Structured to Partner: School District Collaboration with Nonprofit Organizations in Disaster Response

Embed Size (px)

Citation preview

  • 8/9/2019 Structured to Partner: School District Collaboration with Nonprofit Organizations in Disaster Response

    1/35

    Structured to Partner

    0

    TITLE PAGE:

    Structured to Partner: School DistrictCollaboration with Nonprofit

    Organizations in Disaster Response

    Scott E. Robinson*

    Bush School of Government and Public Service

    And

    Angela Bies

    Bush School of Government and Public Service

    * Corresponding author: [email protected]

  • 8/9/2019 Structured to Partner: School District Collaboration with Nonprofit Organizations in Disaster Response

    2/35

    Structured to Partner

    1

    Structured to Partner: School District Collaboration with Nonprofit

    Organizations in Disaster Response

    Abstract:

    Emergency preparedness and response is moving from a specialized circle of

    emergency management professionals and select nonprofit organizations (such as

    the Red Cross and other national relief organizations) to include a broader varietyof organizations not traditionally fulfilling emergency management roles,

    including schools. It is not clear who among these new potential members of

    emergency preparedness networks collaborates with whom. We present the

    results of a survey of Texas public schools and test how structural characteristicsare related to collaboration with nonprofits and relief organizations following a

    local, visible disaster, that of the 2005 Gulf Coast Hurricanes. Our results showthat the propensity to collaborate is related to the size of the districts and its

    degree of centralization, even while controlling for a districts general

    collaborative tendency.

    Keywords: Collaborative Public Management, Nonprofit Management, Emergency Management

  • 8/9/2019 Structured to Partner: School District Collaboration with Nonprofit Organizations in Disaster Response

    3/35

    Structured to Partner

    2

    Introduction

    Numerous and frequently poignant stories emerged after Hurricane Katrina in the fall of

    2005. Most remember the horror stories of the crowds of people neglected at the Superdome.

    Many remember the images of entire neighborhoods being washed away as the flood waters

    overtook the levees in New Orleans or wiped away as the storm scoured coastal communities.

    The result of these perceptions resulted in a negative assessment of how the government

    (defined broadly) handled the disaster. Not all of the stories, however, were so negative. Amid

    the stories of government neglect and incompetence (which sadly, often overshadowed stories of

    the heroism and competence of many civil servants), many contrasting stories emerged about the

    successful efforts of private companies and nonprofit organizations to provide relief. These

    private and nonprofit organizations emerged as important secondary disaster organizations

    organizations who provided services to those affected by a disaster but whose mission was not

    primarily disaster related. Examples of such secondary disaster organizations include local

    churches that provided shelter or food to people in their community who lost their homes;

    schools that enrolled the children of evacuees; or animal protection organizations that provided

    assistance to evacuees with pets. These stories led to a shift in public perception about the role

    and capacity of government to address future disasters adequately, and led many to call for an

    increased role for private and nonprofit organizations, as secondary disaster organizations, in

    governmental planning for emergency response.

    The emergence of private and nonprofit actors as key players in disaster response has

    raised the profile and perceived utility of secondary disaster organizations in emergency

    management. It is now clear that emergency preparedness must include a role for local business

  • 8/9/2019 Structured to Partner: School District Collaboration with Nonprofit Organizations in Disaster Response

    4/35

    Structured to Partner

    3

    organizations, local welfare agencies, local education institutions, local religious institutions, and

    many others. It is increasingly clear that effective disaster response calls for the incorporation of

    organizations for which disasters are not their primary mission. In this paper, we address two

    aspects of the non-governmental components of emergency management. Using a survey of

    Texas public school districts, we investigate how school districts (not primarily disaster

    organizations) connect to nonprofit organizations (including many relief organizations and non-

    disaster nonprofits). In the first part of the paper, we discuss the existing literature on

    collaboration and emergency response. From this literature, we adapt general hypotheses related

    to collaborative propensity between public organizations and nonprofit organizations. The

    hypotheses focus on the role of organizational structure to create capacity for collaboration. We

    then describe a survey that we conducted of Texas school districts to assess their collaborative

    practices in emergency response and its results. Finally, we discuss the most important questions

    remaining about the collaboration between public and nonprofit organizations.

    Collaboration and Disaster Response

    Collaboration has been the subject of great interest in policy and practice discussions

    (e.g. Choi, 2008; Gazley & Brudney, 2007; Graddy & Chen, 2006) and become a popular subject

    of speculation and research in public administration (Foster & Meinhard, 2002; McGuire, 2006;

    OToole, 1997; Waugh & Streib, 2006). Despite such interest, collaboration is defined variously

    and with continuing imprecision. In general, collaboration is considered in terms of organizing

    structures, in which collaboration exists on a continuum from informality to formality, with a

    central concern resting on the issue of at what point a cooperative relationship transforms into a

    collaborative one. Across this continuum, for example, collaboration might span from structures

  • 8/9/2019 Structured to Partner: School District Collaboration with Nonprofit Organizations in Disaster Response

    5/35

    Structured to Partner

    4

    that are informal and episodic, such as collaboration on a task force by business, citizen, and

    nonprofit organizations to address an acute public health concern, to highly formalized contracts

    between governmental agencies and social service providers.

    Collaboration has also been defined in terms of process aspects around a problem about

    which parties that see different aspects of a problem can constructively explore their differences

    and search for solutions that go beyond their own limited visions of what is possible (Gray,

    1989, p. 5). It is clear, however, collaboration suggests something less than authoritative

    coordination and something more than tacit cooperation (Foster, 2002, p. 19). Collaboration for

    our purposes is defined as a formal organization providing assistance in an emergent fashion to

    address a disaster response need, whether as part of an existing collaboration or in the context of

    a new relationship or venture.

