Institute for Policy Research Working Paper
High Stakes Accountability in Urban Elementary Schools:
Challenging or Reproducing Inequality?1
John B. Diamond
James P. Spillane
1 Work on this paper was supported by the Distributed Leadership Project which is funded by research
grants from the National Science Foundation (REC-9873583) and the Spencer Foundation (200000039).
Northwestern University's School of Education and Social Policy and Institute for Policy Research also
supported work on this paper. All inquiries about this research project should be directed to the study’s
Principal Investigator, James Spillane at Northwestern University, 2115 North Campus Drive, Evanston, IL
60208-2615 or firstname.lastname@example.org. All opinions and conclusions expressed in this paper are
those of the authors and do not necessarily reflect the views of any funding agency or institution.
In this paper, the authors use data from interviews and observations in four urban
elementary schools – two high and two low performing – to examine how schools
respond to high stakes accountability policies. The authors argue that school responses to
high stakes accountability depend on school context. In low-performing schools,
responses focus narrowly on complying with policy demands, focusing on improving the
performance of certain students, within benchmark grades, and in certain subject areas.
In contrast, higher performing schools emphasize enhancing the performance of all
students regardless of grade level and across all subject areas. Given the concentration of
poor students and students of color in the lowest performing schools, the authors
conclude that issues of educational equity need to be given careful consideration in the
implementation of high stakes accountability policies.
One of the most consistent findings in educational research is that family
background is linked to children's educational outcomes, attainment, and adult
occupational status (Blau and Duncan, 1967; Coleman et al., 1966; Jencks et al., 1972;
Orfield, 1993). While education is viewed by many as an important mechanism for social
mobility – tied to the “common school” belief that all children should have equal
educational access and opportunity (Cremin, 1951; Muller and Schiller, 2000) – many
scholars argue that schools reproduce social inequality. Over the past decade, policy
makers have mobilized an arsenal of policy instruments in an effort to ensure that all
children receive high quality education. One increasingly popular but controversial
strategy relies on external accountability mechanisms, including high stakes testing, to
transform instructional practices and make teachers and students more accountable for
their performance. Critics argue that these policies will exacerbate inequalities by leading
teachers to marginalize low-performing students (Clotfelter and Ladd, 1996; McDill,
Natriello, and Pallas, 1986) and causing teachers to teach these students only the material
covered on standardized tests. Others contend that such policies are misguided because
limited resources, unprepared teachers, and ineffective instructional practices rather than
incentives are the problem that needs to be addressed (Darling-Hammond, 1994).
Proponents argue that such assessments will reduce gatekeeping processes such as
tracking and low teacher expectations that disadvantage certain students. External
assessments, it is argued, provide objective information for school-based decision-
making and therefore work against more subjective judgements that contribute to
stratification (Muller and Schiller, 2000; Coleman, 1997). For example, supporters argue
that teachers’ assessments of students’ ability as well as decisions about course placement
and grouping arrangements inside classrooms could be based on more objective
information from standardized tests (Muller and Schiller, 2000).
Given the increasing emphasis on external assessment and accountability, and the
strong arguments on both sides of the issue, it is interesting that with a few notable
exceptions (Muller and Schiller, 2000; Roderick, Bryk, Jacob et al., 1999) research on
how these policies unfold in schools and the mechanisms through which they impact
student learning has been limited. Moreover, few studies have explored how the
implementation of these policies may be situated in certain school contexts and how this
may influence their impact on students.
The study that is reported in this paper examines the implementation of high-
stakes testing in high and low performing schools. We seek answers to two related
questions. First, is high stakes accountability policy perceived and implemented
differently in high and low-performing schools? Second, does the implementation of high
stakes testing in these schools suggest that it will reduce social stratification through the
mechanisms outlined by testing proponents?
High Stakes Accountability and Stratification
One aim of accountability policies is to insure that all students receive high
quality instruction and reach a certain level of competence in core subject areas (Muller
and Schiller, 2000). Some districts, like Chicago, have adopted a high stakes version of
these policies that link student performance on examinations to consequences for schools
and, in some instances, students themselves. Opponents of these policies argue that for
these approaches to be fair, instructional changes should precede consequences for
students (Heubert and Hauser, 1999) and that such policies create incentives for
marginalizing low performing students (Clotfelter and Ladd, 1996; McDill, Natriello, and
Pallas, 1986). Neutral observers caution that those implementing such policies must
insure the adequacy of educational resources for the tested students, attend to the
reliability and validity of the exams for their intended purposes, and avoid basing
decisions on one test (AERA 2000).
