the new middle management: intermediary organizations in

Educational Evaluation and Policy AnalysisSpring 2004, Vol. 26, No. 1, pp. 65-87

The New Middle Management: Intermediary Organizations inEducation Policy Implementation

Meredith I. HonigUniversity of Maryland, College Park

Intermediary organizations have become increasingly prominent participants in education policy im-plementation despite limited knowledge about their distinctivefunctions and the conditions that constrainand enable those functions. This article addresses that research-practice gap by drawing on theories oforganizational ecology andfindings from a comparative case study offoiar intermediary organizationsthat helped with collaborative policy implementation in Oakland, California. I define intermediaries asorganizations that operate between policymakers and implementers to affect changes in roles andpractices for both parties and show that such organizations typically vary along at least five dimen-sions. Oakland's intermediary organizations all provided new implementation resources-knowledge,political/social ties, and an administrative infrastructure butfaced different constraining and enablingconditions. Using insights from this strategic case study, this article begins to build theory about inter-mediarv organizations as important participants in contemporarv policy implementation.

Keywords: capacity, district central office, implementation, intermediary organization, Oakland,school-community partnerships

SCHOOL DISTRICT central office administrators,school principals, and other education leadersface contemporary policy demands that exceedtheir traditional capacity for action and, increas-ingly, they call on "intermediary organizations"to help with implementation. For example, pro-fessional development organizations coach class-room teachers in meeting ambitious content andperformance standards and assist central officeadministrators in providing such coaching them-selves (Stein & Brown, 1997; Wechsler &Friedrich, 1997). "Design teams" help schoolsimplement comprehensive school reform mod-els and support central offices in developingcurriculum and other implementation supports(Bodilly, 1998).

Despite their growing number, research and ex-perience teach little about intermediary organiza-tions. These organizations have appeared in thebackground of implementation studies rather thanas the main research focus. Studies and policydocuments occasionally refer to individual inter-mediary organizations without explaining whatmakes an organization an intermediary. Likewise,various organizations self-identify as intermediaryorganizations without qualification and often addthe designation "intermediary" to a repertoire ofother identifiers such as professional developmentorganization, school coach, technical assistanceprovider, and contractor. These trends make it dif-ficult to discern what intermediary organizationsare, what they do, and how they operate. Accord-

The author would like to thank the following people for their helpful comments on previous iterations of this article: Cynthia Cobum,Betty Malen, Morva McDonald, Milbrey McLaughlin, and three anonymous reviewers. Versions of this research were presented atthe Annual Meetings of the Association for Public Policy Analysis and Management (November 4, 2000) and the American Educa-tional Research Association (April 14. 2001) in Seattle, Washington and at the Public Education Network's 13th Annual Conference(November 2003).

65

Honig

ingly, research, policy, and practice provide weakguides for what may be productive and appropri-ate roles for this increasingly prominent partici-pant in education policy implementation.

This article aims to address this knowledge gapby answering three questions: (a) What defines in-termediary organizations as a distinct organiza-tional population? (b) What functions do inter-mediary organizations serve in education policyimplementation? (c) What conditions constrain/enable intermediary organizations in carrying outtheir functions?

Using a brief review of the literature on inter-mediary organizations and organizational ecol-ogy, I define intermediaries as organizations thatoperate between policymakers and policy imple-menters to enable changes in roles and practicesfor both parties. I highlight that organizationsthat fit this definition vary by at least five dimen-sions: the levels of government (or types of orga-nizations) between which they mediate, theirmembership, their geographic location, the scopeof their work, and their funding/revenue sources.I then present findings from a comparative, qual-itative case study of four intermediary organi-zations in Oakland, California that were estab-lished to operate between district central officeadministrators (the policymakers) and leaders ofschool-community partnership sites (the imple-menters). I show that Oakland's intermediaryorganizations primarily provided new resources-knowledge, political/social ties, and an adminis-trative infrastructure-necessary for implemen-tation but traditionally unavailable from schooldistrict central offices or school-community part-nerships and that they faced different constrainingand enabling conditions in carrying out these func-tions. I conclude by drawing on the particulars ofthe Oakland case to suggest initial parameters forbuilding theory about intermediary organizationsin education policy implementation.

Intermediary Organizationsas a Distinct Organizational Population

Many have demonstrated that over at least thepast 15 years district central offices and schoolshave faced demands that far exceed their tradi-tional capacity for action. Such demands includepressures to create systems of standards and as-sessments (Smith & O'Day, 1990), to supportclassroom teachers in delivering complex formsof instruction (Spillane, 1996, 1998), to forge

66

school-community partnerships (Honig & Jehl,2000), to engage citizens in reform (Marsh, 2000;Public Education Network, 2001), to implementwhole school reform strategies (Bodilly, 1998;Datnow & Stringfield, 2000), to manage high-stakes testing (Finnigan & O'Day, 2003), to de-velop pools of highly qualified teachers, to decen-tralize authority (Malen, Ogawa, & Kranz, 1990),and to weather state and mayoral takeover (Kirst& Bulkley, 2000). Perhaps not surprisingly, then,district central office administrators and schoolleaders increasingly have relied on so-called in-termediary organizations to help with variousaspects of implementation.

Policy and advocacy groups have contributedmuch of the documentation of organizations thatcall themselves intermediaries in social policyarenas including education (e.g., Camino, 1998;Kronley & Handley, 2003; Wynn, 2000). How-ever, such documentation typically does not pro-vide a clear definition of intermediary organi-zations and takes for granted that organizationsthat self-identify as intermediaries fit this desig-nation. Further muddying the definition, manyself-identified intermediary organizations add thelabel, intermediary, on to a broad repertoire of otherfunctions and identities. As a result, "intermediaryorganization" has come to refer to a motley collec-tion of organizations otherwise known as techni-cal assistance providers, vendors (Bodilly, 1998),collaboratives (Bodilly, 2001), capacity builders(Camino, 1998; Rodriguez & Pereira, 2000), com-munity development coaches (Urban StrategiesCouncil, 1996), universities (Harkavy & Puckett,1991), resource and referral organizations (Pavetti,Derr, Anderson, Trippe, & Paschal, 2000), exter-nal support providers (Cobum, 2002; Finnigan &O'Day, 2003), professional development organiza-tions (Burch, 2002; Wechsler & Friedrich, 1997),reform support organizations (Kronley & Handley,2003), design teams (Bodilly, 1998), PublicEducation Funds (Public Education Network,2001), regional reform organizations (McDonald,McLaughlin, & Corcoran, 2000), and corporatelaw firms (Suchman, 1995). Organizational cat-egories are not always mutually exclusive. How-ever, the assignment of such varied organizationsto one group in the absence of an analytical defini-tion blurs what may be distinctly intermediaryabout intermediary organizations. Particularly inlight of their proliferation, an examination of inter-mediary organizations seems essential for helping

The New Middle Management: Intermediary Organizations in Education Policy Implementation

policymakers and others to decide whether andhow to involve them in implementation and to holdthem accountable for public funds.

Concepts from theories of organizational ecol-ogy help bring some order to this messy terrain byproviding criteria for distinguishing among orga-nizational populations. These criteria highlightthat a definition should identify intermediary or-ganizations' particular functions and the condi-tions that may enable or constrain them in per-forming their functions. To elaborate, theories oforganizational ecology posit that organizationalpopulations are collections of organizations thatare "alike in some respect" but that do not neces-sarily interact with one another (Scott, 1992, 127).What it means for organizations to be alike hasbeen the subject of some debate but scholars typ-ically agree on two dimensions of likeness.

First organizations in a population have a sim-ilar "technical core"-a similar set of activitiesand competencies essential to an organization'sproduction and survival (Hannan & Freeman,1977; McKelvey, 1982). In other words, organi-zations in the same population perform compara-ble functions essential to their nature and theirbasic maintenance. Accordingly, some define or-ganizational populations as occupying a distinctenvironmental niche (Hannan & Freeman, 1989).This definition means that if an organization other-wise known as a professional development orga-nization also claims to operate as an intermediaryorganization, then observers should be able todistinguish its intermediary functions from itsprofessional development functions. From aneconomic standpoint, the designation, intermedi-ary organization, indicates an added value beyondwhat an organization could or would otherwiseprovide.

Second, the extent to which population mem-bers are able to perform their essential functionsdepends on a distinct set of factors such as struc-tures, capacity, beliefs, meanings, and relation-ships specific to that population (Friedland &Alford, 1991; March & Olsen, 1989). This com-mon dependence means that members rely on sim-ilar resources for support (Scott, 1992) and are vul-nerable to more or less the same external pressures(some of which can result in organizational death)(Hannan & Freeman, 1989). In policy implemen-tation terms, members of an organizational popu-lation face similar constraining and enabling con-ditions when it comes to their ability to performtheir essential functions.

In contrast to the studies and documentation re-ports cited above, Berger and Neuhaus' researchin the 1970s provides an important guide for aninitial definition of intermediary organizations thatfits the parameters outlined by population ecol-ogy. Berger and Neuhaus wrote about "mediatingstructures" which they defined as "those institu-tions standing between the individual and hisprivate life and the large institutions of publiclife" (Berger & Neuhaus, 1977, 2). They foundthat these structures-families, neighborhoods,churches, and voluntary associations-performeddistinct functions: they increased the social con-nectedness of individuals and service delivery sys-tems within the then growing welfare sector andhelped both individuals and systems to adaptand change. Using Berger and Neuhaus' researchand organizational ecology as guides, I defineintermediary organizations as distinct from otherorganizational populations as follows: Intermedi-aries are organizations that occupy the space inbetween at least two other parties. Intermediary

organizations primarilyfunction to mediate or to

manage change in both those parties. Intermedi-

ary organizations operate independently of thesetwo parties and provide distinct value beyond

what the parties alone would be able to develop orto amass by themselves. At the same time, inter-

mediary organizations depend on those parties toperform their essentialfunctions.

