Empowered Or Beleaguered ? Principals ’ Accountability Under New York City ’ s Diverse Provider Regime

By 2008, New York City’s school governing regime contained two market-creation policies. Each reshaped principal incentives. One closed large high schools, replacing them with four-to-eight small schools. Another replaced uniform district-provided services with eleven School Support Organizations (SSOs). Both aimed to empower principals with new discretion. This interview study of a small, stratified random sample of high school principals uses mixed methods to analyze 241 incidents detailing their reactions. Guiding questions include whether principals experienced the policies as empowering. Findings show that two thirds of the principals felt beleaguered rather than empowered; incentives appeared insufficient to provide them with unambiguous direction and confidence in their own decisions. The study concludes by considering what additional resources might be needed to expand the one third who felt empowered into a majority.

focusing on a target educational market (Quint, Smith, Unterman, & Moedano, 2010).Second, to increase their control and willingness to innovate, they were required to purchase educational support services from one of many district-approved vendors (New York City Department of Education, 2007a).Under each policy, school principals are the key decision makers.In the jargon of the current regime, they are empowered (New York City Department of Education, 2007a).
Each of these two policies exists within the context of high-stakes accountability for principals, including both financial rewards and the potential of being fired.In this context, the two policies examined here-hundreds of new small high schools and a market of School Support Organizations-can be understood as essential elements of a particular reform-oriented governing regime.Together they move the district away from a governing regime dominated by what Clarence Stone and colleagues have characterized as bureaucratic compliance to a market-based governing regime designed to enhance efficiency and innovation (Bulkley, Henig, & Levin, 2010; C. N. Stone, Henig, Jones, & Pierannunzi, 2001).
Why study principals' reactions?Hiring and retaining strong principals is widely believed to be among the key policy levers for teacher retention (Boyd et al., 2011), professional growth (Drago-Severson, 2007), school climate (safety, relationships and environment) (Cohen, McCabe, Michelli, & Pikeral, 2009), parental and community involvement (B. S. Simon, 2001), and organizational management (Grissom & Loeb, 29 March, 2011), all indirectly affecting student achievement (Leithwood, Louis, Anderson, & Wahlstrom, 2004).Principals' "role of policy mediator for the entire organization" means that they "determine the conditions under which policy interpretation and implementation will take place" (Leithwood, Louis, Anderson, & Wahlstrom, 2004, p.35).How principals react to policies is crucial to understanding educational policy effects.Research on their reactions is especially important when the policies are novel, interact, and have strong consequences.
The focus of this paper is on understanding how principals make sense of the new incentives in this high-stakes policy environment and whether they interpret the incentives as New York City policy-makers intended.We asked: How do principals experience the accountability implications of new small high schools on the one hand, and of creating a market of school support service providers on the other?Specifically, how do principals leading small and large high schools describe their responsibilities?How are high school principals' choices of School Support Organizations reflected in their responses?Who feels empowered by these policies?
Our analysis shows how the policies interact in complex ways that were not predicted by policy-makers' theoretical (or ideological) assumptions.We found that by identifying the particular School Support Organization engaged, we could predict the amount of conflict the principal would report between district expectations and his or her professional judgment.Although each principal chose his or her SSO, the choices were sometimes influenced by school size and were heavily influenced by the principal's prior on-the-job experience.The other independent variable affecting principal reactions to accountability was the level of academic challenge a school faces.Overall, the principals who described themselves as empowered were a distinct minority; most described themselves as beleaguered or overwhelmed.Although the sample of high school principals interviewed for this study was randomly selected, and mirrored the New York City high school principal population at the time of the interviews, these findings should be considered suggestive because the sample was small (n=18).

Governing Regimes and Forms of Accountability
The research described here rests on two assumptions: First, school districts are governed by a range of distinct governing regimes, each providing essential context for any in-depth study of accountability.Second, school leaders do not simply react to policy makers' expectations; instead principals enact their accountability environments.That is, they attend to some kinds of information and downplay others, thus framing problems in particular ways.A principal's enactment of accountability might weigh personal, professional or political influences more heavily than bureaucratic requirements or market incentives.Thus, it is inaccurate to assume that policy incentives are the only, or even the most important influence on principals.

Governing Regimes
The view that school districts are governed by different types of regimes is based on historical and case research (C.N. Stone, 2005; C. N. Stone, Henig, Jones, & Pierannunzi, 2001).Second generation analysis of this research extends Clarence Stone's distinctions between status quo and activist regimes by distinguishing among different types of activist regimes, each of which has a distinctive logic of accountability (Bulkley, 2007;Henig, 2010;Shipps, 2006Shipps, , 2008)).
Table 1 displays a typology of school governing regimes, identifying the anticipated form of accountability and agent responses. 2The Administrative Regime is based on Clarence Stone's observation that many school governing regimes do not challenge the underlying logic and structure of hierarchical administrator-run district governance (C.N. Stone, Henig, Jones, & Pierannunzi, 2001). 3Such regimes reflect the professional administrative ideal sought and largely attained in the early 20 th Century (Tyack, 1974;Tyack & Hansot, 1982).Today, the Administrative Regime is the bureaucratic status quo alternative to any activist regime, and is characterized by the importance of training, formal credentials, technical routines, and educator cohesion.Compliance with bureaucratic mandates, including certification requirements and established norms, constitutes a full demonstration of accountability in this governing regime.
Table 1 also expands upon Stone's notion of an activist governing regime intent on changing the agenda, the preferred accountability type, and the essential actors, resources, and relationships.It identifies distinctions between types of activist regimes, including Professional Regimes, Empowerment Regimes and Market Regimes.A Professional Regime relies on classroom educators whose expertise in teaching and learning is at the center of decision-making and governance.Its form of accountability relies primarily on professional discretion, constrained by norms of professional practice and influenced by pedagogical values (Darling-Hammond, 1996;Darling-Hammond & Ascher, 1991).Empowerment Regimes enhance the district's political responsiveness to particular interest groups or constituencies.Under some reform agendas the newly empowered may be parents and neighbors seeking community control of the schools (Hess Jr., 1991); other agendas include unions (Kerchner & Koppich, 2000).In either case, accountability rests on demonstrated evidence of political responsiveness to the expectations of the empowered interest groups.Market Regimes rely on performance or market incentives for their accountability, including both competition-driven incentives like waiting lists and efficiency-driven ones like employee incentives.All Market Regimes seek to restructure governance so that customers, contracts and employee incentives drive decisions, rather than political responsiveness, professional discretion or bureaucratic routines.A Corporate Market Regime requires that school principals compete with one another on standardized outcome measures.The goal, borrowed from corporate business models, is to encourage innovation by creating internal competition among managers who take responsibility for turning a reform idea into measurable improvements (Ouchi & Segal, 2003).To the extent that performance targets are pre-specified and standardized, the district's coercive power is heightened.Employee incentives are expected to result in desired outcomes: The highest performers earn financial rewards; poor performers are dismissed.Such policies increase the principal's authority to reject some kinds of instruction and to shape the school's culture by screening staff and students.
This type of structural competition is sometimes referred to as intrapreneuralism because it encourages the behaviors associated with entrepreneurialism within a large organization (Toftoy & Chatterjee, 2004).By 2008, such incentive-based governance reforms became common in states and districts across the nation (Hanawar & Olson, January 10, 2008).Yet satisfied intrapreneurs are uncommon, even in the corporate world.Although structural competition enhances middle level managers' importance to their organizations-they lead internal work groups and link crossdivisional teams-few such leaders feel liberated from bureaucracy.Instead, as Paul Osterman (2009) describes them, corporate middle managers facing structural competition feel "more subject to the control of top executives than in the past," and consequently "less secure."Their "stress levels have gone up sharply", increasing their sense of learned powerlessness.Consequently, many middle managers have also "lost their sense of loyalty to their employers" and become more transient (pp. 7-9).If the same applies to schooling, a Corporate Market Regime predicts many stressed, insecure principals and high turnover.
An Entrepreneurial Market Regime uses competition quite differently.Like Corporate Market Regimes, entrepreneurial ones allow principals to establish organizational norms of behavior, theoretically increasing their ability to block outside distractions.But in Entrepreneurial Market Regimes, principals must compete with one another for segmented markets of clients (families) and for similarly oriented staff.Client competition decreases the district's coercive power because groups of like-minded clients-not bureaucratically enforced standards-set the limits of acceptable behavior (Levin, 2010).When teachers become free agents, markets also decrease principals' coercive power over these subordinates.
Many school systems are adopting regulatory reforms that encourage entrepreneurialism (See examples in Hess, 2008).Yet, for most, Entrepreneurial Market Regimes remain aspirational due to an insufficient supply of new schools and/or service providers (Adams & Hill, 2006) and uncertain regulatory environments (Levin, 2010).Even so, advocates assume that these problems can be addressed with sufficient attention to human and financial capital development and quality control (Hess, 2008).New York City's policy of seeding hundreds of new schools seems to represent one of a few opportunities to test this assumption.Yet its governing regime is not a straightforward empowerment one.
Instead, New York City has a Diverse Provider Regime, which aims to counteract the weaknesses of entrepreneurialism with the benefits of corporatism, and vice-versa.It combines the standardized outcomes and internal competition of the Corporate Market Regime with an increased supply of schools and client choice adopted from Entrepreneurial Market Regimes.In Chancellor Joel Klein's words, "Accountability, in most industries or professions, usually takes two forms…First and foremost markets impose accountability requirements...Second, high performance companies develop internal accountability."Both involve competition because "collaboration is the elixir of the status-quo crowd" (Klein, June 2011, pp.4, 9).
Internal competition among principals for financial rewards (and jobs) is envisioned as a way to tackle the quality control and human capital problems of entrepreneurial regimes; the process is thought to both sort and develop leaders.Vendor competition aims to enhance the innovative forces of intrapreneuralism, while increasing the supply of new service providers.Increased schooling options are meant to address the problems of client responsiveness found in corporate regimes.Standardized measures of school effectiveness were intended to help parents select among the new schools and increase family commitments to the child's school.In these ways and others, New York City aims to combine the best of the two market approaches.For principals, it creates many mixed accountability messages.