Proponents of these policies suggest that they can reduce inequality through
increasing student motivation, creating incentives for teachers to seek improvements in
student outcomes, providing more objective information about students’ performance for
school based decision-making, and increasing academic press in schools – particularly
those serving low-income and minority students (Coleman, 1997; Shouse, 1997; Muller
and Schiller, 2000).
Coleman (1997) advocates the development of output-driven schools, a key
component of which would be external assessment and accountability. He argues that
external assessments such as student performance tests would create new incentives for
school improvement, providing objective information for teachers to assess students and
make course placement decisions, thereby reducing the gatekeeping functions of schools
(Coleman, 1997; Muller and Schiller, 2000). Shouse (1997) argues that designing more
output-driven schools would also increase academic press and have particularly
beneficial consequences for students with lower socioeconomic status. Therefore,
proponents of these policies suggest that three key mechanisms – the creation of new
incentives, the provision of objective information for school decision making, and the
increase of academic press – will combine to reduce schools’ and teachers’ gatekeeping
practices and contribute to a reduction in stratification.
While arguments have been forwarded in support of and in opposition to these
policies, surprisingly little research closely examines how they play out in schools. The
data on outcomes that does exist presents a mixed picture. Data on the implementation of
high stakes testing in Chicago suggests that the percentage of students meeting minimum
competency requirements has increased since the introduction of the policy (Roderick et
al., 1999). However, the policy has differential impacts on students based on their family
background characteristics. For example, African American students were retained at a
much higher rate than their white and Latino/a counterparts because they tend to score
lower than whites and because of the higher proportion of Latino/a students who are in
bilingual programs and therefore exempt from the policy (Roderick et al., 1999). This
likely has a stratifying effect for African American students because grade retention may
be associated with negative long-term outcomes including reduction in self-esteem and
increased likelihood of high school drop-out (Roderick, 1994).
Muller and Schiller (2000) examine how state-level testing policy impacts high
school students’ graduation rates and mathematics course-taking. They show that these
policies equalized students’ academic attainment and reduced the impact of teachers’
gatekeeping through low-expectations, seeming to support the arguments of supporters of
these policies (Muller and Schiller, 2000: 210). However, they also find that when state
tests link students’ performance to consequences for schools it leads to stratification
based on SES, lending support to the arguments of opponents of testing policies.
Therefore, their findings do not strongly support the arguments of proponents of these
policies or their critics. They recommend more research using both qualitative and
quantitative methods to explore the mechanisms through which these policies influence
teachers’ practices and students’ outcomes. Research into the processes of accountability
policy implementation will help inform this discussion.
In this paper, we argue that in order to understand the implications of these
policies it is important to examine how they are understood and implemented in
particular school contexts. In the current paper, we examine schools’ responses to high
stakes accountability policy, paying particular attention to the implications of these
responses for issues of educational equity both within and across institutions. More
specifically, we examine how teachers and administrators in high and low performing
schools respond to high-stakes accountability policy focusing on their responses to
incentive structures, their interpretation and use of test score data, and their subsequent
instructional priorities. We argue that these responses to high stakes accountability are
situated in a school’s status with regard to accountability policy – probation versus high
performance – and argue that school status is correlated with students’ race and social
class. We conclude that differences in responses to accountability policy in different
types of schools may increase rather than reduce educational stratification.
The paper is organized as follows. First, we outline the theoretical tools used to
frame our discussion. Following this, we discuss the methodological approach that
guided the research. We then examine differences in schools’ responses to high stakes
accountability in four elementary schools – two high performing schools1 and two
probation schools. In the final section we discuss these cases, paying particular attention
to the implications of school-level responses to high stakes testing for issues of
stratification. We argue that the picture emerging from these cases suggests that the ways
in which these policies are implemented in particular school context may exacerbate
rather than reduce educational stratification.
Research on the role of family background and educational stratification
demonstrates consistent links between socioeconomic status and students’ outcomes.
Some explanations for this pattern focus on direct effects of family background such as
class-based disparities in parents’ beliefs and involvement patterns (Lareau, 1989; Sewell
and Shah, 1968 (a), Sewell and Shah, 1968 (b); Sewell and Hauser, 1980), family
structure (i.e. number of parents in the home or the number of siblings in the family), and
access to extra-familial resources through parents’ social networks and institutional
affiliations (Coleman, 1988; Coleman and Hoffer, 1987; Wong, 1998; Carbonaro, 1998;
Hao and Bonstead-Bruns, 1998; Hofferth et al., 1998; McNeal, 1999; Wilson, 1987).