This definition has enough analytical power toexclude many self-identified organizations fromthe category, intermediary organizations. Forexample, an organization that has a contract witha district central office to train schoolteachersbut no role related to central office change mightmore appropriately be called a technical assis-tance provider, school coach, vendor or all of theabove. Such an organization does not "inter-mediate" between schools and their central officeto affect change for both parties. Organizationaltypes that work across the country such as someNew American Schools design teams and LocalEducation Funds may function as intermediaryorganizations in some locations but not others.Likewise, the federal government long has reliedon third-party organizations such as commissionsto bring new resources to bear on policy problemsand to help depoliticize certain decisions (e.g.,Salamon, 1981). These relatively familiar andwell-documented organizations would not qual-ify unilaterally as intermediaries because many

67

Honig

have functioned expressly as surrogates for fed-eral agencies in leveraging reform of other orga-nizations rather than as levers for governmentalreform as well. This definition does not suggestthat intermediary organizations cannot performother functions (i.e., that a professional develop-ment organization cannot also be an intermedi-ary organization), but rather offers specific crite-ria to inform the application of the intermediarylabel.

I used this definition to revisit my review ofdocumentation reports purportedly related tointermediary organizations such as those citedabove. I found that organizations that fit this de-scription varying along at least five dimensionsthat provide an initial typology of this organiza-tional population. First, intermediaries operatebetween various levels of government. For ex-ample, an organization might mediate betweenstate and local governments or between schooldistricts and schools. Interestingly, most of the in-termediary organizations identified in my reviewwork between school district central offices andschools. Second, the composition of intermediaryorganizations varies. Some intermediary organi-zations, such as Public Education Funds, typi-cally consist of dedicated staff that do not holdpositions in other organizations while many inter-mediary organizations otherwise called collabora-tives or partnerships derive their membership fromstaff of other organizations. Third, location varies.Some intermediaries such as the Annenberg-funded Bay Area School Reform Collaborativeand the New American Schools school reform de-sign teams tend to be housed physically outsidethe districts within which they work. I call theseorganizations "external intermediaries." Other in-termediary organizations such as many collabo-ratives are based within the geographic areas inwhich they work and accordingly I call them "in-ternal intermediary organizations." Fourth, relatedto the third dimension, the scope of intermedi-aries' work ranges. Some intermediary organiza-tions work with large numbers of policymakersand implementers across multiple jurisdictionswhile others restrict their work within single ju-risdictions. Fifth, intermediary organizations dif-fer in terms of their funding sources. Some arefunded exclusively by public or private sources,which, arguably makes them quasi public or pri-vate organizations, while others receive fundingfrom both public and private sources.

68

While various literatures help elaborate a defi-nition and an initial typology of intermediary or-ganizations the following questions remain: Whatfunctions do intermediary organizations serve incontemporary education policy implementation?What conditions constrain and enable intermedi-ary organizations in carrying out their functions?

Methods

To address these questions, I used a qualitativecase study design because it provides opportuni-ties to describe, define, and analyze little under-stood phenomena such as intermediary organiza-tions overtime (Merton, 1987; Yin, 1989). Basedon the theory-based definition presented above Ideveloped a purposive sample of four interme-diary organizations. As summarized in Table 1,these intermediary organizations operated be-tween the school district central office and school-community partnership sites-formal collabora-tions between schools and other youth-servingorganizations typically in schools' neighborhoods.They were internal, membership organizations,the scope of their work focused on policy imple-mentation within a single school district, and theyreceived public and private funding for their oper-ations. This internal focus allowed me to compareintermediary organizations while holding variouscontextual factors constant. Although not general-izable to all intermediary organizations in all con-texts, findings from such a design may generateinsights and hypotheses to guide theory develop-ment (Hartley, 1994; Merton, 1987) and revealpatterns with such little deviation on a small scalethat they could reasonably represent populations(Phillips & Burbules, 2000).

Oakland's Intermediary Organizationsas a Strategic Case

These four intermediary organizations in Oak-land Unified School District, a midsized urban dis-trict in northern California, provided a strategiccase for building theory about intermediary orga-nizations. First, early data collection suggestedthese organizations fit my theoretical definition ofan intermediary organization. Intermediary mem-bers reported that each organization operated in be-tween policymakers and implementers to leveragechange in both those parties and depended on thosetwo parties to define their essential functions. Allthe intermediary organizations had multi-yearfunding and, in three cases, several revenue streams

0 C) C), .

0 00~~~~~~~~0 -.0.

Ea E! Ct o .

C u CO i CO COC

0-10 C,

-- 0~~~~~~~C) CO C) C)~~~~~~~~~~~~~~~~~~~~~~~~~~~~-

C) - -

CO~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~'

F- O U Cd)FC

69

which suggested that they might operate at leastsomewhat independently. Second, these intermedi-ary organizations formed anew specifically to helpwith the implementation of four policy initiatives,which indicated they might reveal distinctly inter-mediary functions as opposed to organizations thatperformed multiple other functions. Third, becausethese organizations were internal, they promisedopportunities for me to observe their frequent in-teractions with policymakers and implementersand to develop a rich descriptive dataset importantto early stages of theory development.

Fourth, variations in intermediary member-ship offered potentially significant dimensions forinter-intermediary comparisons. For example, twoof the organizations consisted almost entirely ofschool-community site directors and central officeadministrators-the two parties among which theintermediary organizations mediated. One organi-zation primarily included directors of the schooldistrict, city, and county. Citizens representing theMayor's Office and the city council comprised thefourth organization.

Importantly, Oakland's intermediary organiza-tions participated in the implementation of a par-ticular type of education policy-what I call "col-laborative education policies" because they aimedto help schools to establish collaborative part-nerships with other neighborhood-based youth-serving agencies (e.g., human services agenciesand Boys and Girls Clubs) as a primary school im-provement strategy. Collaborative education poli-cies traditionally pose significant implementationchallenges for school systems. I Accordingly, theseinitiatives heightened the likelihood that schoolsand the district central office would look for im-plementation assistance and throw into relief in-termediary organizations' distinctive functions incomplex policy environments. Table 2 summa-rizes Healthy Start, the Oakland Child Health andSafety Initiative, the Oakland Fund for Childrenand Youth, and the Village Center Initiative. Theseinitiatives vary along several dimensions elabo-rated elsewhere (Honig, 2001, 2003). What is im-portant for this inquiry is that all four as designedpromoted school-community partnerships, aimedto reform schools and the district central office toenable implementation, and left the nature of thosereforms up to school and central office discre-tion. Initial funding sources varied but all aimedto leverage ongoing public and private investmentsin implementation.

70

Data Collection

Given the early stage of research on intermedi-ary organizations, I did not seek to evaluate thesuccess or failure of Oakland's intermediary or-ganizations but to better define the nature andscope of their work with the hope that such re-search would help inform future evaluations. Ac-cordingly, my data collection focused on interme-diary organizations' functions and the conditionsthat impacted those functions; these processesserved as my main outcome measures.

I drew primarily on direct, sustained observa-tions and record data from intermediary meetingsand in-depth interviews with intermediary partic-ipants (Barley, 1990). Between November 1998and March 2000, I spent over 100 hours observ-ing formal intermediary meetings and collectedofficial minutes from meetings that predated thestart of my data collection or that I otherwise didnot attend. Altogether I accounted for almost 200hours of intermediary meeting time. I also ob-served 55 hours of city council, school board, andMayor's Education Commission meetings thatconcerned educational improvement. During allobservations I typed almost verbatim transcriptsof conversations.

I conducted semi-structured interviews with in-dividuals affiliated with the intermediary organi-zations including past and present members andstaff. Questions focused on respondents' day-to-day work, their capacity to participate in collabo-rative education policy implementation, and inter-mediary operations. I selected respondents whowere most involved in intermediary operationsbased in part on their attendance at intermediarymeetings. I interviewed 24 individual members atleast once. To better understand the context inwhich intermediary organizations operated, I alsointerviewed nine individuals who did not have for-mal affiliations with intermediary organizationsbut who otherwise affected collaborative educa-tion policy implementation including school dis-trict central office administrators and county andcity agency staff. I selected this group using asnowball sampling technique based on intermedi-ary organization members' reports of individualswho had an impact on intermediary operations. Inall, I conducted 42 interviews. Each lasted be-tween 60 and 150 minutes. Most were audio tapedand transcribed.

I augmented the observations and interviewswith document reviews and informal conversa-

TABLE 2Oakland's Collaborative Education Policies

Oakland's Participation1990-2000Policy Description

Healthy StartInitiated in 1992 and administered by the California Department of Education, thisprogram provided $50,000 planning grants and $400,000 three-year operational grantsto school-community partnerships. Grantees created and implemented their own plansto strengthen youth development and learning through school-community collabora-tion. The state policy design emphasized school-linked case management services.

Oakland Fund for Children & YouthInitiated in 1996 by a successful popular ballot initiative to amend the Oakland CityCharter, this policy required that Oakland set aside 2.5 percent of its city generalfund for programs to strengthen youth development and learning. The estimateddollar value of the Fund at inception was $5.7 million per year for 12 years.

Grantees (various youth serving agencies including schools) created and implementedtheir own improvement plans. School-community collaboration was strongly encour-aged but not required.

Village Center InitiativeInitiated in 1998 with a multi-million dollar grant from the DeWitt-Wallace ReadersDigest Fund Beacons Dissemination initiative, this program funded middle school-community partnerships in Oakland to create and implement their own plans tostrengthen youth development and learning through school-comrnunity collabora-tion. The initiative design emphasized school-linked youth development programsand transforming schools into community centers.

16 schools or 18% ofOakland's publicschools

24 schools citywide or26% of Oakland's publicschools*

15 centers (proposed)including all Oaklandmiddle schools

Oakland Child Health & Safety InitiativeInitiated with a multi-million dollar grant from the Robert Wood Johnson Founda-tion's Urban Health Initiative, Oakland's participation in this initiative between 15 centers (proposed)1998-2000 involved "systems change" at school district, city, and county levels to including all Oaklandsupport Village Center implementation. (See Village Center Initiative above.) Policy middle schoolsdesign emphasized reform of the district central office and other public systems.

