4*RDT4s$s$ Organizations, Structure, and Diverse Clientele: An Examination of Decentralization, Organizational Performance, and the Latino Dropout Problem Holly T. Goerdel Texas A&M University Department of Political Science hollygoerdel@polisci.tamu.edu Prepared for the Project on Equity, Representation, and Governance and the Carlos Cantu Latino Dropout Project, Texas A&M University, Kenneth J. Meier, Director. Decentralization has, not only an administrative value, but also a civic dimension, since it increases the opportunities for citizens to take interest in public affairs; it makes them get accustomed to using freedom. And from the accumulation of these local, active, persnickety freedoms, is born the most efficient counterweight against the claims of the central government, even if it were supported by an impersonal, collective will. -- A. DE TOCQUEVILLE, Democracy in America Abstract Does organizational structure matter to program performance? This study addresses this question by focusing on how decentralization within an organization affects performance, generally. More specifically, it extracts expectations from the general concept to evaluate whether decentralization induces positive (negative) outcomes for organizations serving diverse clientele. Public education provides the context for the investigation. Findings demonstrate that administrative decentralization contributes positively to organizational performance, especially when faced with diversity. These expectations are then extended to evaluate another salient issue in education: the Latino dropout problem. While administrative decentralizion does not significantly decrease the Latino dropout rate (when utilizing a basic education production function), evidence supports the inclusion of such a variable in a more comprehensive model of Latino dropouts. Introduction To what extent does organizational structure affect organizational performance? How does administrative decentralization affect the overall performance of an organization? How about performance when faced with diversity? Organizations can be characterized by their distinctive structural attributes. For example, some organizations are best described as large, complex systems, with rigid rules and centralized hierarchies, while others are better characterized by their diffuse decision-making structures and responsiveness to, and flexibility in dealing with, specialized clientele. Given these (potentially) varying attributes, how can organizations best structure themselves to achieve optimality, in terms of performance? This study explores how the structural construct of decentralization affects organizational performance. The goal is to gain a better understanding of how decentralized structures induce outcomes that reflect positive organizational performance. Equally important, however, is the goal of enabling researchers to see a clearer picture of why structure matters in organizations, more generally. Public education provides the setting for the investigation. More specifically, contextual illustrations of structural concepts are developed using independent school districts as a point of departure. The study is organized in the following manner: In the first section there is a general overview of decentralization. Here, literature is examined on three forms of decentralization: political, fiscal, and administrative. Then, decentralization is dissected analytically, in order to establish how it might induce positive performance outcomes within an organization. The second section focuses on the empirical exploration of the concepts, analytics, and ideas presented in the first section. This includes investigating two hypotheses focusing on both decentralization and organizational performance in general, and in dealing with diverse clientele, in particular. Then, administrative decentralization is mapped onto a salient issue in education: the Latino dropout problem. Finally, there is a discussion on the implications of decentralization (as an administrative strategy) on organizational outcomes and a conclusion addressing why structure matters to performance. Decentralization: Political, Fiscal, and Administrative Political Our current understanding of decentralization as a concept has come from a literature focusing largely on administrative, fiscal, and political federalism. Studies focusing on political decentralization, for example, are often attributed to political scientists who are interested in how decision-making authority is transferred from central governments to local governments or from citizens to their representatives (Macmahon, 1972; Peterson, Rabe, and Wong, 1986). One goal of political decentralization is to give citizens (or their elected representatives) more power in public decision-making. Underpinning this goal, however, is the assumption that decisions made with greater participation will be better informed and more relevant to diverse interests in society than those made only by central political authorities (Hamilton and Wells, 1990). For example, independently organized school districts select their own governing boards, or school boards. This implies that the selection of representatives from local jurisdictions to local boards allows citizens to know better their political representatives and allows elected officials to know better the needs and desires of their constituents. Fiscal Similarly, economists and others focus on another core component of decentralization: fiscal responsibility. In order for organizations to carry out decentralized functions effectively, they must have adequate levels of revenue, as well as the authority to make decisions about expenditures (Oates, 1972; Wenglinsky, 1997). Some organizations might be fully centralized or decentralized, fiscally. However, a growing number of organizations, especially public ones, represent hybrid systems. These systems are indicative of mixed revenue bases and accountability structures. In the case of education, for example, districts receive local revenue through property or sales taxes, as well as intergovernmental transfers that