School Restructuring as a Policy Agenda: Why One Size May Not Fit All

Author(s): Lawrence J. Mussoline and Roger C. Shouse

Source: Sociology of Education, Vol. 74, No. 1 (Jan., 2001), pp. 44-58
Published by: American Sociological Association
Stable URL: http://www.jstor.org/stable/2673144
Accessed: 02/06/2010 20:02
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.
Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
http://www.jstor.org/action/showPublisher?publisherCode=asa.
Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.

American Sociological Association is collaborating with JSTOR to digitize, preserve and extend access to
Sociology of Education.

http://www.jstor.org

School Restructuring as a Policy

Agenda:
Why One Size May Not Fit All
Lawrence J. Mussoline
Pine Grove Area School District

Roger C. Shouse
Pennsylvania State University
Recent studies of school restructuringhave suggested that a school's shift toward
communal norms is likely to be accompanied by particulartypes of technical
reform. The result has been the emergence of a prescriptive agenda of reform
practices in the school-improvement literature,despite evidence that the usefulness and effectiveness of instructional practices vary across contexts. The study
presented here examined the relationship between school restructuring(defined
in terms of the use of "restructured"practices) on mathematics achievement
across categories of school socioeconomic status. The findings raise questions
about the suitabilityof defining restructuringin terms of any specific reformagenda and indicate the need for caution regarding the broad implementation of such
an agenda in disadvantaged schools.

_ he concept of restructuring
generally connotes a school's shift
away from bureaucraticnorms,
structures,and practices,and this shifthas
been sketchedalongtwo distinctand independent dimensions.One dimension,the
socialor normativedimension,conveysthe
idea that schoolscan increasetheir power
as institutionsby reinventingthemselvesas
supportive,collegial,and moralcommunities (Brykand Driscoll1988; Grant1988).
The other, the technicaldimension,conveysthe ideathatschoolscan improvetheir
academiceffectivenessby replacing"traditional"curricular
and instructional
practices
with more cooperative, integrated, and
constructivist approaches (Center on
Organizationand Restructuring
of Schools
1992; Newmann and Wehlage 1995).
Withinthis framework,it made sense, for
U

example,to view Catholicschoolsas models for the normativeshift,even given their

generallytraditionalcore of technicalpractices. At the same time, many public
schoolscould pointto technicalreformsas
evidence of restructuringbut remained
largelyunchangedin termsof theirnormative or moralimpacton students.
Over the past decade, however,some
studieshavetracedan associationbetween
these two dimensions,suggesting that as
schools become less bureaucratic and
more communal,they are also more likely
to revisetheir curricularand instructional
practices. Among the implicationsthat
have been drawn from this assertionare
that evidence of technicalrestructuringis
also evidence of normativerestructuring
and that the overallconcept of restructuring can be operationalizedin terms of a

Sociology of Education 2001, Vol. 74 (January):44-58

44

School Restructuringas a Policy Agenda

school's adoption of particularcurricularand
instructional practices (see Lee and Smith
1995).
Such conceptualizations, however, appear
to sidestep some nagging questions about
school reform and its relationship to school
effectiveness. For instance, may schools that
are shifting toward more communal relationships and are rethinking their instructional
practices be just as likelyto adopt a "Catholic
model"? That is, rather than implement
extensive, sometimes costly and complex
reforms, may they opt for incremental
improvement, perhaps, for example, by
increasing their emphasis on the importance
of traditional academic work for all kinds of
students? Rephrasingthe question, one may
ask, Do schools have to buy in to a particular
agenda of technical reform to be considered
restructured?
Another key question that is often dodged
in the literatureon restructuring,and the one
highlighted in this article, is, To what extent
should one expect the technical reforms
linked to restructuringto produce consistent
achievement effects across all types of
schools? Given the body of research indicating that the relationships between students'
learning and organizational or instructional
practices often vary dramatically,depending
on other school characteristics,such as average socioeconomic status (SES), urbanicity,
and academic press, should one not also
expect restructuringeffects to vary in a similar fashion? In fact, several studies have suggested that low-SES urban schools attain
greater effectiveness and higher achievement
through the use of more traditional types of
organizational and curriculararrangements,
practices, and understandings (Hallinger and
Murphy 1986; Shouse 1996, 1998).
On the basis of this line of evidence, our
study examined the degree to which achievement effects associated with being a restructured school (defined in terms of the use of
specific school practices) vary across lines of
school affluence. Using a sample of public
schools drawn from the National Education
Longitudinal Study of 1988 (NELS:88), we
found these effects to vary in magnitude and
direction, depending on a school's average
socioeconomic level. Specifically,for the pub-

45
lic schools represented in our sample, we
found restructuringto have (1) a significant
positive achievement effect in average SES
schools; (2) no significant achievement effect
in high SESschools; and (3) a significant negative achievement effect in low-SES schools,
especially very low-SES schools. Contrasting
sharply with studies that have reported overall positive restructuringeffects (but that have
overlooked important differences in effects
across categories of school SES),our findings
raise questions about the validity of defining
restructuringin terms of a single list of practices. They also demonstrate the need for
local school policy makers to exercise some
caution when considering the adoption of
restructuring'stechnical reform agenda.