Multiple Accountabilities
Both the theoretical argument for empowerment and empirical research on school leaders suggest that principals do not simply react to such shifts in policy; they make strategic choices.These choices are subject to the limits of bounded rationality (Cyert & March, 1963;H. A. Simon, 1957) and the internal structure of decision-making at any given school.As important, they reflect different types of accountability as school principals experience them (Firestone and Shipps, 2005;Leithwood, Louis, Anderson, & Wahlstrom, 2004).

Table 3 Multiple Accountabilities Framework
Accountability consists of demands for a demonstration of performance claims.As the discussion of governing regimes above suggests, it comes in multiple forms.For example, accountability requires politicians to demonstrate that they have heard and addressed the needs of their constituents.Managerial employees show their bosses that they have used mandated processes to meet performance targets.Market-based entrepreneurs please their customers, or risk losing them to another provider.Lawyers demonstrate that they follow professionally accepted practice or risk ostracism and censure, while religious leaders are expected to live in conformance with the teachings of their faith (Shipps & White, 2009).Table 3 associates each of five accountability types with a governing regime and sketches how each type is expected to influence principal behavior.
Principals face many of the five types of accountability at any given time, a claim that has been initially identified in typologies (Adams & Kirst, 1999;Darling-Hammond, 1989;Firestone & Riehl, 2005), independently substantiated by factor analysis of survey research-decision making by successful principals was found to have multiple components (Leithwood & Jantzi, 2008;Mulford, 2008)-and by research examining principal's interpretation of their own power and influence (Marks & Nance, 2007).
Governance reform aims to reshuffle accountability obligations, promoting some types at the expense of others.However, it seldom eliminates pre-existing obligations.For example, New York City, like many other districts, expects principals to achieve bureaucratically targeted goals first, even if this means putting off other valued educational objectives (Linn, Baker, & Betebenner, 2002;Sleeter, 2007).Yet, the combination of open enrollment and the creation of hundreds of small high schools has also pitted principals against one another in competition for students whose prior performance predicts they can meet outcome expectations (Hemphill, Nauer, Zelon, & Jacobs, 2009;New York City Department of Education, 2008a;Quint, Smith, Unterman, & Moedano, 2010).In such a situation, principals may experience the mandate to meet high outcome targets as being in conflict with the market incentive to differentiate their school's educational services from competitors.Add the principal's awareness of parents clamoring for more voice in school policies (See for example, Commission on School Governance, 2008) and his or her decisions become, at least in theory, very complicated.Therefore, rather than assuming that accountability forces flow directly from regime design, this research asks open-ended questions about how principals experience accountability.
The next section describes how New York City's two empowerment policies-hundreds of new small open-enrollment high schools and a regulated market of school service providers-are intended to affect principals in the context of its high stakes diverse provider regime.Subsequent sections report the data analysis strategies and principals' actual perceptions of accountability after the implementation of these two policies.Finally, I speculate about the policy implications for Diverse Provider Regimes.

Intended Policy Consequences and Rationale
Before clarifying the two policies examined in this paper, a word about the scope of highstakes accountability in New York City: By 2008, New York City principals were individually responsible for meeting outcome targets in two years, as measured by the school's Progress Report.High school Progress Reports use several metrics.A community engagement measure reflects average student attendance (set at 5% of the total school score) and parent, student and teacher satisfaction ratings (10%).Student performance progress (25%) rates the school on its disaggregated graduation rates.A value-added measure of performance growth (60%) overweighs the pass-rates of struggling students on five state-mandated Regents (graduation) exams and their on-time credit accumulation.4Principals were eligible to receive bonuses of up to $25,000 if their scores fell in the top 20%; they could lose their jobs if scores failed to meet targets (New York City Department of Education, 2008b).In the fall of 2008, 15 principals received the top prize, and another 71 received lesser bonuses (Hernandez, 2008a); one year later, 63 principals received bonuses of between $7,000 and $25,000 (Otterman, 2009).Although "the year-to-year volatility of the Progress Report has undermined its credibility," it is already a well-established yardstick for assessing principal performance (Hemphill et al., 2010, p.5).5

Creating a Market of Small Schools
Small high schools are defined as having 550 or fewer students, a cutoff used for several recent New York City studies (See for example, Bloom, Thompson, & Unterman, 2010).The small schools movement began years before Bloomberg's tenure, driven by a logic of "less is more" in which low-performing youth were offered intimate, family-like school options and deeper inquiry into a smaller number of subjects (Sizer, 1984(Sizer, /2004(Sizer, , 1997)).It was inspired by "social justice and democratic movements" (Klonsky & Klonsky, 2008, p.24).Small schools were said to encourage adult cooperation and nurture students who needed individualized support (Ancess, 2003), especially when accompanied by intensely collaborative learning structures (Lee & Smith, 1994).Although small schools have problems-they are many times more likely to experience extreme volatility in their test scores than large schools, and can develop strained social relations over time (Hernandez, 2008b;Kane & Staiger, 2002;Lee, Smerdon, Alfeld-Liro, & Brown, 2000)-by 1995-96, New York City had about 60 small schools and about 99 very large ones.
Small high school openings were stepped up after Bloomberg took control in 2002, this time motivated by the logic of choice and competition.Progress Reports identified "failing" large high schools, which were phased out over three years while 4-8 small schools were located on the same campus (Independent Budget Office, January 25, 2010;Quint, Smith, Unterman, & Moedano, 2010).Each small school is responsible for offering a distinctive path to a common standard of high school graduation, a path that is strongly influenced by at least one external partnering organization.Students are expected to do better in schools they chose than in assigned schools, while school leaders are sensitized to market opportunities by competing for students.
To facilitate the market for small schools, Mayor Bloomberg revamped the choice process by requiring each of the city's 300,000 8 th graders to apply for admittance to up twelve high schools, most receiving one offer in return (Abdulkadiroglu, Tathak, & Roth, 2005).Creating a new supply of small high schools was seen as a structural pre-condition for choice; it increased the number of high school options and therefore the possibilities each student might receive a preferred choice.By 2004-05, this open-enrollment logic expanded the number of small high schools to about 160, alongside 121 large ones.Three years later, small high schools nearly doubled that of large ones, 205 to 114,6 and educated about 18% of students (Quint, Smith, Unterman, & Moedano, 2010).