shift general revenues from taxes collected by the central government (the state, mostly) to local governments for general or specific uses (Wenglinsky, 1997). Along with these sources, school districts have varying capacities to authorize municipal borrowing through loan guarantees (i.e. bond elections). Within the context of education, fiscal decentralization is essential to the (below) discussion of administrative decentralization. Why? Local financing and fiscal authority links decentralization, at the margin, to service provision responsibilities and functions of local districts - so that local politicians (school boards) and public managers (superintendents) can bear the costs of their decisions and deliver on their promises (generally, Tiebout, 1956; Hosek and Levine, 1996). Administrative Finally, there is the primary focus of this study administrative decentralization. Administrative decentralization seeks to redistribute authority, responsibility and financial resources for providing public services among different levels of organization (Farber, 1989; Liner, 1989; Eisinger, 1998). More specifically, this type of decentralization transfers the responsibility of planning, financing, and management of certain (typically explicit) functions from a central decision center (e.g. government) to subordinate units or levels of hierarchy (i.e. multiple decision centers) (see Simon, 1976). In the case of education, this idea can be conceptualized from two perspectives. First, the national government has largely given states discretion over public school organization from planning and financing to management functions. However, within each district, there are variations in structure. Some districts are highly centralized and hierarchical, while others are better characterized as having multiple decision-centers, many of which are located in proximity to front-line professionals (teachers). This concept is especially important when considering the fact that these types of educational organizations are often described as street-level bureaucracies (Lipsky, 1980). As such, there is a perceived benefit in decision-makers being close to street-level bureaucrats namely that the transaction-costs involved in communicating the status of service provision to clients (as either poor, lacking, or needing help, for example) is substantially reduced an idea to be unpacked in subsequent sections. So how might administrative decentralization induce positive outcomes for an organization? That is, how can the benefits of such structural decentralization be turned into positive outcomes for an organization? Rationale for Decentralization Overall Performance As posited throughout, school districts are independent local governments with the power to select their own governing boards (i.e. school boards) and raise revenues via taxation (Meier and OToole, 2001). The organizations themselves follow current decentralization trends and are best described as street-level bureaucracies, whereby front-line professionals (teachers) are given a great deal of discretion (OToole and Meier, 2002). However, teachers often find themselves removed from the decision making process. In fact, they often depend on school administrators to make decisions in the best interest of their clients, or students. This dependence either benefits or harms students, depending on whether or not administrators are centrally structured. Unfortunately, some school districts centrally locate their administrators. This makes it less likely for administrators to have the necessary knowledge of local school conditions. As a result, they often cannot appropriately provide, or make decisions concerning, educational services. In fact, administrators working in centralized environments may engage in activities that are not necessarily driven by demands coming from the street level. Moreover, as administrators become more removed from the street level, both in proximity and authority, the transaction costs involved in communicating from below (where client contact exists) potentially increases. On the other hand, decentralized structures, which place administrators in proximity to teachers, potentially decrease transaction-costs associated with communicating from below. Equally important, however, is the inherent opportunity of interaction available under a decentralized structure, between decision-centers (administrators, i.e. principals) and street level bureaucrats. When exploited efficiently, this medium provides a way for those close to the students (teachers, for example) to convey organizational needs to those with authority to affect change. Diverse Clientele In the present case, then, decentralization represents a mechanism for governance within the organization that also works in conjunction with needs that arise from diversity in clientele. Research addressing whether or not decentralization benefits diverse populations (i.e. minorities) is ambiguous. In this literature, clientele are often assumed to be relatively homogenous in terms of ethnicity and preferences, see for instance Hoxby (1996), Alesina and Spolaore (1997), and Besley and Coate (1999). At best, systems designed to incorporate the ideas, concerns, and initiatives of multiple actors, or sets of actors, in an organization increase the potential for minority interests to be represented, or at least communicated. At worst, however, these concerns and ideas are not translated into performance gains for minorities (Staeheli, Kodras, and Flint, 1997). Modeling the impact of administrative decentralization on organizational performance for diverse (and overall) clientele, will contribute to an understanding of what happens when organizational difficulties (or challenges) are turned into performance gains. Modeling the Impact of Decentralization In considering the impact of decentralizationand structure more generallyupon the performance of an organization, a model is developed to specify the influence of administrative structure (controlled by resources and constraints in the environment, as well as fiscal decentralization) on program performance: Ot = 1 t +  2(Da) t +  3(Df) t +  4(X)t +  t [1] where Ot is some measure of organizational outcome, Da is a measure of administrative decentralization, Df is a measure of fiscal decentralization, X is a vector of environmental forces,  is the error term, estimable parameters are 1,  2,  3, and  4, and the other subscripts denote time periods. While it is expected that various types of decentralization (i.e. political, administrative, and fiscal) matter to organizations, the focus in the present study is on administrative decentralization, or Da. The above model of structure and performance sets up an education production function whereby student performance is a function of inputs to the organization. A measure of administrative decentralization is added to these in order to determine if structure matters, given controls for other factors that affect performance (environmental forces). Initial analysis covers how decentralization affects overall performance. Then, attention is turned to an equally important set of analyses focusing on decentralization and performance of diverse clientele. This part of the study is addressed within the context of a salient educational issue: the Latino dropout problem. Methods Units of Analysis Data used in this study are drawn from independent school districts in Texas. It is maintained in this study that independent means that the school district is not subordinate to another unit such as a city. Independent districts have their own elected board, have the ability to tax and set budgets, and acquire bonding authority by a vote of the residents (Meier and OToole, 2001). More specifically, data are drawn from school districts with more than 500 students. In order to capture whether or not decentralization matters to organizations with diverse clientele, the analysis is further limited to districts with more than 10 percent but no more than 90 percent Anglo students. Data on performance and control (environmental) variables are pooled for six years (school years 1996/1997-2001/2002) producing a total of 2773 possible cases for analysis. Measures The measures used in this study can be discussed in terms of model [1]: administrative decentralization (Da); fiscal decentralization (Df); the vector of environmental forces (Xt); and current performance (Ot). The following discussion addresses these items in this order. Administrative Decentralization. The administrative structure of an organization can be centralized or decentralized. In the present study, administrative decentralization within a district is measured as the percentage of total administrators located in on-campus locations. This measure works well because it considers the potential benefits gained from placing administrative decision-makers near street level bureaucrats. While the measure only considers whether or not administrators are located in proximity to the street (or implementation) level, it offers a good starting point to see how districts who are decentralized in this manner differ from those who are not. At a minimum, if it is found that decentralization matters in this limited sense, future results using better measures could only prove more compelling. As such, administrative decentralization is expected to contribute positively to organizational performance. Fiscal Decentralization. School districts represent hybrid fiscal systems. While they have the power to tax property (and sales, in some instances), many districts receive state funds. Research on how source of aid affects educational performance is equally ambiguous. Meier, Wrinkle, and Polinard (1999) find that increases in state aid benefits minority students, but find no relationship between state aid and nonminorities. While more state aid coming into a district might increase overall performance by helping to equalize the tax base, it is unclear whether more local revenue (versus state aid) allows for more flexibility in responding to local needs. In order to address this specifically, fiscal decentralization is measured as the percentage of total revenue coming from local sources. A positive relationship is expected between fiscal decentralization and performance. Environmental variables. To ensure that any relationships found between structure and organizational performance are not spurious, a variety of other factors that influence education performance are included in the model. As mentioned previously, these environmental forces are controls to the education production function. Moreover, due to the fact that this study is primarily interested in testing aspects of organizational structure, the discussion of these variables is brief. Since production functions typically include measures of environmental constraints and resources, the process of analyzing them can be simplified by looking to recent literature on public education policy. Scholars suggest that environmental variables be categorized as either task difficulties or program resources ( Hedges and Greenwald, 1996; Meier and OToole, 2001). We learn from Jencks and Phillips (1998), for example, that racial inequalities and income disparities are negatively correlated with educational performance, particularly when focusing on standardized testing. Evidence like this leads scholars to believe that measures of race and poverty should be included in models concerned with constraints from the environment and their influence on program performance. Not including them would be misspecifiying the model. This study considers three measures of race and poverty, namely the percentages of African American, Latino, and poor students in a given school district. In terms of resources, a growing literature (primarily focused on longitudinal studies) is beginning to confirm the basic tenet that schools with more resources generally fare better (Wenglinsky, 1997). Three measures of resources are included in this study. Average teacher salary, total instructional expenditures per student, and average years of teaching experience are related to both the financial and human resources of the school district. Teachers, principals, and