LB
2806.22
.G78
{2001
IGSL
UCB

 

I

 

PACE WORKING PAPER SERIES 01—1

 

Straw Into Gold,
Resources lhtO RGSUltS:

T ’ Spinning Out the

Implications of the

"New" School Finance

April 2001

INSTITUTE OF GOVERNMENTAL
STUDIES LIBRARY

FEB 03 2017
UNIVERSITY OF CALIFORNIA

POIICV Analysis for California Education
University Of California, Berkeley and Stanford University

 

 

RESEARCH SERIES 01-1

Straw Into Gold,

Resources Into Results:
Spinning Out the Implications
of- the "New" School Finance

April 2001

w. Norton Grubb
Luis A. Huerta

  

Policy Analysis for

California Education

 

PACE

 

”153%: m 3

 

 

cknowledgments

a thank Policy Analysis for California Education (PACE) and the William and Flora
wlett Foundation for supporting this research. We also thank Bruce Fuller, Allan
den, and Catherine Clark who made helpful comments on an earlier draft. Please
1d comments to:

w. Norton Grubb

David Gardner Chair in Higher Education

University of California, Berkeley / Graduate School of Education
3659 Tolman Hall

Berkeley, CA 94720-1670

(Phone) 510-642-0709

(Fax) 510-642-4803

e-mail: wngrubb@uclink4.berkeley.edu.

Luis A. Huerta

Policy Analysis for California Education (PACE)

University of California, Berkeley / Graduate School of Education
5653 Tolman Hall

Berkeley, CA 94270-1670

(Phone) 510-642-7223

(Fax) 510-642-9148

e-mail: lhuerta@uclink4.berkeley.edu

 

Table Of Contents

Executive Summary

 

Introduction

 

I. The Political Economy of Resource Use: The Diversion of Dollars
ll. Converting Resources to Results: Opening the Black Box

 

Educational Production Functions

 

School Effectiveness Research

 

Identifying Effective Teaching Conditions

 

III. Implications for Research

 

Natural Experiments

 

Evaluating Self-conscious Reforms

 

Returning to Effective Schools

 

Estimating Structural Equations

 

IV. Implications for Educators and Administrators

 

V. Implications for Public Policy and School Reform:
Creating Complementary Reforms

 

VI.Toward a New Narrative for Spending

 

Table 1

 

Table 2

 

Endnotes

 

References

 

’10
”I”!
’16
’18
23
25
25
27
28
3C

 

Executive Summary

Educational resources have been the
subject of endless political battles, including
efforts to expand resources and to equalize them.
However, the connections between resources and
outcomes remain obscure: real spending per
pupil has increased steadily without any obvious
effect on learning, reform efforts often cost more
without any corollary effects, and there is a great
deal of evidence that additional resources do not
have substantial effects on learning except under
special conditions.

This paper presents the “new”school
ﬁnance, one that—in contrast to most
discussions about funding that have concentrated
on spending patterns only—asks how resources
are used within schools and classrooms, and
whether they are used in ways that can enhance
educational outcomes. This approach is not
particularly new, since many researchers have
called for more careful investigation of how
resources are spent when their analyses failed to
reveal clear relations between spending and
outcomes. However the “new” school finance
has not been the subject of consistent
investigation, and there has been little progress in
understanding the conditions under which
spending will (or will not) enhance learning.
This paper reviews a variety of literatures, from
different areas of education, in order to clarify
the implications of this perspective for
researchers, for practitioners (like principals),
and for policy makers.

While the “old” school finance
perspective has usually assumed that additional
resources are self—evidently valuable, it’s clear

that there are many ways to spend resources to

little effect. The ﬁrst section of this paper details
what might be called the political economy of
waste: the political and organizational features of
schools that lead to resources being spent with
no potential effects on outcomes. The purpose of
such a conceptualization is to clarify how
difficult it might be to spend additional resources
and enhance educational outcomes, however
measured. It also leads to a series of hypotheses
for what one might ﬁnd in examining, the effects
of any spending increases, suggesting the
research strategy of asking “where the money
has gone,” being careful to trace the effects of
funding to the school and classroom levels.
Section 11 reviews several areas which
have, contrary to the “old” school ﬁnance, tried
to move beyond discussions of spending to
more detailed analyses. The effective school
literature is one such area, though it tended not to
examine resource use and generally did not
examine classrooms. The literature on
educational production functions is another, with
its tendency to conclude that “spending doesn’t
make a difference.” But it generally fails to
specify how resources are used within the
schools, and therefore crucial variables—teacher
experience, for example, or class size—may have
either positive or negative effects on outcomes
like test scores. The solution in both cases is to
trace resources more carefully to the ways in
which they are used within the classrooms and
schools. In the case of educational production
functions, this leads to more complex equations;
teaching characteristics and student ability to
beneﬁt from instruction are two crucial variables,
with resources inﬂuencing outcomes by
improving either of these two. In addition, the

interactions among different aspects of

 

practice—for example, between teachers’
practices and students’ motivation, between
teachers’ practices and school policies, between
school practices and district or state
policies—mean that various resources interact in
complex and unknown ways, contrary to the
simple linear additive speciﬁcation of most
production functions. The section ends with
various suggestions, drawn from the reform
literature, about how to measure teaching
characteristics and student ability to learn.

The following sections detail the
implications for research, for educators, and
administrators, and policy-makers. The
potentially valuable directions for research
include analyzing natural experiments where
spending increases suddenly, to understand how
such windfalls are actually used; analyzing self-
conscious reform efforts that use resources in
speciﬁc (and varying) ways; returning to the
effective schools strategy in somewhat modiﬁed
ways; and estimating more complex versions of
educational production functions. The
implications for practitioners include the need to
determine what might be effective practice, and
then the search for ways to fund it—a bottom—up
or site-base approach to spending priorities.

For policy makers, the implication of the
“new” school finance is that spending may be
necessary but not sufficient to enhance
outcomes. The challenge in policy is therefore to
determine what might be necessary in
addition—potentially, the restriction of spending
in the manner of categorical funding, or the
enactment of complementary reforms, or the
development of leaders (like principals and the

superintendents) with the requisite reform vision.

Various recent failures in state and federal

ii

policies, like the likely ineffectiveness of
reducing class sizes, provide some clues for
alternative policy directions.

The “new” school finance is a difficult
subject because it demands that funding be
treated not as political spoils, nor as self-
evidently effective, but as only one of the
resources necessary to make schools effective.
Developing the associated agendas for research,
practice and policy will take sustained effort and
a reformulation of thinking about resources. The
alternative is to continue the current patterns in
which expenditures in education keep expanding

with little to show for them.

 

Introduction

It’s impossible to deny the importance of
resources to education. Generations of
reformers have come along, each with a new
formula for reforming schools, each needing
more money—“spending again and again and
again,” to paraphrase Cuban’s (1990) review
of reform efforts. Advocates for poor children
have rediscovered disparities in spending nearly
every decade, from Ellwood Cubberly’s (1905)
complaints about reliance on local revenues at
the turn of the century to Jonathan’s Kozol’s
(1993) latest attack on “savage inequalities.”
And in a political system dominated by interest
group liberalism, debates over school resources
often dominate those about teaching and
learning—battles over inputs rather than
outcomes, means rather than ends—despite
pleas to “put education above politics.”

But the inadequacy of debates over
resources—of the debates, not necessarily the
resources themselves—has become apparent to
many participants. In K—12 education, real
resources per pupil in public schools have risen
constantly throughout this century, as Table 1
indicates. Although spending has been
relatively stable (in real terms) during the
19905, spending per pupil has doubled since
the late 19605, and it rose particularly fast
during the 19605, 19705, and 19805. In terms
of more comprehensive measures of public
resources in education, Table 2 clariﬁes that we
spend more as a fraction of Gross Domestic
Product than virtually any other developed
country except Canada, partly because of our

extensive enrollments in higher education. (Our

spending on elementary-secondary education is
closer to the average of these countries.) It’s
difﬁcult to argue in any simple way for more
spending, if only because the past thirty years
of higher spending have not resolved
educational problems nor brought about the
reforms we might need. Those states that have
litigated school ﬁnance issues successfully have
seen disparities in spending between rich and
poor districts decline markedly (Hickrod et al.,
1997), without disparities in achievement
among students decreasing in any obvious way.
And while some urban districts still have low
levels of spending, other cities have spending
levels well above the national average—for
example, Washington, DC with $8,290,
Hartford with $10,017, Boston with $8,225,
Newark with $10,925, Kansas City with
$9,4361—and still have not escaped the
patterns of limited offerings, high dropouts, low
test scores, low rates of movement into post
secondary education, high turnover among
teachers and administrators, and political
turmoil that defines the urban condition.