Creating a Market of School Support Organizations
In January 2007, Mayor Bloomberg announced that all school principals would be required to purchase support services from one of eleven School Support Organizations, or SSOs.School discretionary budgets were increased by an average of about $170,000 to pay for these services.Principals' SSO choices were to remain in effect for two years (New York City Department of Education, 2007a).Each SSO offered professional development, training in the use of the DOE's tools for measuring performance, assistance with special education, English Language Learners (ELLs), and the school's worst performing students.SSOs were supported by a district-wide data analysis system intended to help them target school service needs, but the DOE no longer provided such services directly to schools.In 2010, the DOE consolidated the number of authorized SSO entities from eleven to five, but the underlying rationale remains the same.See (Hemphill et al., 2010, p.14).
The DOE's rationale followed an argument made by analysts endorsing a change from a "school system" to a "system of schools" (Hill, Foster, & Glendler, 1990), which is loosely based on the model of the multi-divisional corporation (Ouchi & Segal, 2003).The district would no longer set norms for teaching processes, but would remain the standard-setter, data gatherer, and evaluator of school progress.Chancellor Klein claimed that by creating a market of SSOs, "We are transforming a school system based on compliance and top-down decision-making to one that empowers principals to make key decisions about what's best for students and their school communities" (New York City Department of Education, 2007a, p.1).By April of 2007 the district had created five SSO options.The city's non-profit education sector added another six.7Each SSO was asked to differentiate its services.As Table 4 suggests, SSOs came in three distinct types.The ESO was based on the premise that principals know best what services they need.ESO principals organized themselves into 22 networks of 10-27 schools and purchased the services of a shared, 5-person team: network team leader or general manager, business manager, achievement coach, instructional specialist, special services manager.All other services were purchased on the open market.The ESO charged the least for its services, gave some rebates and grew slightly in two years.In 2010, the DOE significantly expanded this model of SSO by mandating that most principals self-select into ESO-like networks, renamed Children First Networks. 8et many principals wanted outside assistance in identifying critical needs, practical help with special populations, and ready-made professional development.Consequently, in 2007 the district asked four superintendents to each develop a Learning Support Organization (LSO).The Integrated Curriculum and Instruction (ICI) LSO stressed differentiated and customized support to prepare students for the 21st-Century global economy and attracting many large high school principals.The Community LSO emphasized school-family-community interactions and the notion that schools were community resources.The Leadership LSO self-consciously focused on developing leadership skills across the staff.The Knowledge Network LSO offered a "content-rich" common core curriculum. 9In 2010, the district consolidated these four options in favor of a greater number of self-selected and self-sufficient "networks" under the ESO formula.Our research reflects these four LSO options, in existence when the interviews were conducted.
The third category of SSOs was the non-district Partnership Support Organizations (PSOs) developed and approved through an RFP process.They brought unique resources (e.g., a specialized curriculum, a model of coaching, access to university professors and courses) to a small group of schools.The largest of the sponsoring organizations was the foundation-supported New Visions for Public Schools, which has long championed small high schools.Many of the new small school principals chose NVPSO as their SSO, accounting for much of the growth.Other privatized options included the Fordham and City (CUNY) Universities in New York City and the city's Public Education Association (CED-PEA).Two non-profits with for-profit affiliates also made the cut. 10These external partnering organizations were retained in 2010 since they had district contracts.This new market of internal and external SSOs was expected to spur school-level innovation.Yet LSOs were undeveloped and PSOs were not approved until six weeks before selections were made.Consequently, principals reported choosing an SSO that was staffed by people they already knew, or knew of.Principal Eliot explained, "We chose [our SSO] because [of]…relationships that I've developed over the five years as principal…Those are the same people I knew before and the same people I'll continue to pick up the phone to call if I have an issue or a question."Principal Swinburne, a first-year principal concurred, "I inherited [my] network.I landed here the day the school opened.I had no idea I was going to be here…but let me be very, very careful -clearabout this.If I had to choose my [network], it would have been those exact same…people."Principal Bronte concurred, although he was unclear about whether another might be better: "I think I know what I want for my school.I think in my early years, I basically have been stuck with the same people who were [in] my regional support system, and I think in many years I learned a lot from them, and I think I did rely on them.I think I rely on them less now because I think I'm more, I more have an idea of what I want and what I think needs to be done."Under such circumstances, it is not surprising that during the second open-enrollment period in spring of 2009, less than 6% of schools (88 of 1474) recorded a new choice of SSO.
Therefore, this analysis characterizes the SSO choice as a reflection of principal predisposition and experience, rather than an exogenous causal influence.Like administrator experience, SSO choice was expected to influence how principals conceptualize their decisions only to the extent that it reflected some aspect of their prior professional experience.Because of this selfselection, it was anticipated that principals in different SSOs would respond to district-wide accountability pressures in distinct ways, which would be revealed in patterned responses.

Research Methods
The 18 randomly selected principals whose lengthy (avg.=1.5 hours) structured interviews analyzed here took place in winter and spring of 2007-08, following implementation of the two 9 SSO products and services varied too.For example, the ICILSO aimed to develop schools as pathways for academic and economic success by tapping into the latest brain and developmental research and by focusing on meta-level skills like critical thinking, synthesizing and analysis.To accomplish this, ICILSO affiliated with 14 external providers of professional development including Harvard's "American's Choice" and offered professional development courses like "Mendelian Genetics," "DNA Transformation and Protein Isolation," "Developing Mathematical Ideas," and "Mathematical Patterns and Functions," all of which aim to enhance the content knowledge of teachers in large comprehensive high schools. 10In 2007, PSOs' active solicitation of schools was limited to those with which they had developed a prior relationship.market-creation policies described above.11Informants were stratified to include equal numbers of principals from each of the six SSOs chosen by at least ten high school principals (i.e., ESO, ICILSO, CLSO, LLSO, KNLSO, and NVPSO).They were also stratified to ensure equal numbers of large and small school principals.12Permission to enter the schools was secured using required college and DOE institutional review processes.All interviews were conducted confidentially; names used are fictitious and some descriptive data is omitted in Table 5 to preserve anonymity.
All interviews followed a semi-structured protocol developed by the author and two colleagues (Razik, 2006;White, 2006), which adapted critical incident technique (Flanagan, 1954).CIT has proven useful for identifying actors' perceptions about the behaviors critical to their own performance in environments where individual discretion is high and decisions have serious consequences (Chell, 1998).The lengthy interviews consisted of seven open-ended questions structured to elicit incidents crucial to each principal's experience of accountability, resulting in a total of 241 incidents.Principals were asked to provide an explanatory example, or typical story, about each type of the five types of accountability identified in Table 3.Multiple probes ensured that incidents were described in detail (e.g., Did you consult with anyone?What was their advice?How did you finally decide?).Although solicited by questions identifying a particular type of accountability, incidents did not always conform to the type sought.Nor did principals limit themselves to one incident for each type of accountability.The average number of incidents per informant was 13.4.Interviews were tape-recorded, transcribed and qualitatively coded Using NVivo 8.0.A large subset (14%) was complex, and nearly 4% created one-off codes (See Reducing Critical Incidents below).Coded incidents were correlated with five explanatory variables and chi-squares were conducted on grouped variable counts to identify patterns using PASWStatistics 18.0 (See Analysis below). 13s table 5 shows, a large proportion of the principals interviewed proved transient.Five principals left the job within two years, nearly a 30% transiency rate.Those with the highest transiency rate led small schools with relatively high achieving students (75%) although they had the same average tenure at interview as all other principals (4.5 years).This pattern reflects a recognized trend in the contemporary era (Gootman, 2006), and suggests a combination of reasons.Other research on New York City has found that small high school principals leave early in their careers from burnout (Hemphill, Nauer, Zelon, & Jacobs, 2009).Floundering principals sometimes point to the lack of opportunity to learn from their local counterparts precisely because their support networks are not geographically based, while accountability tools based on gain scores are quick to reward (and punish) low-achieving schools (Hemphill et al., 2010).Our findings shed some additional light on this transiency (See Policy Implications below).

Critical Incidents
An incident is a description of a decision situation in which the principal was required to act.It is typically related as a story, with a beginning, middle, and end, heroes and villains.Alternatively, it was told as an argument explaining why a principal's view of causes and consequences led to her response.A critical incident is any situation involving a principal decision that, in hindsight, seemed consequential to her leadership and the school's performance.An incident took, on average, seven minutes to relate, and most involved responses to several probing questions.

Reducing Critical Incidents
All critical incidents were coded according to the two essential dimensions.The first was drawn from the multiple accountabilities framework in Table 3: the type of accountability experienced-bureaucratic, political, professional, market, or moral.For example, a critical incident the principal characterized as a demand from parents requiring him to suppress, negotiate or mobilize external resources was coded as a "political" incident.An incident characterized by the principal as requiring mandated compliance to a particular DOE rule or outcome target was coded "bureaucratic."When a principal spoke about a decision situation where she was in competition with another school, requiring her to be innovative to gain a competitive advantage, it was coded as "market", and so on.
The second essential dimension of a critical incident is the relationship between accountability types when more than one is mentioned. 14The principal's assertion of a single accountability type, requiring little deliberation, was coded as "simple."The characterization of the critical incident as 14 My thanks to Monica White who first brought the relationship component to my attention.
requiring a choice between two conflicting accountability types, each encouraging a different decision, was coded as "conflicted."Two or more accountability types that reinforced one another in a single incident, leading to an over-determined response, was coded as "aligned."Three accountability types in a complex relationship of conflict and support, in which the principal reported balancing many interests and alliances was coded as "complex."Thus, each incident could be coded as a combination of its accountability type and relationship, a process that created 30 distinct incident codes.Because critical incidents were most often described as reflections of multiple accountability types, either because several forces were perceived to be in conflict or reinforcing one another, our counts of accountability codes are weighted. 15he most common incident involved conflict between the principal's professional judgment and how she understood the district's rules.For example, if in the professional judgment of the principal, a chronically failing and truant student required access to a different type of school, but the district's procedural rules required that the school continue to serve the student, her recounting might describe a clash between two types of accountability-professional accountability and bureaucratic accountability.In Table 6 this incident would be coded as bur vs. prof, which includes both the types of accountability mentioned and the conflicted relationship between them.This accountability coding was near ubiquitous; only one first-year principal failed to report any such incident.Indeed, other research suggests that most principals routinely make difficult judgment calls about whether, when, and how to follow their professional training and instincts, and when to abide by the rules (Firestone & Rhiel, 2005).But the findings of this study go beyond this most common type of accountability conflict by mapping the range of incidents critical to principal accountability in New York City.
To take another example, a different principal might describe the same two influences described above-the principal's professional judgment about who can best serve a student and the district's rules about continuing students-as supporting one another, because the principal felt his school was a good fit, and he bolstered his pedagogical arguments by referring to district policy.In this case, the incident would be coded as bureaucratic supports professional (bur supp prof).Table 6 shows all 30 distinct accountability codes that the reduction process produced.
Inter-rater reliability for incident coding was assured by asking an experienced colleague to code a randomly chosen subset of the 241 critical incidents.The author and colleague initially achieved 75% agreement, a Cohen's Kappa of .41,and "moderate" consensus according to Stemler (2001).