administrators (i.e. superintendents) are in constant search of resources in order to meet the diverse needs of students in their district. More specifically, administrators are concerned with securing fiscal resources to update structural and technological facilities, provide competitive teaching salaries, and provide special needs services. However, since most spending in education pays for salaries of teachers and other staff (Meier, Wrinkle, and Polinard, 1999), the environment can be described as personnel intensive. As such, it can be exploited as a resource. For example, Hanuschek and Pace (1995) evaluate how economic incentives, like higher salaries, attract better qualified people to a profession. Superintendents and principals also value the resource of an experienced workforce of teachers, and often work to obtain employees with such credentials (Meier, Wrinkle, and Polinard, 1999: OToole and Meier, 2001). Along with these, school size, as a structural component is also important. School size can act as a constraint or resource. School size is typically measured as the average daily student attendance (Bidwell and Kasarda, 1975). Generally speaking, the structural dimensions of size has an inverse relationship with performance (Marriot, 1949; Thomas, 1959; Indik and Seashore, 1961; Katzell, Barett, and Parker, 1961). Coleman and Hoffer (1987) find that smaller schools promote social interaction, which creates a form of social capital that facilitates the work of the school. Here, school size is measured as average daily student attendance in a district. Taking all these factors into consideration, teaching experience, total instructional spending per student, and teacher salary are all expected to positively contribute to organizational performance. Conversely, school size is expected to have a negative relationship with organizational performance. Dependent variables. Several performance indicators will be used as dependent variables. For the first set of analyses, which considers overall organizational performance, the dependent variable of interest is the percentage of overall students in each school district passing all parts of the state-wide standardized test each year the Texas Assessment of Academic Skill tests, commonly known as the TAAS test. This test, given to all students in grades 3 through 8 and 10, covers topics in reading, writing, and mathematics. The TAAS exam is a high-stakes test (in grade 10) that receives regular media attention. The Texas Education Agency also publicly scrutinizes (and sanctions against) districts that perform poorly on this exam and praises districts that perform satisfactorily. This makes it a satisfying measure of overall organizational performance. Equally important to this study, however, is the performance of minorities (African Americans and Latinos) on the TAAS test. In order to see how administrative decentralization affects minorities in a highly salient performance area, percentage of Latinos and African Americans dropping out is also a performance (dependent) variable of interest. Hypotheses As posited earlier, administrative decentralization brings together decision makers and street level bureaucrats, in an effort to create, or foster, an environment that is more responsive to the needs of clients in general, and diverse clients, in particular. Since administrative decentralization reflects the expected behavior of mid-level managers to engage with street level bureaucrats, it follows that H1: When controlling for resources and constraints in the environment, as administrative decentralization (Da) increases, the performance of all students increases. Likewise, in order to test the diversity/decentralization hypothesis, it follows that: H2: When controlling for resources and constraints in the environment, as administrative decentralization increases, performance of minority students increases. Testing the Models The following is an empirical exploration of the effects of structure, namely administrative decentralization, on organizational performance. Models representing overall performance, as well as those focusing on performance of minorities, are included. Particular attention should be given to the statistical and substantive impacts of administrative decentralization. The first model is drawn from equation [1] and is primarily addressing the first hypothesis. It is additive and non-autoregressive. The results are presented in Table 1. The key coefficient for administrative decentralization is positive and significant, indicating that structure, Da, contributes positively to organizational performance. When controlling for resources and constraints in the environment, as organizations decentralize administrators from central locations to on-campus cites, the percentage of overall students who pass the TAAS exam increases, supporting the first hypothesis (H1). [TABLE 1 ABOUT HERE] Other findings in Table 1 suggest that average teacher salary, instructional spending per student, and teacher experience all contribute positively to program performance, as expected. This demonstrates the ability of districts to turn both financial and human resource capital into performance gains. In addition, all three task difficulty measures (percentage of African American, Latino, and low income students in a district) are significantly related to performance in a negative manner, as expected. Finally, school size and fiscal decentralization are not found to be significantly (negatively) associated with performance. The second model, reported in Table 1, incorporates an autoregressive term to account for organizational inertia in delivering educational services. This is done by including a lagged dependent variable. Picking up the impact of the independent variable of interest in the presence of a lagged dependent variable is difficult (OToole and Meier, 2002). However, if administrative decentralization continues to contribute to performance, results will further support the importance of structure. The autoregressive model simply adds a lagged dependent variable to equation [1]: Ot = 1 t +  2(Da) t +  3(Df) t +  4(X)t +  5 (Ot-1) +  t [1] Results are presented in the second column of Table 1. The key coefficient for administrative decentralization, 1.57, remains positive and significant --indicating that, even after incorporating an autoregressive term into the educational production function, structure still matters to performance. Other findings from the autoregressive model suggest similar relationships among control variables and performance. Diverse Clientele and Organizational Performance Next, Tables 2 and 3 evaluate the influence of administrative decentralization on minority performance. The models used in these analyses are equations [1] and [2]. Following the process above, Table 2 reports minority performance on the TAAS test resulting from both non-autoregressive and autoregressive models. Administrative decentralization is expected to be positively associated with minority performance on the TAAS test (H2). Results are presented in Table 2. [TABLE 2 ABOUT HERE] The key coefficient for administrative decentralization, 9.09, is positive and significant. When controlling for resources and constraints in the environment, as organizations decentralize administrators from central locations to (street-level) on-campus cites, the percentage of minority students passing the TAAS test increases. Other findings in Table 2 suggest that human and financial resources contribute positively to organizational performance. Conversely, environmental constraints, particularly the percentage of African American and low income students, negatively influence performance, as expected. Turning attention to the autoregressive results found in the second column of Table 2, the coefficient for administrative decentralization, 3.93, remains positive and significant. While the impact from this measure of structure is diminished by the inclusion of a measure of past performance (autoregressive term), the coefficient remains large. The most provocative finding across Tables 1 and 2 is that administrative decentralization continues to contribute positively and significantly to overall and minority performance. The evidence is compelling enough to assert that there is something different about districts who situate administrators in proximity to street-level bureaucrats, versus those that do not in terms of performance. Comparing the impact of the Da coefficient across overall, minority, and Anglo pass rates reveals another interesting finding. [Table 3 about here] On closer examination, it is apparent that districts that are more decentralized (administratively) experience higher performance among minorities, compared to overall and Anglo performance. For example, in Table 3, the impact of Da on performance is greater for minorities, than for Anglos, or the overall group. This gives support to the contention that administrative decentralization is a mechanism for converting needs that arise from diversity (in clientele), into performance gains. On balance, these findings strongly support both H1 and H2. More specifically, administrative decentralization contributes positively to organizational performance, when controlling for resources and constraints in the environment. In the following section, the influence of structure on minority performance is mapped onto another salient issue in education: the Latino dropout problem. Administrative Decentralization and Latino Dropouts Another measure of performance for schools is the dropout rate. A decade ago, dropout studies focused on individual background characteristics as determinants for dropping out (Hess and Lauber, 1985). Characteristics such as race, economic status, and household resources are a few examples. The idea that schools, as organizations, may play a role in determining the dropout rate was not introduced until the late eighties (Raudenbush and Bryk, 1986; Wehlage and Rutter, 1986; Fitzpatrick and Yoels, 1992). Since that time, researchers have studied high school dropout using three clusters of factors: 1) individual factors, 2) family factors, and 3) structural factors (Velez and Saenz, 2001). The present study investigates, empirically, how decentralized administrative structures help (harm) the dropout rate among one minority group: Latinos. Unlike individual social characteristics, the structure of a school is under the control of educators and is amenable to change in order to achieve desired outcomes. Schools can intervene more effectively in the dropout process by considering the broad organizational context in which school failure occurs (Baker, et al., 2001). This requires knowledge of the structural determinants of dropping out. As it stands, there is only one study on the influence of school structure on the Latino drop out rate. Velez and Saenz (2001) explore factors that (may) lead to a comprehensive model of the drop out rate among Latinos. They find individual, family, and structural factors to be foundational components of such a model. Structural factors such as school practices, relative size of ethnic group, and community economic contexts are associated with dropping out. Individual factors such as oppositional behavior (also known as adversarial subculture), academic expectations and performance, role taking, generational status and acculturation, Spanish language use, and ethnic group membership also seem to be associated with dropping out. Finally, family factors associated with Latino dropout include family structure, economic situation, and neighborhood of residence. To be clear, the object of this study is not to quantify and empirically explore a comprehensive model of the dropout rate among Latinos. Instead, the goal here is to capture some theoretical gains by exploring how school structure, particularly administrative decentralization, affects the performance of diverse clientele in this case, Latinos. Keeping this in mind, the previous performance-education production function model [1] is tested. Remember, it is additive and non-autoregressive. The results are presented in Table 4. The key slope of interest is administrative decentralization. The slope is appropriately