And so it’s not enough to debate the
resources in schools, as many advocates have
come to realize. Money may be necessary for
improved learning, however deﬁned, but it
cannot be sufﬁcient. Instead it’s critical in
addition to ascertain how these resources are
spent—whether they are spent on activities and
practices within schools and classrooms that
enhance teaching and learning, or whether they
are spent ineffectively.2 This idea has been
stated in several different vocabularies: for
example, the concern over “more bang for the

buck” in educational politics and journalism

 

(Boyd and Hartman, 1988); the emphasis on
the efﬁciency of school spending, in the
vocabulary of economists (e. g., Levin, 1994);
and the focus on the improvement of teaching,
among those who have been principally
concerned with teachers and pedagogy (e. g.,
Miles and Darling-Hammond, 1998). And then,
as everyone knows, students bring their own
resources to the classroom —motivation,
willingness to learn, certain basic attitudes and
literacies that schools generally take for
granted. Without these student resources,
school spending may be quite ineffective—or,
alternatively, it may be necessary to spend
public resources on mechanisms to enhance
student preparedness, like early childhood
programs, family literacy efforts, or parent
participation. While the range of these other
factors necessary to make school spending
effective is quite large, as we will clarify in
Section II, the logic underlying them is similar:
school spending can be effective in enhancing
learning only in conjunction with specific
practices—in schools and classrooms, in
families and communities—that themselves
contribute to learning.

The central insight of what might be
called the “new” school finance—in contrast
to the “old” school finance that has
concentrated only on spending patterns, and has
sometimes neglected how resources are used
within classrooms and schools—is that the
effective use of resources is a two—stage
process. It’s first necessary to ascertain those
practices and teaching conditions within
schools and classrooms3 that enhance learning.

Then it’s necessary to allocate resources to

2

those practices, rather than to other ineffective
uses. (We will formalize the distinctions
between resources and practices in Section H,
when we present a small extension of the
education production function approach.) This
insight is not particularly new, since others
frustrated with the limits of the “old” school
ﬁnance have come to the same insight.4 In
addition, the current efforts in school ﬁnance
litigation have been trying to move beyond
equity measured by funding to conceptions of
adequacy based on the ability of resources to
achieve certain levels of performance (e. g.,
Clune, 1994; Minorini and Sugarman, 1999).
But even if the “new” school finance is not
new, its perspectives are not yet widespread,
either in research, or in legislation and policy-
making, or in the practices of administrators
and school reformers. Analyses of school
resources still concentrate on the dollars spent,
rather than how these resources are used (e. g.,
Ladd, Chalk, and Hansen, 1999, from the
National Research Council’s Committee on
Education Finance). Some reforms like class
size reduction continue to spend huge sums
with little regard for how changes will affect
Classrooms. Principals and other school leaders
seem to pay little attention to the educational
efficacy of their spending decisions, despite
school—based management and other changes
that give them (some) greater power (Boyd and
Hartman, 1988). So it’s worth continuing to
articulate the perspectives of the “new” school
finance since it will not become the dominant
perspective or narrative for examining school
resources until educators, policy-makers, and

researchers all embrace it.5

 

In this paper we ﬁrst examine in Section I
the inadequacy of the “old” school
ﬁnance—the view that concern with spending is
sufﬁcient. This critique readily leads to a deeper
understanding of why spending seems so often
to be derailed into ineffective practices,
particularly in urban school districts. This kind
of conceptual approach is useful since if it were
possible to identify effective practices, it would
still be necessary to channel resources to
them—and to understand the political and
organizational difﬁculties of doing so.

In Section II we examine several efforts
to examine effectiveness. One is the production
function literature relating outputs (usually test
scores) to inputs like pupil-teacher ratios,
teacher experience, and teacher ability. This
literature has—unfortunately, and despite
substantial effort—failed to provide enough
insights about how resources should be used
within schools and classrooms. Its general
analytic framework could be elaborated to do
so, and we provide several examples of how to
do so. However, we suspect that these formal
models are more useful as metaphors of how
schools work than as guidance for statistical
analysis. When we examine current ideas about
effective practices, we can easily develop a long
list of educational resources that cannot be
easily measured—and so the estimation of an
adequate production function is likely to be a
long way in the future.

A second literature on efﬁcacy is that of
the school—effectiveness literature, with its effort
to ﬁnd out what distinguished high—performing
schools. However, many of these

characteristics—for example, strong educational

leadership, or an orderly school climate—do
not necessarily cost much (even though they
may require substantial effort), reinforcing the
conclusion that “money doesn’t matter.” This
literature did not look carefully into classrooms,
and often neglected resources issues; like the
educational production function research, it
doesn’t lead in any obvious way to
recommendations for spending. We end this
section with a listing of the kinds of teaching
conditions (at both school and classroom
levels) and student attributes that have been the
focus of various researchers and reformers,
suggesting the kinds of inﬂuences that need to
be considered.

If the perspectives of the “new” school
ﬁnance are to be useful, then they must provide
some guidance for new forms of research, and
beyond that for practice and policy. In Section
III we outline the implications for
research—including the accounting exercises
that are common in investigating equity—that
might, over a longer period of time, yield such
answers. In Section IV we outline the
implications for administrators, particularly
principals operating under the greater school-
level control allowed by site—based management
or charter schools. Finally, in Section V we
outline the implications for legislation and
policy, both in speciﬁc cases (e. g., the
legislative efforts to reduce class size) and in
general, by distinguishing several ways to get
resources to schools in more effective ways.

The perspectives of the “new” school
ﬁnance are ways of bridging one of the great
divides in education. Macro perspectives—-

often drawn from economics, political science,

 

history, sociology, statistics, and other social
sciences—have focused on the large economic,
apolitical, historical, and cultural inﬂuences on
schools, but usually without entering the
classroom. Micro or classroom perspectives,
often drawing on psychology, have emphasized
the interactions among teachers and students,
but have often neglected the macro forces that
have structured the classroom, and the roles of
teachers and students, in particular ways. Often
the twain do not meet—as in the “old” school
ﬁnance, which has investigated the determinants
of spending without asking how it inﬂuences
classrooms, or in the educational production
function literature that has left the classroom as
a black box. But the “new” school finance
needs to ask how resources are related to
school and classroom practices—and so it
might lead to a more complete analysis of
education than the great divide has often
allowed.6

Many educators don’t like to think much
about money—it’s dross, or straw, or ﬁlthy
lucre than impedes loftier goals like educating
all children to the limits of their potential. But if
it’s tainted, it’s also a necessary evil, necessary
for producing the varied educational results that
educators, and parents, and policy-makers want.
The conversion of resources into results should
not be like Rumpelstiltskin’s magic that
enabled the miller’s daughter to spin straw into
gold, and the “new” school finance should
move away from alchemy toward a clearer
understanding of the requirements for effective

school spending.

I. The Political Economy
of Resource Use:
The Diversion of Dollars

The boundless concern, among educators
and politicians alike, about the level and
allocation of dollars seems to assume that
ﬁnancial resources are the currency of the
realm—the only kinds of resources worth
discussing. But there are infinite ways for these
resources to be misspent, without making
changes in classrooms and schools that might
affect learning, or (in the language of Cohen,
Raudenbusch, and Ball, 1999) to be inert rather

than active. We need mention just a few:

I Resources can be embezzled, or spent to
hire friends and relatives of little ability.

I Resources can be spent on what economists
call rents—for example, on increases in
salaries that do not call forth greater efforts,
a larger supply of prospective teachers, or
reduced tumover—~—so that teachers,
administrators, or suppliers beneﬁt at the
expense of taxpayers but conditions in
school do not change and students do not
benefit. This may happen in justifiable
circumstances, for example when teachers
are under—paid relative to comparable
professionals. If such salary increases
reduce turnover, then they might be effective
over the long run—but if turnover is caused
more by teaching conditions than salaries,
then even justiﬁable increases are simply

rents.