Analysis
Analysis omitted the nine one-off accountability codes at the bottom of Table 6.Counts of the top 21 accountability codes (each reflecting two or more incidents) were correlated with five variables.Two variables estimated the market creation policies being examined-school size and choice of support network.Three variables estimated either the principal's leadership experience (i.e., years of principal experience) or the school's educational challenge (i.e., school diversity index, school index of student performance or "peer" index).As briefly described below, the three nonpolicy variables estimate conditions that literature suggests should have an independent influence on principal decision-making.
Only those incidents whose count led to significant correlations with either of the two policy variables or the three non-policy variables are pursued in depth.For these incidents, patterns were sought in content (e.g., teacher selection, parental complaints, mandated accountability tools) and principal tone (e.g.confident, frustrated, puzzled).The patterns help explain why the correlations and chi square findings are significant, but also how they are consistent.
In addition to correlations, incident counts were aggregated separately by accountability type and accountability relationship to determine if either dimension was separately significant in understanding variation.Chi-square tests of significance were conducted to evaluate the null hypothesis of no association between each of five variables and the two aggregated incident dimensions.Due to the relatively small sample size, all variables were categorized and nonparametric tests of significance were used. 16Below are brief descriptions of the three non-policy context variables.
Experience.Common sense and some research suggest that a principal's years on the job is related to decision-making (Hamilton, Ross, Steinbach, & Leithwood, 1996;Leithwood, Steinbach, & Begley, 1992).On-the-job experience is expected to affect the mental models that leaders bring to organizing the school and to be highly dependant on the situational contexts.As Kenneth Leithwood and colleagues put it: "Like tools, concepts can only be fully understood through experience with their use and the refined appreciations (including tacit knowledge) that occur as a result of feedback from such use" (Leithwood, Louis, Anderson, & Wahlstrom, 2004, p.69)The relationship between experience and success may not be linear, however.For example, one study found that more job experience was positively correlated with principals' use of less-efficacious, external forms of power (i.e., coercion, rewards, and hierarchical authority), while the internal forms of power associated with effective schools (expertise and referent power) were more commonly used by principals with less experience (Lyons & Murphy, April 8, 1994).Our study categorized the number of years as a principal into two groups, and found significant correlations and qualitative differences between early career and experienced principals' reactions to New York City's high stakes accountability.
Prior performance.Prior student achievement is a significant predictor of school performance that justifies many reform efforts (See for example, McDermott, 2003).New York City asserts that its measure of prior student performance, the Peer Index, is the best predictor of future student performance because it provides a baseline for judging schools on a growth model (New York City Department of Education, 2007b;Roderick, Jacob, & Bryk, 2002).The Peer Index also acknowledges that principals of schools with high proportions of low-achieving students (i.e., low peer index) face a more difficult graduation rate challenge than those whose incoming students have higher test scores.It is a weighted average of the students' 8 th grade English-Language Arts and math exam rankings that range from 1 to 4.5.The higher the number, the higher scoring the school's entering students. 17Our study adopted the Peer Index as one of two measures of the school's educational challenge.It was assumed that a principal's perceptions of accountability would vary based on the Peer Index of the school.Both significant correlations and qualitative differences were found.
Diversity.It is a truism that the educational task is more difficult in highly diverse schools than in predominantly white schools (Jencks & Phillips, 1998;Teddlie, Stringfield, & Reynolds, 2000).The diversity index used in this study is RAND's school race/ethnicity diversity index, standardized on a 0-1 scale: "the probability that two members [students] of the [school] population chosen at random will be of different subpopulations" (White quoted in Gates et al., 2003, p.75). 18his variable allows for examination of differences in principals' views of accountability based on student, our second measure of educational challenge.In our sample, the diversity index is also inversely correlated with the percentage of black students (-.870, p =.01) reflecting New York City's complex racial segregation.In RAND's study of national survey data, diversity values above .7 were associated with a marked increase in principals' perceptions of poverty-related instructional problems and conflict among parents and community members (p.123).Despite expectations that we might find similar results, no significant correlation between the diversity index (or its inverse, percent black) and any incident codes was found.As intriguing as we found the lack of statistical significance, this article does not pursue it further.

Methodological Limitations
Correlation and chi square analyses were used to identify the particular subset of incidents where it would be fruitful to examine differences in content and tone.Qualitative analysis helped explain why the differences occurred and how principals experienced them.Although triangulating results between quantitative and qualitative analysis improves reliability, there are also limitations in this methodological approach.
One limitation is selection bias involved in seeking volunteers for in-depth interviews.To test the hypotheses that differences by school size, student diversity, incoming student performance, principal experience or gender might have influenced the number of randomized redraws required to identify 18 willing participants, and hence the sample's reliability, we correlated all these variables.Only the timing of the interview request had a statistically significant correlation with the number of random redraws required to find a willing participant.Later in the school year was more difficult to schedule.Nevertheless, it may be that these randomly chosen 18 principals are unrepresentative of the larger group of New York City high school principals in some untested way.
Another limitation involves the large amount of data collected when using CIT.Because multiple incidents are sought, and none rejected, voluble principals may have disproportionate impact on the results.Reporting the proportion of the identically coded incidents in the findings section mitigates this possibility.In addition, all one-off incidents were excluded from analysis, focusing the qualitative analysis on patterns rather than outliers.
A third limitation of using CIT is its reliance on self-reports.We do not know whether principals' reports accurately reflect what they did.The principal's self-report does not exhaust other perspectives on the same situation.Nor are their accounts reliable statements of what it takes to improve school performance.Leaders may feel certain that a particular decision had a self-evident, universal impetus, or that change in their working environment obviously shifted the consequences they faced, even when their intentions and/or shifted incentives did not have a causal relationship to outcomes.Instead, this choice of interview technique is intended to capture how principals make sense of and decide to act upon their accountability by eliciting the stories that they tell to justify, explain or rationalize their behavior (D.Stone, 1998).

Key Findings
While our research must be considered suggestive because the sample is relatively small, the findings are internally consistent and help illuminate how high school principals react to the complicated accountability environments found in Diverse Provider Regimes.The results of correlation and chi square analyses show that both policy and non-policy variables were systematically related to principals' views of accountability.
The main finding concerns the interaction of the two market-creating policies.Large school principals (n=9) relied significantly more on outside political resources to support their professional decisions than did small school principals (n=9).This pattern became more nuanced, when incidents were grouped for chi square analysis.Six principals, one third of the sample, drove much of the observed difference.They reported less conflict and evinced more confidence in their decisions than did the twelve principals who chose the other four SSOs.The six had each chosen one of two SSOs, NVPSO or ICILSO.NVSPO specialized in working with small high school principals (n=3), while ICILSO specialized in working with large high school principals (n=3).This evidence suggests that the choice of SSO was sometimes influenced by the school's size through the mechanism of SSO specialization and market segmentation.
Close analysis of the incidents in question revealed that their perceptions of accountability coherence and feelings of confidence came about for different reasons depending on whether the principal headed a small or a large school.Principals leading large schools reported less conflict because they experienced a great deal of alignment in their accountability environments: Political, moral, and market accountability all reinforced their professional judgments more than was typical.Principals leading small schools reported less conflict in their accountability environments because they felt each type of accountability stood alone, uncontested and uncomplicated by rival claims.Below I further explain the significant correlations in Table 7, and the significant chi-square in Table 8 by analyzing the content and tone of the incidents involved.