signed, though not significant. [Table 4 about here] Other findings in Table 4 suggest that instructional spending per student significantly reduces dropout rates for Latinos, while the percentage of minority students and school size contribute positively to the dropout rate. Owing to the fact that this is not a comprehensive Latino dropout model, the overall explanatory power of the model is less than 3 percent. Given these results, how does the influence of administrative decentralization on dropping out compare for other minority groups? Table 5 reveals similar findings for African Americans. [Table 5 about here] Administrative decentralization is appropriately signed, though not significant. Also, though not reported here, the environmental resource and constraint variables follow Table 4 closely. While it is unclear as to whether administrative decentralization is a determinate of (not) dropping out among minorities, it is interesting that the coefficient (for the impact) of Da is substantially smaller for Latinos than African Americans, and overall groups. Table 5 compares the slopes of administrative decentralization across overall, African American, and Latino dropout rates. Again, while each of the three slopes are signed appropriately, they are not significant. Perhaps decentralizing administrative functions is not the primary mechanism used by those advocating minority interests (particularly in case of Latinos) from the street level, to affect the dropout rate. On the other hand , this could reflect the inability of street-level bureaucrats, as well as administrators, to turn difficulties associated with diverse clientele into performance gains, when it comes to dropping out. In addition, this may indicate that the manipulation of structural mechanisms is not the most effective strategy for achieving lower dropout rates. A more likely explanation is presented in the following discussion. Discussion This study attempts to evaluate the influence of structure on organizational performance. Previous analyses focused on statistical results from various models that incorporated a measure of structure, or administrative decentralization. The first set of findings (Tables 1,2, and 3) overwhelmingly support a positive relationship between administrative decentralization and overall performance. Equally important, is the positive relationship between administrative decentralization and minorities, when considering TAAS test performance. Each table supports the general hypothesis that structure matters to organizational performance. More importantly, these gains are induced by the creation of organizational environments that are conducive to low transaction cost communication, as well as (potentially) meaningful engagement of decision makers at the street level. It also reflects the ability of street-level and mid-level bureaucrats to use decentralization as a tool for achieving performance gains. Taking this into account, it makes sense to see how administrative decentralization might influence another highly publicized educational issue: the Latino dropout problem. Judging by the relationships established between administrative decentralization and minority performance, it is expected that administrative decentralization has the same influence on the Latino dropout rate. Unfortunately, that relationship is not empirically supported. In fact, it appears that the (substantive, not statistical) impact of administrative decentralization on the Latino dropout rate is quite small, compared to overall and African American dropout rates. An explanation for this is grounded in the use of a deliberately underspecified model for estimation, namely the basic education production function. As mentioned in the aforementioned dropout literature review, three sets of factors contribute to a comprehensive model of the Latino dropout problem: individual, family, and structural factors. Therefore, it is recommended that administrative decentralization be included in a comprehensive Latino dropout model that incorporates appropriate measures of these factors, as well as other possible determinants, like macro-economic conditions and school discipline practices. Conclusion Does organizational structure matter to program performance? Yes. This study demonstrates how administrative decentralization, as a type of structure, induces positive organizational outcomes. In fact, it specifically demonstrates how administrative decentralization can benefit clientele, in general, and diverse clientele, in particular. While administrative decentralization did not significantly reduce the rate of Latino dropouts (using a basic education production function), initial evidence (found in Tables 1,2, and 3) is compelling enough to recommend the incorporation of the measure, Da, into a fully specified model of Latino dropouts. One fact is clear, structure matters to organizational performance. More than that, school districts that decentralize their administrators experience higher program performance than those that do not decentralize, all else equal. As such, school districts should evaluate structural decentralization strategies as they continue to provide educational services. Table 1. The Impact of Administrative Decentralization on Performance: Standardized Tests Dependent Variable = Overall TAAS Pass Rate Independent Variable Slope t-score Slope t-score Structure Administrative Decentralization 3.31* 3.83 1.57* 3.16 Controls Fiscal Decentralization -.120n -.22 -.086* -.28 Teachers Salaries (K) .0003* 5.01 .0006* 1.58 Instruction Spending (K) .0023* 9.88 .0004* 2.74 African American Students % -.237* -21.52 -.031* -4.53 Latino Students % -.105* -12.80 -.009* -1.81 Low Income Students % -.137* -11.89 -.023* -3.32 School Size -.000007 -.68 -.00001* -2.21 Teacher Experience .673* 10.60 -.060* 1.63 Lagged Pass Rate -- -- .779* 78.50 R-Square .55 .85 Standard Error 5.77 5.77 F 290.44 1100.12 N 3322 2707 Coefficients for annual dummy variables omitted. * significant, p < .10 Table 2. The Impact of Administrative Decentralization on Minority Performance: Standardized Tests Dependent