I Resources can be spent without changing

practices—as when staff development fails

 

to change how teachers teach—or without
getting reforms “right,” as when a school
makes a change incompletely or
ineffectively. This is part of the
implementation problem, and many studies
of implementation can be read as analyses
of how resources are misspent.

I Resources can be spent on well—intentioned
but ineffective practices—adopting
“technology” without knowing how to use
it, or following the reform du jour that has
no effect on learning. Sometimes resources
are spent on symbolic practices—perhaps a
new program to enhance retention, or a
change in superintendents to assure parents
that everything possible is being done—that
don’t have much chance of affecting

leaming.

I Education requires a great deal of spending
for buildings, administration, transportation,
safety measures, sometimes food and other
social services that are necessary before
even one child enters the classroom; these
non—instructional expenditures average
about 34 percent of overall K- 12 budgets.
While necessary, such resources (and
increases in them) do not contribute in any

direct way to enhanced learning.

I Resources may be spent on changes that
are necessary for learning to improve, but
are not by themselves sufﬁcient—spending
on computers without staff training, on
reducing class size without worrying about
the supply of qualified teachers, or
alleviating the commotion within a school

without improving teaching. A perfect

example involving class size came in Austin,
Texas, where 15 schools received an
additional $300,000 per year for five years.
In thirteen cases achievement didn’t change
despite reductions in class size; as the
district superintendent explained, “They
didn’t change the way they were doing
things . . . All they did was take that
support, lower pupil-teacher ratios, still use
the same curriculum, still use the same
instructional methods” (Murnane and Levy,
1996, p. 94). Often schools with windfalls
of money have to spend a great deal on
deferred maintenance, new buildings, or up-
to-date textbooks—all surely necessary but

unlikely by themselves to improve learning

I Resources may be spent on changes whose
effects are distant—like improving school
climate and orderliness, involving parents or
investing in family literacy programs, or
developing a cadre of committed teacher-
leaders able to spearhead reform. Then if
change takes place—a new principal or
superintendent, a different reform du jour, a
shift in the governance structure—the
resources spent in earlier changes are
effectively lost. The problem of rapidly
changing reforms is part of this issue since
developments in one direction are then

reversed.

While there are surely more ways of
converting money into ineffective or inert
resources, these are enough to clarify our point.
Furthermore, these reasons tend to fall into one
of two categories: some of them (the ﬁrst three,

as it happens) undermine any changes taking

 

place in schools and classrooms; others (like
the last four) do change practices, but the
practices are in the end ineffective. The trick is
both changing practices in schools and making
changes that matter to learning, however
learning is deﬁned or assessed.

What’s troubling about public schools is
not that they occasionally misspend resources,
but that they seem to be structured to do so. It’s
simple to develop a “theory” of resource use
that identiﬁes the structural conditions——
political, organizations, and economic—that
direct resources away from enhanced learning.
Consider these widely-accepted features of
public education, at different levels of the
schooling system:

Conventional politics as practiced in the
US. usually involves interest groups in the
struggle for scarce public resources, where the
power of interest groups derives from their
cohesion and numbers and not from the
rightness or effectiveness of their causes. The
constituency for jobs is often more powerful
than the constituency for improved educational
performance, and resource—starved communities
are most likely to battle over resources as
sources of jobs and salaries, not as inputs to
improve education. The voices for improved
performance (like those of parents) tend to be
fragmented into class and racial groups anyway,
particularly in urban districts. The best chance
for interest group politics to improve the quality
of education probably exists where parents with
homogeneous preferences and organizational
competence can exercise their voice—like

middle-class parents in small suburbs.

In addition, education seems especially
prone to symbolic politics, which by deﬁnition
engages in battles that are not about the
effectiveness of educational practices but about
the image and authority of grown—ups. The
current debates about eliminating social
promotion, the development of “naming and
shaming” tactics in district and state policy,7
and much of the current standards movement
strike us as largely symbolic, more concerned
with an image of educational toughness than
with any careful analysis of what might help the
performance of students. (Or, putting this in a
kinder light, demands for enhanced
performance are only half the problem, with the
other half being the ability of schools to
respond—but politics now seems to gravitate
toward the tough demands.) Symbolic politics
are likely to be especially acute in urban
districts because the depth of problems makes
symbolism attractive, because racial politics
often gets converted into symbols (like the race
of the superintendent and his or her staff), and
because hardball politics—«say, Guiliani in New
York or Daley in Chicago—discourage
considered decisions, consensus and
compromise.

Several much—discussed characteristics of
schools as organizations—loose coupling,
organizational inertia, the lack of slack
resources—exacerbate the implementation
problem, so that resources for reform are often
spent without changing much of anything. In
addition, the technical difﬁculties of developing
effective practices, which (we argue below)
require a series of jointly necessary and

individually insufﬁcient conditions—or, to

 

revert to slogans, whole—school (or whole-
district) change rather than piecemeal
reform—are substantial in any event. But they
are particularly difﬁcult under the conditions in
many schools—including disagreements over
goals and pedagogies, instability in personnel,
inconsistencies in conceptions of roles, the
inevitable lack of resources. Instability and
turnover—~among teachers and administrators,
as well as among students—make changes
requiring stability much more difﬁcult. All
these conditions appear to be worse in urban
districts where the lack of consensus, poor
personal relationships (Payne, 1998), the
instability of virtually all personnel, and the lack
of slack resources all thwart change in any
form.

Within schools, many classes seem to
have developed implicit bargains or treaties:
teachers pretend to teach, and students pretend
to learn (e.g., Powell, Farrar, and Cohen, 1985).
The day of reckoning, when students leave
school for a competitive world, is too far off, or
poorly understood; the discipline necessary on
all sides to continue learning is too difﬁcult,
particularly when teachers feel that they work
under difﬁcult conditions, or students are
distracted by family issues, or the siren calls of
an entertainment culture are too strong. As in so
many other areas of social life, a high-quality
(or high—learning) equilibrium requires real
effort to sustain it; once these conditions fail,
schools and classrooms slide into a low—quality
(or low-learning) equilibrium, where
participants reinforce each other’s ineffective
behavior. Under these conditions all the

resources in the world won’t make much

difference. Politicians and reformers may
deride low expectations and low motivation all
they want, but rhetoric is hardly sufﬁcient to
reverse the situation. The standards movement
might have a way out of the low-quality trap, by
providing all participants the same motivation
simultaneously, but only when students and
teachers alike accept the legitimacy of these
standards and have the facility to respond to
them.

Finally, school ﬁnance litigation, which
has been such a powerful inﬂuence on
spending issues, is necessarily a crude
instrument of policy. Courts can forbid
practices but cannot (or will not) usually
specify what should take place, and their
remedies are usually fashioned in the simplest
of terms. This has reinforced a tendency to see
the problem as one of spending levels rather
than the use of resources, since courts have
usually been concerned with spending. Only
recently, with conceptions of adequacy based
not merely on spending but instead on efforts
to achieve certain minimum levels of
performance (e.g., Clune, 1994; Minorini and
Sugarman, 1999), is there any possibility for
the instructional concerns to enter litigation
strategies.

So it isn’t hard to develop a long list of
structural conditions, rooted in the nature of
politics, in the organizational features and
economic conditions of public education, in the
relationship between legislative and judicial
decisions, that indicate why resources may be
poorly spent rather than spent in ways that
might enhance learning. Furthermore, our

working hypothesis is that these conditions are

7

 

much worse in urban and poor districts (e. g.,
Ballou, 1998; Payne, 1998), where
improvements in learning are most needed,
compared to afﬂuent and suburban districts.
From this perspective the problem in urban
schools is not necessarily that resources are
inadequate (though that may be true in some
districts) nor that the conventional urban
“problems” like students without parental
support, unmotivated students, too many LEP
and special education students are so serious
(though that too may be true), but rather that
structural conditions make urban districts
unable to address their own issues with the
resources they have. Ineffectiveness in
spending and inert resources come to dominate,
and it’s small wonder that “resources don’t
make a difference.”