Note. Measures of school involvement in theOakland Fund forChildren and Youth were only available for2001-2002. Staffes-timate that these figures have remained fairly consistent since 1997. These counts indicate programs provided on a school site notother school-linkages (e.g., recruiting from schools, collaborations with schools not involving the provision of programs in aschool facility) and therefore comprise conservative estimates of school involvement.

tions. Documents included written reports andelectronic copies of articles from local newspa-per archives (reviewed between 1980 and 2000)that related to intermediary organizations andOakland education more broadly. I used thesedocuments to better understand intermediary or-ganizations' functions and circumstances overtime and the use of documentation in intermedi-ary operations. I spent countless hours in infor-mal conversations with intermediary members inbetween meetings about Oakland schools and in-termediary organizations.

Coding and Data Analysis

I used NUD*IST software to code the interviewand observation transcripts and electronic record

data. Data coding was largely inductive and itera-tive since theory provided guides for identifyingintermediary organizations (i.e., sampling) but notintermediary functions or constraining/enablingconditions (Phillips & Burbules, 2000). 1 codeddata in three phases. First, as part of a broaderstudy of collaborative policy implementation, Iused concepts from organizational learning theoryto define demands these policies placed on districtcentral offices and school-community partner-ships. In this phase, I separated data related specif-ically to intermediary organizations. (For a de-tailed description of this first phase, please seeHonig, 2003.) Second, I coded the data regardingintermediary organizations using broad codes thatrequired little interpretation including "reported

71

Honig

and observed intermediary functions" and "en-abling and constraining conditions" related tothose functions. Within the dataset on functions, Ideveloped codes related to functions using con-stant comparative methods to distill patterns (Miles& Huberman, 1994). I excluded the functions thatdid not appear over the course of at least three ofthe four intermediary organizations' operations.Third, I returned to the data related to conditions. Iidentified inductively and through constant com-parison a set of broad categories related to thesefactors and main sources of their variation.

The design of this case study did not allow meto conclude how implementation would haveproceeded in Oakland without intermediary or-ganizations. To sharpen my view of intermediaryorganizations' distinctive contributions, I com-pared data from 1990 through 1995, when onlyone of the four intermediary organizations oper-ated with data from 1996 through 2000 whenOakland boasted four active intermediary orga-nizations with their own budgets and politicalsupport. Also, in interviews I asked respondentsto describe implementation opportunities and ob-stacles, to identify all their possible implementa-tion resources, and to discuss when and why theycalled on intermediary organizations for assis-tance. These resource maps helped me pinpointhow respondents viewed intermediary organiza-tions' added value.

Oakland's experience with intermediary orga-nizations generated a dizzying array of organiza-tional and professional names. To simplify mypresentation of data and to reinforce salient as-pects of my findings, I use the following terms inmy report of findings.

* Central ofce administrators: Staff of theschool district central office including: (a) Front-line central office administrators-Central officestaff with primary responsibility for supportingschool-community partnerships; these adminis-trators had such titles as "Director of Village Cen-ters." (b) Senior central office administrators: Di-rectors of the district central office with primaryresponsibility for the overall policy of the cen-tral office such as assistant superintendents andsuperintendents.

* Sites: Partnerships between schools and otheryouth-serving organizations designed locally bythe participating school-level organizations thatreceived implementation funding or other supportfrom one of the four focal policies.

72

* Site directors: School principals, non-profitagency directors, and others formally responsi-ble for developing and implementing individualschool-community partnerships.

* Intermediary conveners: Employees of non-profit organizations that managed the intermedi-ary organizations. One citywide non-profit orga-nization convened the Staff Groups as part of itsbroader operations. Another citywide non-profitorganization staffed the Directors and CitizensGroup as its sole function.

To highlight membership differences thatproved consequential in this case, I refer to the in-termediary organizations by the following names,also indicated in Table 1:

* Staff Groups: Two intermediary organiza-tions whose central office membership came pri-marily from the central office frontlines.

* Directors Group: The intermediary com-prised mainly of agency directors including seniorcentral office administrators.

* Citizens Group: The organization with citizenrepresentatives appointed by city council mem-bers and the mayor and managed by the DirectorsGroup's staff.

Intermediary Organizations' Functions

Data related to the early to mid 1990s indicatedthat Oakland's four intermediary organizationshad been established mainly to provide collectiveaccountability for implementation funds and to in-crease local support for implementation and thatthey did serve these intended purposes. For exam-ple, the state department of education required thateach school district participating in Healthy Startcreate an intermediary organization-what thestate called a district-level collaborative-to dis-burse public funds for Healthy Start implementa-tion.2 Accordingly, Oakland formed an organiza-tion that I call a Staff Group by drawing membersfrom the district central office and sites as well asthe county offices of education, public health, andsocial services and a county-level collaborationfor children. This group met monthly for most ofthe 1990s and, by late 1999, had distributed ap-proximately half a million dollars to support im-plementation of school-community partnerships.

The other Staff Group formed out of an agree-ment between Oakland leaders and a national phil-anthropic foundation to award grants to school-community partnerships through the Village


Center Initiative (called Beacon Schools nation-ally). Frontline central office administrators andsite directors attended this intermediary organiza-tion's weekly meetings regularly between Sep-tember 1998 and early 2000 during which time theStaff Group distributed approximately $1 millionin grant funds for site implementation. The Direc-tors Group provided what many senior centraloffice administrators called a "first opportunity"or "catalyst" for county, city, and school agencyleaders and elected officials to engage in cross-agency planning for youth development and learn-ing. Documents and conversations with the fun-der, a national philanthropic foundation, indicatedthat such boards were meant to symbolize andto realize high-level political support for "sys-tems change" to support positive youth outcomes.Oakland's implementation plan involved the redi-rection of core county, city, and school districtcentral office resources for Village Center im-plementation districtwide. The Citizen's Groupcame together in 1998 under an amendment to theOakland City Charter mandating that Oaklandcreate a new broad-based organization to disbursean annual 2.5% set-aside of the city general funddedicated to youth development and learning.

While these activities were important aspectsof intermediary organizations' operations, Oak-land's intermediary organizations primarily func-tioned to provide resources-knowledge of sitesand policy systems, social/political ties to sites andpolicy systems, and an administrative infrastruc-ture -necessary for implementation of collabora-tive education policy but traditionally unavailablein the district central office or sites. The absenceof these resources helped define the primary func-tions of Oakland's intermediary organizations.Accordingly, in each the following sub-sectionsI make three points: first, I explain briefly thatcollaborative education policies demanded thatthe district central office and sites have accessto particular resources3 ; second, I highlight thatthese resources are typically in short supply anddifficult to develop in district central offices andsites; third, I demonstrate that all of Oakland's in-termediary organizations to some degree providedthese resources.

Knowledge of Sites and Policy Systems

Collaborative education policy designs call forsites to develop goals and strategies appropriateto their local needs and resources and for district

central office administrators to enable sites tomake and implement their own decisions; sites'decisions and experiences are to serve as keyguides for central office resource allocation andrules as opposed to mainly more traditional policydesigns in which central office decisions direct siteimplementation (Honig, 2003). Accordingly, sitedirectors and central office administrators mustdevelop site knowledge-familiarity with sites'goals, strategies, and experiences-and systemsknowledge- an understanding of central office re-source allocation processes and other proceduresthat might be aligned to sites' decisions.

One respondent explained that without siteknowledge, central office administrators andothers imposed their own priorities on sites andotherwise impeded implementation:

Unless there is some ownership and unless some-one is [sites are] telling us what they need, thenits bureaucrats-and I include non-profit pro-viders in that-sustaining their employment. It's[central office administrators what they] think forwhatever reason-some maybe good reasonssome maybe not -that people need.

Regarding systems knowledge, one centraloffice administrator emphasized that his abilityto support site implementation depended on hisknowledge of how central office resources mightbe directed to support sites' decisions:

Quite frankly, you need to know all that stuff....What are the ... [school board's] administrativebulletins that are in place? You need to know that... What is our policy around sharing informa-tion? . .. Contracting, things with vendors, con-sultants, the nature of things. You need to knowthat it is out there and kind of know where to go... Who you need to talk to.

Another central office administrator confirmed,"That individual [participant in implementa-tion] . . . really needs to be hooked into someonewhether it is a mentor or key supervisor but some-one .. . that has a firm understanding of how ourdistrict [central office] operates [to re-direct centraloffice resources]."

Central office administrators and site direc-tors, in reflecting on implementation in the early1990s, typically reported that site and systemsknowledge were in short supply. According toone long time central office administrator:

All we had to do was a little basic arithmetic tofind out . .. how many [staffl can they outstation

73

Honig

if this thing grows to 10 schools.... I don'tthink that kind of basic work [documentationof sites' goals and outcomes] was ever reallyaccomplished to prove the case.

Central office administrators reported that theirlimited time and sites' under-developed goals andstrategies impeded knowledge development. Re-garding the latter, several central office adminis-trators acknowledged that early in implementationthey had funded sites to begin implementationbased on their students' high needs for additionalservices not sites' "readiness" or their developmentof goals and strategies and that consequently, siteknowledge was essentially unavailable (OaklandSchool Linked Services Work Group, 1999).

In part to address site knowledge gaps, seniorcentral office administrators reported that theyhired new staff into frontline central office posi-tions to serve as site liaisons. Several frontlinecentral office administrators indicated that theybelieved they had more site knowledge than typ-ical central office administrators and that in theirfrontline positions they had greater access to sitesthan if they were "higher up" in the central officehierarchy. According to one:

The thinking was you can't just have a district[central office] person ... because they generalgenerally won't have that other [site] experi-ence. That is not the culture. [When they hiredme], they wanted someone that came, that hadexperience in working with [sites] in general,had an understanding about ... collaborationsand working with diverse groups. They felt itbetter to go on the outside [of the central office].

However, these first-time central office adminis-trators had little familiarity with central officepolicies or how to change them in ways thatmight enable implementation. One intermediaryconvener captured the downsides of hiring cen-tral office staff with limited systems knowledge:

If you build a house you get an electrician to doelectrical work. You don't expect the electricianto know how to do the plumbing.... In the sameway, you don't expect the service provider[frontline central office administrator or site di-rector] to be able to articulate the stuff in policyterms.