Variable = Overall TAAS Pass Rate Independent Variable Slope t-score Slope t-score Structure Administrative Decentralization 9.09* 3.04 3.93* 1.75 Controls Fiscal Decentralization -8.88* -4.57 -1.96 -1.35 Teachers Salaries (K) .0018* 7.66 .00005* 3.01 Instruction Spending (K) .0045* 5.15 .0014* 2.08 African American Students % -.219* -5.82 -.063* -2.23 Latino Students % -.011* -.39* .007 .32 Low Income Students % -.307* -7.53 -.084* -2.69 School Size -.00007* -2.09 -.00004* -1.70 Teacher Experience .8003* 3.62 -.060* 1.63 Lagged Pass Rate -- -- .686 48.11 R-Square .38 .66 Standard Error 18.20 11.99 F 121.76 310.41 N 2773 2212 Coefficients for annual dummy variables omitted. * significant, p <.10 Table 3. Comparison of Overall vs Minority Performance: Standardized Tests All Pass Minority Pass Anglo Pass    (t-score) (t-score) (t-score) Model 1 3.31 9.09 2.11 (3.83) (3.04) (2.75) Model 2 1.57 3.93 1.64 (3.16) (1.75) (3.35) Model 1 is not autoregressive. Model 2 is autoregressive. Coefficients are from the variable of interest, administrative decentralization. Both models control for resources and constraints in the environment. These include fiscal decentralization, teacher salary, instructional spending per student, school size, teacher experience, and percentage of African American, Latino, and low income students in a district. Past performance is controlled for in Model 2. Table 4. The Impact of Administrative Decentralization on Latino Dropouts Dependent Variable = Latino Dropout Rate Independent Variable Slope t-score Structure Administrative Decentralization -.098 -.28 Controls Fiscal Decentralization .123 .54 Teachers Salaries (K) .00001 .53 Instruction Spending (K) -.00002* -2.67 African American Students % .008 1.80 Latino Students % .0008* 2.34 Low Income Students % -.0014 -.30 Class Size .000005 1.04 Teacher Experience .049* 1.91 R-Square .02 Standard Error 2.17 F 4.02 N 2433 Coefficients for annual dummy variables omitted. * significant, p <.10 Table 5. Administrative Decentralization and Dropping Out All Latino African-American Dropouts Dropouts Dropouts    (t-score) (t-score) (t-score) -.355 -.098 -.397 (-1.47) (1.11) (-0.81) Coefficients are from the variable of interest, administrative decentralization. The model controls for resources and constraints in the environment. These include fiscal decentralization, teacher salary, instructional spending per student, school size, teacher experience, and percentage of African American, Latino, and low income students in a district. References Alesina, A., and E. Spolaore. 1997. On the Number and Size of Nations, Quarterly Journal of Economics 112: m 1027-1056. Baker, J., et. al. 2001. The Flip Side of the Coin: Understanding the Schools Contribution to Dropout and Completion. School Psychology Quarterly 16:4 406-426. Besley, T., and S. Coate. 1999. Centralized versus Decentralized Provision of Local Public Goods: A Political Economy Analysis, NBER Working Paper, Number 7084. Bidwell, Charles and John Kasarda. 1975. School District Organization and Student Achievement. American Sociological Review 40: 55-70. Coleman, J.S., and Hoffer, T. 1987. Public and Private High Schools: Impact on Communities. New York: Basic Books. Eisinger, Peter. 1998. City Politics in an Era of Federal Devolution. Urban Affairs Review 33:3 308-325. Farber, Stephen B. 1989. Federalism and State-Local Relations. In A Decade of Devolution: Perspectives on State-Local Relations. Liner, E. Blaine, ed. Washington: The Urban Institute Press. Fitzpatrick, K.M., and W.C. Yoels. 1992. Policy, School Structure, and Sociodemographic Effects on Statewide High School Dropout Rates. Sociology of Education 65:766-793. Hamilton, Christopher and Donald T. Wells. 1990. Federalism, Power, and Political Economy. Prentice Hall: Englewoods Cliffs, NJ. Hedges, Larry V. and Rob Greenwald. 1996. Have Times changed? The Relation Between School Resources and Student Performance. In Does Money Matter? Gary Burtless, ed. Washington, D.C.: Brookings. Hess, G.A., and Lauber, D. 1985. Dropouts from the Chicago Public Schools. Chicago, IL: Chicago Panel on Public School Policy and Finance. Hosek, James and Robert Levine, eds. 1996. The New Fiscal Federalism and the Society Safety Net: A View from California. Rand Report. Hoxby, C. M. 1996. Are Efficiency and Equity in School Substitutes or Complements?, Journal of Economic Perspectives 10: 51-72. Indik, B.P., and Seashore, S.F. 1961. Effects of Organization Size on Member Attitudes and Behavior. Ann Arbor: Survey Research Center of the Institute of Social Research, University of Michigan. Jencks, Christopher, and Meredith Phillips, eds. 1998. The Black-White Test Score Gap. Washington, D.C.: The Brookings Institution. Katzell, R.A., Barrett, R.S., and Parker. 1961.T.C. Job Satisfaction, Job Performance, and Situational Characteristics. Journal of Applied Psychology 45: 65-72. Liner, E. Blaine, ed. 1989. A Decade of Devolution: Perspectives on State-Local Relations. Washington: The Urban Institute Press. Lipsky, Michael. 1980. Street-level bureaucracy: Dilemmas of the Individual in Public Services. New York: Russell Sage Foundation. Marriot, R. 1949.Size of Working Group and Output. Occupational Psychology 23: 47- 57. Macmahon, Arthur W. Administering Federalism in a Democracy Oxford University Press: New York. 1972 Meier, Kenneth J. and Laurence J. OToole, Jr. 2001. Managerial Strategies and Behavior in Networks: A Model with Evidence from U.S. Public Education. Journal of Public Administration Research and Theory 11:3 271-93. Meier, Kenneth J., Robert D. Wrinkle, J.L. Polinard. 1999. Representative Bureaucracy and Distributional Equity: Addressing the Hard Question. The Journal of Politics, Vol. 61, No. 4. (Nov., 1999), pp. 1025-1039. Niskanen, W. A., Jr. 1971. Bureaucracy and Representative Government. Aldine-Atherton, Oates, Wallace E. 1972. Fiscal Federalism. New York: Harcourt Brace Javanovich. OToole, Laurence J. and Kenneth J. Meier. 