Each of the structural conditions we have
outlined contains the idea for its solution.
Eliminating symbolic politics in favor of
reasoned debates about what would be best for
children can happen within individual schools,
and certainly does when a community, a
principal, and some dedicated teachers coalesce
around educational goals (e.g., Meier, 1995;
Comer, 1996). Replacing interest group politics
with more principled debates over education
and resources can certainly take place with new
approach to teacher unionism and less
confrontational politics (e. g., Kerschner and
Koppich, 1993; Peterson and Chamey, 1999).
Reducing many forms of instability also
happens in successful schools that have
generated loyalty among staff and consistent
attendance among students. So the purpose of

articulating the political economy of resource

8

use is not to be discouraging, but rather to
clarify the conditions necessary to make more
effective use of the resources schools already
have.

Much of what we’ve discussed so far
involves the problems of using resources to
change schools and classrooms, rather than to
pump more money in and leave the essential
conditions of teaching and learning unchanged.
But the second necessary question is what
changes are effective. What practices within
schools and classrooms will really enhance
learning? If we knew the answer to this
question, then we could concentrate our efforts
on spending money wisely, on clearing away
the organizational and political and economic
barriers to effective spending. This is the
direction of the Obey-Porter legislation
providing federal funds for “proven practices”;
of categorical funds for speciﬁc practices
“known” to work, ranging from Head Start to
school lunches; of foundation efforts to
replicate successful models of practice; of
reformers creating networks of schools
replicating their promising practices, like the
Coalition of Essential Schools, the Accelerated
Schools movement, schools following Success
for All or the Comer model of community
involvement of efforts to develop small schools.
This is also the strategy of the effective schools
literature, which tried to identify the conditions
in schools that were “known” (through
statistical analysis) or thought (by reputation) to
be especially effective.

One problem is that it’s nearly
impossible to come up with a list of “proven
practices.” The American Institutes for

 

Research inventory of 24 school wide reforms
(AIR, 1999) claims that only three of the 24
have “substantial” evidence for their
success—and then the detail (which few except
researchers will read) reveals that one of the
three (High Schools That Work) has no
comparative research; another (Direct
Instruction) has been evaluated by standardized
tests well-suited to its emphasis on specific
skills but not to broader conceptions of
learning.8 The effectiveness of several major
federal categorical prograrns—particularly
Chapter I, and (for political as well as technical
reasons) bilingual education—has been
extremely controversial. The networks of
reforming schools all acknowledge that the
ﬁdelity of the reform varies enormously, which
makes it difﬁcult to distinguish the
effectiveness of the reform idea itself from its
implementation. And controlled experiments
have always been anathema in education (Cook,
1999) and so—in contrast to the job training
world, where they have been much more
common9—hard—nosed analysts can always
claim that some aspect of selection or self—
selection is responsible for any positive results.
Therefore it’s hard to come up with a deﬁnitive
list of “proven practices,” and educators
continue to be blasted for not having the
evidence that other ﬁelds—medicine is the most
common comparison—presumably have.

The second problem is that, as we shift
from issues of implementation to questions of
what practices we want to implement, the
unavoidably political issue arises of what we
want education to be. Do we, for example, want

to measure outcomes with standardized tests of

grammar facts and decontextualized time-rate—
distance problems, or do we want authentic
writing exercises and real problem-solving?
Where do we stand in the reading wars and the
math wars, between the advocates of
conventional teacher-directed, behaviorist
instruction with content drawn from school
versions of the conventional disciplines and the
advocates of more constructivist, student-
centered, interdisciplinary and project—oriented
instruction?lo Where do we stand on various
versions of bilingual education, some that
simply help non-native children pass
conventional multiple—choice English tests and
some that value and help maintain important
aspects of their original culture? Sometimes an
educational idea is elastic enough that many
different practices can march under its
banner—like bilingual education, or the
Coalition of Essential Schools that seems to
include both highly traditional schools drawing
on the principle of “using one’s mind well” as
well as constructivist schools drawn to its use
of exhibitions—but this only exacerbates the
problem of deciding what the reform is. And so
the notion of “proven” or “promising”
practices that enhance learning can’t be
disentangled from discussions about what
education is all about. In contrast, most tests of
“proven practices” and virtually all the state
and district efforts to improve schools through
“naming and shaming” transacts this
discussion by relying on conventional multiple
choice test scores as the only measure of
effectiveness.

In the absence of any deﬁnitive evidence

about what enhances learning or any political

 

consensus about what learning is, we often
place resources in the hands of people who
might know. Sometimes these are principals,
who have been granted greater powers in some
version of site-based management; in other
forms, school councils including parents,
administrators, and teachers have the power to
decide. Sometimes, as in voucher and choice
mechanisms, parents are given the right to
decide what enhances learning. Conventionally,
local school boards have been given this
authority—and district policies and practices
presumably express their decisions about
effective practices. More recently states have
played more active roles; many recent policies
are more concerned with the quality of
instruction than the allocation of funds has
been—state tests for students or teachers, state
standards or curriculum guides, staff
development and pre-service education
requirements, pilot projects of various kinds.
And the federal government has resources too,
and in some legislation—Obey—Porter, the
expansion of Chapter I to allow whole-school
changes and practices other than pull—out
programs, the requirement to integrate academic
and vocational education in the Perkins
Amendments—Congress expresses its
conceptions of effective use of resources.

Each of these implicit decisions about
who is best able to decide the effective use of
resources has its own limitations. Principals are
often unengaged in much thinking about
educational issues (as we will outline in Section
IV), teachers may continue with business as
usual, parents have limited information about

educational alternatives, school boards and

10

legislatures are concerned more with
conventional politics than with effective
practices, and so on. In many circumstances,
asking who has the power over resource
allocation—and therefore over explicit or
implicit decisions about effective
practices—simply returns us to the political and
organizational issues we raised earlier. But in
the absence of much certainty about “proven
practices,” we might at least ask who has the
budgetary authority to determine how resources
are spent and what their engagement with
educational issues is.

And so both the major questions of the
“new” school finance—how are resources
spent at the school and classroom levels, and
whether they are spent on practices that might
enhance learning, however measured—are
difﬁcult, each in its own way. And while many
of the answers may be obvious and
discouraging, there’s no substitute for
confronting the conditions under which

resources are now going for naught.

II. Converting Resources
to Results: Opening
the Black Box

Several areas of research have wrestled
with the ways resources are spent, and could
therefore provide some guidance about how
best to spend available resources.
Unfortunately, these investigations have often
not gone far enough into the school and the
classroom (even though other lines of research

have), and as a consequence their results are

 

often unhelpful. In thinking about more
productive approaches to research, in Section
[[1, it’s helpful to review these lines of research,
particularly the efforts to link outputs to inputs
through educational production functions and
the attempts to identify the characteristics of
effective schools. Finally we will review a series
of claims of what practices might be effective,
from reformers of different stripes.

Educational Production Functions

Educational production functions, like
production functions in economics, attempt to
link inputs to outputs, but without being precise
about the nature of production itself—that is,
about the nature of education. The most
common production function can be simply
represented as

(1) SO = f(R, FB) + u

where SO represents school outcomes, R
includes information about resources, FB
measures the effects of family background, and
u is an error term. In theory SO could include
any kinds of educational outcomes, including
competencies measured by new and
“authentic” assessments as well as
conventional test scores, and including changes
in various abilities as well as levels of
competencies. In practice outcomes have
invariably been measured by levels of
conventional test scores. The school resources
R are generally those that can be readily
measured—spending per pupil, pupil/teacher
ratios, teacher experience, sometimes measures
of teacher “ability” like test scores, and other
school resources like library books and science

labs. In theory resources R could include those

accumulated over time, in a series of schools
from kindergarten through the time when
outcomes are measures, though a static and
therefore incomplete measure of resources is
conventionally used (except in Krueger, 1997).
Measures of parental occupation, or education
levels, or income levels are normally included to
correct for the effects of family background,
and occasionally more detailed information is
available—for example, about reading material
and other practices in the home—to capture
other resources available to students. Such
equations have been estimated for many
different data sources, with different units of
analysis—sometimes individual students,
sometimes classrooms, sometimes schools or
districts. There’s also a large literature on third-
world countries (reviewed in Fuller and Clark,
1994) which departs from the American results
where there are truly enormous differences
among countries—for example, in the
availability of textbooks—but which often
reproduces many of the American conclusions.
The ﬁnding that the effects of resources
are, more often than not, statistically
insigniﬁcant (Hanushek, 1986; Hanushek et al.,
1994; see also the symposium on Primary and
Secondary Education in Journal of Economic
Perspectives, Fall 1996) has often been
interpreted as showing that “spending doesn’t
make a difference” because of the relatively
small and variable effects of school resources
compared to the powerful and consistent effects
of family background.11 One way to challenge
this negative interpretation has been the
technical critique that Hanushek’s summary of

the literature was not properly carried out, and

11

 

that a formal meta-analysis (rather than the
counting exercises used by Hanushek) yields
somewhat more positive results (e. g., Hedges,
Laine, and Greenwald, 1994). The upshot of
this debate about how best to summarize a
literature with wide—ranging results has been a
relatively weak statement: resources might
matter under some conditions (Hanushek,
1997; Hanushek et a1. 1994)—though it isn’t
clear what these conditions might be.