Adding to these challenges, most central officeadministrators, site directors, and intermediaryconveners interviewed described collaborativepolicy implementation as "complicated," "diffi-

74

cult" or "ambiguous" and maintained that moresite and systems knowledge did not necessarilymean that they understood how to use it to advanceimplementation. One long-standing site directorsuggested that site directors themselves lacked ex-perience communicating their plans and experi-ences to central office administrators "produc-tively" and that miscommunications occasionallycontributed to the challenge.

Oakland's intermediary organizations strength-ened site directors' and central office administra-tors' site and systems knowledge by conveningregular meetings, documenting and disseminatinginformation, simplifying information about siteexperience, and establishing knowledge buildingas an important, ongoing process.

Regular meetings

Each intermediary organization convened meet-ings that helped central office administrators andsite representatives build such knowledge. Centraloffice administrators and site directors partici-pated directly on the Staff Groups where site im-plementation reports consumed the lion's share ofStaff Group meeting agendas. Members of the Di-rectors and Citizens Groups learned about sitesthrough their full-time staff who had direct contactwith sites in part through their participation on theStaff Groups and who reported on site implemen-tation process at their regular meetings. How tosupport site implementation appeared as the sec-ond most frequently discussed topic at both theDirectors and Citizens Group meetings, accord-ing to observations and official meeting minutes.4

Some central office administrators commented andobservations confirmed the importance of inter-mediary organizations' meetings to their siteknowledge. "I'll find out about [sites' budgets] onThursday [at a Staff Group meeting]," one cen-tral office administrator explained when askedhow she kept abreast of sites' progress with fundraising. Another central office administratorconfirmed:

I used to be able to go to sites on a regular basis.I was there all the time. Now the [Staff Group]has become important for that. My time haschanged. It helps me to know that on the weeksI can't get there [to sites] that I will still connect.

These meetings also helped site directors andcentral office administrators to access systemsknowledge. For example, one Staff Group con-

vened central office business managers early inimplementation to inform sites and frontline cen-tral office administrators about rules regardingschool facilities and to teach business managersabout potential conflicts between sites' implemen-tation plans and central office policy. As one StaffGroup convener explained:

We [one of my staff people and I] have both beenin school reform work for some time and wereanticipating that there needed to be a way for peo-ple to trouble shoot the issues and not get stalleddown in the bureaucracy. So we started havingconversations [with business managers].... Wewere anticipating issues related to access tofacilities. We brainstormed everything from hav-ing a religious group that wants to meet on a reg-ular basis in a classroom at a school. What wouldit mean to have a utility pay station set up in theschool? We were just trying to think of every-thing that would pose ... issues that schools gethung up on.

Intermediary members who knew about thesemeetings unanimously attributed early implemen-tation successes to the "business managers meet-ings." Several site directors reported that beforethere were intermediary meetings they couldask questions about central office policy eitherby phoning central office administrators or by at-tending school board meetings. However, phonecalls were not always returned and school boardmeetings typically provided them only three min-utes during periods of public comment to shareinformation and no dedicated time for boardmembers' responses.

Intermediary organizations' meetings were wellattended. One of the Staff Groups met for at leasttwo hours each week and observations and offi-cial meeting minutes confirmed that rarely weresites or the central office not represented. TheCitizens Group met semi-monthly for at least twohours. The other Staff Group and the DirectorsGroup met for at least two hours each month.Such high attendance rates appear striking giventhat intermediary organizations met relatively fre-quently and that frontline central office adminis-trators in particular faced significant competingdemands on their time. When asked in interviewswhy they regularly attended meetings, central of-fice administrators and site directors reported thatmissed meetings meant missed opportunities tohelp chart the course of collaborative policyimplementation.

Documentation and dissemination

of information

Each intermediary organization also increasedcentral office administrators' and site directors'knowledge by operating as a primary documenter,disseminator, and consumer of information aboutimplementation at site and systems levels. Inter-mediary organizations' various written records ofimplementation included formal implementationreports, case studies, and meeting minutes, whichwere not only available but, only proactively used.One Staff Group hired a former central office ad-ministrator to write a ten-year history of Oakland'sexperience with school-community partnershipsthat they later used to brief an incoming centraloffice superintendent among others about theirprogress with implementation (Oakland SchoolLinked Services Work Group, 1999). The Direc-tors Group kept detailed minutes of meetings thattheir staff frequently reviewed at the start of sub-sequent meetings to remind themselves and otherswhere they had been and where they were headed.

Simplified information about experience

Intermediary organizations strengthened cen-tral office administrators' and site directors' site-and systems-knowledge by helping them managethe sheer amount of complex information avail-able about site-and systems-level implementation.In particular, intermediary organizations "simpli-fied experience"-they translated sites' plans andexperiences into discrete terms that central officeadministrators and sites more easily understood(Hatch, 1997; March, 1994). As one participant inthree intermediary organizations explained:

Partnership is easier around specific shared goalswithout having to be in sync around everythingelse. Like if we know we want the district[central office] to be more community-friendlywe [the intermediary organization] can workthrough that together with the district and wecan move the district on being more open tocommunity involvement and other broad themesof the initiative.... Being able to work throughthose negotiation issues on a small level aroundfour [sites] is doable. You know, forcing the dis-trict to have to be more community friendly ingeneral-I mean how would you do that?

Simplifications of experience provided small, con-crete steps that central office administrators andsites directors had the capacity to take immediately

75

and successfully-what Weick called "smallwins" (Weick, 1984). As one intermediary mem-ber from a county office described:

And if you do A-if you can pin point the oneor two things we can do differently immedi-ately, then we can build the broader things. Ifpeople don't believe anything can change, youdon't spend a year building a theory of change.There is some low hanging fruit.... We [poli-cymakers] start making some incremental deci-sions. We start sharing data. Then gradually wewill be doing other things that look more likereal systems reform.

Ongoing knowledge building processes

Observations and interviews suggested that in-termediary organizations helped build centraloffice and site knowledge by creating regularopportunities for central office administrators andsite directors to revisit past decisions and to viewimplementation as a process of continuous knowl-edge building. As one central office administratorexplained, thanks in large part to the structure ofengagement with sites that one intermediary orga-nization provided, he had come to understand col-laborative policy implementation as involving on-going knowledge development:

[I learned] That you can't set something in stoneand say this is the way it is going to be forever. Itjust doesn't work like that. You have to continueto revisit the process and the guidelines that youset forth. Does this still work for us based on ourcircumstances now? Is this still real for us? Thenyou modify and change it.... Assessment has tobe ongoing. This [implementation] is a very fluidsort of operation.

Certain intermediary organizations set asidemeeting time specifically to revisit past deci-sions. The Citizens Group, for exarnple, annuallyreviewed their grant-making guidelines in a pub-lic forum. Central office administrators and sitesreported that they knew they had these opportuni-ties to lobby for changes in these guidelines andseveral pointed to specific changes in the annualcall for proposals as evidence that the revisionprocess was genuine.

Social and Political Ties to Sites andPolicy Systems

Collaborative education policies demand closecentral office-site ties to enable the kinds of knowl-edge building highlighted above. Site directors re-

76

ported that they tended to share information withindividual central office admninistrators with whomthey had developed direct relationships over timethrough past work rather than with the school dis-trict writ large. Central office administrators typi-cally pointed out that applications for fundingand sites' funding contracts provided informationabout what sites planned to accomplish but thatthe "real deal" about actual implementation deci-sions, successes, and pitfalls came through phonecalls, site visits, and other opportunities for direct,ongoing relationships with sites. Frontline cen-tral office administrators indicated that they alsoneeded systems ties-relationships within the cen-tral office-to develop their systems knowledgeand, ultimately, to marshal central office resourcesfor site implementation. According to one centraloffice administrator:

It is our relationship [with senior central officeadministrators]. That is wvhy I was able to presentarguments as to why that . . . needs to be there[why specific resources needed to be allocated tosites].... If you are not at the table, it makes itkind of hard to do that.

Long-standing site directors indicated that priorto their involvement with particular intermediaryorganizations they generally had weak ties withthe district central office. Frequent turnover ofschool principals challenged the continuity ofsuch ties (Oakland School Linked Services WorkGroup, 1999). Most central office administratorsreported that they lacked the time necessary toestablish such relationships. According to one:

Zero. Well, actually maybe five percent....That was about as deep as I was able to [spendwith sites] last week even though I needed toget deeper into it.... The thing is I am going tomeetings [regarding my other responsibilities].They might start at 7:30 [a.m.]. My workdaydoesn't start until 8:30. My first meeting is at7:30. The next one is at 9:00. The next one is at11:00. Then there is one at 1:30, one at 3:00 andthen when do I get to come in here [to the of-ficel? Maybe at 4:15 1 get to come in here andthen you have voice mail messages and it takesme 30 minutes to listen to that and then I havegot to return those phone calls and that doesn'tgive me a chance to get to the work [with sitesl.

Site directors attributed weak central office-siteties to their own mistrust of the district central of-fice after years of the central office's mismanage-


ment and broken promises (both alleged and sub-stantiated) (Coburn & Riley, 2000; Gewertz,2000; May, 2000). One central office administra-tor referred to a mid-1990s teachers strike as apoint when "we [the district central office] werestruggling" to establish strong site relationships.The strike did not impede sites' work directly, butit fed long-standing tensions about central officeperformance that carried over into collaborativepolicy implementation. Site directors explainedthat in Oakland, historically, only "squeaky wheelsget new textbook money or portables [buildings]or program money" and that at first they viewedthe intermediary organizations as new opportuni-ties to demand assistance. Site directors reportedthat they intended "to throw a hissy [fit]" at meet-ings or that implementation would "really breakapart" if the meetings ended without the centraloffice administrators allocating resources to sites.Budget crises, superintendent turnover (fourbetween 1990 and 2000), and threats of statetakeover were not infrequent in Oakland (FiscalCrisis and Management Assistance Team, 2000).Frontline central office administrators reportedtheir willingness to be responsive to sites but thatthey struggled to manage the frustrations suchcrises generated among site directors. One con-versation at a Staff Group meeting illustrates suchtensions:

Former Central Office Administrator: [arguingthat particular sites can weather a recent centraloffice decision not to provide promised fund-ing]: I think depending on your vision and whatyou are doing costs can vary greatly. I knowsome places would die for your budgets....I'm saying it varies widely....