    In terms of cross-sector collaboration, the definition focuses on the parties to the

    collaboration. Bryson, Crosby and Stone define cross-sector collaboration as partnerships

    involving government, business, nonprofits and philanthropies, communities, and/or the public

    as a whole (2006, p. 44). The specific policy domains in which people have investigated

    collaborative public management have ranged from environmental regulation (Koontz &

    Thomas, 2006), social service delivery (Provan & Milward, 1995), and the focus of this paper

    emergency management (Waugh & Streib, 2006; Simo & Bies, 2007). In each of these areas,

    writers have been optimistic that collaborative practices will improve outcomes and efficiency,

    as well as increase the democratic inclusion of stakeholders. However, there are still many

    standing questions in relation to policy networks and collaborative public management

    (Robinson, 2006).

  • 8/9/2019 Structured to Partner: School District Collaboration with Nonprofit Organizations in Disaster Response

    6/35

    Structured to Partner

    5

    One gap in our knowledge that is particularly salient to collaboration and emergency

    response is the proclivity for secondary disaster organizations to engage in collaborative public

    management practices, and more specifically, the proclivity for collaboration with nonprofit

    partners. Why is it that some secondary disaster organizations connect with other organizations

    in responding to or planning for disasters while others do not?

    In the sections that follow, we explore the dynamics of collaboration with special

    attention to issues related to collaboration in disaster response contexts. We review the

    controversies over the meaning of collaboration as well as the forces that are hypothesized to

    predispose some organizations to collaborate. The section concludes with general hypotheses

    reflecting our literature-grounded expectations of the role of organizational structure in the

    decision whether to collaborate.

    Choosing To Collaborate

    A key area requiring investigation is how partners choose to link into collaborative

    networks. For the most part, research (especially the large N quantitative literature) has focused

    on the characteristics of collaborations in operation. The issue of how participants decide to

    participate is central to the democratic potential of collaborative networks. Fung (2006) warned

    of the illusory nature of some forms of participation. Even large numbers of participants in

    administrative processes may not be particularly democratic if the participants are similar and

    fail to represent the diversity of the effected population. A key mistake, Fung noted, was to

    assume that self-selected participation systems would likely result in diverse participants. While

    the solutions to the potentially skewed distribution of participants are not obvious or inexpensive,

  • 8/9/2019 Structured to Partner: School District Collaboration with Nonprofit Organizations in Disaster Response

    7/35

  • 8/9/2019 Structured to Partner: School District Collaboration with Nonprofit Organizations in Disaster Response

    8/35

    Structured to Partner

    7

    Collaboration in Emergency Management and Disaster Response

    Scholars and practitioners in emergency management have been among those most active

    in promoting the importance of collaborative public management. By their very nature,

    emergency management taxes the capacity of any organization or community. As has been

    evident in the aftermath of recent emergencies ranging from terrorist attacks to natural disasters,

    no one organizationor even municipality or statecan deal with large-scale emergencies

    alone. It is nearly impossible, then, to effectively prepare for or respond to emergencies without

    extensive collaboration.

    Louise Comfort has been a pioneer in researching collaborative public management in the

    area of emergency response. Her work on complexity theory and the management of emergency

    response networks has made it clear to researchers and practitioners alike that emergency

    response requires collaborative networks and that these networks raise a number of new

    questions about the governance of these arrangements (Comfort,1994, 2007). The most basic

    question raised in these studies was who is it that participates in disaster response?

    Recent research into large scale disaster response has noted the prominent role of

    nonprofit organizations in providing response and relief service. For example, research on the

    response to Hurricane Katrina revealed the importance of nonprofit organizations in community

    response to disaster. Robinson, Berrett, and Stone (2006) described the emergence of many new

    partnerships between nonprofit organizations and local government officials in the Dallas/Fort

    Worth area, as well as among networks of nonprofit organizations. This early research showed

    how the picture of emergency response following the hurricanes of 2005 would have been

    woefully incomplete without including the nonprofit collaborations. Similarly, Simo and Bies

  • 8/9/2019 Structured to Partner: School District Collaboration with Nonprofit Organizations in Disaster Response

    9/35

    Structured to Partner

    8

    (2007) documented the diversity of nonprofit collaborations along the Gulf Coast generally

    during the same response period.

    Mirroring the general literature on collaboration, the research into nonprofit

    collaborations in disaster response have documented that there are many such collaborations.

    What is lacking is a theory of why some collaborative relationships emerge while many other

    potential collaboration fail to materialize. Why do some organizations have a diverse set of

    collaborative partners while others collaborate with few (or no) external partners? We seek to

    address this question. Before we present our research, we want to take a moment to discuss

    collaboration from the side of the nonprofit organization.

    Collaborations with Nonprofits

    Nonprofits, like public agencies, are increasingly engaged in collaborative activity to

    achieve public purposes. Such collaboration routinely cuts across across government, business,

    and nonprofit sectors (Milne, Iyer, & Gooding-Williams, 1996), especially as nonprofit-public

    sector collaborations become a prominent feature of public service delivery in the United States

    and abroad (Gidron, Kramer, & Salamon, 1992).

    In their research on government-nonprofit collaborations,Gazley and Brudney (2007)

    argued that a desire for resource stability facilitates collaboration. The authors found that the

    motivation by nonprofits to collaborate is driven by an organizations need to secure scarce

    resources, with funding for nonprofits and its expertise and capacity for government as

    predictors (p. 340). Government and nonprofit respondents provided remarkably similar

    responses to questions about the goals of the collaboration explaining collaborations as

    valuable in jointly addressing problems, improving community access to a service, improving

  • 8/9/2019 Structured to Partner: School District Collaboration with Nonprofit Organizations in Disaster Response

    10/35

    Structured to Partner

    9

    community relations, and promoting shared goals (p. 398). Their research also suggested that

    nonprofits differ in some respects from government in terms of motivations to partner, with

    nonprofits (relative to government) more likely to emphasize using collaborations to build

    relationships to presumably help them gain resources (p. 398).