A different response to the conclusion
that “spending doesn’t make a difference” has
been to rely on those few studies that do
confirm a relation between resources and
outcomes. Project START, the Tennessee
experiment in class size reduction, randomly
assigned students to larger and smaller class
and to larger classes with aides, and found
substantial gains in learning, lasting at least
until sixth grade, with especially high gains
among black students (Mosteller, 1995;
Krueger, 1997; Nye et al. 1993). Because the
Tennessee results were presumably achieved
with random assignment,12 they have been
especially widely cited. In addition, Ferguson’s
(1991) analysis of Texas districts found
signiﬁcant positive effects of teacher scores on
a statewide test, students per teacher, teacher
experience, and teaching with master’s degrees;
Ferguson and Ladd (1996) found substantial
effects of school resources in Alabama. Most
recently Payne and Biddle (1999) critiqued the
methodologies of conventional production
functions and then went on to estimate an
exceedingly simple and aggregate (district-
level) equation, claiming to ﬁnd a strong effect
of spending per pupil on math achievement.

12

One problem with this response is that it
ignores the many serious research efforts which
have failed to ﬁnd signiﬁcant effects—the
uncomfortable fact that Hanushek has tried to
remind us of. In the realm of random
assignment studies, for example, the frequent
citations to the Tennessee experiments usually
fail to mention an earlier experiment in Toronto,
with a greater range of class sizes, a more
transparent randomization procedure, a much
richer variety of outcomes measured, and a
more lucid explanation of the results, but that
failed to find effects of resources on any of six
test scores except for math concepts (Shapson,
Wright, Eason, and Fitzgerald, 1980). A second
problem is that even those studies that have
found certain resources to be effective have
acknowledged that they cannot tell why
resources might make a difference. For
example, Ferguson’s (1991) analysis of Texas
districts acknowledged that “we can only
speculate what teachers with high scores do
differently from teachers with low scores” (p.
477). Similarly, the Tennessee experiments
could not clarify why smaller classes made a
difference. One of the studies suggested that
greater teacher morale, more frequent teacher—
student interaction, or a greater variety and
extent of student participation might be
responsible (Finn and Achilles, 1990). On the
other hand, Mosteller (1995) inferred from
teachers and administrators that small classes
might be better for socializing young children
to school, and Krueger’s (1997) results are also
consistent with a socializing mechanism—an
interpretation that suggests that small classes
would have declining effects as students

 

become older. In a re-analysis of these data,
Goldstein and Blatchford (1998) noted the high
among-school variation in reading
scores—implying that the effects of class size
reduction might have differed among schools,
possibly because of teacher expectations of
students who had been in smaller classes. Once
again, they called for greater attention to the
processes within the classroom in order to
illuminate the causal process.

A substantive or educational critique of
this literature—that is, one that pays attention to
the conditions of teaching and learning—is that
it treats the educational process as a black box,
and fails to specify how resources are used.
(These studies also fail to specify precisely how
socio—economic status and other aspects of
family background affect educational outcomes
like test scores, though this point has been less
widely noted.) From the perspective of the
“new” school finance, spending per pupil may
increase, but without knowing more precisely
how resources are used, it’s inappropriate to
expect that increased revenues will increase test
scores or any other outcome. Similarly, teacher
experience might enhance learning as
experienced teachers develop more effective
practices through trial and error, through staff
development, or through collaboration with their
peers; but given the inherent weaknesses in
these forms of improvement and the possibility
that experienced teachers are burned out rather
than skilled, there’s no a priori reason to
expect greater experience by itself to increase
test scores or any other outcomes. If teachers
with more education tend to receive more

degrees in their disciplines and become more

oriented to coverage of the discipline, rather
than learning more about teaching strategies,
then additional formal schooling might not
increase student learning. If pupil—teacher ratios
are reduced but teachers continue lecturing in
the same old ways, then again an expensive
reform will fail to improve learning in any form.
If resources like library books and computers
are available but go unused—or, even worse, are
used in pull-out sessions with librarians or
computer specialists who interrupt regular
classes—the availability of such materials might
not enhance learning and might even reduce it.

To know more precisely how resources
are used, direct observations of educational
practices would probably be necessary to see,
for example, whether teachers teach differently
when they have smaller classes, or whether
experienced teachers show signs of bum—out
rather than increased facility.l3 Shapson et al.
(1980), with their collection of information
about teaching practices through classroom
observations, provides conﬁrmation of this
approach: they found that, while teachers felt
that classroom conditions were improved in
small classes, they did not change their teaching
practices in many ways, and so a reduction in
class sizes without an attempt to change
practices was ineffective.

One way to summarize this critique of
conventional production functions is to
elaborate the formal model used. The simplest
approach is simply to recognize that resources
can be used to enhance various kinds of
teaching conditions TC, and that these teaching
conditions in turn enhance learning and
outcomes of various kinds. Formally,

13

 

(2) TC = f(R) + e

(3) SO = g(TC, FB) + u

The ﬁrst of these equations describes the
ways in which resources are (or are not)
translated into the classroom and school
conditions related to learning—for example,
teachers who have mastered and practiced a
range of pedagogies, schools with a collegial
atmosphere in which teachers provide sustained
support to one another, a school atmosphere
that is purposive and orderly.14 The second
equation describes the effects on these teaching
conditions on valued outcomes of schools, both
cognitive and non—cognitive. In contrast to this
formulation, the conventional production
function (equation 1) is a reduced form version
of (2) and (3), conﬂating two very different
processes.

We can continue to elaborate this model.
One further approach (particularly emphasized
by Cohen, Raudenbusch, and Ball, 1999)
recognizes that students come to school with
very different abilities to beneﬁt from
conventional schooling (call this student ability
to beneﬁt, or SA). This reﬂects differences
among students in their cognitive preparation
for schooling, in their motivation, in the
expectations of their parents, in the resources
and discipline provided by their parents and
others around them. The variations in students’
ability to beneﬁt from instruction provides a
speciﬁc way for family background to inﬂuence
school outcomes, but this ability can also be
enhanced by public resources—by early
childhood programs, the efforts of teachers in
the early grades to socialize children (as in the
Tennessee results), family literacy efforts,

14

programs to increase parent participation,
guidance and counseling, mentoring efforts, and
the like. Conversely, it may be undermined by
conditions leading to student resistance (Willis,
1977), or black students’ notion that school
success is too “white” and therefore
reprehensible (as Fuller and Clark, 1994,
emphasize in their description of the
“classroom culturalists”). Furthermore,
teaching conditions and a student’s ability to
benefit from schooling surely inﬂuence one
another. For example, teachers may respond
positively to motivated students and negatively
to those who are disruptive; schools provide
different levels of resources through tracking or
teacher assignments to students perceived to
have different levels of preparation—sometimes
more and sometimes less (Gamoran, 1988;
Brown, 1988). And student motivation may
increase as teachers shift to more constructivist
practices, or to project—based learning, or as
schools create more orderly and learning—
centered cultures. Formally, this means that TC
and SA affect one another, or:

(4) TC = f(SA, R, . . .) +e

(5) SA = g(TC, R, FB, . . .) + u

(6) SO = h(TC, SA, FB, . . .) + v

Compared to equations (2) and (3), this
model places greater emphasis on the many
effects—both school-based and family- or
community—based—of student engagement and
motivation. However, the reduced—form
equation from this model is still equation
(l)—clarifying once again that conventional
production functions ignore the multiple
processes internal to schools by which

outcomes are generated. If, for example, a

 

school allocates experienced teachers to
unmotivated students and thereby enhances
their engagement, but not by enough to
eliminate the differences among motivated and
unmotivated students, then a conventional
production function will show that teacher
experience reduces outcomes when it has in fact
narrowed these differences.