Site Director: You have never even been tomy site. You don't even know our budgets. Youmake us crazy every time you talk like this. Idon't care what my total budget looks like toyou. If I can't make my payroll because I maynot receive the check you promised this is goingto hit the fan. We need to follow up on this. Thisis really contentious.

Intermediary organizations helped managesuch long-standing tensions between sites and thedistrict central office. Site directors in particularcounted such management as among intermedi-aries' central contributions to implementation. Asone site director highlighted:

When the district [central office] is obligated toprovide services, who makes sure that hap-

pens? The broader policy issue is how do wetake heat off sites so you don't have to spendrelationship capital to get things resolved thatshould not require negotiations. It's a role for[the intermediary].

Another site director corroborated the impor-tant role for intermediary organizations as man-agers of central office-site ties. When one inter-mediary convener proposed that the intermediaryorganization play a lesser role in implementa-tion, the site director reacted:

I don't want [the intermediary to provide] a listof people in the district [I should contact for sup-port]. What I don't want is for someone to tellme to talk to [an Assistant Superintendent]. It'sa question of are you [the intermediary] mar-shalling the process? Are you taking it on?

Sustained observations revealed that "taking iton" involved two specific strategies: translatingsites' implementation experiences into terms onwhich central office administrators could/wouldtake action and buffering sites from central officenon-responsiveness.

Translation of sites' demandsinto actionable terms

Intermediary organizations helped sites trans-late their implementation experiences into requeststo which central office administrators believedthey could respond, similar to the simplificationsof experience highlighted above. For example, di-rectors of one site attributed early implementationdifficulties to persistent safety problems beforeand after school. The site directors, seasoned com-munity organizers, had made frequent presenta-tions at school board and city council meetings tocall for radical reform of Oakland's police depart-ment, but neither the school district nor the citypolice department had responded. Frontline cen-tral office administrators almost unanimously re-ported that they were aware of the site's safetyconcerns but that they had not taken stems to ad-dress them. At a series of Staff Group meetingsduring their first years of operation, the site direc-tors and other intermediary members translatedthe site directors' original demands into a specific,de-politicized request for changed police sched-ules at a particular street intersection that theyidentified as a "tipping point" for neighborhoodsafety problems. Staff Group members (also full-time staff to the Directors Group) submitted this

77

proposal to the City Manager's Office and, subse-,quently, the Chief of Police, central office adrmin-istrators, and site directors agreed to a series ofmeetings to explore both the specific safety pro-posal and whether the sites' neighborhoods wouldprovide useful testing grounds for broader policereforms.

Buffers for sites

Intermediary organizations also helped man-age long-standing, strained relationships betweensites and the central office by buffering sites fromcentral office non-responsiveness. For example,the Staff Groups occasionally made special grantsto sites out of their own funds to help sites meettheir payroll demands and otherwise weather de-lays in the central office's processing of payments.The Citizens Group continued to fund sites, re-gardless of central agency budgets, thanks to the12-year amendment to the city charter that allo-cated a set percentage of the city general fund astheir annual budget.

Intermediary organizations also buffered sitesfrom unexpected central office policy changes byserving as a sort of advance warning team thatcommunicated central office rules and proceduresto sites, explained the rationale for existing poli-cies and policy changes, and offered suggestionsfor how to manage such changes. Staff Groupsoften deployed members to sites to provide suchupdates. For example, when the central office tem-porarily suspended promised funding to sites, oneStaff Group sent a letter to all site directors ex-plaining the situation, taking responsibility for notleveraging the promised central office funds, andmaking themselves available to sites to discuss al-ternatives. As one member of both Staff Groupsexplained when reflecting on the role of interme-diary organizations in such instances, "We need tobe the conscience of the process and the constantface in the face of change."

Administrative Infrastructure

Collaborative education policy designs call forsites to develop goals and strategies specific to theneeds and strengths of their students; in turn, cen-tral office administrator shift from distributing rel-atively uniform resources across all sites to pro-viding resources specific to individual site goalsand strategies, including the deployment of staffto work with sites individually. I call these de-mands "site-by-site support." One long time cen-

78

tral office administrator captured respondents'consensus regarding the importance of site-by-sitesupport:

I think the thing in my experience that reallymakes [site implementation] work is somebodyhaving the resources to really staff and move itforward. Actually that's probably the singlemost important thing .. . individualized supportfor it. 'Cause I think it just takes more time andenergy to work with each school . .. about whatthey need and what . .. things cost. So much re-ally depends. It's really a case-by-case, school-by-school basis.

The district central office, like other govern-ment bureaucracies, generally did not have anadministrative apparatus-budget, evaluation, orstaffing systems-for site-by-site support. A com-prehensive state audit found the districts' budget-ing and accountability systems incongruous withschool site-based decision making let alone col-laborative site-based decision making (Fiscal Cri-sis and Management Assistance Team, 2000). Theinterim superintendent observed that traditionalcentral office administrator training had left hiscentral office and school staff ill prepared to im-plement policy strategies involving site-baseddecision-making:

Because of the diversity of the community and ofth i s district and of the i ndividual people workinghere there has always been the feeling that no onesize fits all and those decisions are best left toschool site personnel including parents and com-munity members. Problem is, none of these folks[central office administrators or school princi-pals] have had training in supporting that kind ofdecision-making.

One frontline central office administrator indi-cated that barriers to site-by-site support in-cluded lack of time for central office administra-tors' to work with sites individually:

Once [implementation] ... grew to four to fiveschools .. .it got to the point where we realized,wow, now I had so much else on my plate thatwe really need to bring in somebody. So we ...hired S- . . . the same thing happened to him.Then we got L- . . . and pretty soon it just, thequicksand starts sucking you down. And that'sjust the way it is in a bureaucracy.... It's in-evitable that it will happen.

In addition, central office administrators andsite directors did not always agree about how to


define site-by-site support and these disagreementsimpeded implementation. For example, during aparticularly contentious debate at a series of StaffGroup meetings, central office administrators ar-gued that certain sites had just received significantfederal grants and, with annual operating budgetsover $1 million in some cases, did not need thesame level of assistance from the district centraloffice as sites without such grants. The grant-winning sites objected on the grounds that the cen-tral office should not penalize them for their hard-won success at attracting additional funds and thatthe size of their programs meant that their annualbudgets barely covered their costs. Both centraloffice administrators and site directors advocatedfor distribution of resources on a site-by-site basis.However, central office administrators assessedsite-by-site support by comparing sites' budgetswhereas certain site directors defined site-by-sitesupport as funding commensurate with their par-ticular expenses.

Intermediary organizations addressed thesechallenges by helping to establish resource allo-cation systems appropriate to site-by-site support,augmenting staff time available for site-by-sitesupport, and developing new standards for re-source allocation and site accountability.

Site and central office systemsfor resource allocation

Intermediary organizations promoted the use ofmemoranda of understanding (MOUs) for theallocation of resources and other supports on asite-by-site basis. Unlike some other governmentcontracts that delineated services a vendor wouldprovide in return for payment, Oakland's MOUsestablished the contributions various parties wouldmake toward achieving specific, shared imple-mentation outcomes. The three intermediary or-ganizations that disbursed funding directly to sitesused MOUs to clarify how the intermediary orga-nizations would assist sites with implementationand each site's reciprocal obligations to raise ad-ditional funds, develop a collaborative planningprocess, and improve specific student outcomes.The intermediary organizations helped centraloffice administrators craft their own MOUs withsites; this strategy enabled the central office toenter into site-by-site agreements without tacklingthe total reinvention of the central offices' bud-geting and staffing systems in the short term. Asone former central office administrator reported,

MOUs provided a basic structure for site-by-sitesupport:

MOUs really were a Healthy Start thing. Thestate [Department of Education] suggestedMOUs but the [Staff Group] really gave shape tothem. Then when we had [funds for sites] com-ing in through [the Staff Group] we used MOUsto give the money out and to establish sort ofbasic agreements about the money, because eachsite would use it differently. Then we usedMOUs for the Village Centers, for the VillageCenter funds, to give out those funds to help startVillage Centers.... Well by then MOUs werehow we did things but there was no template. Sowe worked to develop a template for [central of-fice] MOUs and now that's pretty much, prettymuch how it's done.

Staff time

Intermediary organizations increased the num-ber of staff hours dedicated to site-by-site supportthrough several strategies. First, interviews and ob-servations confirmed that the intermediary meet-ings provided regular, structured opportunities forcentral office administrators to focus on individ-ual site issues. As discussed earlier, the meetingstypically featured reports about sites' implemen-tation progress and provided other opportunities tobring information about sites' plans and experi-ences to central office administrators' attention. Inaddition, each intermediary organization hired ordedicated staff that increased the sheer number ofperson hours spent responding to sites' requestsfor support. For example, a citywide non-profit or-ganization managed the Staff Groups and helpedsites develop their collaborative partnerships,evaluate their programs, and raise funds. The Di-rectors Group hired staff experienced with site-by-site coaching who were not necessarily eligible foror interested in formal positions within govern-mental agencies. These staff provided a range ofassistance including convening sites to networkwith their peers.

Standards and accountability

Intermediary organizations also managed dis-putes about site-by-site support by developing andmaintaining clear standards and accountabilityprocesses for the provision of support and by pro-viding regular opportunities to revisit both. Theseprocesses and opportunities for revisions helpedestablish expectations that rules and resource allo-

79

cations would vary across sites and overtime, andsite directors, central office administrators, andothers could participate formally in establishingthe terms of resource allocation decisions. For ex-ample, the Citizens Group operated under a pro-vision in the Oakland City Charter to develop afive-year strategic plan and to issue applicationsfor site funding annually based on that plan (Plan-ning and Oversight Committee of the OaklandFund for Children and Youth, 1997). This struc-ture provided two different opportunities for set-ting and revisiting standards: the strategic planoutlined the broad parameters for grant makingand a schedule for revising those parameters everyfive years; the Request for Applications was re-vised annually in ways consistent with the five-year strategic plan. When a city council memberlobbied to decrease site funding one year to covercosts for a new housing program unrelated to col-laborative policy implementation, the CitizensGroup pointed to the strategic plan as the basis forrejecting this request and successfully avoided thedisplacement of funds. Intermediary organizationswithout legal mandates regarding regular reviewof rules created their own review schedules byclarifying their decision rules after the fact andestablishing explicit lines of accountability. OneStaff Group, for example, publicly took responsi-bility for its decision not to provide promisedfunding to particular sites and communicated thisdecision and its rationale widely at meetings andthrough letters to site directors.