    Gazley and Brudney also explored factors that explain non-collaboration. The authors

    found that concerns about internal capacity and mission, rather than external factors such as

    statutory pressure, appear to provide the strongest rationale for entering or avoiding public-

    private partnerships (p. 412). Jang and Feiock (2007) also suggested that the costs of

    collaboration and subsequent influence of nonprofits financial stakeholders are associated with

    collaborative behavior.

    Similarly, recent research on nonprofit-school partnerships provides a nuanced view of

    institutional and environmental factors on collaboration, and indicated that organizational

    characteristics and local policy environments may have a significant effect on the form and

    intensity of collaborative efforts (Sowa, 2008, p. 317). In her study of interagency collaborations

    to deliver early childhood services between school districts and community-based organizations,

    Sowa (2008) emphasized the importance of examining the type of resources shared between

    organizations in a collaborative relationship, and beyond this, the commitment of collaborative

    partners, because varying degrees of commitment may produce significant variation in the

    implementation of these collaborative service delivery mechanisms (p. 317). To the extent that

    nonprofit organizations operate on different resource bases than organizations in other sectors,

    Sowas argument suggests that there may be inherent difficulties in cross-sector collaboration.

  • 8/9/2019 Structured to Partner: School District Collaboration with Nonprofit Organizations in Disaster Response

    11/35

    Structured to Partner

    10

    This research illustrates the interest in better understanding the complexities of

    collaboration that crosses sector boundaries between government and nonprofit organizations.

    Given the prominence of these collaborations and the complexities involved in crossing sector

    boundaries, it is important to test propositions about why some government organizations choose

    to collaborate with nonprofits while others do not. The next section develops general

    propositions about when organizations are likely to seek collaborative partners grounded in the

    theory of inter-organizational relations.

    Theoretical Expectations

    Based on the existing literature reviewed above, we have expectations about the sorts of

    factors that would increase the probability that a public agency would seek to collaborate with a

    nonprofit organization. It would be prohibitively complex to assess all of the possible factors that

    can influence the propensity to collaborate with nonprofits. We have chosen to focus on how

    organizational structure affects the propensity to collaborate. We will control for other potential

    factors in the empirical test, but our theoretical focus will be on structure in this section.

    First, we expect larger organizations to possess the slack resources and the opportunities

    for structural specialization to facilitate collaboration of any type. As a general effect, this should

    affect organizations propensities to collaborate with nonprofit organizations. This expectation is

    defended in detail in Thompsons classic text Organizations in Action (2003[1967]).

    Organizations generally seek to protect their core operations from turbulence in the environment.

    Managers dont want the people who perform functions close to the core mission of the

    organization to worry about the supply of resources or other potentially uncertain components of

    the environment. As a result, they create specialized units to monitor and interact with the

  • 8/9/2019 Structured to Partner: School District Collaboration with Nonprofit Organizations in Disaster Response

    12/35

    Structured to Partner

    11

    uncertain environment. Only larger organizations, though, have the economies of scale to devote

    resources to these sorts of specialized units. Larger organizations can have separate press

    relations units, analysis departments, or the like. All of these units serve to keep people

    performing the core operations of the organization in their daily tasks.

    Recent empirical work confirms some of theoretical explanations for nonprofit

    collaboration reviewed above. Specifically, resource-dependence and environmental uncertainty

    explanations for collaboration have been supported in the nonprofit literature (e.g., Gazley &

    Brudney, 2007; Guo & Acar, 2005; Foster & Meinhard, 2002; Singer & Yankey, 1991). The

    diversity of the evidence convinces us that this phenomenon is general across sectors. As a

    result, we offer the following proposition.

    Proposition One: Larger organizations are more likely to collaborate than smaller ones.

    In addition to organizational size, the arrangement of resources within an organizations

    structure is also likely to be important. Using the same Thompsonian logic as above, one expects

    to see greater investment in boundary spanning operations in organizations where resources are

    sufficiently centralized to realize economies of scale. If resources are decentralized, each unit

    divides the potential base for such boundary spanning organization. Where these resources are

    instead centralized, it is more likely that an organization can overcome the threshold start-up

    costs needed to create a boundary spanning relationship. As a result we also propose the

    following:

    Proposition Two: Organizations with a greater degree of centralization are more likely to

    collaborate than organizations with a lower degree of centralization.

  • 8/9/2019 Structured to Partner: School District Collaboration with Nonprofit Organizations in Disaster Response

    13/35

    Structured to Partner

    12

    We outline our approach to testing these propositions and the results of our tests in the next

    section.

    METHODS AND RESULTS

    These propositions call for extensive data on the collaborative behavior of organizations

    and a testable model of these behaviors. In this section, we review the data collection strategy

    and our approach to operationalization, a nested regression model with which we simultaneously

    test hypotheses derived from propositions one and two, and the results of the analysis.

    Data Collection

    As reviewed in a previous section, scholars have studied collaborative behaviors in a

    variety of policy arenas. We chose to test these propositions within the context of emergency

    management and disaster response. Specifically, this choice allows us to test the propositions

    related to collaboration with nonprofit organizations in ways that would not have been possible

    in other policy arenas. We chose to survey Texas public school superintendents about their

    collaborative behaviors for a variety of reasons related to the strength of this sampling frame and

    the saliency of the context.

    First, this sampling frame is well studied and provides us a background against which to

    compare our survey methodologyreducing the probability of results that are the result of

    unpredicted characteristics of an untested sampling frame. Second, the sampling frame is one

    where repeated surveys have resulted in relatively high response rates. This reduced the chance

    that the survey would result in a data set that was too small for statistical analysis. Third,

  • 8/9/2019 Structured to Partner: School District Collaboration with Nonprofit Organizations in Disaster Response

    14/35

    Structured to Partner

    13

    previous research on the sample frame allows us to leverage years of previous research to test

    propositions about structural and financial characteristics of the agencies. Fourth, the subject area

    of collaboration (emergency management and disaster response) was salient, allowing for survey

    measurement at a time when the respondents were most likely to be deliberately choosing

    collaborative strategiesrather than operating under unconscious forces of structural inertia.