This example clariﬁes another problem
with conventional production functions:
There’s no reason to think that patterns of
allocating resources and of generating teaching
conditions are the same in all schools. The
conversion of resources into teaching
conditions, in equation (4), is a process that
principals under site-based management, or
parent or school—level councils, can in theory
inﬂuence. Similarly, the use of resources to
affect student motivation and ability to learn, in
equation (5), may vary from school to school,
or at least from district to district as different
programs to motivate students are attempted.
Indeed, as Brown and Saks (1980, 1982) have
pointed out, individual teachers make resources
decisions within their classrooms, as they
allocate more or less time to low-performing
students, to certain kinds of students (girls, or
Latinos, or LEP students, for example). The
attempt to estimate reduced-form production
functions like equation (1) assumes that the
underlying processes within districts, schools,
and classrooms are uniform. If this is not the
case, then the reduced—form coefficients will be
weighted averages of the coefﬁcients for
individual districts or schools, and could readily
average positive and negative coefﬁcients and

ﬁnd effects to be close to zero. Thus one point

of elaborating the processes by which
schooling outcomes are achieved is to focus our
attention on the different ways schools (or
individual teachers, for that matter) can use the
resources they are given.

A second problem with conventional
production functions, particularly if we use
them as guides for statistical analysis using
conventional linear regression, is that they don’t
help researchers and educators think about the
interactions among different variables. As we
argued above, many conditions necessary for
effective learning may be necessary but not by
themselves sufficient. For example, from
equation (6), teachers skilled in constructivist
methods (one element of TC) will not be
effective if students have been prepared only in
behaviorist classrooms (part of SA) and do not
work well under the less obviously disciplined
conditions of a student—centered classroom;
either teachers have to re-socialize their
students, or they may revert to more
conventional teaching. If there needs to be an
equilibrium between teachers’ approaches and
students’ expectations,” then neither is
effective without the other. Similarly, reduced
class size might required staff development in
order to enable teachers to change their
approaches; to use computers effectively,
teachers almost certainly require more staff
development (President’s Committee, 1997);
and many reforms intended to reshape teaching
fail because of the lack of staff development, or
principal support, or stability, or assessments
that are consistent with their goals. A
congruence among teaching staffs, principals
and other administrators, and district policies is

15

 

often necessary, so that reforms can fall apart
when a school takes one route while the district
tries to impose another. But the statistical
techniques used by researchers are not good at
detecting the inﬂuences of variables that are
necessary but not sufﬁcient, and so
conventional production functions are also
liable to misstate what happens within
schools.16

Obviously there are many data problems
in thinking about how to estimate equations (2)
and (3), or alternatively (4), (5), and (6).
Outcomes are usually measured by
conventional tests scores rather than more
varied measures of learning; gain scores, or
changes in learning, are rarely available;
measures of teaching conditions TC and
students’ ability to beneﬁt from instruction SA
are not widely available; few data sets provide
any information on the cumulative experiences
and resources of students. Therefore these
equations are for the moment more useful as
metaphors for conceptualizations of the
processes underlying learning, to focus the
attention of researcher and educators on the
important issues, rather than as equations that

could be statistically estimated.

School Effectiveness Research

The literature on effective schools
developed in response to findings from early
production functions that “schools don’t make
a difference.” The research began to examine
the characteristics of effective schools, some of
them selected as outliers in statistical studies of
test scores controlling for the socio-economic
status of students, and others selected by

16

reputation. Various studies came to roughly the
same conclusions, often summarized as the
“five-factor model”: effective schools are those
with strong administrative leadership, high
expectations for student achievement, an orderly
atmosphere conducive to learning, an emphasis
on the acquisition of basic academic skills, and
frequent monitoring of student progress
(Edmonds, 1979a and b; Austin, 1981; Clark,
1980). Of course, the number of factors can be
expanded; for example, Sammons, Hillman, and
Mortimore (1996) have articulated an 11—factor
model relying somewhat more on British and
European findings—though all the factors cited
in these reviews tend to overlap. Some of them
(e. g., strong leadership, an orderly and learning-
on'ented environment) emphasize characteristics
of schools, while others (like high expectations,
an emphasis on purposeful teaching and
learning) focus on characteristics of individual
classrooms and teachers, though
understanding that some characteristics of
effective classes are made much easier if
conditions within the school are supportive.
Quickly, however, effective school
research came under attack (e. g., Purkey and
Smith, 1981; Rowan, Bossert, and Dweyer,
1983; Cuban, 1984; Cohen, 1983). One
problem, present also in educational production
functions, is the measurement of effectiveness
by conventional test scores; then the ﬁnding
that effective schools stress conventional
academic skills—the only competencies
covered in conventional tests—and monitor
students constantly (with the same standardized
tests) is almost a tautology since it says that
effective schools as measured by conventional

 

tests stress the competencies required by
conventional tests. The methodologies of
identifying exemplary schools have also been
critiqued, and the case study methods used to
investigate schools were often unstandardized
and incomplete—though the consistency of
ﬁndings in the ﬁve—factor model tend to offset
this criticism. Finally, the results of this
literature have often been cited and used in a
formulaic fashion, as if the ﬁve—factor model
could be considered a recipe for success; in
practice, however, the implementation of these
factors is considerably more complex, with
some of them (e. g., the need for strong
leadership) frequently misinterpreted or used in
simple ways that did not consider the
interactions among the factors inﬂuencing
learning. As with educational production
functions, the effective schools literature did not
have a way of thinking about conditions that
might be necessary but not sufﬁcient.

For our purposes, two aspects of the
school effectiveness literature are particularly
important. One is that, with their emphasis on
the characteristics of schools, most studies
failed to enter the classroom to see what
teachers might be doing differently in these
schools (Tedlie and Stringﬁeld, 1993, Chapter.
10 and Conclusions). Thus much of the school
remained a black box, as it continues to be in
educational production functions.

Second, this literature has little to say
about the allocation of resources. Indeed, many
of the characteristics of effective schools in the
ﬁve—factor model can be implemented without
additional resources—though perhaps different

(and more expensive) staff development might

be useful to enhance expectations and develop
programs for monitoring student progress. But
in general these studies were either silent about
resources, or failed to find any systematic
relationship between effectiveness and the
resources provided to schools. One could read
the school effectiveness literature as reinforcing
the simple conclusion of educational production
functions that “resources don’t make a
difference.”

However, a different reading suggests
instead that the ways resources are used, rather
than the level of resources, has made a
difference to effective schools—and the study
of effective schools at least began to identify
what ways of spending money might be
effective. In addition, some authors have noted
that some minimum spending level might be
necessary but not sufﬁcient to guarantee
effectiveness; as Gray (1990) concluded, in
language similar to our own, “adequate levels
of resources seem to be necessary but not a
sufficient condition for a school to be effective.
In twenty years of reading research on the
characteristics of effective schools I have only
once come across a record of an ‘excellent’
school where the physical environment left
something to be desired” (p. 213).

Thus both educational production
functions and the school effectiveness literature
suffer from some similar flaws, including
reliance on conventional test scores, a simple
model of causality, and an inattention to the
“black box” of the classroom. The case
studies of effective schools did at least enter
schools and ask what was going on within
them; the methodology remains an attractive

l7

 

one, to which we return in Section III, and the
conclusion that resources might be necessary
but not sufﬁcient is consistent with our
argument throughout this paper—and may lead,
as we will see in Section V, to a plan for
different kinds of funding mechanisms. But
neither literature has been particularly
successful in linking resources to outcomes
through the mechanisms of schools and
classrooms.

Identifying Effective Teaching Conditions

In order to think abut schooling in this
more complex way, we need to distinguish
among the resources R that schools and
districts have; the teaching conditions TC that
might inﬂuence learning outcomes, some of
which require resources and some of which
may not; and the student ability to benefit from
instruction, which may come from school
efforts including resources, from parent and
community inﬂuences, from social or cultural
inﬂuences, and from other sources. Resources,
which are the most readily measured of these
three kinds of variables, can be described by
expenditure per pupil, or (as in conventional
production functions) can be disaggregated into
components like the teacher-pupil ratio T/P,
average teacher salaries S, and other
components of expenditure like administration
costs per pupil A, materials costs per pupil M,
capital outlays per pupil K, and so on, leading
to the obvious identity:

(7) R_ (T/P)S+A+M+K+...