Two Views of Restructuring

Though generally conceived as representing a
"purposeful change" in patterns of rules,
roles, and relationships in schools (Corbett
1990), the concept of restructuringhas been
more recently cast with greater precision and
at multiple organizational levels. Forexample,
as indicated in the opening paragraph, some
researchers have highlighted restructuringin
terms of changes in school norms, particularly those governing school social relationships.
Restructured schools are said to evidence
decentralized forms of decision making that
are informed by a heightened sense of collegiality among teachers (Darling-Hammond
1995; Murphy 1991). Teachers in restructured schools are said to take greater responsibility for students' learning, and the school
culture is said to be marked by more personalistic relationships and a stronger ethos of
caring (Louis, Marks,and Kruse1996). In this
sense, studies by Brykand Driscoll(1988) and
Grant (1988) have been important in laying
out the theoretical foundation for the normative dimension of school restructuring.
In contrast to this conceptualization, however, more recent scholars have viewed
restructuringin terms of a change in instructional focus and technique. In restructured
schools, they have argued, classroom activities and teachers' evaluations of students
become more cooperative, "authentic," and
aimed more at developing "thinking skills"

46

Mussoline and Shouse

than at accumulatingspecificcontent knowl- problem, at least in secondary studies that

edge (Elmore 1990; Murphy 1991; have used large-scale national data that
Newmann 1996). Although such ideas are include items representing specific instrucnot new (they fit well withinthe classicideal tional practices deemed to be "restructured."
of progressiveschooling), they received a Here, however, studies of the direct impact of
good deal of impetus from the Carnegie such practices have produced mixed effects
Councilon AdolescentDevelopment's(1989) at best.
report on middle school education, which
Lee and Smith (1 995) attempted to pull
calledfor such innovationsas team teaching, the technical and social dimensions of
cooperativelearning,and the eliminationof restructuring together in a single analytic
framework. At the heart of their analysis resttracking.
Althoughthere is some overlap between ed the assumption that normative change
these two understandingsof restructuring, tends to accompany technical change. Lee
we viewthem as clearlydistinct.Forexample, and Smith thus viewed a school's adoption of
increased collegialityamong teachers may "restructured practices" as evidence of its
lead to team teaching.On the other hand, it shift toward more collegial, communal, and
can cer- caring relationships. After identifying indicamay not, and school administrators
tainlyimpose team teachingon their profes- tors of 12 such practices in the NELS:881Othsionalstaffs-perhaps to increasecollegiality, grade data set, they categorized schools as
perhapsto demonstrateauthorityor confor- restructured if the schools had implemented
mance with current popular educational at least 3 of the 12 items (as indicated pritrends.The distinctivenessof the two dimen- marilyin principals'reports-see Table 1 for a
sions is also revealedin the various"effective list of these items).1 Among the results of
schools" studies of the 1980s, which called their analyses was the finding that students
not only for the expansion of traditional who attended restructured schools attained
school practices and characteristics(more significantly higher levels of achievement
stringent course requirements, "direct than did those who attended more traditioninstruction,"and the like),but for morecom- al types of schools. Moreover, the gaps in
munal features (such as having a common achievement between students from different
academic curriculumfor all students and a socioeconomic backgrounds were significantstronger"senseof community"in the school) ly smaller in the restructuredschools.
Most of the 12 practices listed in Table 1
(Good and Brophy 1986; Hallinger and
relate to the core technology of schooling
Murphy1986).
Both understandingsof restructuringrest (that is, they involve organizational or curricon the belief that it can help raisethe level ular arrangements for delivering instruction
and qualityof academicachievementamong to students). Some relate more to the estabAmericanstudents (Newmann 1992, 1996). lishment of teachers' collegiality and empowTestingthis belief, however,requiresa satis- erment or students' and parents' attachment
of the concept, so to the school. And though it is plausible that
factoryoperationalization
its achievement effects can be empirically several of the practices may be connected to
measured. Bryk and Driscoll (1988), for students' achievement, empirical evidence of
example,operationalizedcommunalityalong their impact continues to be either mixed or
three dimensions(common beliefs and val- contested. It is important to understand,
ues, a common agenda of activities,and a however, that Lee and Smith's (1995) argupervasiveethos of caring)using an arrayof ments about restructuringdo not depend on
items from the High School and Beyond the direct achievement effects of any practice
study. Devising an index to measure the listed in Table 1. Nor did Lee and Smith claim
degree of communalitywithin the sampled that schools can improve themselves by
schools,they reporteda positiverelationship adopting any three of these practices. Rather,
betweencommunalityand students'achieve- they suggested an indirect relationship, argument. In terms of the technical view of ing that the presence of some reasonable
restructuring,operationalizationis less of a number of these practices signifies a school's

School Restructuringas a Policy Agenda

47

Table 1. NELS:88First Follow-up Survey Items Representing Restructured Practices

Students keep the same homeroom throughout high school.
Emphasisis placed on the staff solving school problems.
Parentsvolunteer in school.
Interdisciplinaryteaching teams are used.
There is independent study in mathematics and science.
There is independent study in Englishand social studies.
There are mixed-abilityclasses in mathematics and science.
There is a focus on cooperative learning.
Students' evaluations of teachers are important.
The school-within-a-school organizational design is used.
Teacher teams have a common planning time.
There is flexible time for classes.
Source: Lee and Smith (1995).

shift toward a more communal form of orga- pal-teacherrelations,decision making, and

nization.
parentalinvolvement.Principalsin low-SES
Although we believe that this assumption schoolstended to be "taskoriented,"exercisrequires further examination, the use of this ing tight controls over curriculum and
or any list of practices to operationalize the instruction,whereas principalsin high-SES
concept of restructuring is problematic for schools tended to be more "relationship
oriother reasons. From a practical standpoint, it ented," allowingteachersgreaterautonomy
allows a number of what may be academical- in exercisingprofessionaljudgments. While
ly neutral (or even academically questionable) principals of high-SESschools sought to
practices to ride piggyback into the arena of establish strong ties to their surrounding
discourse on school improvement. Team community, their low-SES counterparts
teaching across content areas, for example, worked hard to insulatetheir schools from
represents a considerably complex technolo- externalinfluencesthat might interferewith
gy, the academic impact of which has not the attainmentof academic goals. In short,
been empirically demonstrated.2 Packaged leadershipin effectivelow-SESschoolstended
within the restructuring agenda, however, to be more authoritativeand traditional;in
the idea acquires a degree of normative effectivehigh-SESschools, it appearedto be
and collegial.
momentum. A bandwagon effect can devel- more interpersonal
We believeit would be incorrect,however,
op that may overshadow the influence of
sound empirical evidence on local school or to concludethat the high-SES
schoolsstudied
district policy making.
by Hallingerand Murphy(1986) were more