    We conducted the survey in the immediate aftermath of Hurricanes Katrina and Rita

    (about 60 days after Katrina). One should not discount the importance of this salience. Salience

    makes the survey a better test of conscious strategies, but may misrepresent more typical

    strategiesbut this is a question for future investigation. While this salience does create an easy

    test of collaboration it tests for collaborative behaviors when these behaviors are most likely

    it is appropriate to use an easy test strategy when studying relatively unstudied behaviors (like

    the decision of public agencies to collaborate with nonprofits in emergency management).

    Finally, the sampling frame includes an array of organizations that varied on all of our variables

    (large and small districts, districts with recent or distant emergency experience, etc.).

    The survey went to all public school district superintendents in Texas with two follow-up

    waves. The final response rate was approximately 50%. Comparisons of the respondents to

    known characteristics of the non-respondents suggest that the respondents were slightly (though

    not statistically significantly) more affluent than non-respondents and significantly larger on

    average (though of a relatively small magnitude). With these comparisons, we are confidents that

    the results of the survey provided reasonable approximations of the population values.

    Operationalization

  • 8/9/2019 Structured to Partner: School District Collaboration with Nonprofit Organizations in Disaster Response

    15/35

  • 8/9/2019 Structured to Partner: School District Collaboration with Nonprofit Organizations in Disaster Response

    16/35

    Structured to Partner

    15

    months of 2005. We have greater faith in patterns observed across tests using multiple

    operationalizations of collaboration.

    Each proposition also includes an independent variable. The first proposition is related to

    organizational size. We use state data on each school district to measure each districts size. For

    each district, we included the logged total number of full time equivalent employees as the

    measure of size. This measure is highly correlated (at around .997, depending on the year) with

    other popular measures of district size such as total enrollment. We logged the value to reduce

    what was a significant skew. This measure of size represents organizational size and not the

    geographical area served by the district. This limitation will be important as we discuss the future

    directions for research in the area.

    The second proposition deals with the structural centralization within the district. We

    operationalized this concept in two ways. First, we included a measure of the percentage of the

    district budget dedicated to central administration. High values indicate a leaders general

    preference for centralized offices rather than campus-based capacity. This measure is similar to

    prior measures of administrative centralization and bureaucratization (Robinson, 2004; Robinson

    et al. 2007). Second, we asked specifically about the districts strategy toward delegating

    preparedness activities to the individual campus level. The specificity of the question should

    balance the pressures toward reporting centralized emergency preparedness with the general

    predisposition among education administrators (and the programs that certify them) to pursue

  • 8/9/2019 Structured to Partner: School District Collaboration with Nonprofit Organizations in Disaster Response

    17/35

    Structured to Partner

    16

    campus-based policymaking. The result was a response pattern that suggests variety in district

    delegation strategies (ranging from delegation to centralization).1

    While we expect that these measures of size and centralization to explain some of the

    variation in observed collaborations between school districts and nonprofit organizations, we do

    not expect that these propositions will explain all of the variation. We take various steps to

    ensure that spurious forces are not contaminating our inferences related to propositions one and

    two. To control for the differences in disaster vulnerability and recent disaster experience, we

    included variables that measure reported impact of recent disasters as well as the perceived

    likelihood that the district will experience a disaster in the near future.

    Even with the control for disaster vulnerability, there are a variety of factors that could be

    spuriously related to both organizational structure and collaboration with nonprofit

    organizations. The set of variables omits a variety of important concepts in collaborative public

    management that we simply could not simultaneously measure. To assess whether our model is

    greatly affected by this under-specification, we tested two models. The first is as described

    above. The second specification includes an additional variable of how many collaborations,

    other than those reported with nonprofit and relief organizations, the superintendent reported.

    This serves as a control for general collaborative predisposition and will account for all of the

    unmeasured factors related to both general collaboration and nonprofit collaboration. Any

    reported effects in the models that include the general collaborative tendency measure indicate

    the influence of the variable on the specific probability of collaborating with nonprofit and relief

    1These two measures of centralization do not present multi-collinearity problems in that they are uncorrelated

    with each other.

  • 8/9/2019 Structured to Partner: School District Collaboration with Nonprofit Organizations in Disaster Response

    18/35

    Structured to Partner

    17

    organizations beyond the effect that the variable has on general collaborative tendencies. This

    represents a hard test for the relationships considered.

    Regression Model

    We test our two propositions using a probit regression model. The dependent variable

    was given a value of 1 if the superintendent reported that the district had collaborated with

    nonprofit or relief organizations (or had regularly scheduled meetings with them in the third

    model). If the superintendent reported that they had not collaborated with nonprofit

    organizations, the dependent variable was given a value of 0. Non-responses were omitted from

    the analysis.

    Besides logging the size and centralization variables (as described above), we assume

    simple linear relationships between all of the independent variables and the dependent variables.

    While nonlinear relationships are potentially interesting, we reserve those possibilities for future

    research. The result is the following regression equation.

    EQ.1 - P (Nonprofit Collaboration) = f [0 + 1 (Size)) + 2 (General Delegation) + 3

    (Emergency Planning Delegation) + 4(Likelihood) + 5 (Recency) + 6(Impact) + ]

    We also tested models that added the general collaborative predisposition. The result is equation

    2.

    EQ. 2 - P(Nonprofit Collaboration) = f [0 + 1 (Size)) + 2 (General Delegation) + 3

    (Emergency Planning Delegation) + 4(Likelihood) + 5 (Recency) + 6(Impact) + 7 (General

    Collaborative Tendency.)+ ]

    For each of the models, we employ robust estimators to produce robust standard errors.