Given a conventional salary structure, the
average teacher salary reﬂects the proportion of

teachers with greater amounts of experience and

18

higher credentials together with the salaries paid
to each category of teacher. Therefore an
identity like (7) includes almost all the resource
variables conventionally included in production
functions like equation (1).

It’s more difficult to specify measures of
teaching conditions that might enhance student
outcomes, because there’s been such varied
research relating such teaching conditions to
outcomes (equation 3 or 6)—~—this is the
problem of coming up with “proven
practices.” However, even if there’s not much
definitive research, it’s worth summarizing what
various researchers and advocates have said
about effective teaching conditions because the
list clarifies the kind of information that
researchers could collect and the goals that
reformers might pursue. Here’s a list,
necessarily partial, of what to look for in order
to measure teaching characteristics TC and
student ability to benefit SA:

1. Teacher characteristics: Teachers with
regular credentials, rather than emergency
credentials; teachers with degrees in their
ﬁeld of specialization, rather than teachers
teaching out of their ﬁelds (Raudenbusch,
Fotiu, & Cheong, 1999; Monk, 1994);
teachers with course work in pedagogy
(Monk, 1994); teachers with high verbal
scores or other measures of ability.17 The
efforts to recruit more Black and Latino
teachers assumes that the race or ethnicity
of teachers matters, at least to some
students or to some forms of learning
(such as those about race), but this
Shibboleth has rarely been examined

empirically.

 

2. Staff development: Many reforms

require staff development, and many
reforms have failed for lack of staff
development. However, given the
widespread perception that most staff
development is ineffective—particularly
Friday afternoon workshops and other
“one-shot” efforts—the form of staff
development may be more important than
the amount, with continuous efforts with
peers a potentially more effective
approach (Darling-Hammond and
McLaughlin, 1996; Lieberman, 1996). A
related condition requiring resources is
organizational “slack,” or time and
energy that can be devoted to
improvement; schools in which teachers
are frantically busy (as most schools are)
don’t have any resources to put into
either diagnosing or correcting their

problems.

3. Interactions among teachers: Common

4

planning time (Miles and Darling-
Hammond, 1998), or school structures
like “houses,” schools—within-schools,
Academies (Stern, Raby, and Dayton,
1992), majors (Grubb, 1995), or joint
classes that facilitate or require common
planning. It’s clear that such structures
are necessary but not sufﬁcient, since
some schools have houses or Academies
without having them do anything; they

exist on paper only.

Interactions between teachers and
students: Many commentators,
particularly those stressing student-

centered teaching, have emphasized the

respectful and considerate treatment of
students, and many horror stories
(including tales of racist and sexist
practices) describe instances of abusive or
uncaring teachers. Miles and Darling-
Hammond (1998) emphasize teachers
knowing students well, and go on to
describe some practices that can enhance
that: creating smaller schools, or smaller
units (like houses or Academies) within
large schools; reducing the number of
aides and specialist teachers so that
teaching resources are spent on
classroom teachers; teaching in longer
blocks of time so that teachers teach the
same amount of time per day but face
fewer students. Class size reduction is
another method of enhancing the
interactions between teachers and
students; however, if these reductions are
relatively small‘8 or if teachers continue
lecturing as they always have, then an
expensive reform may not inﬂuence

learning at all.

5. Effective teaching practices: There is

an extensive literature investigating
effective teaching practices which has
identiﬁed practices like time on task and
on new material, high expectations,
frequent monitoring of progress, a warm
yet disciplined classroom environment,
certain patterns of questioning and
reinforcement, and many other
characteristics that are no doubt necessary
in a variety of settings (e.g., Brophy and
Good, 1986; more generally Wittrock,
1986; Waxman and Walberg, 1993). The
19

 

20

problem is not that the conceptions and
results about effective teaching are
missing; indeed, there are almost too
many results, with different studies
emphasizing different aspects of teacher
behavior. But in particular data sets—«for
example, the NELS data set examined in
Goldhaber and Brewer (1996), or the
NAEP data analyzed in Raudenbusch,
Fotiu, and Cheong (1998—only a few
practices are measured. Only when data
have been experimentally collected—for
example, by the Toronto experiment
analyzed in Shapson et a. (1980) or some
of the international studies summarized in
Fuller and Clark (l994)—has it been
possible to include a variety of teacher

practices.

. Constructivist teaching practices:

Within the past 15 years there’s been
much more attention to constructivist
teaching, and there’s even some evidence
that such practices are more effective.19
Constructivist practices remain
controversial, however, and the “reading
wars” and “math wars” pit advocates for
traditional skills-oriented teaching against
constructivists. A judicious middle
ground—indeed, a Deweyan
synthesis—would argue for a hybrid
approach to teaching combining both
behaviorist and constructivist practices,
and recent reviews by the National
Research Council have supported such a
synthesis (Snow, Burns, and Grifﬁn,
1998).

7. The use of time: Time on task is one of

the most consistently-mentioned
resources. While the absolute amount and
proportion of time devoted to instruction
may be important, others have noted that
blocks of uninterrupted time may be as
(or more) important, encouraging the shift
to block scheduling. The National
Commission on Time and Learning
argued for greater ﬂexibility in the use of
time, allocating time to ﬁt instructional
requirements rather than the other way
around (National Commission, 1994).
However, instructional time is a good
example of a teaching condition that is
necessary but not sufficient. More time
devoted to inept teaching is hardly likely
to improve the quality of learning, even
though simple recommendations to
extend the school day or school year are

common.

8. School offerings: Raudenbusch, Fortiu,
& Cheong (1999) have documented the
effect of offering “advanced” courses
(measured in math by algebra I, which
doesn’t seem particularly advanced). One
reform in Southern states, High Schools
That Work, has emphasized the
replacement of general track offerings——
general math or general science—by
more demanding academic or integrated
courses. In various states including
California, access to the courses required
for university admission is an issue since
some urban schools don’t offer these
courses. Others have stressed the

importance of extensive writing

 

opportunities in all classes, and the
National Writing Project also stresses the
importance of both the amount and the
kind of writing—valuable not only to
learn to write well but also as an adjunct
to reﬂection and constructivist learning.
Evidently many schools—especially
urban high schools—don’t provide
advanced enough offerings to challenge
their students.

9. School climate: an orderly or disciplined

climate (see generally the school
effectiveness literature), or a climate in
which learning is central. The emphasis
on school conditions shifts the emphasis
from the classroom level, and clariﬁes that
what a teacher does in the classroom is
affected by what happens in other
classrooms. This concern starts to move
toward student attitudes toward learning,
since these are presumably shaped by
other teachers—and a student’s entire
history of experiences in schools—as
well as by family and community

inﬂuences.

10. School vision: A related condition is a

vision for a school as a whole, with
consistency of vision between the
principal and teachers a related
characteristic. Many reform efforts—for
example, the precepts of the Coalition of
Essential Schools, or the avoidance of
“remedial” pedagogies in Accelerated
Schools, or “education through
occupations”—have a vision at their
center, one that is more effective when it

is consistently articulated and practiced

11.

12.

by all the members of a school
community. The literature on effective
schools has also stressed the importance
of “strong” principals, where one
characteristic of strength is a vision that is
communicated throughout the school. As
with school climate, the emphasis on a
consistent vision and whole school
reform clariﬁes the importance to
individual classrooms of what happens in
other classrooms.

Stability: The absence of stability is
often mentioned as detrimental to reform,
and instability in student
participation—from moving around
schools or attending sporadically
(Rumberger and Larson, 1998)——is
detrimental to individual students, and
perhaps to classes and schools as well.
Stability is another resource that is
necessary but not sufﬁcient since a
school may be stable but mediocre—so
reform efforts and stability may be jointly
necessary.

Student behavior and ability to
beneﬁt from schooling: These
characteristics can be directly measured
by such variables as self—reported interest
in school, time spent on homework, and
attendance (e.g., Taylor, 1998, using
NELS data). Other measures would focus
on levels of disruption versus cooperation
in the classroom. This kind of variable
provides one explanation for the effects
of family background (or socio-economic
status, or class, or income) on school

21

 

13.