Intermediary Organizations' Constrainingand Enabling Conditions

The section above identified functions of allfour focal intermediary organizations. In thissection I examine the extent to which each inter-mediary organization performed these functionsand the conditions that constrained or enabledthese operations. My comparative analysis re-vealed two broad sets of findings. First, con-ditions that enabled site knowledge and ties -i.e. having frontline central office administratorsas members-constrained systems knowledgeand ties; conditions that enabled systems knowl-edge and ties-specifically, having senior centraloffice administrators as members or close allies -contributed to weak site knowledge and ties. Inpractice, this meant Staff Groups generally hadsite knowledge and ties to inform central officepolicy changes but weak systems knowledge and

80

ties to marshal those changes; the Citizens andDirectors Groups had the systems knowledgeand ties to marshal policy changes but not thesite knowledge on which to base them or the tiesfor strengthening their site knowledge.5 Second,over time three intermediary organizations facedfiscal constraints that weakened their indepen-dence from the central office and other funders.Eventually, these constraints curtailed their corefunctions and prompted the three organizationsto cease intermediary operations or completelydisband.

Resource Tradeoffs:Site Versus Systems Knowledge and Ties

Staff Groups typically had stronger site knowl-edge and ties but weaker systems knowledgeand ties than either the Citizens or the DirectorsGroups. As elaborated above, Staff Groups in-cluded site directors who shared site infornationdirectly as well as frontline central office admin-istrators who had relatively more site knowledgeand ties than other policymakers. An analysis ofmeeting minutes and transcripts between 1998and 2000 revealed that site directors share signif-icantly more information about their implementa-tion plans and progress with each Staff Group thanwith the Directors and Citizens Groups combined.

Observations also uncovered several instancesin which Staff Groups members demonstrated alack of systems knowledge and ties. For example,in the late 1990s site directors requested informa-tion from one Staff Group about district central of-fice procedures for hiring adults to work on schoolcampuses. State and federal law delineated spe-cific hiring procedures in these cases. However,participating frontline central office administratorswere unfamiliar with those rules and suggestedthat the rules were ambiguous (as opposed to justunclear to them) when they asked the districts'legal staff to clarify existing procedures. Weeks ofdebate ensued among Staff Group members asthough they had discretion to develop their ownprocedures. Eventually, the legal department re-turned a "judgment" -essentially a copy of federaland state laws-that contradicted sites' wishes.By that time, Staff Group members had spentweeks in debate rather than negotiating for policychanges.

On the flipside, the Directors and CitizensGroups had precisely the systems knowledge andties that Staff Groups lacked. Members of the


Directors Group essentially were "the system" -the interim superintendent, school board members,and their counter parts from other sectors with theauthority to set their agencies' official policies.The Directors Group's mandate also included thepower to affect systems changes in support of siteimplementation. The Citizens Group memberswere handpicked to represent the city council andthe mayor. Importantly, the Directors Group'sfull-time staff also managed the Citizens Groupand reported that their main responsibilities in-cluded support for the flow of information and re-lationships between the Citizens Group and vari-ous public agencies including the district centraloffice.

However, interviews and conversations withsite directors confirmed the analysis of meetingminutes I highlighted-that the Directors and Cit-izens Groups had significantly less knowledge ofand ties to sites than the Staff Groups. Site direc-tors often reported that they believed the citizenmembers of the Citizens Group lacked the exper-tise necessary to assist them with implementa-tion and that they did not seek such support fromthem.6 Frequent comments from site directors inmeetings and interviews indicated that the Direc-tors Group members "already [had] all the moneyand the power" and that sites should be carefulabout including them in conversations about im-plementation process. This caution played out inpractice. For example, at the end of the 1990s, theinterim superintendent requested that DirectorsGroup staff work with the central office to surveymiddle schools to increase the Directors Group'sknowledge of the community partnerships inwhich each school engaged. The lead project staffreported that this labor-intensive effort mainlyyielded incomplete or unretumed surveys. Whenasked to explain this response, staff suggested thatschool principals likely did not see the value incompleting a survey associated with the DirectorsGroup because the Directors Group lacked "thepull" with schools, that school principals did notknow "what's in it for them" to work with the Di-rectors Group, and that many were "suspicious" ofthe Directors Group because they claimed to sup-port site implementation but did not provide fund-ing directly to sites.

Importantly, Oakland's participants did notseem to view these resource tradeoffs as funda-mentally problematic but as constraints to bemanaged. During frequent conversations at Staff

Group meetings and two Directors Group meet-ings, participants noted the intermediary organi-zations' various strengths and weaknesses andpursued plans to create new relationships amongthe intermediary organizations that built on thestrengths of each. For example, one plan calledfor the creation of a new intermediary infrastruc-ture in which the Staff and Citizens Groups oper-ated as committees of the Directors' Group. Inthis configuration, the Directors Group wouldmediate between "the system" and the other in-termediaries while the other intermediaries wouldmediate between sites and the Directors Group.

Fiscal Constraints andDiminishing Independence Over Time

Oakland' s intermediary organizations managedthese tradeoffs and participated centrally in im-plementation through much of the mid to late1990s. Buy by the end of that decade, each facedfiscal; constraints that jeopardized its independentstatus, and, ultimately, its functions. Specifically,at the end of the 1990s one of the Staff Groupsspent approximately six months developing a planto fund additional sites. Subsequently, the interimschool district superintendent announced thatthe school district would not contribute promisedfunds for site expansion in the short term and hetransferred Staff Group responsibility to a newfrontline central office administrator who had mul-tiple other programs and initiatives to manage.The interim superintendent indicated in an inter-view and at two intermediary meetings that a sig-nificant budget shortfall and negative state auditleft him little choice but to freeze resources inmultiple areas including collaborative policy im-plementation and that he hoped the budget freezewould be short-lived (Fiscal Crisis and Manage-ment Assistance Team, 2000). Rather than drawon their reserves or other funding sources, StaffGroup members suspended their site expansionplan. Transcripts of interviews and several StaffGroup meetings suggested that Staff Group mem-bers generally believed they could not proceedwithout the school district central office's finan-cial and political support from the school districtcentral office. The Staff Group's original fundinghad expired and the Staff Group's non-profit con-veners expressed reluctance to dedicate funding tosite expansion themselves when, according to one,their "primary systems players weren't doing theirpart." Attendance at Staff Group meetings waned

81

among site directors and central office admin-istrators and soon thereafter the Staff Groupdisbanded.

By the late 1990s, the other Staff Group had abudget of approximately $300,000 in new stateand federal revenue to allocate to sites. When thelate 1990s budget shortfall hit, the interim super-intendent requested that the Staff Group commit80% to cover the district's costs of providing man-dated health services. The interim superintendentargued the one-time commitment of funds wouldhelp alleviate some of the districts' fiscal strainsand that district health staff helped generate thatrevenue in the first place.7 After significant debateand a change in superintendent, the Staff Groupagreed to surrender the 80%. Several Staff Groupmembers commented that they believed the "fundswere lost anyway" and at least the "voluntary"contribution could cast them as supportive of thesuperintendent and thereby put them in a poten-tially stronger bargaining position in the future.Shortly thereafter, the non-profit convener of theStaff Group resigned for a post in another organi-zation, and without funding to allocate the StaffGroup suspended further meetings at least for theshort term.

Meanwhile, the Directors Group faced increas-ingly negative pressure from its national funder tocease support for collaborative policy imple-mentation. According to correspondence, publi-cations, and conversations, the foundation hadtaken a position that school-comrnunity partner-ships would not yield the kinds of demonstrableimprovements in the status of youth that the foun-dation's initiative required. Members of the Di-rectors Group spent approximately one and a halfyears rebutting the funder's concerns and debat-ing internally whether the group should proceedwith its own plans and surrender the foundationfunds. Several Directors Group members pointedout that as heads of major county and city agen-cies collectively they could construct a budget forthe intermediary organization using local funds.However, Directors Group members ultimatelydecided that the value of participating in the highprofile national initiative outweighed the limita-tions. They abandoned the school-communitypartnership strategy at least in the short term,launched a new policy initiative to provide tar-geted services to juvenile offenders, and changedtheir name to amplify that they had become a neworganization.

82

The Citizens Group ran significantly less risk ofredirection. The amendment to the city charter thatauthorized both the Citizens Group and its fund-ing operated for twelve years or longer if a sub-sequent popular ballot initiative enabled extension.However, the Citizens Group depended on the Di-rectors Group for its full time staff. As part of itsown reinvention, the Directors Group decided thatstaff support to the Citizens Group was incon-sistent with its new juvenile justice focus. Sub-sequently, a local community foundation agreedto provide staff support to the Citizens Group. Theeffects of this change on the Citizens Groups' in-termediary role could not be observed within theperiod of study. However, given the significantrole that Citizens Group staff played in helpingthis organization operate as an intermediary orga-nization rather than simply as a surrogate grantmaker for the city, this change did not bode wellfor the Citizen's Group's operation as an organi-zation with distinct, intermediary functions.

In sum, Oakland's intermediary organizationsfaced several bumps in the road during collabora-tive policy implementation. No one intermedi-ary organization had the capacity to address allthe site-central office resource gaps at equallyhigh levels-a constraint many reported as achallenge to be managed rather than a fatal ob-stacle. By contrast, fiscal constraints over timeultimately jeopardized the intermediary organi-zations' survival.

Summary, Conclusions, and Implications

This article draws on findings from a strategicresearch site to begin to build theory about inter-mediary organizations as distinct actors in edu-cation policy implementation. Without such aknowledge base, policymakers of various stripesmay misjudge their available implementationresources and overlook opportunities to enablechange. Intermediary organizations themselvesmay fail to make the best use of their distinctiveinstitutional resources.