    Results

  • 8/9/2019 Structured to Partner: School District Collaboration with Nonprofit Organizations in Disaster Response

    19/35

    Structured to Partner

    18

    The results of the three models are presented in Table 1. Model 1 performs relatively

    well, successfully predicting 68% of collaborations and for a proportionate reduction of error of

    37%. The model follows previous results on generalized collaborative behaviors in suggesting

    that organizational size and centralization both make collaboration with nonprofit organizations

    more likely. These three variables (including the two measures of centralization) are in the

    predicted direction (larger organizations and greater centralization induce greater probability of

    collaborating with nonprofits) and were strongly significant.

    Models 2 and 3 provide tests on the robustness of these initial findings. Model 2 adds the

    general collaborative tendency as an independent variable. This robust model performs better

    overall with a predictive accuracy of 76% with a proportionate reduction in error of 52%. This

    additional variable changes the nature of the estimated values. As one would predict, the general

    collaborative tendency is strongly significant. The effect size is also quite large.

    Figure 1 illustrates the size of the effect of general collaborative tendency. Using the

    simulation techniques described in Appendix 1, we estimated the predicted probability of

    collaboration for each level of general collaborative tendency while holding all other variables at

    their mean (for continuous variables) or median (for ordered variables). For purposes of

    comparison, note that 48% of superintendents reported collaboration overall. As Figure 1

    illustrates, the expected value of an otherwise typical district with no other collaborations is less

    than 20% likely to collaborate with a nonprofit or relief organization. The line in the center of

    each box plot is the median prediction with the grey and outer enclosed areas representing the

    uncertainty around the estimated median. An organization with five other collaborative partner

    types (the maximum possible) is over 80% likely to collaborate with a nonprofit or relief

  • 8/9/2019 Structured to Partner: School District Collaboration with Nonprofit Organizations in Disaster Response

    20/35

    Structured to Partner

    19

    organization. These predicted probabilities give an indication of the large effect size of

    generalized collaborative tendency, which is not surprising.

    The introduction of such a baseline variable, much like adding a lagged dependent

    variable in a time series model, reduced the magnitude and significance of the other components

    in the model. The size of the organization is still significant though the effect size is reduced by

    25%. Still, the variables representing organizational size and centralization were significant in

    both models.

    Model 2 serves as a hard test for any variable, so we focus our analysis on the results of

    this hard test. Figure 2 illustrates the effect of organizational size on the probability of

  • 8/9/2019 Structured to Partner: School District Collaboration with Nonprofit Organizations in Disaster Response

    21/35

    Structured to Partner

    20

    collaboration with nonprofit or relief organizations. The figure reports the estimated probability

    of collaboration for an organization at each decile within the sample (e.g., an organization that is

    larger than only 10% of other districts, an organization larger than only 20% of other districts,

    etc.).

    As seen in Figure 2, when all other variables were held at typical values larger

    organizations are much more likely to collaborate with nonprofit and relief organizations.

    Districts larger than only 10% of other districts were about 20% likely to have reported

    collaboration. Districts larger than 90% of other districts were about 80% likely to report

    collaboration.

    Figures 3 and 4 report the effect of our two measures of centralization. Both general

    centralization and emergency management centralization are significant factors in the prediction

  • 8/9/2019 Structured to Partner: School District Collaboration with Nonprofit Organizations in Disaster Response

    22/35

    Structured to Partner

    21

    of collaborations with nonprofit and relief organizations. Figure 3 shows that the effect of

    general centralization can increase the probability of collaboration from 25% to 70% as one

    moves from the first to the ninth decile of centralization. Similarly, Figure 4 shows that as one

    moves from planning that is reported as entirely or mostly campus-based to entirely district-

    based, the expected probability of collaboration with nonprofit or relief organization increases

    from 34 to 54%.

  • 8/9/2019 Structured to Partner: School District Collaboration with Nonprofit Organizations in Disaster Response

    23/35

    Structured to Partner

    22

    Model 3 changes the dependent variable to regularly scheduled meetings with nonprofit

    and relief organizations. The baseline probability of such collaborations is much lower. In the

    survey sample, only 15% of districts reported holding regularly scheduled meetings with

    nonprofit or other relief organizations. The relative rarity of the collaborations makes estimating

    a model predicting the relatively rare events difficult. The predictive power of the model seems

    strong at 85%, but this represents only a 3% proportionate reduction of error because the nave

    model of simply always predicting that a district will hold no regularly scheduled meetings will

    be correct only 84% of the time. Proportionate reduction of errors is typically low in the presence

    of such relatively rare events, as in this case.

    Given the importance of the generalized collaboration variable in model 2, we have

    included it here as well. Here again, generalized collaborative tendency is the most important

  • 8/9/2019 Structured to Partner: School District Collaboration with Nonprofit Organizations in Disaster Response

    24/35

    Structured to Partner

    23

    explanatory variable. Figure 5 illustrates the effect of general collaborative tendencies, all other

    variables held at typical values. It is important to notice that the scale of the figure has been

    changed. Although the figures for collaboration reported simulated probability ranging to 100%,

    the figures for regularly scheduled meetings only go to 50% because of the rarity of the

    prediction.

    Figure 6 illustrates the impact of organizational size on regularly scheduled meetings. Again, the

    variable is significant but the effect is only dramatic at large district sizes.

  • 8/9/2019 Structured to Partner: School District Collaboration with Nonprofit Organizations in Disaster Response

    25/35

    Structured to Partner

    24

    Figure 7 illustrates the impact of general centralization and emergency management

    centralization. General centralization has the same muted effect seen for size and general

    collaborative tendency as shown in Figures 5 and 6. Interestingly emergency management

    centralization is not significant in this model. The implication is that the tendency to schedule

    regular meetings with nonprofit and relief organizations is not related to reported tendencies for

    internal emergency planning, but only to general district policies regarding centralization.