Differential Paths to Effectiveness

This sort of "agendizing" of the restructuring
concept is also troublesome in the way it
clashes with some substantial theory and evidence that organizational characteristics or
instructional practices that are linked to
school effectiveness in one context may be
ineffective or even counterproductive in others. For example, Hallinger and Murphy
(1986) found that effective high- and low-SES
schools function differently in terms of princi-

"communal"thantheirlow-SEScounterparts.
In fact, both these organizationalstyles bear
some elements of the "schoolsas communities" ideal. The low-SESschools seemed to
operate somewhat in line with the Catholic
model;that is, they reliedon an authoritative
and clearly articulatedvision and on welldefined organizationalboundaries. In the
high-SESschools, communalityappearedto
be a more secular phenomenon; with its
visionderivedfromconsensusand localcommunity support and expectation,there was
less of a need for rigidorganizationalbound-

48
aries. It thus seems questionable whether
many of the practices listed in Table 1 could
equally represent both styles of communality.
Also noteworthy, when coupled with
Hallinger and Murphy's study (1986), is evidence from the so-called effective schools
studies of the 1980s. The "effective practices"
that emerged from that literature not only
conveyed a relatively traditional and intensified approach to teaching and learning
(Elmore 1990), but were typically based on
evidence from urban elementary schools
(Good and Brophy 1986). Such evidence led
us to suspect that achievement effects associated with some kinds of restructuring practices would differ across categories of school
urbanicityor SES.
Furtherevidence to support this suspicion
can be found in studies of the achievement
effects of high school academic press and
communality. Shouse (1 996), for example,
found the pattern of achievement effects associated with these characteristicsto differ dramatically across categories of school SES. In
his study, academic press was defined as the
degree to which school organizationalculture
was driven by traditionalachievement-oriented values, norms, and goals. Communal organization was defined along lines similar to
those in Bryk and Driscoll (1988). Among
moderate and high-SESschools, higher levels
of communality were associated with higher
mathematics achievement. Among low-SES
schools, however, the relationship between
communality and achievement hinged on the
strength of a school's academic press. When
academic press was weak, higher levels of
communality were associated with lower levels of mathematics achievement. When academic press was high, communality wielded a
significant positive achievement effect. Partof
Shouse's argument in this regard was that for
schools serving disadvantaged or disaffected
youths, the objective and universalistic elements of academic press served as a check on
the tendency for subjective, personalistic,
and/or therapeutic forms of caring to take
center stage in place of a universal expectation of rigorous academic activity.
A subsequent study of middle and junior
high schools (Shouse 1998) not only revealed
a similarlystrong connection between acade-

Mussoline and Shouse

mic press and achievement in urban schools,
but reported additional findings that run
counter to the claims of some advocates of
technical reform. Specifically,the study found
that the use of cooperative learning and middle school grade configuration were linked to
significantly lower levels of mathematics
achievement in urban schools. It also found
significant positive achievement effects associated with urban schools' use of more traditional practices, such as ability grouping in
mathematics and the presence of formal policies emphasizing the importance of daily
homework. Highlighting the value of certain
traditionaltypes of social understandings and
technical practices, particularlywithin disadvantaged schools, the study suggested that
some kinds of restructuring initiatives may
not produce the same academic benefits for
students in low-SESurban schools as for their
more affluent peers.

Technological Complexity and

Resource Scarcity
Beneath this empirical evidence, what theoretical basis exists for suspecting that contextual differences may alter the impact of
restructuring practices in disadvantaged
schools? In our view, the problem is essentially related to the deficits in human and social
resources with which low-SES schools must
often contend, as well as to the high complexity associated with many restructuring
practices. juxtaposed against an organizational task laden with difficultyand uncertainty, that of raising academic achievement
among socially disadvantaged students, the
scarcity of resources and technological complexity constrain the ability of these schools
to implement many types of restructuring
reforms successfully. We briefly trace the
argument here and treat it more fully in the
Discussion section.
In terms of resources, several studies of
school achievement have emphasized the differential availabilityof social and culturalcapital across boundaries of school affluence
(Coleman and Hoffer 1986; Hirsch 1996;
Shouse 1996). These studies have indicated
the ways in which students from advantaged
communities tend to have greater access to

49

School Restructuringas a Policy Agenda

the type of background knowledge and social
support structures that are necessary for
school success. Schools that serve such communities can tap in to this advantage in two
ways. First,their students' broader knowledge
base increases the likelihood that some types
of restructured classroom practices will be
effective (such as those that emphasize constructivist or student-centered learning and
the acquisition of higher-order knowledge).
Second, the availabilityof denser, more academically oriented social support structures in
advantaged communities helps reinforce and
facilitate schools' academic goals. In other
words, a "safety net" is available to students
in advantaged schools that reduces the
potential risksof flawed, poorly executed, or
highly complex instructionalpractices.
In addition to lacking these social
resources, disadvantaged schools suffer from
a dearth of certain kinds of human resources
(such as high-quality teaching staffs and low
rates of turnover and absenteeism among
teachers) that are crucial for the successful
implementation of many of the complex
technical reforms often recommended by
advocates of restructuring.Fordisadvantaged
schools, then, implementing the types of systemic reform represented in the schoolrestructuringliteraturemay be a much riskier
venture than for their more affluent counterparts. Disadvantaged schools may reduce or
avoid the risk by focusing on incremental
reform:improving their more traditional, but
less complex practices, rather than substituting them with more radical reforms. The
quandary they face seems well captured by
Graham's (1985:12) metaphor comparing
progressive and traditionalschooling:
Let us considerAmericanprogressiveeducation as porkchops, and traditionaleducation
as apples.