The Interactive Effects of School Size and SSO Choice on Accountability
Small high school principals appear to lack a resource that many large high school principals have in relative abundance: the support of community and political leaders.Large school principals voiced many more incidents in which outside organizations-politicians or community-based organizations-supported their professionalism, sometimes providing the ballast to withstand other opposing community forces (.547, p =. 05).The many programs large school principals offer, their relatively deep history as neighborhood institutions, and the diversity of their external constituencies are all reasons they reached out for support.Principals of large schools sometimes even expressed a comradeship with local politicians.
Half of the incidents in which political resources supported professional goals referred to community-based organizations or government agencies (4 incidents from 4 principals).For example, Principal Salinger told us that community-based organizations came to her aid when she experienced parental pushback over student internships.Parents complained that "I work, I don't get off till 4:30-5:00.I'm concerned that my child is now getting out of school a little bit earlier [to commute to an internship site].These are mean, evil streets.My child will be even home alone a little bit longer now if say, for instance, my child doesn't go to the internship or doesn't have [an internship] program on that particular day."So Salinger asked her partnering organizations for flexibility: ask "Could you start your program a little bit earlier on a Tuesday?"Uniformly, the reply she heard was "Sure, what do you want?" Pleased and also a bit surprised she surmised; "The community [groups] worked with me because they knew what I was trying to achieve on many levels.I have such a good relationship with all my partners [such] that no one said, you know, 'what are you doing?You know, why are you messing up the schedule?"Moreover, she often "reached out" to them.Then she explained why she thought this didn't always happen for principals of small high schools: "It is hard work; it's a lot of meetings.It's a lot of networking; it's a lot of begging.'Please?Can you?'"The time needed for this intensive community engagement was justified for the leader of a big high school serving thousands of nearby students, less so for the leader of a small school serving a few hundred students drawn from across the city.
Most other incidents of external support for the principal's professional goals referenced local politicians (38%, 3 incidents from 3 principals).Principal Ellison reflected on the day that "[My city councilman] had come in to introduce himself," mentioning that he was a former student."There was just something so likeable and honest about him that we established a relationship [right] off the bat… I had a particular problem several years ago, when I first became a principal.I felt that it would have been a hindrance to my appointment here… I went to someone I felt I could trust, because he had a vested interest…I went to somebody I felt could help me.And he did.[Our] conversations, then went from the councilman to people at the district office to express his concerns about what was going on."Principal Ellison concluded: "I don't know if I could have [resolved the situation] without his support and his assistance." Principal Marquez also described the importance of politicians' assistance to large schools with many disparate needs and deep community roots.She further explained that the favors were reciprocal."I also have testified in front of the borough president, and they both have been my guest speakers… I have testified in front of the city council… the politicians recognize the need, and besides, they are all really, really nice people.[Politicians] are all interested in the community…They are my strongest supporters" (Emphasis in original).Far fewer small school principals reported external political support for their professional judgments, and none specifically referenced the support of politicians.Recent analysis has suggested that successful superintendents manage upwards to the school board and orchestrate a network of external actors that control key resources (Meier, 2003;Meier & O'Toole, 2001).When their authority is enhanced, orchestrating external organizations and partners may be as important to principals of large high schools as managing relationships with board members and community leaders is to superintendents.
Chi square analyses also revealed distinctions between accountability environments as experienced by SSO choice, which complicates these small and large school principal differences.Principals affiliated with the NVPSO-all leaders of small schools-and those affiliated with ICILSOall leaders of large schools-told fewer stories of conflict than would be predicted.Examining all of the non-conflict incidents from these six principals revealed two distinct patterns: ICILSO principals reported large proportions of the incidents in which market accountability, political, and moral accountability supported their professional judgments.NVPSO principals identified the largest proportion of single-unconflicted and unaligned-accountability incidents.For NVPSO principals, accountability was made up of a series of largely uncontested and straightforward decisions, each of which reflected a single source of influence.It was their job to meet these disparate demands, or at least appear to do so.
Like most large school principals, those in the ICILSO relied on political support as described above.ICILSO principals also reported 50% of the incidents (3 incidents from 2 principals) in which their own sense of moral accountability supported their professional judgments.Principal Salinger described a common view of this alignment when she spoke of her commitment to small classes.
My moral accountability is [summarized as] the least amount of kids that I can put in a class., as much as I can afford it and can do it.I get very agitated when I have to have more than 28 kids in the class, personally, as the principal.Alignment came too from ICILSO principals' embrace of competition as an adjunct to their professional judgments (29%, 2 incidents from 2 principals).Principal Eliot conveyed it this way: "When I talk with parents, I'll bring up…the three other [competing] schools I really focus on.So I always print our [Progress Report] and those three schools' [Progress Reports], and 90% of the time we're better in every category.So that's why I'm happy to show it to all the constituencies and talk about it with all the constituencies… we scored higher than [one of the city's selective schools] did…They made a big issue of it because they didn't do that well and so I made sure I posted that article."He spoke of parents as constituents in recognition that his role was political, but public relations was also instrumental to his professional credibility.He actively constructed the public perception of a strategic market advantage over his school's main competition.Principal Swinburne also spoke at length about how competition for students was a key activity that engaged school leaders and discover his school's competitive advantage."It's a competitive market right now.And really we have to make Jamaica as saleable or more saleable than some of the other choices that [parents] were making… What it is right now is just a matter of making this school as attractive for those parents and those kids as some of the [small] schools that they've been diverted away to" But this was not in conflict with his sense of professionalism, and he found a number of ways to align the two: I'm working with my feeder schools… We have regular dialogues.My kids go to their schools, and we do push-ins with the kids about the high school experience.And they do a little bit of PR for [Swinburne] High School.We're doing some bridge collaboratives through the summer with their eighth graders coming to us… I'm looking at being able to offer programs, academies, smaller learning communities, within the school program here that are thematic in such a way that we'll draw kids, because this is a area of interest, and the school next door, down the block, or somewhere else in [our area], is not offering it.
NVPSO principals appeared empowered in a different way.Compared to the principals choosing other SSOs, they more often identified bureaucratic, market, and political accountability as separate, independent influences on their decision-making.They mentioned 63% of all the incidents (n=10) in which uncontested bureaucratic accountability drove principal's decisions.Two of them mentioned 46% of the incidents uncontested political accountability (n=6), as well as 24% of the incidents (n=4) in which uncontested market accountability was the only influence they experienced.Although each of these 20 incidents involved a problem requiring an executive decision, they were recounted as uncomplicated ones, usually resolved by the principal's unilateral action, occasionally through collaboration with, at most, one organization.NVPSO principals were also among the most self-satisfied.
NVPSO principals typically related uncontested bureaucratic obligations as if their sole responsibility was to implement district regulations within the limits of the school's resources.For example, Principal Morrison responded to the district requirement that his small school meet specific graduation rate targets by personally recording every student's attempt at a Regents examination and the highest test score attained.He then derived a weekly assessment of the student's odds of meeting the graduation standard.This revealed a distinct NVPSO pattern: He personally supervised most of the school-developed assessments and mandated multiple testpreparation strategies.
I do this for a reason.To piggyback on what the guidance department is doing, and I have three counselors here, which is a lot for a small school.I have a kind of a unique model but I must do this in order to ensure that nothing is missed.{For example] I know this kid already qualifies for a Regents diploma.So now my next step with this student is trying to get her to the Advanced Regents.She'll sit for the Spanish Regents… I meet with the guidance staff weekly in looking at what we need to do to get the kids to the next level…We review prior exams, obviously… In the regular class we try to piggyback other areas in order to reiterate and to review things that are supposed to be-content that should be known for exams.We have a slew of Regents review books that the teachers use all the time.And in fact, the teachers know that in the 'do-now' assignment [in every class] and homework assignment and exams I expect, not Regents questions, but Regents-type questions...We [require] mock Regents exams.Many instances, six weeks to three months before [the actual test].Principal Morrison had internalized this bureaucratic target as an essential managerial task, and personally made sure it was accomplished.In turn, he expected his teachers and guidance counselors to internalize the tasks of test-preparation and diagnostic testing as he mandated them.
Market and political accountabilities provide other examples of uncontested demands being personally met by NVPSO principals.Principal Hemingway, for example, saw his market accountability as requiring that he compete for the best students.But he differed from the ICILSO principals.His market accountability was not enacted as a way to support his professional judgments, but rather in reaction to a mental model drawn from the corporate world.He explained: [T]he way marketing works in the [school] system is not the same way as it works in business, okay?Because I think in business, barriers between the different markets are a lot greater.There's a much more distinct demarcation between markets in a line of business.Whatever market a kid might be in; let's say the kid is interested in architecture and he might want to go to architecture school.'" Principal Hemingway concluded that his segment of the student market required one-to-one recruitment of students because the competition was able to blur market barriers.Like other NVPSO principals, he took on these marketing tasks, construing his job as a personal competition between rival principals.Principal Shelly described an example of uncontested political accountability in which his reaction was uncomplicated by other pressures, but also required his time-consuming attention.In speaking of his school's district-mandated School Leadership Team (SLT)-a combination of community members, parents, and school staff that provides an avenue for parental and community involvement-he said " I intentionally tried to create opportunities for the community to actually look at things and make decisions.So, for example, when we got ready to start second trimester, we wanted to change the bell schedule because of issues with physical education, that sort of thing.We presented that to the school leadership team and allowed them to make the decision, which they did.They agreed with what I had suggested.Recently, we were asked to participate in the new incentive program for students, the cell phone incentive program.And that I felt was a great opportunity for the leadership team to make a decision, so we presented that to them and they made a decision about that, as well…I presented [these decisions] not as 'We've decided this', but [as] 'This is a proposal that we had'.And I had a bell schedule with the word 'draft' on it.In this case, the principal construed his political accountability to the SLT as making sure they felt involved, even though the decision was already made.Similarly, he asked his community partner to staff a desk in his office.He justified it as a way to give his partnering organization a sense of ownership for and commitment to administrative decisions emanating from his office.
Thus, all six principals in ICILSO or NVPSO reported low to moderate amounts of conflict in their accountability environment.The small school principals among them seemed to have engineered this by ignoring some incentives and information as unwanted distractions, compartmentalizing decision situations and acting as if their choices were simple.The large school principals among them imputed common interests and motives to all parties, reframing their decisions as having multiple sources of influence, all reinforcing their own professional judgments.These principals were the most sanguine about their ability to thrive in the face of bureaucratic constraints by making active use of markets and political support.They revealed a capacity to see support and alignment where others saw conflict and stress.In their different ways, all six can be said to be "empowered," even though the small school principals among them could also be seen as personally over-committed.
The other four SSOs attracted principals who held more skeptical views of the regime's empowering policies.These principals led an equal number of large and small schools.The accountability they describe is one of heightened awareness of their obligations to the school's external constituents and their own professional goals, often in conflict with the DOE's expectations.For example, small school Principal Tan (ESO) spoke at length about how his school is distinctive, and unusual within the system.
I think I need to explain our governance system, how we work here.This is not a top down organization.I don't make all the decisions and then everybody has to do what I say…I believe that teachers know their students and know the curriculum better than I do.They're with their students every day.I'm not…It's a harder way to run the school.Some decisions take a little longer to make than others because they're talked out a lot and people have to really agree.Consensus is a hard thing to do.You might have a dissenting opinion and if the consensus goes the opposite way, you need to agree to buy into that even though you don't believe in it or even though you don't think it's right…It's a harder way to operate; it's a lot more stressful.It's much easier to say, 'This is what I want you to do, now you need to do it.'It's very simple to do that.But I believe that that's not a good process because you don't have buy-in, you have people that might become very unhappy about what they need to do and how they want to educate their students.He provided critical incidents that demonstrated this distinctiveness, including deep concern about pedagogical directives from DOE that conflicted with the schools' own hard-won vision of best practice.Rather than believing he was given the tools to shape his own school's image and pedagogical orientation, he felt hemmed in by useless bureaucratic requirements.
I am not a political person.However, when you become a principal you learn to become one, because you need to, very quickly.It's numbers, it's all about numbers, and I think that's what's happening now in New York City with all these requirements and assessments.What I was very insulted about as an educator was the fact that all of a sudden New York City decided that they needed to do interim assessments and periodic assessments.Well you know what?That's what teachers do.We always periodically assess our students.I mean how else can you do it?But now, it has to be a standardized assessment.Why? Why?I want to know why?Because the powers that be need to show that there's improvement because of mayoral take over!I really believe that.
Unlike the large school principals in ICILSO or the small ones in NVPSO, Tan and others in this conflicted group argued that their identification with their school's broader community didn't translate into meaningful advocacy.It sometimes left them feeling powerless.As large school Principal Angelo (ESO) put it: "I do feel a huge sense of responsibility to my constituency because I feel like they have no voice.I feel like they're misunderstood in the City of New York.I feel like they're seen as a threat because they're boys of color."She was especially upset by the lack of an opportunity to hash out her pedagogical and philosophical disagreements with district officials."Part of the thing that's so scary and so disgusting about the progress model is that there's no fight moment, right?There's no fight moment." Principal Cather was especially thoughtful in the ways that he described his disillusion.Mindful that his school was founded with a great deal of district fanfare and external support, he put his conflicts this way: [Cather High School] was founded under this new wave of those of us who were trying to come back into the system to break up the large under-performing schools in [this neighborhood] to find alternative ways of approaching pedagogy, approaching practice, to develop and define new best practices in order to move the kids where we need to go… However, with this new administration there's this emphasis on data more particularly… So I've had to adjust a little bit… And what I've had to do is I've had to decide what's more important.To hold on to what I believe?No.
When we founded the school we were based on portfolio.We did a lot of projectbased activities.And the Regents [examinations] was something that loomed large.We didn't have them until sophomore, junior year.We really believed that we could change the attitude, the academics, of our incoming students and if we gave them a wonderful project-based, small-group, class size, type of curriculum that was heavily into the humanities that our students would automatically do well on Regents.We found out that they not always performed as well.It wasn't automatic.So we've had to make some corrections.We've also found over the last couple of years that we've had to do a little bit more test prep.Not as a focus, but learning how to sit for the exam.Have the Regents taken on a little bit of perhaps a disproportionate amount of role in our lives?Sure.But that's the [new] reality.
Unlike the confident, do-it-yourself principals leading small NVPSO schools or ICILSO principals who used market signals to support their professional judgments, these large and small school principals felt disadvantaged by the expectation that they should compete for students, teachers and DOE rewards.Small school Principal Steinbeck (KNLSO) spoke for many when he said, "The competition, in some ways, [is] not really fair, it's not on a fair playing field…when we first opened the school, I did a lot of recruiting by myself."Like NVPSO principals, she took on the work herself, "I was the everything recruiter.I put days and nights and hours into recruiting."But unlike them, she was demoralized by the exercise."I found out who was my main competition and why.There's a school not too far from us… and they had been in existence for years prior to us, but why they were a big lure is they gave laptops to incoming freshmen.So, they were, they were big competition… Now that's a competition that I'm recently running against now, and we didn't have this problem [before]."She believed she lacked the budget to compete effectively.
Small school Principal Ishiguro (LLSO) also found that her market accountability was in conflict with the high standards she wanted for the students in her school."I feel the pressure to compete.I compete for students, and I compete for teachers…because of the theme of my school [requires] somebody who has an appreciation for the [our theme] and somebody who can pretty much deliver and execute a lesson."But she went on to say that as free agents teacher were hard to keep in her low-income neighborhood.So she was forced to take many who she knew in advance would be transient."Those alternative certification folks, it's a hit or miss.Because they have a Peace Corps mentality.I mean I'm parochial school trained so [for me, teaching] is a vocation.It's a calling.The [alternative certification applicants], they're career changers" [Emphasis in original].As a result of competing for teachers, she felt forced to lower her standards."How about some common sense in the classroom?Someone who commands respect…Someone that children can look up to who doesn't come in with piercing and tattoos.I mean I'm dealing with minority children that haven't even been out of the borough to travel.How about that?" In a manner similar to that reported in studies of putatively "empowered" corporate middle managers, this 60% of principals in our study reported a high incidence of complex accountability that revealed a lack of trust in the district.Large school Principal Bronte (KNLSO), for example, felt manipulated."[The bureaucrats in DOE] are very clever.They have completely figured out how they are going to say 'We're going to empower you, but by the way, you better have gotten A, B, C, D, E, F, G done.' Of course, you know there is only one way to get all those things done.So they have created the assessment tool, which guides what you have to do…So, they are very clever people." Large school Principal Ellison (CLSO) described the principalship in defeated terms: "I feel as though I'm sitting in the firing line, the only one in the middle, and everybody else is shielded.It's a very lonely position.I've accepted the challenge, and I've walked into it knowing that it has its own reality.And yet, at the same time, I feel as though I have 17 different bosses."Some were explicit about feeling personally disrespected.Large school Principal Murdoch (CLSO) told me "I think the one thing that many of us resent…[is] that it's not often respected that we have put the time into the system, and we have done it for the right reason.And there is no effort on any level other than the people that we interact with closely that we are ever given an opportunity to be looked at as people who have dedicated their careers to the system and to the students in the system.I know that I am not the only one that feels that." Many believed that in such an environment, some schools and principals would win, but others would lose over the long term.To avoid feeling helpless, Large School Principal Murdoch (CLSO) cultivated a vision of the principalship as a relatively short part of one's educational career "In yesteryear, you were a veteran principal when you had ten, 15 years [as a principal] in the system.I hardly think that five years in the system makes me a veteran principal.Sadly in their eyes now at six years I am a veteran principal".She went on: When they ask me now 'Why did you do this?', the answer for me is truthful.I never wanted to be a principal, but when the opportunity was presented to me to be the principal of the school that gave me the tools where I am today, I had no choice but to say yes.Because for me, coming back here to lead a school that I totally believe in, that's a full circle in your career…Would I do it now over again knowing what I know now?I don't think so."