  • 8/9/2019 Structured to Partner: School District Collaboration with Nonprofit Organizations in Disaster Response

    26/35

    Structured to Partner

    25

    Discussion and Directions for Further Research

    Our results just scratch the surface of the issues related to collaboration with nonprofit

    organizations. These initial results suggest that there is much to learn about why collaborations

    emerge. Structural factors provide some leverage in explaining the collaborative behaviors of

    these civil servants. The mix of influences depends on the nature of the collaboration. The results

    do more to reveal the limitations of the extant conceptual toolbox related to collaborative public

    management than to shed light on the collaborations themselves.

  • 8/9/2019 Structured to Partner: School District Collaboration with Nonprofit Organizations in Disaster Response

    27/35

    Structured to Partner

    26

    Consider this example. The findings depend on whether collaboration is defined openly

    (in this case, allowing survey respondents to define it for themselves) or whether a strict

    definition requiring regularly scheduled meetings is used. The evidence suggests that these two

    different activities are quite distinct. This raises the question of whether the data represent two

    different types of collaboration, two extremes of one essential type of collaboration, or whether

    only one of these is bona fide collaboration. We do not currently possess the conceptual

    framework to distinguish between these possibilities nor is there sufficient consensus on the

    definition of collaboration to distinguish between different types. This sort of conceptual work is

    a precondition for compelling interpretations of the data.

    While conceptual limitations affect our ability to interpret the results of the analysis,

    significant data limitations are also present. The survey used here went to public managers and

    provides insight into the government side of the cross-sector collaboration. What we do not have

    is data on the nonprofit side of the collaboration. The review of the literature revealed some

    important similarities between the theories of public sector collaboration and cross-sector

    collaborations; in the present research we were not able to test propositions about the nonprofit

    side of the relationship.

    The key data we still need to develop is on the supply side of the cross-sector

    collaborations. Scholars have generally hypothesized that nonprofit supply is expressed as a

    function of the needs and resources of the region (Bielefeld & Murdoch 2004, p. 222). Studies

    (e.g., Gronbjerg & Paarlberg, 2001; Wolch & Geiger, 1983) have confirmed the robustness of

    this assumption. More specifically, population growth (Bielefeld, 2000), community wealth

    (Wolch & Geiger, 1983; Wolpert, 1993; Corbin, 1999), and proportion of older residents

    (Gronbjerg & Paarlberg) have been identified as determinants of the nonprofit supply. Peck

  • 8/9/2019 Structured to Partner: School District Collaboration with Nonprofit Organizations in Disaster Response

    28/35

    Structured to Partner

    27

    (2008) summarizes the literature on the nonprofit location as tending to frame the determinants

    of nonprofit location as a function of neighborhood characteristics, hypothesized as:

    measures of need (such as poverty density), popula tion characteristics (such as

    education levels, racial distribution, age, families' structure), community

    characteristics (such as heterogeneity, incorporation date, urbanicity), resources (such

    as government expenditures, property tax base, housing values), and other

    organizations' locations (both for-profit and nonprofit organizations.(p. 139)

    Although nonprofit supply is somewhat implicit in the environmental and structural

    explanations for nonprofit motivation to collaborate, it has generally not been explicit and has

    typically been defined in non-spatial terms. Measurement challenges may partially account for

    this gap, as may the fact that the collaboration and supply literature streams have developed

    concurrently, but for somewhat distinct purposes. Service provision by nonprofits has been used

    as a measure of nonprofit sector size, as has annual expenditures by nonprofits. Spatial analyses

    are increasingly promising, with new methods allowing multidimensional considerations to

    space, such as nonprofit income in relationship to nonprofit concentration and population

    characteristics. To the extent that larger school districts are in locations with a greater supply of

    partners, some of the role of size in the statistical models may be picking up factors related to

    nonprofit partner supply. The inclusion of the general collaborative propensity addresses this

    problem to a great extent. To the extent that nonprofit supply is related to the supply of other

    sorts of actors (general partner density), the general collaboration propensity control variable will

    partial out these factors. All that will remain are the elements unique to nonprofit supply and not

    correlated with the supply of other actorsa much more limited problem.

    It is not surprising that there are few data on the supply of nonprofit organizations. There

    are considerable conceptual difficulties in developing such a measure. What is the relevant range

  • 8/9/2019 Structured to Partner: School District Collaboration with Nonprofit Organizations in Disaster Response

    29/35

    Structured to Partner

    28

    of a spatial measure of nonprofit supply? Should we only count nonprofit organizations located

    within the geographic boundaries of, in this case, a school district? What prevents a nonprofit

    from nearby assisting? How broad should we draw that circle? Our intuition is to bring the

    measure up to the level of a metropolitan standard area but this measure is itself difficult to

    implement and would dramatically reduce our sample to school districts within MSAs. Despite

    these difficulties, future work that directly measures nonprofit partner supply would be a great

    addition to the model. Further, to be able to parse nonprofit partner supply by primary or

    secondary mission focus on emergency and disaster response could assist understanding of cross-

    sector collaboration in the specific policy domain of emergency response and could have

    implications for disaster planning, particularly in deliberative efforts to improve cross-sectoral

    network formation.

    This research leaves us wanting to press forward in three directions. First, we want to

    elaborate the theory of collaboration to understand better which factors should be related to

    which types of collaboration. Second, we would like to collect data (beyond the available IRS

    Form 990 data) on the nonprofit side of these cross-sector collaborations to see how structural

    elements on the partners side can influence the emergence of partnerships. Third, we would like

    to develop measures of the supply of nonprofit organizations available to public agencies as

    partners. These developments should provide considerable insight into cross-sector collaborative

    networks.