. A pork chop at its best is

absolutelywonderfuland is enhancedby the

additionof a bakedapple. Butthe porkchop
half done, half cooked . . . will give you trichi-

nosis.Thebakedapplehalfcookedis not very

imaginative,but is stillverynutritious.
At the risk of belaboring the analogy, we
suggest that for disadvantaged schools, the
situation boils down to one in which they are
able to serve students a more nutritious pro-

gram by focusing on basic recipes (we return

to this point in our final discussion). If we are
correct, then we would expect to find some
negative consequences for disadvantaged
schools that attempt to deliver a more elaborate or skill-intensivestyle of cuisine.

METHODOLOGY
Source of the Data
Our analyses are based on data from the
NELS:88first follow-up (10th-grade) survey.
Conducted by the National Center for
Education Statistics (NCES), NELS:88examined students' educational experiences from
the 8th grade through high school and into
college, highlighting family, community,
school, and classroom factors that influence
educational success.
Beginning with a base-year (8th-grade)
survey, NELS:88 also included first (1 Othgrade) and second (1 2th-grade) follow-up
surveys. To obtain a representative sample of
American public and private school students,
the base-year survey began with a representative sample of 1,035 8th-grade public and
private schools, stratified by region, urbanicity, and minority enrollment. A random sample of students was obtained in each school,
resulting in an overall sample of approximately 24,599 students. In addition to a
questionnaire, each student in the base-year
survey completed cognitive tests in history,
and
science.
mathematics,
reading,
Questionnaires were also administered to a
parent, the principal, and two teachers (one
from science or mathematics, the other from
Englishor social studies) of each student.
The NELS:88first follow-up survey (conducted in 1990) had the same basic components and student sample. However, in following students in the base-year sample to
their 1Oth-gradeschools, researcherswho are
examining school effects must grapple with
three methodological issues. First, the students no longer represent a random sample
in each high school. Second, the schools they
attended no longer constitute a probability
sample of schools similarto that of the baseyear study. Finally,in many schools, the sizes

SO
of the student and teacher samples no longer
permit a comprehensive analysis of school
effects.
To address these issues, this analysis used a
subsample of 371 first follow-up public
schools with at least 15 NELS-sampledstudents (resulting in a total student sample of
6,994 with an average of 19 NELS-sampled
students per school).3 This filtering helped
increase the reliabilityof both our measure of
school-level SESand of the hierarchicalmodels produced in our analysis.We chose to limit
our analysis to public schools because it is on
these schools that reform efforts tend to
focus. In addition, because our interest centered on "typical" comprehensive high
schools, this subsample excluded vocational
schools and schools in which 30 percent or
more of their students were placed in remedial reading or "alternative"programs.
Besides the reduction in size, the primary
result of our school-sampling strategy was
that compared to the overall NELSsample,
ours overrepresented public schools with
large student enrollments. This overrepresentation is not reallyundesirable, however, since
the problem of increasing school effectiveness is of particularconcern for large public
schools (Bryk,Lee, and Holland 1993). If, for
instance, our findings differed from those
obtained from some theoretically representative sample of American high schools, they
would still indicate a differential pattern of
effect for the larger public high schools represented in our subsample.

Mussoline and Shouse

adjusted for the specified student-level variables. The individualschool means that make
up this grand mean serve as the dependent
variable of the school-level equation (which
links school-level variables to variation in
achievement levels across schools).
In our analysis, which focused primarilyon
the achievement effects of school-level characteristics, the student-level model served
mainly as a set of controls, that is, to reduce
the likelihood that any observed school-level
effects might actually result from differences
in student composition across schools. For
example, our student-level model included
variables representing student SES,race, high
school program (academic, general, or vocational), prior course work in mathematics,
and prior academic ability. (See Table 2 for
descriptive statistics of all variables used in
this analysis.)
Our school-level models included several
types of variables. First,a categorical variable
(labeled RESTRUCT)
was included to indicate
whether a school was classified as restructured. In classifying schools as "restructured,"
we used a criterionsimilarto that used by Lee
and Smith (1995). That is, schools were considered restructuredif their principals reported the use of at least three "restructuring
practices" (see Table 1) and nonrestructured
if they did not (see note 1). Next, a continuous measure of school-average SES was
included (based on the average SES of the
NELS-sampled students in each school) to
control for the impact of school affluence.
Labeled MEANSES,this continuous measure
was
then used to assign schools to one of
General Analytic Strategy
three categories of average SESbased on cut
We examined the achievement effects of points of 1 standard deviation above and
restructuringwithin categories of school SES below its mean. This process resulted in a
using hierarchical linear modeling (HLM), a breakdown of 55 high-SES schools (with
type of multiple regression that is useful in 1,102 students), 263 middle-SES schools
examining student and school factors associ- (with 5,081 students), and 53 low-SES
ated with variation in achievement. Using schools (with 1,028 students).
HLM,the total variation in student achieveBecause the first round of HLM analysis
ment can be separated into its student-level revealed negative restructuringeffects in lowand school-level components. Separate SES schools, an additional category of "very
regression equations (or "models") are then low-SES" schools was created, consisting of
specified to explain the variationat each level. 32 schools (14 of which were "restructured")
The intercept of the student-level equation falling more than 11/4 of a standard deviation
represents the estimated grand mean below the overall mean of MEANSES.These
achievement across the sampled schools, categorical indicators of school affluence