Other scholars focus on the interaction between these background characteristics and
school practices. In these accounts, schools impact students though micro-political
processes such as low teacher expectations (Roscigno, 1998; Brophy and Good, 1973;
Rist, 1970, 1977; Rosenthal and Jacobson, 1968), tracking (Oakes, 1995), and cultural
reproduction processes (Bourdieu, 1979; Bourdieu and Passeron, 1977).
A third approach – institutional stratification– highlights the implications of inter-
organizational processes for social stratification (Roscigno, 2000). Because race and
social class shape school attendance patterns and contribute to the creation of highly
segregated school contexts (Orfield, Bachmeier, James, and Eitle, 1997), family
background can contribute to stratification through the distinctly different characteristics
of the schools students attend (Roscigno, 2000). Differences in schools’ monetary
resources (Elliot, 1998; Greenwald, Hedges, and Laine, 1994; Hedges and Greenwald,
1996; Kozol, 1991), instructional quality (Smith, Lee, & Newman, 2001), presentation of
valued knowledge (Anyon, 1991) course offerings (Ayalon, 1994), and social
organization (Bowles and Gintis, 1976), all contribute to the maintenance of social
Roscigno’s institutional stratification perspective captures these inter-
organizational dynamics, emphasizing the “multi-level and inter-institutional nature of
racial educational disadvantage” (Roscigno, 2000:271). Roscigno writes:
Arguably, the most important of these inter-institutional linkages in relation to
race/class reproduction in education has to do with family background inequalities
and their consequences for achievement through the character and resources of
the schools one attends. What this means, more straightforwardly, is that family
background shapes residential options. Where one resides, in turn, has a large
impact on the school one attends and, consequently, achievement [emphasis in
original] (Roscigno, 2000:271).
He concludes that residential segregation leads to indirect effects of family background
through enrollment patterns in public versus private schools, the race and social class
composition of schools, monetary expenditures, and school climate (Roscigno, 2000).
Other work demonstrates that the concentration of low-income and African American
students in certain schools may have detrimental implications for student outcomes apart
from the individual characteristics of students (Bankston and Caldas, 1996). Taken
together, this work demonstrates that educational stratification based on race and social
class is at least partially maintained through race and class-linked institutional processes.
Building on this work, we argue that the implications of policies like high stakes
accountability are also shaped by institutional stratification processes. In contemporary
urban contexts there are several types of schools including private schools (both religious
and non-religious) which are often thought to be the highest quality, magnet schools
which are often considered the “elite” public schools, and neighborhood schools which
can be further divided into high and low quality categories. Social class and race are
important in patterning the schools that children attend, with the more highly valued
settings being most accessible to middle- and upper-class children. The different types of
public schools are likely to implement the policy differently. Therefore, if students are
concentrated in different types of schools based on race and social class, they will be
impacted by the policy in distinct ways.
Taking the Chicago Public Schools as an example, data2 shows that schools on
academic probation have a higher percentage of African American and low-income
students, on average, than the typical Chicago Public School. African Americans make
up 52% of the district’s student population but 83% of students attending probation
schools. Likewise the district average of low-income students is 84% while the average
for probation schools has 92% low-income students. Perhaps most interesting, however,
is the fact that while white students make up 10% of the district student population, they
make up less than 1% of the students attending elementary schools on probation.
These figures are more striking when compared to data from Chicago magnet
schools which consistently rank among the district’s highest performing schools.
Elementary magnet schools contain 55% low-income students (compared to 92% for
probation schools) and 53% Black students (compared to 83% for probation schools).
White students make up 17% of elementary magnet school students (as compared to less
than 1% in probation schools).3 These data demonstrate that structural processes related
to family background shape students’ access to schools of different quality. African
American and low-income students are more likely to be found in the lowest performing
schools while white and middle income students are more likely to be found in higher
performing magnet schools.4 Table 1 presents the mean percentage of students by race
(black/white) and social class in the Chicago district as a whole as well as its magnet,
high performing, and probation elementary schools.
[insert table 1 here]
The multiple factors that contribute to this process are beyond the scope of this
paper, however, as we shall see, the observed patterns likely have important
consequences for students’ educational experiences.5 In the discussion that follows, we
focus on school status in relation to high stakes accountability – probation versus high
performance – as an important factor that shapes schools’ responses to accountability
policy and students’ access to educational equity.
This paper is based on data from the Distributed Leadership Project, a four-year
longitudinal study of elementary school leadership funded by the National Science
Foundation and the Spencer Foundation. The project began with a six-month pilot phase
during the Winter and Spring of 1999 involving seven Chicago elementary schools, four
interview only sites and three schools where we conducted interviews and extensive
fieldwork. The first full year of data collection (Phase 1) began in September 1999 and
focused on eight Chicago elementary schools, two of which were also part of the study’s