    With the increased importance of secondary disaster organizations, it becomes all the

    more important to understand the dynamics of collaboration. In the cases studied here,

    secondary disaster organizations (school districts) sought collaboration in relation to emergency

    management and disaster response with nonprofits to varying degrees. The data suggest that

  • 8/9/2019 Structured to Partner: School District Collaboration with Nonprofit Organizations in Disaster Response

    30/35

    Structured to Partner

    29

    these secondary organizations may be limited by their organizational structures as to the level of

    collaboration they are likely to seek. However, these secondary organizations are those that most

    need to seek additional resources (including expertise) from partners. Attempts to cultivate

    collaboration between secondary disaster organizations and others in matters of emergency

    management and disaster response needs to be sensitive to the importance of structure and the

    barriers to collaboration that poorly suited structures may represent.

  • 8/9/2019 Structured to Partner: School District Collaboration with Nonprofit Organizations in Disaster Response

    31/35

    Structured to Partner

    30

    Table 1.Regression of collaboration with nonprofit organizations

    MODEL 1 - THE BASELINE COLLABORATION MODEL

    MODEL 2 THE BASELINE COLLABORATION MODEL WITH THE GENERALIZED

    COLLABORATION VARIABLE ADDED FOR ROBUSTNESS

    MODEL 3 THE REGULAR MEETING COLLABORATION MODEL WITH THE

    GENERALIZED COLLABORATION VARIABLE

    ______________________________________________________________________________

    Model (1) (2) (3)

    Organization Size .45 (4.4) .44 (4.5) .43 (4.3)

    Planning Delegation .16 (2.0) .18 (2.1) .13 (.14)

    General Centralization .68 (3.5) .84 (3.8) 1.1 (4.3)

    General Collaborative ---------- .44 (4.5) .12 (2.2)

    Propensity

    N 435 435 470

    Percent Correctly Predicted 68% 76% 85%

    Percent Reduction in Error 37% 52% 3%

    Note: The absolute values of the Z-statistics are in parentheses.

    Controls for perceived likelihood of disasters and the impact of recent disasters were also

    included in all models.

    Coefficients significant at greater than .05 are in bold.*

  • 8/9/2019 Structured to Partner: School District Collaboration with Nonprofit Organizations in Disaster Response

    32/35

    Structured to Partner

    31

    REFERENCES

    Alexander, J., & Nank, R. (2009). Publicnonprofit partnership: Realizing the new public

    service.Administration and Society,41(3), 364-386.

    Bailey, D. & Koney, E. M. (1996). Interorganizational community-based collaboratives: A

    strategic response to shape the social work agenda. Social Work,41(6), 602-611.

    Bardach, E. (1998). Getting agencies to work together: The practice and theory of managerial

    craftsmanship. Washington, DC: Brookings Institution Press.

    Benjamin, L. (2008). Bearing more risk for results: Performance accountability and nonprofit

    relational work.Administration & Society, 39(8), 959-983.

    Bielefeld, W. (2000). Metropolitan nonprofit sectors: Finding from NCCS data. Nonprofit and

    Voluntary Sector Quarterly, 29(2), 297-314.

    Bielefeld, W., & Murdoch, J. (2004). The locations of nonprofit organizations and their for-profit

    counterparts: An exploratory analysis.Nonprofit and Voluntary Sector Quarterly, 33(2),

    221-246.

    Brinkerhoff, J. (2002). Government-nonprofit partnership: a defining framework. Public

    Administration and Development, 22(1), 19-30.

    Brown, T.L. & Potoski, M. (2003). Contract-management capacity in municipal and county

    governments. Public Administration Review, 63(2), 153-164.

    Bryson, J. M., Crosby, B. C., & Stone, M. M. (2006). The design and implementation of cross-

    sector collaborations: Propositions from the literature. Public Administration Review,

    66(Suppl. to Issue 6), 44-45.

    Corbin, J. J. (1999). A study of factors influencing the growth of nonprofits in social services.

    Nonprofit and Voluntary Sector Quarterly, 28, 296-314.

    Comfort, L. K. (1994). Self-organization in complex systems.Journal of Public Administration

    Research and Theory, 4(3), 383-410.

    Comfort, L. K. (2007). Crisis management in hindsight: Cognition, communication,

    coordination, and control. Public Administration Review, 67(Suppl. to Issue 6), 189-197.

    Dicke, L. (2001). Ensuring accountability in human services contracting: Can stewardship theory

    fill the bill?American Review of Public Administration, 32(4), 455-470.

  • 8/9/2019 Structured to Partner: School District Collaboration with Nonprofit Organizations in Disaster Response

    33/35

    Structured to Partner

    32

    Ebrahim, A. (2003). Making sense of accountability: Conceptual perspectives for northern and

    southern nonprofits.Nonprofit Management & Leadership, 14(2), 191-212.

    Foster, M. K., & Meinhard, A. G. (2002). A regression model explaining predisposition to

    collaborate.Nonprofit and Voluntary Sector Quarterly, 31(4), 549-564.

    Fung, A. (2006). Varieties of participation in complex governance. Public Administration

    Review, 66(Suppl. to Issue 6), 65-75.

    Galaskiewicz, J. (1985). Professional networks and the institutionalization of a single mind set.

    American Sociological Review, 50(5), 639-658.

    Gazley, B. (2008a). Why not partner with local government? Nonprofit managerial perceptions

    of collaborative disadvantage.Nonprofit and Voluntary Sector Quarterly, OnlineFirst, no

    volume number yet assigned.

    Gazley, B. (2008b). Beyond the contract: The scope and nature of informal government

    nonprofit partnerships. Public Administration Review, 68(1), 141-154.

    Gazley, B., & Brudney, J. L. (2007). The purpose (and perils) of government-nonprofit

    partnership.Nonprofit and Voluntary Sector Quarterly, 36(3), 389-415.

    Gidron, B., Kramer, R., & Salamon, L. (Eds.). (1992). Government and the third sector:

    Emerging relationships in welfare states. San Francisco: Jossey-Bass Publishers.