as a PolicyAgenda
SchoolRestructuring

LU

C~~

ON
0

Ln

00
"t

ON

00
0

[NcYn

Ln

000
'.0

ON

\.0

00

L(n

cf

L(n

ON

rl-I

\.0

00
0

Lfn

Lf

ON

L/n

ON

51
ON
\.0

rN-

N
000

00

1-

r-

"t

\N'00

\0

\.0

\.0

00

L(n

00

rn

LU

C~~
o

00

VI Ln
0

[N--

ON

L(n-

ON

00
N

6~~~~~~~0
60 6

L(n

L(n~

00

00

L(

0~

0L

cu

-j

Ec

m
E

>2

*'~~~~~
~~~0
cu

4-j<

*'

E0

iB~~~~~

EUU
~~L1 ~~~~~1

E
cu~~
LU

c 00
LU

V
I~~~~~~~~~~~~)~~
u
~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Ln

-'

CL0-

4--j

cfn

00

'

U~~~~~~~~

"t

ON

,,~~~~~~~~~~ccmE40

00

C4E

00

~~~~~E
0~
~~~~) 0u

u
V)

L-

ON

L(n

LN00

~
0

000

Lu

"t
N

~
-

r-

[N~

L(

L(n

~~~~~~'

Ln

000

0_

L1~~0

<

I~~~~~~~~~~~~~~~L

o
L

CU

-C<

52
were, in turn, used to construct dummy interaction terms that were designed to tease out
the differentialimpact of restructuringacross
categories of school affluence.
Finally,given the conceptual and empirical
connections established in previous studies
among school restructuring,communal organization, and academic press, we deemed it
important to distinguish and control for the
influence of the latter two characteristics on
students' achievement. Doing so also enabled
us to gauge the extent to which restructuring
effects were attributable to either of these
two factors. Variables developed previously
(Shouse 1996; see also Bryk and Driscoll
1988) were used to represent school academic press and communality (See the
Appendix for details on these two variables.)

Dependent Variable
Student achievement is represented by
NELS:88mathematics IRT(item response theory) scores (see NCES 1994). Although
NELS:88 also includes reading, history, and
science scores, the mathematics scores represent the most valid indicator of school effectiveness. For one thing, because the mathematics test contains the greatest number of
items and ability levels, it is most immune to
floor and ceiling effects. In addition, mathematics scores are more likely to reflect inschool learning than are scores in other subjects (Haney 1996; Karweit and Ricciuti
1997). Fortechnical information on the validity and reliability of NELS:88, see Rock,
Pollack,and Quinn (1 995).4

RESULTS
Table 3 displays the results of our analysis. As
is customary in presenting HLM results,
school-level effects are presented in the top
panel and student-level effects in the bottom
panel. In the HLManalysis reported here, student-level variables are "centered" around a
mean of 0.5 This centering does not affect the
way coefficients are interpreted, and it allows
the school effects reported in the top panel to
reflect the experience of an average generaltrack student. At the school level, continuous

Mussolineand Shouse
measures (such as MEANSES, APRESS,
COMM, and the dependent variable representing students' mathematics achievement)
were rescaled to a mean of 0 and a standard
deviation of 1. Thus, for these variables, the
coefficients reported in Table 3 are standardized; that is, they represent the difference in
mathematics achievement (measured in standard deviations) associated with a difference
of 1 standard deviation in the particularindependent variable. The dummy variables
LOWSES,VLOWSES,HISES,and
(RESTRUCT,
the interaction terms created from them)
have not been standardized. Theircoefficients
thus represent additive or diminutive effects
associated with being a particular type of
school.
Table 3 presents two HLM models, each
one designed to tease out the interaction
between the effects of restructuring and
school SES.6 The model represented in the
first column highlights the impact of restructuring in low-SES schools; the second does
the same for very low-SES schools. As was
noted earlier, each model controls for the
effects of school average SESand school levels of communality and academic press.
The equation represented in the first column of Table 3 reveals that the base effect
associated with being a restructuredschool is
small (.03). Although this effect does not
meet a 95 percent level of confidence, it suggests that within this sample of public
schools, restructured schools outperform
nonrestructured schools in mathematics
achievement by an average of 3 percent of a
standard deviation. (It should also be pointed
out that school effects are generally considered "substantive" if their standardized coefficient is at least .10.) A scan down the first
column, however, reveals that the effect of
restructuring in low-SES schools is negative
-.08). These effects are addi(LOWSES*RESTR,
tive. Thus, achievement levels in low-SES
restructured schools average 5 percent of a
standard deviation below those in other
restructuredschools and 8 percent of a standard deviation below those in nonrestructured schools.
On the basis of this pattern of effect, we
focused on the impact of restructuring
among schools that were even more disad-

School Restructuringas a Policy Agenda

53

Table 3. Restructuring Effects on Mathematics Achievement (standardized HLMcoefficients, standard errors in parentheses)
Variable

Description

II

School Level

RESTRUCT

Restructuredschool

.03
(.02)

.03
(.02)

MEANSES

School mean SES

.04*
(.02)

.02
(.02)

LOWSES

Low-SESschool

.02
(.04)

LOWSES*RESTRUCT Low-SESrestructuredschool
VLOWSES

-.08*
(.04)
.01
(.05)

Verylow-SESschool

-.12*
(.06)

VLOWSES*RESTRUCT Very low-SESrestructuredschool

HISES

High-SESschool

.05
(.05)

.05
(.05)

HISES*RESTR

High-SESrestructuredschool

-.06
(.05)

-.06
(.05)

COMM

School communality

.01
(.01)

.01
(.01)

APRESS

School academic press

.04*
(.01)

.04*
(.01)

SEX

Student's sex

.02*
(.01)

.02*
(.01)

Fl SES

Student's SES

.02*
(.01)

.02*
(.01)

Fl S22SUM

Number of semesters in mainstream

.08*
mathematics since the 8th grade

Student Level

.08*

(.00)

(.00)

VOTRACK

Vocational program

-.08*
(.02)

.08
(.02)

ACTRACK

Academic program

.10*
(.01)

.10*
(.01)

MINOR

Blackor Hispanicstudent

-.09*
(.02)

-.09*
(.02)

BY2XMIRR

8th-grade achievement

.74*
(.01)

.74*
(.01)

Coefficient at least twice its standard error.