Non-Policy Factors: Principal Experience and Student Performance
On-the-job experience and the school's education challenge also affected principals' reactions to accountability.Overall, experienced principals (n=9) reported more alignment between DOE obligations and their own professional judgments than did early career principals (n=9).But principals of schools teaching 3 rd quartile Peer Index students (n=8), including experienced ones, reported more conflict between DOE requirements and their professional judgment than did schools with students in the bottom or 2 nd quartiles of the Peer Index (n=10).Below, content and tone analysis of the incidents in question shows how these two correlations illuminate the role of experience and educational challenge in principals' reactions to accountability.
Experienced principals (5,6 yrs.) reported 82% of all the incidents in which bureaucratic regulations supported the principal's professional commitments (.701, p = .01),either because principals found the district's new data tools valuable, or because they had attracted special favor from the district as the result of being known.In contrast, early career principals (1-4 yrs.) reported significant amounts of conflict with district directives, even when outside agencies-the state, a partnering CBO, an SSO network leader-supported their professional judgments (-.486, p=. 05).
Inquiry teams were a popular topic among the experienced principals who reported bureaucratic and professional alignment (24%; 4 incidents from 4 principals). 19These mandated school-based teams of educators were intended to support principals in expanding the school's "sphere of success," by serving as "a vehicle for distributing school leadership broadly" (Talbert et al., 2009, p.5).The DOE refers to inquiry teams as one of each principal's accountability tools.The inquiry team concept is disseminated and supported by SSOs.
According to experienced principals in our study, inquiry teams helped to spread out the responsibility for meeting the ambitious graduation rate targets that were crucial to each principal's career because principals' performance was being measured by such targets.Principal Morrison explained how he responded to this incentive: "[My] inquiry team work[s] with a group …of students who have been chosen based on the fact that they are not on track [to graduate] but should be.
[W]e're trying to figure out what it is that's keeping them off track."His strategy was to explain "what [the inquiry team] was about and [ask] them if they'd be interested to submit a letter to me indicating they're interested and set up a meeting…[P]eople would rather get paid for something they want to do rather than something they're not that invested in.And I was fortunate that the two people that wanted to do this were very gung-ho."For Morrison, the crucial task was to identify teachers "really interested in finding the reasons that these kids are having these problems" so that 19 Inquiry teams were developed as part of a process for school improvement funded by New Visions for Public Schools, the sponsor of the NVPSO PSO.The DOE streamlined and adopted the process as a requirement for all schools in 2007-08 (Robinson et al., 2008).Even so, of the four principals whose incidents make up this sub-group, only one was affiliated with NVPSO.
he could count on them to help him meet the DOE's targets.Smaller numbers of experienced principals mentioned other SSO-supported accountability tools as bureaucratic support for their professional judgments including student-level data bases, building councils, and School Quality Reviews. 20nother group of experienced principals (18%; 3 incidents from 3 principals) addressed the special favors that can be had when district administrators know a principal well.Principal Eliot described how 'relationships' can facilitate luck and how a good school becomes better as a result, "The superintendent at that time from [my] region was approached by [a woman] who runs [a business training institute] and she said, 'I have a program ending at a school.Would you like to have it at one of your schools?'…He turned to her…in the elevator and said, 'Yes.Let's put it at that school.'…Andjust from that conversation in an elevator [we have] dramatically altered the lives of some of these kids.They've gone to Austria.They've gone to California…And they have won the [borough] Championship three years in a row now." Many early career principals reacted quite differently to DOE rules and regulations.They were suspicious of the DOE motives and unprepared to use accountability tools to the same extent.Other research has speculated that since new high schools are opened every year and principals with scant teaching experience are frequently hired to run them, these principals simply do not know what is needed, or how to make use of the DOE's mandated accountability tools (Hemphill et al., 2010).Our research shows that these early career principals may have tried to circumvent DOE's rules and tools when conflicts arose between them.
Principal Salinger gave a typical incident (86%, 6 incidents from 5 principals) in which conflict with the DOE was only resolved by finding support outside of the district (-.486, p=.05).
Aiming to offer internships to potential dropouts so they might stay in school, she grew frustrated by DOE's seat-time rules.She found herself "talking to other principals" asking, "You do internships…How is it that you [are allowed to] let your kids go every Wednesday?How do you do that?"One replied, "Oh, I wrote to the state [for a waiver]."Given this simple workaround, Principal Salinger wondered, "Why is this [waiver] a secret?No one [from the DOE or SSO] ever told us this.There's no manual for principals.You really have to do all the grunt work yourself.Nobody tries to help you" More than querulous, the experience made her mistrustful even after she got the waiver."As a matter of fact, I didn't even inform my superintendent." The average achievement level of the school's student body also influenced how a principal's job experience affected his perception of conflict with DOE.The school's Peer Index helped to explain a large proportion of the conflict that even experienced principals had with DOE directives.The more accomplished the students, the more likely its principal was to perceive conflicts between DOE's requirements and his professional judgments.For principals leading schools with Peer Indices in the 3 rd quartile, the unerring focus on meeting minimum test score targets and graduating all students, as well as mandates to use the accountability tools, put unacknowledged limits on expectations for higher performing students.
Principals of schools with 3 rd quartile Peer Indices recounted significantly more conflicts between their bureaucratic obligations to DOE and their professional commitments (.508, p = .05)and also three-way conflict between bureaucratic, professional and political influences (.510, p=. 05) than did principals leading schools with students who tested lower on the Peer Index scale.The largest group of such incidents dealt with a clash between the DOE's many expectations and principals' time and resource constraints (38%; 11 incidents, 7 principals).For example, experienced Principal Tan would have preferred a mandate that gave him a high achieving "bottom line, but otherwise left him alone.""I wonder why they just don't say to me, 'You're accountable for graduating your students…You do whatever you feel you need to do to get that job done.And that's how we'll hold you accountable.'"In his view, "That really is the crux of education.It's not all these assessments.It's not all these compliance issues…Leave me alone.[If left alone,] I can be in classrooms.There can be more fruitful discussions about instruction.All this junk takes you away from that….giving me a 'B', giving me a 'C' [on the Progress Report]…[This type of] accountability is not for my purposes; it's for their purposes.It's for whatever data or statistics they want to show."Parents of high-achieving students, and even the media, were also construed as a political pressure to provide programs and services that DOE did not reward.As Principal Ellison put it "Sometimes I feel as though I'm sitting in the firing line, the only one in the middle, and everybody else is shielded.And it's a very lonely position." Another source of conflict was the clash between the principals' developmental goals for students or teachers and the DOE's emphasis on outcomes (21%, 6 incidents from 5 principals).Principal Bronte's explanation of the tension was typical."You have to make decisions between how much test prep you're going to do versus how much instruction you do…[that] may not necessarily benefit the students for that particular test, but you feel it is the best thing educationally." 21 Helping to make the case that the school's educational challenge interacts with other perceived deficits, Table 6 shows that principals whose schools were in the bottom or 2 nd quartiles on the Peer Index told us significantly more incidents in which the DOE's rules supported their professional judgment, but only when it was external actors who caused political problems for the schools (-.476, p=.05).In one example of how this complex interaction worked (5 incidents from 4 principals), early career Principal Ishiguro recounted how it took three years for her to make an important change in her external partner, although it delayed needed improvements in her lowperforming school."While sitting as a principal for three years, I knew that [my partnering organization] wasn't delivering services the way I felt that they should…Their [staff] attrition was relatively high and [each new employee] …knew less than the prior person.But she pursued a change only when DOE created an open enrollment opportunity, validating her instincts."[When the DOE's] RFP process opened up and we were on the list to pick [a new partner]… I already knew who I had in mind.I had five top contenders and I went for it." As the patterning of these incidents shows, experienced principals of high achieving schools conceptualized their job more broadly than did the DOE, chafing at the limitations imposed on their discretion that were implied by high-stakes performance targets aimed at low-achieving students.But when experienced principals led low achieving schools, such conflicts nearly disappeared; instead they valued the data tools that the district provided to meet its targets because these tools helped them enlist teachers and other staff in the work of diagnosing and supporting students.
On the other hand, early career principals often felt they lacked the resources and support they needed.They also lacked the confidence to use the accountability tools as readily as their more experienced counterparts.They may have had insufficient experience to sort through the 21 Principals leading schools with relatively high performing students also told us more complex incidents in which bureaucratic accountability conflicted with their professional judgments and simultaneously with what some parents wanted from the school (.510, p =.05).These six incidents from five 3 rd quartile Peer Index principals each represented one-off content.
responsibilities of managing relationships with parents and other external actors without the active intervention of DOE.These findings are especially troublesome because early career principals led twice as many of the low-achieving schools as did experienced principals in this sample, and they led somewhat more small schools.