    Graddy, E. A., & Chen, B. (2006). Influences on the size and scope of networks for socialservice delivery.Journal of Public Administration and Theory, 16, 533-552.

    Gronbjerg, K., & Paarlberg, L. (2001). Community variations in the size and scope of the

    nonprofit sector: Theory and preliminary findings.Nonprofit and Voluntary Sector

    Quarterly, 30(4), 684-706.

    Guo, C., & Acar, M. (2005). Understanding collaboration among nonprofit organizations:

    Combining resource dependency, institutional, and network perspectives.Nonprofit and

    Voluntary Sector Quarterly, 34(3), 340-361.

    Jang, H. S., & Feiock, R. (2007). Public versus private funding of nonprofitorganizations. Public Performance and Management Review , 31(2), 174-190.

    Jamal, T., & Getz, D. (1995). Collaboration theory and community tourism planning.Annals of

    Tourism Research, 22(1), 186-204.

  • 8/9/2019 Structured to Partner: School District Collaboration with Nonprofit Organizations in Disaster Response

    34/35

    Structured to Partner

    33

    Jang, H. S., & Feiock, R. C. (2007). Public versus private funding of nonprofit organizations:

    Implications for collaboration. Public Performance & Management Review, 31(2), 174-

    190.

    McGuire, M. (2006). Collaborative public management: Assessing what we know and how we

    know it. Public Administration Review, Special Issue, 33-43.

    Milne, G. R., Iyer, E. S., & Gooding-Williams, S. (1996). Environmental organization alliance

    relationships within and across nonprofit, business, and government sectors.Journal of

    Public Policy & Marketing, 15(2), 203-215.

    OFlynn, J. (2009). The cult of collaboration in public policy. The Australian Journal of Public

    Administration, 68(1), 112-116.

    OToole, L. J. Jr. (1997). Treating networks seriously: Practical and research-based agendas in

    public administration. Public Administration Review, 57(1), 45-52.

    Peck, L. R. (2008). Do antipoverty nonprofits locate where people need them? Evidence from a

    spatial analysis of Phoenix.Nonprofit and Voluntary Sector Quarterly, 37(1), 138-151.

    Pfeffer, J., & Salancik, G. R. (1978). The external control of organizations: A resource

    dependence perspective. New York: Harper & Row.

    Provan, K. G., & Milward, H. B. (1995). A preliminary theory of network effectiveness: A

    comparative study of four community mental health systems.Administrative Science

    Quarterly, 40(1), 1-33.

    Robinson, S.. E. (2006). A decade of treating networks seriously. Policy Studies Journal, 34(4),

    589-598.

    Robinson, S. E., Berrett, B., & Stone, K. (2006). The development of collaboration of response

    to Hurricane Katrina in the Dallas area. Public Works Management and Policy, 10(4),

    325-327.

    Robinson, S. E., & Gettis, B. (2007). Seeing past parallel play: Survey measures of

    collaboration in disaster situations. (Bush School Working Paper #391). Retrieved

    February 17, 2010, fromhttp://bush.tamu.edu/research/workingpapers/

    Salamon, L. M. (2002). The tools of government: A guide to the new governance . New York:

    Oxford University Press.

    Singer, M. I., & Yankey, J. A. (1991). Organizational metamorphosis: A study of eighteen

    nonprofit mergers, acquisitions, and consolidations.Nonprofit Management and

    Leadership, 1(4), 357-369.

    http://bush.tamu.edu/research/workingpapers/http://bush.tamu.edu/research/workingpapers/http://bush.tamu.edu/research/workingpapers/http://bush.tamu.edu/research/workingpapers/
  • 8/9/2019 Structured to Partner: School District Collaboration with Nonprofit Organizations in Disaster Response

    35/35

    Structured to Partner

    Simo, G., & Bies, A. L. (2007). The role of nonprofits in disaster response: An expanded model

    of cross-sector collaboration. Public Administration Review, 67(Suppl. to Issue 6), 125-

    142.

    Sowa, J. E. (2008). Implementing interagency collaborations: Exploring variation in

    collaborative ventures in human service organizations.Administration & Society, 40(3),298-323.

    Smith, S.R. & Lipsky, M. 1993.Non-profits for hire: the welfare state in the age of contracting.

    Cambridge, MA: Harvard University Press.

    Thompson, J. D. 2003[1967]. Organizations in Action: Social Sciences Bases of Administrative

    Theory. Edison, NJ: Transaction Press.

    Van Slyke, D.M. (2003). The mythology of privatization in contracting for social services.

    Public Administration Review, 63(3), 296-315.

    Van Slyke, D. M. (2006). Agents or stewards: Using theory to understand the government-

    nonprofit social service contracting relationship.Journal of Public Administration

    Research and Theory, 17(2), 157-187.

    Waugh, W. L. Jr., & Streib, G. (2006). Collaboration and leadership for effective emergency

    management. Public Administration Review, Special Issue, 131-140.

    Williamson, O. E. (1975).Markets and hierarchies: Analysis and antitrust implications. New

    York: Free Press.

    Williamson, O. E. (1985). The economic institutions of capitalism. New York: Free Press.

    Williamson, O. E. (1991). Comparative economic organizations: The analysis of discrete

    structural alternatives.Administrative Science Quarterly, 36(2), 269-296.

    Wolch, J. R., & Geiger, R. K. (1983). The distribution of urban voluntary resources: An

    exploratory analysis.Environment and Planning, 15, 1067-1082.

    Wolpert, J. (1993). Decentralization and equity in public and nonprofit sectors. Nonprofit and

    Voluntary Sector Quarterly, 22, 281-296.

    Word, J. (2006).A structural examination of collaborative relations between nonprofitorganizations in the greater Jacksonville area. Unpublished doctoral dissertation, Florida

    State University, Tallahassee.