54
vantaged, those falling into the very low-SES
category (11/4 of a standard deviation below
the mean). The equation presented in the
second column of Table 3 differs from that
presented in the first column in one important way. In the first model, we compared
restructuring effects in high- and low-SES
schools with those in mid-level SES schools.
The second model offers a similar type of
comparison, but the variables that previously
represented low-SES schools were replaced
with variables representing very low-SES
schools. In other words, the base category of
mid-level SES schools was expanded downward to include 21 additional schools that
were previously included in the low-SEScategory (those with MEANSES
values between -1
and -1.25). The restructuringeffects for highSESschools remain separated in this model so
that contrasts may be drawn between
restructuringeffects in the middle and lowest
school-SEScategories.
The point of key interest in the second column of Table 3 is the negative restructuring
effect for very low-SES schools (-.1 2). The
effect indicates that when the effects of the
other school variables are held constant,
mathematics achievement levels in restructured schools in the very low-SES category
average 12 percent of a standard deviation
lower than in the restructured mid-level SES
schools. In addition, subtracting the negative
effect for very low-SESschools from the main
restructuring effect reveals that among
schools in the lowest SES category, nonrestructured schools outperform their restructured counterparts by 9 percent of a standard
deviation (again, with other school-level
effects held constant).
Aside from the restructuring effects highlighted in Table 3, one also notes a significant,
albeit small, effect associated with the measure of school academic press and a virtual
noneffect associated with the measure of
school communality. It should be noted,
however, (based on exploratory analyses not
tabulated here) that the inclusion of these
variables had virtually no influence on other
effects reported in the two models presented
in Table 3.

Mussoline and Shouse

DISCUSSION
Our findings validate the concerns raised earlier about the claims of various studies of
school restructuring. First,we found that for
large public schools of the type examined in
this study, the magnitude and direction of
restructuringeffects are not consistent across
all levels of school affluence. Restructuring
effects are largely nonsignificant for schools
that primarily serve students of average or
above-average affluence and are significantly
negative for those that mainly serve disadvantaged students. These results reinforceour
concern that local school policy makers need
to maintain some healthy skepticism about
the usefulness of any broadly prescribed systemic reform agenda. In particular, leaders
and decision makers in schools serving predominantly low-income communities may
need to exercise caution in adopting complex
educational practices.
We suspect that the technological complexity associated with certain restructuring
practices (for example, team teaching, cooperative learning, and heterogeneous grouping) accounts for at least some of the pattern
of effects revealed in our analyses.7As instructional practices become more complex, they
not only stretch the capacities of teaching
staffs, but become riskierin terms of increasing students' achievement. To illustrate,consider students' achievement as a function of
(1) the availability of academically oriented
social capital outside school and (2) teachers'
instructional skill. Compared to students in
disadvantaged communities, students in
more affluent communities tend to have
greater access to academically oriented social
capital, and their teachers tend to be more
highly skilled.Thus, not only do their teachers
tend to be better prepared to implement
restructured practices effectively, but, even if
they are not, the availabilityof an "academic
safety net" outside the schools will also buffer
students' achievement from teachers' professional
mistakes
and
shortcomings.
Unfortunately, for large low-SES urban
schools, the situation tends to be reversed.
The teachers in these schools tend to be less
prepared to implement complex practices
successfully, and the consequences of their

School Restructuring as a Polky Agenda

mistakes more directly affect their students'
academic achievement.
Having pointed out the difficulties that
particulartypes of restructuringreforms pose
for disadvantaged schools, we think it is
important to emphasize that our analysis is
not intended as a criticism of any particular
restructuringreform or of efforts by disadvantaged schools to experiment with alternative
organizationalor curriculararrangements as a
means of improving instruction. But it seems
naturalto us that leaders and decision makers
in such schools would tend to "restructure"
their schools in different ways, depending on
their perceptions of organizational needs and
constraints. In some schools, restructuring
may involve striving for greater professional
collegiality or decentralized decision making.
In others, it may mean adopting, modifying,
and/or improving upon traditional instructional practices, starting a schoolwide discussion about core values, or finding ways to
increase academic press.
This kind of activity, we believe, is incongruent with externally devised checklists or
templates for school improvement. Ironically,
such external designs may actually encourage
local or district administrators to impose
"restructured"arrangements on their professional staffs. In one case, an urban high
school principal, frustrated over teachers'
skepticism regarding a district-proposed local
school site-management plan, stated angrily
to her staff, "People, understand this! We will
become an empowered school!" (quoted in
Boyd and Shouse 1997:154).
Again, we emphasize that the point here is
not to denigrate the various practices included in the restructuring agenda or to cast
undue praise on those they are intended to
replace. Instead, our major point is that if
school restructuring is to be meaningful,
researchers, administrators, and teachers
must take great care that it is not transformed
into another policy bandwagon that all must
ride. We argue-and our study indicatesthat there are multiple avenues to school
effectiveness and that "tradition" and
"restructuring"are not signposts on opposite
ends of the road. As Cuban (1 998) noted,
school "goodness" can be found in both traditional and progressive settings. Our evi-

55
dence echoes his view and suggests the need
for researchers, administrators, and teachers-particularly those who are concerned
about or working in disadvantaged schoolsto think more critically, incrementally, and
intuitively about school restructuring and
reform.