I show that Oakland's intermediary organiza-tions operated between the district central officeand sites to leverage changes for both parties.These intermediary organizations depended oncentral office administrators and site directors/staff to define their functions and central officeand site participation helped shape the extent towhich intermediary organizations executed thosefunctions. At the same time, the intermediary or-


ganizations added distinct value to these parties-in this case, site and systems knowledge, site andsystems ties, and administrative tools otherwiseunavailable from the district central office or sites.Accordingly, the empirical findings from thisOakland case study support the theoretical defini-tion of intermediary organizations posited above:Intermediary organizations occupy the space be-

tween at least two other parties and primarily

function to mediate or manage change in both

those parties. These organizations operate inde-

pendently ofthese two parties andprovide distinct

value beyond that which the parties alone would

be able to develop or to amass themselves. These

organizations also depend on those parties to

perform their essentialfunctions.The research presented here also suggests sev-

eral additions to this basic definition that are im-portant guides for further inquiry. First, the litera-ture review at the start of this article revealed:Intermediarv organizations vary along at least

five dimensions-the parties between which they

mediate as well as their composition, location,

scope of work, andfunding sources. Second, as

the display of case study data suggests, it may bevirtually impossible to understand the value thatintermediary organizations add in policy imple-mentation without an analysis of policymakers'and implementers' implementation demands andtheir limitations in meeting them. Accordingly:Intermediary organizations' functions and their

abilities to perform those functions are context

specific-contingent on given policy demands and

policymakers' and implementers' capacity to meet

those demands themselves.The Oakland case also shows that intermediary

organizations' constraining and enabling condi-tions may vary even when they operate in thesame location and essentially focus on similar im-plementation challenges. In this study, site andsystems knowledge and ties appeared as tradeoffs.Respondents' comments and the proposed inter-mediary reorganization plan suggested that thefact that no one intermediary organization had allthe necessary resources to fill the resources gaps isnot inherently problematic. The more fundamen-tal implementation question is do intermediariesand other organizations comprise a web of rela-tionships and operate in connected and coordi-nated ways to fill gaps between policymakers andimplementers? These observations suggest an-other corollary to the working definition of inter-

mediary organization: Intermediary organizationscan augment their own capacity to carry out theircore functions as needs change and lessons arelearned in part by participating in a web of inter-mediary supports that build on the strength ofeach organization.

The Oakland case also suggests that long-termfiscal independence may befundamental to thesurvival of intermediary organizations. Recallthat Oakland's intermediary organizations did notseem to risk over dependence on the district cen-tral office or its other funders throughout most ofthe 1990s. In fact, the Staff Groups in particularhelped sites weather the somewhat regular centraloffice fiscal constraints over the course of the1990s. But by the end of the 1990s, the interme-diary organizations had begun to exhaust theirown discretionary resources for sustaining bothsites and themselves, the central office suspendedits promised support, and, ultimately, three inter-mediary organizations essentially folded.

Implicationsfor Practice and Research

This study examined a single albeit strategiccase, which limits its power for providing directlessons for practice. Nonetheless, this study raisesseveral questions for policymakers to consider. Forone, should policymakers seek to involve interme-diary organizations in implementation? The Oak-land case does not offer a prescription but rathershows that in some circumstances intermediary or-ganizations can help augment resources necessaryfor implementation. Intermediary organizations -especially intermediary organizations that drawmembers from other organizations-may be animplementation support strategy that suited Oak-land particularly well. Oakland's site directorsgenerally had the wherewithal to hold their own(and then some) in potentially intimidating inter-mediary meetings with central office administra-tors and other traditional site monitors and funders.Central office administrators were active mem-bers and representatives from Oakland's well-developed non-profit sector played key roles. Sim-ilar intermediary organizations in districts withoutsuch strong site leadership, central office partic-ipation, and non-profit experience may not fareas well as Oakland's. Accordingly, the questionother policymakers might consider is not how toestablish intermediary organizations just likeOakland's but what intermediary arrangementsmight be appropriate to their particular localcontexts.

83

Honig

Second, how can the participation of interme-diary organizations be strengthened to improve thequality of education policy implementation? Theimportance of fiscal independence suggests thatpolicymakers could build dedicated resourcesfor intermediary operations into policy design.This recommendation comes with two cautions:(a) dedicated funding can lead to unproductiveover reliance on the funder, particularly if the fun-der and intermediary organization disagreeabout intermediary functions as was the casewith the Directors Group and, by extension, theCitizens Group; (b) dedicated funding for inter-mediary operations may create intermediary con-stituencies with investments in perpetuating theintermediary organizations even after they haveoutlived their use. However, the Oakland casedemonstrates a well-worn lesson of organizationalchange: any shift-particularly an institutionalshift-requires an infusion of new resources tofacilitate reform. Policymakers might considerproviding short-term funding for intermediary or-ganizations to establish their infrastructure and de-creased funding over time as intermediary organi-zations generate other revenue or exhaust theirpurpose and disband.

Policymakers and others may strengthen indi-vidual intermediary organizations also by support-ing multiple intermediary organizations. Whenviewed together Oakland's four intermediary or-ganizations had the makings of a system of sup-port linking various sites with different levels ofthe district central office (and other policy sys-tems). Given significant gulfs between policy-makers and implementers in public school dis-tricts and other arenas, arguably it is unreasonableto expect any one intermediary organization aloneto provide a sufficient bridge.

This case also suggests avenues for future pol-icy research. For one, findings come from a singlecase of a particular type of intermediary organiza-tion. Do the propositions presented here hold forsimilar intermediary organizations in other con-texts? Do.they hold for intermediary organizationswith different features? In particular, Oakland's in-termediary organizations seemed to have specialopportunities to connect with the central office andsites in part because of their geographic proximity.External intermediary organizations may fare wellin other districts or struggle from lack of regular ac-cess to policymakers and implementers. Future re-search may illuminate the pros and cons of differ-ent intermediary organization across contexts.

84

Second, Oakland's intermediary organizationswere established anew primarily to serve inter-mediary functions. However, as discussed earlier,many self-identified intermediary organizationsadd that descriptor on to a repertoire of other tasks.Does adding on intermediary functions raise dif-ferent opportunities and obstacles than those pre-sented here?

Third, intermediary organizations operate incomplex policy arenas that include not only thepolicymakers and implementers between whomthey mediate but their various funders as well. Thetrajectory of the Directors Group suggests thatfunders can significantly curb intermediary orga-nizations' operations when funders and interme-diary members disagree about goals and strategies.An exploration of funders' roles seems essential,particularly in cases in which intermediary organi-zations operate without their own revenue streams.

Fourth, how do intermediary organizations fareover the long term? Will intermediary organiza-tions evolve, change, and possibly die over time assites and central offices enhance their own re-sources and change their practice or will sites andcentral offices become over-reliant on intermedi-ary organizations and avoid building the necessaryknowledge, ties, and administrative infrastructurethemselves? Will intermediary organizations tendto perpetuate their own existence even after theyhave outlasted their use? Will dedicated fundingover the long term help or make matters worse?These questions seem imperative given interme-diary organizations' essential role as an enabler ofchanges in other organizations.

Finally, as intermediary organizations becomemore prominent participants in education policyimplementation and future research elaboratestheir processes, researchers must attend to ques-tions of success and failure. Specifically, on whatbasis shall the success and failure of intermediaryorganizations be judged? This study providessome initial parameters and propositions that canhelp frame such an inquiry. In particular, thisstudy emphasizes that intermediary organizationsinherently are dependent on other parties for thedefinition and execution of their core functions-a dependency that will challenge researchers whoaim to develop discrete measures of intermediaryorganizations' specific added value in various con-texts. Regardless of future evaluation findings,ongoing research on intermediary organizationspromises to shed new light on the essential spaces


in between branches of formal policy systemswhere implementation unfolds.

Notes

See for example: Cibulka & Kritek (1996);Mawhinney & Smrekar (1996); Smylie & Crowson(1996); Smylie, Crowson, Chou, and Levin (1994);Marsh (2003).

2This intermediary organization oversaw funds asso-ciated with the Local Educational Agency Medi-CalBilling Option. This federal program reimbursed schooldistricts for their share of the cost of providing requiredhealth services to students. Before the Billing Option,school districts (i.e., Local Educational Agencies orLEAs) and states split the costs of these services. Underthe Billing Option, the former federal Health Care Fi-nance Administration reimbursed LEAs for their shareof the cost. The Califomia Department of Education re-quired that districts participating in this program makedecisions about allocating the new revenue through acollaborative process and that funding priority go toHealthy Start implementation.

I3 elaborate on these demands in two separate pa-pers (Honig, 2002, 2003).

4Filling member vacancies appeared as the most fre-quently discussed topic for the Citizens Group. As dis-cussed later, the most frequent topic for the DirectorsGroup concerned how to negotiate their mission withtheir national benefactor.

I Elsewhere I call this tradeoff the "paradox of theperiphery" (Honig, 2003).

6 Respondents' most frequent comment about theCitizens Group concerned members' inexperiencewith site implementation. Several respondents notedand reviews of meeting minutes and group proceduresconfirmed that city laws regarding conflicts of interestbarred individuals from serving on the Citizens Groupif they had personal or professional ties to implement-ing sites or other investments in the futures of their po-tential grant recipients.

7See note 3 for an explanation of this intermediaryorganization's revenue stream.

References

Barley, S. R. (1990). Images of imaging: Notes ondoing longitudinal field work. Organization Sci-ence, 1(3), 78-108.

Berger, P. L., & Neuhaus, R. J. (1977). To empowerpeople: The role of mediating structures in publicpolicy. Washington DC: American Enterprise Insti-tute for Public Policy Research.

Bodilly, S. J. (1998). Lessons from New AmericanSchools' Scale-Up Phase: Prospects for bringingdesigns to multiple schools. Washington, DC: TheRAND Corporation.

Bodilly, S. J. (2001, April). Philanthropic efforts atcreating instructional reform through intermedi-aries. Paper presented at the Annual Meeting ofthe American Educational Research Association,Seattle, WA.