Policy Implications
In theory, all New York City high school principals were to be advantaged by the small school and SSO policies intended to implement the system's Diverse Provider Regime.Principals were targeted as politically powerful actors.Whether new or experienced, principals were envisioned as sufficiently competent to benefit from the additional discretion given them.Student outcomes were assumed to result from their choices, albeit steered by a set of sharp consequences.In political terms, the district enhanced principals' blocking power22 by making them responsible for establishing the instructional, disciplinary and cultural norms of the school, and reinforced that power by allowing them to screen students and select a like-minded School Support Organization (Schneider & Sidney, 2009).
Yet the evidence from this study suggests that most principals did not react to the incentives built into these policies as intended.Most principals stuck to known support systems rather than taking risks with their new powers.They felt beleaguered rather than empowered.Some of the riskaverse principals voiced concerns that New York City's high stakes accountability context could cut short their careers as likely as it might reward them.But another source of their reactions was Mayor Bloomberg's efforts to establish a new activist governing regime of diverse educational providerssmall schools and SSOs-which focused on raising the bottom of the achievement spectrum.
New York City's high school principals balanced an array of demands and expectations emanating from the district office, parents, teachers, collaborating organizations, their training, professional experience and moral commitments as well as competition from other schools.Among the principals interviewed here, these multiple accountabilities were about twice as likely to be in conflict with one another as to be unopposed or reinforced.Our analysis shows too that while accountability conflict was widely experienced, it was most acute among early career principals leading low achieving small schools.This remains a growing category in the city because new small schools are created to supplant large schools with low-achieving students.
Conflict among differing types of accountability was the most evident indicator of beleaguered principal perceptions; tone was a second indicator.While experienced and early career principals both reported conflict when they competed for students and teachers, choosing their own SSO was a more difficult task for early career principals leading new small open-enrollment schools than it was for experienced principals leading large, politically embedded neighborhood schools.Small school principals were often heading schools with no established clientele and unknown teacher quality; and small schools proved more vulnerable to test score volatility.Their principals found it difficult to anticipate the educational services they would need.
To complicate matters further, many principals understood that high-stakes rewards focused on the bottom of the achievement distribution, thus frustrating those who chose to service the learning needs of higher achieving students, whether in small or large schools.Disillusion set in among some who felt they were not rewarded for their efforts to improve education for higher achieving students.
For such reasons, the majority of principals in this study (60%) described New York City's Diverse Provider Regime as fraught with conflicting accountabilities.The combination of variable sized schools, and multiple service providers in a high-stakes accountability environment created an abundance of no-win decision situations in which a series of bad choices had to be made.
As anticipated by research on mid-level managers (Osterman, 2009), this beleaguered group also responded to queries about their accountability with stress, insecurity, and mistrust.Phrases like "scary and disgusting", "threat", "on the firing line", "very lonely position", "not often respected" peppered their descriptions of their job and the choices they were asked to make.For these principals, empowerment required more than freedom to make decisions, and rewards for good ones.Many describe the need for district support services and external resources, which are not yet available.
The high principal transience found in this study and by others may be partly explained by these findings.Our research suggests that being the principal of a small high school with students chosen from throughout the city may diminish a principal's political resources, when compared to neighborhood schools that have large student populations and local constituencies.These principals may quickly become overwhelmed by conflicting demands and feel isolated -even targeted-when they discover they have few political resources in the school community with which to push back at district policy.
Should our research proves reliable when replicated, we anticipate that Diverse Provider Regimes like New York City's will systematically disadvantage early career principals who lead small schools, setting them up for disillusion, stress, mistrust, and extreme transience.At the same time, experienced and inexperienced principals alike may become disillusioned because they are no longer rewarded for providing a rich educational environment to higher achieving students.Under this set of circumstances, the Diverse Provider Regime could be said to provide too many choices while offering insufficient support.Like market systems elsewhere, Diverse Provider Regimes require a delicate balance of regulation, information, professional services and risk-taking.Empowerment policies such as those studied here appear insufficient to the task of improving the confidence and competence of school leaders, even as they are identified as the key decision makers on behalf of students.
But what can we learn from those who did report being empowered?Should they not be the policy focus?Whether leading large or small schools, the principals who chose to affiliate with either NVPSO or the now defunct ICILSO communicated a sense of confidence, control and creativity.Those principals conveyed their decisions as over-determined, rather than conflicted.Although the small school principals among them relied less on their professional judgments as a touchstone for their decision-making than did the large school principals, making choices about how to proceed was not difficult because most incidents were construed either as the alignment of multiple sources of accountability or simply uncontested, legitimate demands.Either way of framing critical incidents led to a single decision in which these principals voiced great confidence.In their separate ways, both groups fit the expectations of empowered professionals unperturbed by either the demands of the district or market competition.
Notwithstanding these principals relatively high transience (30%), advocates of the New York City Diverse Provider Regime might expect this group to become model intrapreneurs; the core leaders of a decentralized system of schools.Given the high interest in principal preparation and mentoring (The Wallace Foundation, 23 August, 2011), should not these principals guide our preparation of others?Should we teach new principals to recognize and compartmentalize multiple, conflicting accountabilities until they have sufficient on-the-job experience to make good use of the resources provided?Might it be possible for beleaguered, but experienced, principals to learn self-confidence and the ability to reframe problems into opportunities from their empowered large school counterparts?If these are the tasks required to expand the empowered minority of New York City principals into a majority, they are daunting, and will require outsized resources of their own.
But that appear doubtful now.Not only are the theoretical benefits of market governance hard to capture, markets involve well-known tradeoffs.Market formation as a policy instrument, and market-based governing regimes in general, have the benefits of being responsive to local needs and preferences, flexible and open to innovation while (often) promoting local community spirit and solidarity.But their weaknesses are also well known.Chief among them is their inequality for service recipients, and sometimes for service providers as is evident among the principals studied here (Howlett & Ramlesh, 1995).
Markets also have other, less discussed, weaknesses that would hamper efforts to create a large pool of empowered principals in New York City.Market-based systems require high levels of trusted information to operate efficiently and effectively.For example, principals would need to know a great deal about their students' competencies and their teachers' skills in order to predict the resources needed to improve both.They also need much information about the range of services available, their relative efficacy and the trade-offs in costs.Parents would need to know a great deal about how principals make decisions and how data is used to make effective choices about the right schools for their children.District officials would need to use reliable and valid measures of school performance that do not skew principal decisions in inappropriate ways, and they need to apply their sanctions and rewards transparently so that principals and the public both trust the outcomes.Moreover, markets are frequently more costly because of the loss of economies of scale brought about by local responsiveness and decentralized purchasing (Howlett & Ramlesh, 1995).Scaling up to the level of mass service provision is something markets do poorly without large infusions of wealth and well-founded public trust.Given enough trust and enough public wealth, markets can operate efficiently, but neither wealth nor trust seems likely to grow under the current politics of austerity, budget cuts and public mistrust of big institutions.
In sum, market-based governing regimes have theoretical assumptions that can be thwarted by their own weaknesses.New York City's Diverse Provider Regime may be one case in point.If it is to avoid this fate, our research suggests that much more needs to be done to provide most, if not all, high school principals with the attitudes and skills currently held by a minority.