NOTES
1. In addition to identifying "restructured"
schools, Lee and Smith (1 995) also identified
what they referred to as "moderate" and
"unrestructured" schools, based on the
schools' use of other types of practices and
reforms.
2. For an excellent case-study analysis of
the complexity and resource demand of thematic team teaching, see Meister (1 997).
3. This subsample was developed by
Shouse (1994) to examine the impact of
school academic press and communality on
students' achievement.
4. In our analyses, we opted to use 1Othgrade mathematics achievement scores as
our dependent variable while controlling for
8th-grade mathematics achievement. This
design contrasts with the design used by Lee
and Smith (1 995), which used 8th- to 1Othgrade mathematics achievement gain as a
dependent variable. In Lee and Smith's
design, controls were included for 8th-grade
ability in reading, history, and science, but
not in mathematics. On the basis of our
exploratory analyses (using a NELSsubsample
similarin size to that used by Lee and Smith),
we suspect that their design unintentionally
obscured the impact of school affluence on
students' achievement by not fully controlling
for the possibility of ceiling and floor effects.
Specifically, although our variable representing school mean SES is a significant predictor of 1Oth-grade mathematics achievement (after the effects of 8th-grade mathematics achievement are controlled), it does
not significantly predict 8th- to 1Oth-grade
mathematics achievement gain unless a control is included for 8th-grade mathematics
achievement. In fact, when this variable is
included, higher 8th-grade mathematics
achievement scores are associated with signif-

56

Mussoline and Shouse

icantly lower 8th- to 1Oth-grademathematics

achievement gain. We suspect that this association results from residual floor and ceiling
effects associated with the NELS:88 mathematics test. (We say "residual"because NELS
includes design elements that are intended to
mitigate the impact of such effects.) In other
words, without controls for prior mathematics ability, the association between school
mean SESand achievement gain appears misleadingly weak because it is easier for students who start at lower levels of achievement to show larger gains.
5. Student-level variables in this analysis
were "fixed" and grand-mean centered. See
Bryk and Raudenbush (1992:202-03) or
Shouse (1994:102-04).
6. Relevantvariance statistics for the models reported in Table 3 are as follows: (1) 85
percent of the variation in the dependent

variable is within schools and 15 percent is

between schools, (2) the six student-level
controls account for 77 percent of the withinschool variance, and (3) 73 percent of the
school-level variation is accounted for by
school-mean SES (MEANSES)and an additional 5 percent of the school-level variance is
accounted for by the remaining school-level
variables.
7. Though our study did not trace significant negative effects to any specific restructuring practice, we found that 9 of the 14
very low-SES schools had adopted either
team teaching or cooperative learning. In
addition, exploratory analyses revealed that
compared to their low-SESand moderate-SES
counterparts, these very low-SESschools were
significantly more likelyto have adopted one
or the other of these two practices.

APPENDIX
Indicators of School Academic Press and Communality
The following are general descriptions of the NELS:88items used to represent academic press and communality. Some characteristicslisted here were represented by more than one NELSitem. See Shouse
(1996) for a more detailed description.
Indicators of School Academic Press
Academic Climate
*

Principals'reportregardingacademic climate (class activitiesare highly structured,students expected to do

homework, students encouraged to compete for grades, students encouraged to take academic classes)
* Principal'sreport of the extent to which the school honors and publicizes students' achievement
* Semester requirements in mathematics and foreign languages
* Percentage of teachers in the school with at least a master's degree
* The extent of "mainstream"course taking in science and the humanities
* Students' perception of the overall school academic demand
Disciplinary Climate
*
*
*

School policies on absenteeism, misbehavior,and parental notification

Teachers'and students' perceptions of the disciplinarypolicy and climate
School policies regarding students' absence from school

Teachers' Instructional Practices and Emphases

*

Emphasison an absolute level of achievement in determining students' grades

Emphasis on covering the curriculum, developing students' understanding, and creating enjoyable
lessons as instructionalgoals
* Teachers' homework policies: amount assigned and regularityof feedback
* Teachers' reports of having contacted parents of poorly performing students
* Time spent by teachers in planning and preparing out of school
* Students' reports of classroom instructionalquality and academic demand

57

School Restructuringas a Policy Agenda

Indicators of School Communality
Shared Values
* Teachers'agreement about school and classroom goals
* Teachers'consensus on core beliefs and values
* Teachers' belief that students can learn
* Principal'sand teachers' agreement on school policies
* Principal'sand teachers' agreement on attitudes toward students
CommonAgenda of Activities
*

Track similarity

*
*
*

Course-takingsimilarityin mathematics, science, English,social studies, and foreign languages

Proportionof students in extracurricularactivities
Proportionof students in leadership roles

Ethos of Caring and Collegiality

Teachers' perception of faculty cooperation and collegiality
Percentage of teachers reporting that they seek help from colleagues in and out of their own department
* Teachers' reports of time spent on cooperative projects
* Staff commitment to evaluation
* Teachers' perception that the school seems like "one big happy family"
* Students' perception that teachers show interest in them as people
* Students' perception that teachers really listen to what they say
*
*

REFERENCES
Boyd, WilliamL., and Roger C. Shouse. 1997. "The
Problems and Promise of Urban Schools." Pp.
141-65 in Childrenand Youth:Interdisciplinary
Perspectives,edited by H. Walberg, 0. Reyes,
and R. Weissberg. Thousand Oaks, CA:Sage.
Bryk,Anthony S., and Mary E. Driscoll. 1988. The
School as Community:TheoreticalFoundations,
Contextual Influences, and Consequences for
Students and Teachers.Chicago: Universityof
Chicago Benton Center for Curriculum and
Instruction.
Bryk, Anthony S., Valerie E. Lee, and Peter B.
Holland. 1993. Catholic Schools and the
Common Good. Cambridge, MA: Harvard
UniversityPress.
Bryk, Anthony S., and Stephen W. Raudenbush.
1992. HierarchicalLinearModels: Applications
and Data AnalysisMethods.Newbury Park,CA:
Sage.
Carnegie Council on Adolescent Development.
1989. TurningPoints:PreparingAmericanYouth
for the 21st Century. New York: Carnegie
Corporation.
Center on Organization and Restructuring of
Schools. 1992. "Estimating the Extent of
School Restructuring."Issues in Restructuring
Schools, Fall, 1992. Madison, WI:Author.
Coleman, James S., and Thomas Hoffer. 1986.
Publicand PrivateHigh Schools: the Impact of
Communities.New York:Basic Books.