2002. Plus ca Change: Public Management, Personnel Stability, and Organizational Performance, Paper for the Research Workshop on the Empirical Study of Governance, Management and Performance, The Bush School of Government and Public Service, Texas A&M University, February 21-23. [Im sure this is published I cant find where.] Peterson, Paul E., Barry G Rabe and Kenneth K. Wong. 1986. When Federalism Works. Washington D.C.: The Brookings Institution Raudenbush, S.W., and A.S. Bryk. 1986. A Hierarchical Model for Studying School Effects. Sociology of Education 59: 1-17. Simon, H.A. 1976. Administrative Behavior: A Study of Decision-making Processes in Administrative Organizations (3rd edition.) New York: Free Press. Staeheli, Lynn A, Janet E. Kodras and Colin Flint, eds. 1997. State Devolution in America: Implications for a Diverse Society. Thousand Oaks, CA: Sage Publications. Thomas, E.J. 1959. Role Conceptions and Organizational Size. American Sociological Review 24: 201-219. Tiebout, Charles M. 1956. A Pure Theory of Local Expenditures. Journal of Political Economy 64:416-424. Velez, W. and R. Saenz. 2001 Toward a Comprehensive Model of the School Leaving Process among Latinos. School Psychology Quarterly 16: 445-467. Whelage, G. and Robert Rutter. 1986. Dropping Out: How Much do Schools Contribute to the Problems? in Gary Natriello (ed.), School dropouts: Patterns and Policies New York: Teachers College Press. Wenglinsky, Harold. 1997. How Educational Expenditures Improve Student Performance and How The Dont. Princeton, NJ: Educational Testing Service.  To be sure, the perspective focusing on decentralization within districts is of primary interest.  This happens, of course, within limits. It may be that teachers communicate organizational needs to mid-level managers who do not have the discretion to decide such matters. In cases like this, however, mid-level managers can be valuable resources for where to direct communication.  At this time, fiscal decentralization is evaluated as a type of control. Primary concern rests with administrative decentralization, or Da.  This does not exclude hybrid systems. It can be interpreted as tending toward centralization or decentralization.  Empirical evidence concerning this measure should be interpreted with caution. Results could be biased toward rich districts.  Collected and reported by the Texas Education Agency.  This work is qualitative in nature.  For example, tracking (i.e. curricular placement), grade retention, and push-out behavior. PAGE  PAGE 32  q fUz!!%%(((..000085a52747>7D7N7R7V7X7Z7^7f7j7n7p7r7v7~7777777788z8~899999999::::{==@@hAiAAAAAAAB0BBBDDEEH* j0JU6 56CJ565\>~$da$d>~ !"#$~ q r s t u v w x y z fz!!#'(((+.0085`50778x88&9V99 : :7<{===` !"#$%~ q r s t u v $dha$$a$$`a$$da$v w x y z fz!!#'(((+$da$d`ddd`$a$+.0085`50778x88&9V99 : :7<{===@@Bd`$da$dh d^`$da$dd`=@@BBEE6I7IKNPTWGXHX]]]^^V_W____S`T`g`a8d9dNdOdf ijjllmnnooRpkqrTttttv[x\xx{v}Dwx΅uIJyzИK^BBEE6I7IKNPTWGXHX]]]^^V_W____$^a$ 0d^`0d`$da$d`dEFFHH7IOIWWGX\X[[]] ]]^^____T`h`bb3d4dOdfij jjjj"j$j&j*j2j6j:jjBjJjNjTjVj^j`jhjjjxj|jjlmnnqtrtuuw www{'{||Q~R~~~DE$1LM5CJOJQJ5CJCJH*5 j0JU6Z_S`T`g`a8d9dNdOdf ijjllmnnoo$da$d`$a$ d^`$a$dhpdh^pd`d$^a$$a$oRpkqrTttttv[x\xx{v}Dwx΅d$da$d`uIJyz$d&dP^a$ $d^a$$d&dP^a$ $d^a$$^a$$^a$d`d$a$d`JhϜ؜M*.>@FHL&>efcnhM8:46#l7ˬݬOkӭ5Jޮ*655CJOJQJ CJOJQJCJ CJOJQJ5CJOJQJ CJOJQJOJQJ 5OJQJOИKL}~͙345jkl $d^a$KL}~͙345jklКњ8RhDEϜМڜۜJKLΝϝН>?@vwxߞ7M~&(*,nefdКњ8RhDEϜМڜۜ $d^a$$d&dP^a$$d&dP^a$ $d^a$ۜJKLΝϝН>?@vwx$d&dP^a$ $d^a$ߞ7M~&(*0d&dP^`0d^$d&dP^a$ $d^a$$d&dP^a$*,nef$d&dP^a$ $d^a$45cdnoABTrst¥ CVhLM246 d45cdno$d&dP^a$ $d^a$$d&dP^a$ $d^a$ABTrst¥ CVhd^$d&dP^a$ $d^a$LM246d&dP^ d^  d^`d^ $d^a$ $d^a$ #w$d^`a$$^a$$^a$ $d^a$#w>GHAwxȭDZ[gͯ!RSq./ {|Գ`,- [rsŶPQ)*ӸԸ+NOPκϺ!LMNANbcdb>GHAwxȭDZ[gͯ!RS d^`d`^`*R԰9a|ѳ~D Gs BP)o+Aû#d8Vwܾ'TUVWaefhi`a &'(*+10J j0JUCJH*CJ j0JU5H*6mH 5mH mH 66B*OJQJphB*OJQJphLq./ {|Գ`,- ^`[$\$`^ [rsŶPQ)*ӸԸ+NOPκϺ!LMN^`NANbcdL¾UVhdd`^`L¾UV*24679: `()*6789:&`#$ 12456:0J0JmH j0JU# 01h/ =!"#$%# 01h/ =!"#$% i(@( NormalCJmH <@< Heading 2$$d@&a$56@6 Heading 3$d@&6<@B< Heading 4dd@&[$\$5<A@< Default Paragraph Font.>@. Title$da$5:^@: Normal (Web)dd[$\$BC@B Body Text Indentd`8&@!8 Footnote ReferenceH*LS@2L Body Text Indent 3$d^a$&)@A& Page Number:P@R: Body Text 2$da$CJ2@b2 Footnote TextCJ, @r, Footer  !*2:@O`ttd Bi    dd  BBBBBB  %   z z z z z z z z z z z z z z z z z z z z z z z z z z z z z z z zzgz!'C.28?ELSgX^djqw}LZ̘›x $  5  AsDZf6cJ:Y E*1:emqv +B_oۜ*N:fhijlnoprtuvwyz{|~=K:gksx} !!4 _Hlt45341349 _Hlt45341350 _Hlt45341391 _Hlt453413920000@@@@eeeepqyRYaf ,,,,,,,,,,....".00000000001 1<1>122f9g9i9999999999>>\?i?IFWFGGGJJ%K.KVL^LIMJMQMVM]MMMNN NN2N8NWWWZZ aaaa!a*a+a-ajjkkZo`oHpIpSpXp\pdpkpppvpppR}T}Z\ĕƕ=EQYMTgnMUǢעLT cjltܪ@CZ\'1˭ϭ/9JNEM`fW^X_ouƱϱ",ĴƴKKMM___6g@gG~J~:gq=G 33333333 hollygoerde!H:\PERG\Cantu\CantuRelease[1].doc hollygoerde!H:\PERG\Cantu\CantuRelease[1].doc hollygoerde!H:\PERG\Cantu\CantuRelease[1].doc hollygoerde!H:\PERG\Cantu\CantuRelease[1].doc hollygoerde!H:\PERG\Cantu\CantuRelease[1].doc hollygoerde!H:\PERG\Cantu\CantuRelease[1].doc hollygoerde!H:\PERG\Cantu\CantuRelease[1].doc hollygoerde!H:\PERG\Cantu\CantuRelease[1].doc hollygoerde!H:\PERG\Cantu\CantuRelease[1].doc Nick Theobaldhome:web:reports:report025.docv26@|%hRN:L:EYQd x h   ^ `OJQJo(h ^`OJQJo(oh ^`OJQJo(h | | ^| `OJQJo(h LL^L`OJQJo(oh ^`OJQJo(h ^`OJQJo(h ^`OJQJo(oh ^`OJQJo(^`5OJPJQJo( ^`OJQJo(o pp^p`OJQJo( @ @ ^@ `OJQJo( ^`OJQJo(o ^`OJQJo( ^`OJQJo( ^`OJQJo(o PP^P`OJQJo(h ^`OJQJo(h ^`OJQJo(oh pp^p`OJQJo(h @ @ ^@ `OJQJo(h ^`OJQJo(oh ^`OJQJo(h ^`OJQJo(h ^`OJQJo(oh PP^P`OJQJo(h^`.h^`.hpLp^p`L.h@ @ ^@ `.h^`.hL^`L.h^`.h^`.hPLP^P`L.hh^h`.N:|%v26EY x6        8USXȴ~@l{ |012ab``` @`6`8`t@`j`@``D@``T@GTimes New Roman5Symbol3 Arial? Courier NewY CG TimesTimes New Roman;Wingdings"1hN&N&N&[ K: #02Q= Organizations, Structure, and Diverse Clientele: hollygoerde Nick Theobald Oh+'0 0 HT p |  '> Organizations, Structure, and Diverse Clientele: e  hollygoerdeollollNormaleNick Theobaldga2ckMicrosoft Word 10.1@F#@k@k@k [ ՜.+,08 hp  'Texas A&M Universityga:K > Organizations, Structure, and Diverse Clientele: Title  !"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_`abcdefghijklmnopqrstuvwxyz{|}~Root Entry F1Tables$WordDocumentJ SummaryInformation(DocumentSummaryInformation8CompObjXObjectPool FMicrosoft Word DocumentNB6WWord.Document.8