Table 1
Four Ideal Types of Governing Regimes and Their Theoretical CharacteristicsSince taking control of New York City's schools in 2002 Mayor Michael Bloomberg has created a particular type of Market Regime.Table2illustrates how Bloomberg's version of a Diverse Provider Regime borrows from both corporate-style market regimes and entrepreneurial ones.

Table 2
Variations in Market Regimes

Table 5
Description of the Sample

Table 6 Critical
Incidents by Principal

Table 8
Chi Square Tests of Independence Between Five Explanatory Variables and Two Dimensions of Accountability a Unweighted types produce similar results; * p<.05 There have been times when I've had to go over [that number].There are times when I've had to give a teacher 34 kids and I go to that teacher and say, 'I'm going to do everything that I possibly can to reduce the numbers in your class.Please bear with me; it may be like this for a week or two until I can figure out the numbers.'Andthey're like, 'Okay, [Principal Salinger] because we know that just because you can you don't pack out classes.'She was well aware that her ethical commitment helped students indirectly, by improving teachers' professional lives."Because you want teachers to be able to get that hands-on experience with [each] child, to sit and help, and go and circulate [in] the classroom.How can you do that with 34 kids in a class?...Sure, you can get away with it.It couldn't be grieved [by the union].But morally is this what's right for kids?No, it's not."