Burch, P. E. (2002). Constraints and opportunities inchanging policy environments: Intermediary or-ganizations' response to complex district con-texts. In A. Hightower & M. Knapp & J. Marsh &M. McLaughlin (Eds.), School districts and in-structional renewal (pp. 111-126). New York,NY: Teachers College Press.

Camino, L. A. (1998). Building local infrastructurefor youth development: The added value of capacity-building intermediary organizations. Washington,DC: Center for Youth Development and PolicyResearch, Academy for Educational Development.

Cibulka, J. G., & Kritek, W. J. (Eds.). (1996). Coordi-nation among schools, families, and communities.Albany, NY: State University of New York Press.

Coburn, Cynthia E. (2002, April). The role of non-system actors in the relationship between policy andpractice: The case of reading instruction in Califor-nia. Paper presented at the Annual Meeting of theAmerican Education Research Association, NewOrleans, LA.

Coburn, K. G., & Riley, P. A. (2000). Failing grade:Crisis and reform in the Oakland Unified School Dis-trict. San Francisco, CA: Pacific Research Institute.

Datnow, A., & Stringfield, S. (2000). Working to-gether for reliable school reform. Journal of Edu-cation for Students Placed at Risk, 5(1,2), 183-204.

Finnigan, K., & O'Day, J. (2003). External support toschools on probation: Getting a leg up? Philadelphia,PA: Consortium for Policy Research in Education.

Fiscal Crisis and Management Assistance Team.(2000). Oakland Unified School District Assessmentand Recovery Plans. Sacramento, CA: Author.

Friedland, R., & Alford, R. R. (1991). Bringing soci-ety back in: Symbols, practices, and institutionalcontradictions. In W. W. Powell & P. J. DiMaggio(Eds.), The new institutionalism in organizationalanalysis (pp. 232-263). Chicago, IL: The Univer-sity of Chicago Press.

Gewertz, C. (2000, February 9). Calif. audit cites litanyof troubles in Oakland schools. Education Week,p. 11.

Hannan, M. T., & Freeman, J. (1977). The populationecology of organizations. American Journal of So-ciology, 82, 929-964.

Hannan, M. T., & Freeman, J. (1989). Organizationalecology. Cambridge, MA: Harvard University Press.

Harkavy, I., & Puckett, J. L. (1991). The role of medi-ating structures in university and community revi-talization: The University of Pennsylvania and WestPhiladelphia as a case study. Journal of Researchand Development in Education, 25(1), 10-25.

85

Honig

Hartley, J. F. (1994). Case studies in organizational re-search. In C. Cassell & G. Symon (Eds.), Qualita-tive methods in organizational research: A practi-cal guide (pp. 208-229). Thousand Oaks, CA: Sage.

Hatch, M. J. (1997). Jazzing up the theory of organi-zational improvisation. Advances in Strategic Man-agement, 14, 181-191.

Honig, M. I. (2001). Managing ambiguity: The imple-mentation of complex education policy. Unpub-lished Doctoral Dissertation, Stanford University,Stanford, CA.

Honig, M. l. (In Press). Where's the "up" in bottom-upreform? Educational Policy.

Honig, M. 1. (2003). Building policy from practice:Central office administrators' roles and capacity incollaborative policy implementation. EducationalAdministration Quarterly, 39(3), 292-338.

Honig, M. I., & Jehl, J. D. (2000). Enhancing federalsupport for connecting educational improvementstrategies and collaborative services. In M. C. Wang& W. L. Boyd (Eds.), Improving resultsfor childrenandfamilies: Linking collaborative services withschool reform efforts. Greenwich, CT: InformationAge Publishing.

Kirst, M., & Bulkley, K. (2000). 'New, improved'mayors take over city schools. Phi Delta Kappan,81(7), 538-546.

Kronley, R. A., & Handley, C. (2003). Reforming rela-tionships: School districts, external organizations,and systemic change. Providence, RI: The AnnenbergInstitute for School Reform at Brown University.

Malen, B., Ogawa, R., & Kranz, K. (1990). What do weknow about school based management? A case studyof the literature-A call for research. In W. Clune &J. Witte (Eds.), Choice and control in American edu-cation, Volume 2: Decentralization and school re-structuring. Philadelphia, PA: Falmer Press.

March, J. G. (1994). A primer on decision making.New York, NY: The Free Press.

March, J. G., & Olsen, J. P. (1989). Rediscovering in-stitutions: The organizational basis of politics. NewYork, NY: The Free Press.

Marsh, J. (2000). Connecting districts to the policy di-alogue: A review of literature on the relationship ofdistricts with states, schools, and communities.Seattle, WA: Center for the Study of Teaching andPolicy, University of Washington.

Marsh, J. A. (2003, April). Understanding joint work:District-community partnerships for educationalimprovement. Paper presented at the Annual Meet-ing of the American Educational Research Associ-ation, Chicago, IL.

Mawhinney, H. B., & Smrekar, C. (1996). Institu-tional constraints to advocacy in collaborative ser-vices. Educational Policy, 10(4), 480-501.

May, M. (2000, January 22). Scathing report on Oak-land schools, district earns dismal marks in stateaudit. San Francisco Chronicle, p. A-I5.

McDonald, J. P., McLaughlin, M. W., & Corcoran, T.(2000, April). Agents of reform: The role andfunc-tion of intermediary organizations in the AnnenbergChallenge. Paper presented at the American Educa-tional Research Association, New Orleans, LA.

McKelvey, B. (1982). Organizational systematics.Berkeley, CA: University of California Press.

Merton, R. K. (1987). Three fragments from a sociol-ogist's notebooks: Establishing the phenomenon,specified ignorance, and strategic research materials.Annual Review of Sociology, 13, 1-28.

Miles, M. B., & Huberman, A. M. (1994). Qualitativedata analysis: An expanded sourcebook (2nd ed.).Thousand Oaks, CA: Sage.

Oakland School Linked Services Work Group. (1999).Partnershipfor change II. Oakland, CA: Author.

Pavetti. L., Derr, M., Anderson, J., Trippe, C., &Paschal, S. (2000). The role of intermediary orga-nizations in linking TANF recipients with jobs(MPR Reference No: 8543-400). Washington, DC:Mathematica Policy Research, Inc.

Phillips, D. C., & Burbules, N. C. (2000). Postposi-tivism and educational research. Lanham, MA:Rowman & Littlefield Publishers, Inc.

Planning and Oversight Committee of the OaklandFund for Children and Youth. (1997). StrategicPlan, Oakland Fundfor Children and Youth. Oak-land, CA: City of Oakland.

Public Education Network. (2001). Communities atwork: A guidebook of strategic interventionsfor com-munity change. Washington, DC: Author.

Rodriguez, A., & Pereira, J. A. (2000, October). Therole of intermediaries in organizational capacitybuilding. Paper presented at the Annual meeting ofthe Association for Public Policy Analysis andManagement, Washington, DC.

Salamon, L. M. (1981). Rethinking public manage-ment: Third-party government and the changingforms of government action. Public Policy, 29(3),255-275.

Scott, W. R. (1992). Organizations: Rational, natural,and open systems (3rd ed.). Englewood Cliffs, NJ:Prentice Hall.

Smith, M. S., & O'Day, J. (1990). Systemic school re-form, Yearbook of the Politics of Education Asso-ciation (pp. 233-267). New York, NY: FalmerPress.

Smylie, M. A., & Crowson, R. L. (1996). Workingwithin the scripts: Building institutional infrastruc-ture for children's service coordination in schools.Educational Policy, 10(t), 3-21.

Smylie, M. A., Crowson, R. L.. Chou, V.. & Levin,R. A. (1994). The principal and community-school

86


connections in Chicago's radical reform. Educa-tional Administration Quarterly, 30(3), 342-364.

Spillane, J. P. (1996). School districts matter: Localeducational authorities and state instructional pol-icy. Educational Policy, 10(1), 63-87.

Spillane, J. P. (1998). State policy and the non-monolithic nature of the local school district: Orga-nizational and professional considerations. AmericanEducational Research Journal, 35(33-63).

Stein, M. K., & Brown, C. A. (1997). Teacher learningin a social context: Integrating collaborative and in-stitutional processes with the study of teacherchange. In E. Fennema & B. S. Nelson (Eds.), Math-ematics teachers in transition (pp. 155-19 1). Mah-wah, NJ: Erlbaum.

Suchman, M. C. (1995). Localism and globalism in in-stitutional analysis: The emergence of contractualnorms in venture finance. In W. R. Scott & S. Chris-tensen (Eds.), The institutional construction of or-ganizations (pp. 3943). Thousand Oaks, CA: SagePublications.

Urban Strategies Council. (1996). What is a commu-nity building intermediary? What work does it do?What makes it effective? Oakland. CA: Author.

Wechsler, M. E., & Friedrich, L. D. (1997). The roleof mediating organizations for school reform: Inde-pendent agents or district dependents. Journal ofEducation Policy, 12(5), 385-401.

Weick, K. (1984). Small wins: Redefining the scaleof social problems. American Psvchologist, 39(1),40-49.

Wynn, J. R. (2000). The role of local intermediary or-ganizations in the youth developmentfield. Chicago,IL: Chapin Hall Center for Children.

Yin, R. K. (1989). Case study research: Design andmethod. Thousand Oaks, CA: Sage Publications, Inc.

Author

MEREDITH 1. HONIG is Assistant Professor, Ed-ucational Policy and Leadership, University of Mary-land, College Park, 2202 Benjamin Building, CollegePark, MD 20742; [email protected]. Her areasof specialization include policy design and implemen-tation, organizational change in cities, and school-community connections.

Manuscript received February 10, 2003Revision received August 20. 2003

Accepted January 20. 2004

87

COPYRIGHT INFORMATION

TITLE: The New Middle Management: IntermediaryOrganizations in Education Policy Implementation

SOURCE: Educ Eval Policy Anal 26 no1 Spr 2004WN: 0410803465004

The magazine publisher is the copyright holder of this article and itis reproduced with permission. Further reproduction of this article inviolation of the copyright is prohibited. To contact the publisher:http://www.aera.net/

Copyright 1982-2004 The H.W. Wilson Company. All rights reserved.