Corbett, H. D. 1990. On the Meaning of

Restructuring.Philadelphia:Researchfor Better
Schools.
Cuban, Larry.1998, January 28. "A Tale of Two
Schools." EducationWeek,p. 48.
Darling-Hammond, Linda. 1995. "Restructuring
Schools for Student Success." Daedalus
124:153-62.
Elmore, Richard F. 1990. "On Changing the
Structure of Public Schools." Pp. 1-28 in
RestructuringSchools: The Next Generationof
Educational Reform, edited by Richard F.
Elmore.San Francisco:Josey-Bass.
Good, Thomas L., and Jere E. Brophy. 1986.
"School Effects."Pp. 570-604 in Handbookof
Researchon Teaching,edited by M. Wittrock.
New York:Macmillan.
Graham, PatriciaA. 1985. "Schools: the Raw, the
Cooked, and the Half-baked. Massachusetts
Association of School Committees Journal
18:12-1 7.
Grant, Gerald 1988. The World We Created at
Hamilton High. Cambridge, MA: Harvard
UniversityPress.
Hallinger,Philip,and Joseph F.Murphy.1986. "The
Social Context of EffectiveSchools." American
Journalof Education94:328-55.
Haney, Walt. 1996. "Reflecting on Authentic
Assessment and School Reform." Teachers
CollegeRecord98:329-45.
Hirsch,E. D. 1996. TheSchools WeNeed. New York:
Doubleday.
Karweit, Nancy L., and Anne E. Ricciuti 1997.

58

Mussoline and Shouse

Summer Learning: Achievement Profiles of

Restructuring.Reston, VA:National Association
ProspectsFirstGraders.Washington, DC: U.S.
of Secondary School Principals.
Department of Education, Planning and
1996. "Introduction:the School Restructuring
EvaluationService.
Study." In Authentic Achievement. San
Lee, ValerieE., and JuliaB. Smith. 1995. "Effectsof
Francisco:Josey-Bass.
High School Restructuringand Size on Early Newmann, Fred M., and Gary C. Wehlage. 1995.
Gains in Achievement and Engagement."
Successful School Restructuring.Madison, WI:
Sociologyof Education68:241-70.
Center on Organization and Restructuringof
Louis,KarenS., Helen M. Marks,and Sharon Kruse.
Schools.
1996. "Teachers' Professional Community in Rock, Donald A., Judith M. Pollack, and Peggy
RestructuringSchools." AmericanEducational
Quinn. 1995. Psychometric Report for the
ResearchJournal33:757-98.
NELS:88Base YearThroughSecond Follow-up.
Meister, Denise G. 1997. "Implementing Change
Washington, DC: U.S. Department of
from
Discipline-based
Teaching
to
Education, Office of EducationalResearchand
Interdisciplinary
Teaming:A Case Study of Five
Improvement.
High School Teachers."Unpublished doctoral
Shouse, Roger C. 1994. "Academic Orientation in
dissertation,PennsylvaniaState University.
School Society: the Meaning and Importance
Murphy, Joseph F. 1991. RestructuringSchools:
of Academic Press."Unpublished doctoral disCapturingand Assessing the Phenomena. New
sertation, Universityof Chicago.
York:Basic Books.
1996. "Academic Press and Sense of
National Center for Education Statistics. 1994.
Community: Conflict and Congruence in
Second Follow-up:StudentComponentData File
American High Schools." Researchin Sociology
User's Manual. Washington, DC: U.S.
of Educationand Socialization11:1 73-202.
Department of Education, Office of
1998. "Restructuring's Impact on Student
EducationalResearchand Improvement.
Achievement:
Contrasts
School
by
Newmann, Fred M. 1992. "A Developing Design."
Urbanicity." Educational Administration
A
In A Leader'sGuide to School Restructuring:
Quarterly34:677-99.
Special Report of the NASSP Commissionon

LawrenceJ. Mussoline, Ph.D., is Superintendentof Schools, Pine GroveArea School District,Pine

Grove,Pennsylvania.His main fields of interest are school improvementand curricularreform.He
is preparingto teach a course on educationaladministrationas a new adjunctprofessorat Temple
University.
Roger C. Shouse, Ph.D., is AssociateProfessor,Departmentof EducationPolicyStudies, Collegeof
Education,PennsylvaniaState University,UniversityPark.Hismain fieldsof interestare school effectiveness and school culture. He is currentlyexamining the "school as community"metaphor, to
understandhow schools convey a sense of democraticspiritto students.
Thisresearchwas supportedby a grant fromthe AmericanEducationalResearchAssociation,which
receivesfundsforits AERA
GrantsProgramfromthe National ScienceFoundation,and the National
Center for EducationStatistics and the Office of EducationalResearchand Improvement(U.S.
Departmentof Education)under NSF Grant RED-9452861. Opinionsreflect those of the authors
and do not necessarilyreflectthose of the granting agencies. The authors'names are listed alphabeticallybecause each contributedequally to this work.Addressall correspondenceto Dr. RogerC.
Shouse, Department of Education Policy Studies, College of Education, Pennsylvania State
University,302B RackleyBuilding,UniversityPark,PA 16802; E-mail:rcs8@psu.edu.