22

performance, though this explanation
would surely be controversial; often, the
reasons behind the effects of family
background are not explicit. These kinds
of explanations have led to efforts to
reshape family and community culture
and make them more consistent with
conventional school values—for example,
through early childhood education to
enhance school readiness, parent
education, family literacy efforts to
facilitate literacy in the home, efforts to
increase parental involvement, and
guidance and counseling to clarify to
students the importance of schooling.
Most of these efforts have been quite
marginal, and some have been
undermined by the class and racial biases
embedded in them, but the tactic of
reshaping families and their inﬂuences
directly—or at least getting them to
reinforce the mission of the

schools—remains attractive to reformers.

Peer effects and students as resources:
It’s clear that students of higher socio-
economic status are a resource to other
students (e. g., Gamoran, 1988); that is,
students do better when they are in
schools with higher SES students, even
controlling for their own family
background. Parents consistently try to
have their children attend schools with
higher SES students—or avoid schools
with low-SES students (as white and
Asian parents avoid schools with black
and Latino students). But precisely what

explains the effects of peers remains

ambiguous, although aspirations or
expectations, compliant versus inattentive
or disruptive behavior, and ancillary
knowledge are among the possible
explanations. Like the shift from the
classroom to the school, the recognition
of peer effects clarifies that the resources
inﬂuencing teaching conditions include
the entire social composition of the
school, and resources devoted to (or
withdrawn from) some students will

inﬂuence other students.

There are several things to notice about
these various teaching characteristics and
dimensions of student ability to beneﬁt from
instruction. One conclusion is that many of
them are necessary but not sufficient, so that
single-dimension reform efforts are likely to be
ineffective. As a result, it’s possible to spend
substantial sums on reforms but have little to
show for them. This seems to be the case with
class size reduction in California, for example,
where there have so far been few changes in
teaching practices in reduced-size classrooms
(CSR Research Consortium, 1999).20

Another conclusion is that, while some of
them (like the staff development efforts
required for reforms, or the introduction of
computers) require additional resources, others
require that existing resources be used in
different ways (staff development is a clear
example). Some of them—like the mobility of
low-income students—probably cannot be
changed for any sum of money, and similarly
it’s hard to imagine how to increase the stability
of teachers and administrators in urban districts

 

under current conditions except by paying truly
huge sums of money. In other cases, changes
might be achieved either though resource—
intensive or cost less ways; for example, the
standard recommendation to have strong
principals might be attained by increasing
salaries to have a larger pool of individuals
from which to choose, or by restructuring the
job so that it is attractive to more teachers, but it
might be possible instead to reshape the
programs that prepare principals so that they
pay more attention to educational rather than to
administrative aspects of the job. A vision for
education is in some sense cost-free, though
developing such visions has usually been the
business of reformers who have devoted
substantial resources to spreading their vision.

A ﬁnal conclusion is that, as we consider
these measures of effective teaching conditions
or student resources, each becomes more
complex. Staff development is not enough; the
kind of staff development, and no doubt its
consistency with a particular vision of reform,
must be known. Time on task may not be
enough; blocks of time and the ﬂexibility of its
use may be necessary. And so elaborating the
conception of resources leads us to still further
elaboration.

As a result, the implications for resources
of trying to enhance any of the teaching
conditions we have just outlined are
ambiguous—at least in our current state of
knowledge. And so, for the moment, we are
forced to come back to the conclusions of the
production function literature: resources might
affect learning under some conditions, but we

still aren’t sure what those conditions are.

III. Implications for Research

From the perspective of the “new”
school finance, the task for research is to
determine the practices within schools and
classrooms that enhance learning, however
deﬁned, and to determine the resources they
require. Discussions of resources should never
be divorced from how they are spent, and the
school and classroom should never be left as a
black box. Now, this is a tall order, since it
requires different types of information in any
one piece of research. However, there seem to
us at least four promising directions for
research. Some of them—like returning to
effective schools research with a slightly
different focus or trying to ﬁnd data sets with
certain variables—retum to previous research
strategies; others, like exploiting natural
experiments, may be somewhat different. But
all of them try to probe more deeply into
schools and classrooms, to understand the
effects of resources “at the coal face,” as the
Aussies say—in the crucial interactions among
teachers and students.

Natural Experiments

Occasionally, schools experience
substantial infusions of money. This has
happened, for example, in school ﬁnance cases
where states have increased their aid to districts,
as in Kentucky, New Jersey, and Texas; this
happened in the Chicago schools in 1990, as
part of the Chicago School Reform Act. In
other cases pilot projects have increased
resources to a select group of schools, as
California’s AB 1274 reforms did; currently a
set of “underperforming “ schools will be

23

 

provided grants up to $168 per student, with the
intent of allowing them to increase their test
scores. These “natural experiments” provide
opportunities to see ﬁrst whether and how the
additional resources are spent to enhance
teaching conditions, and second to ask whether
these changes might improve learning and other
schooling outcomes. The first of these
questions—what do additional resources
buy—needs to be answered not only at the level
of the district and not only by detailing the
effects on gross measures of resources like
class size, pupil/teacher ratios, and
administrative expenses, but ideally at the level
of the school and classroom, by examining the
effects on the teaching characteristics
mentioned in the previous section. There is now
a literature asking “where does the money go,”
tracking the effects of reforms in New Jersey
(Firestone, Goertz, Nagle, and Smelkinson,
1995), New York (Lankford and Wyckoff,
1995), Chicago (Hess, 1999), and nation-wide
(Monk, Nabib, Odden, and Picus, 1995; see
also Picus and Wattenbarger, 1996). Goertz
and Natriello (1999) have summarized the
effects of spending increases in three states,
concluding that patterns of spending by
functional areas (instruction, administration,
etc.) are relatively unchanged after ﬁnance
reforms. But these research efforts tend to
confine themselves to the effects on gross
measures of resources—on administrative
costs, average teacher salaries (rather than what
such salary increases might buy in terms of
teacher ability or stability), and spending for
special education, for example. Another

generation of studies needs to trace the path of

24

spending from the initial increases, to effects on
gross resources, to the effects of these
resources on classroom and school practices.

For example, in his examination of the
reforms in Chicago, Hess (1999) clariﬁed that
schools used their discretionary
funding—which grew from $197 per pupil per
year to $763 over a ﬁve-year period—in very
different ways. Schools with improving
achievement tended to enrich their offerings by
adding computers, music, art, science labs, and
physical education. Those with declining
achievement spent more on resource teachers in
math and reading, classrooms aides, reduced
class sizes, discipline, counseling, and truancy
programs. Arguably, then, the more effective
schools spent their resources on enriching the
curriculum, while the ineffective schools
engaged in “remedial” practices and spending
on personnel other than classroom
teachers—who, according to Miles and
Darling-Hammond (1999) and Krueger
(1997)——-may be ineffective.

An examination of reforms in California
(Little et al. 1999), which provided a select
number of schools with an additional $155
extra per student per year (or about four
percent) over five years, found that resources
were largely spent on “basics” or “more of
the same”—staff time, computer equipment,
and other conventional instructional
resources—because most California schools
are starved for resources (the state ranks 47th in
spending per pupil). Very little was spent on
staff development— which might have led to
changes in teaching and learning—and
spending had little to do with restructuring

 

different schools. A central question remains
whether resources were spent in ways
consistent with a reform vision, and this was
difficult to detect without considering the full
range of funding and reforms.21 In both
Chicago and California reforms, therefore,
schools used additional resources in myriad
ways, the effectiveness of spending can be
undermined in myriad ways, and so the
conditions necessary for additional resources to
enhance learning are both complex and varied.
These kinds of studies would over time provide
a clearer image of the preconditions necessary
for school reform and increased spending to be
more effective.

The approach of learning “where the
money goes” would be most valuable in
understanding the political economy of
resource allocation introduced in Section 1,
including the political and organizational forces
that often undermine the effective use of
resources. In the terms of Section II, this kind
of research is more concerned about the
resource allocation of equation 2 (or equations
4 and 5), and can say relatively little (at least
directly) about the efﬁcacy of these resources
(equations 3 and 6)—though Hess’ effort to
distinguish schools by increasing or declining
achievement is one rough way to approach
efficacy. But at least this approach could help
us learn more about the allocation of resources
to the school and classroom characteristics that
might matter—that are necessary (if not
sufﬁcient) preconditions for more effective
schools.

Evaluating Self-Conscious Reforms

A related research tactic would be to
investigate how self—conscious reforms spend
additional resources, and how they do (or do
not) use these resources to make changes in
practices that might matter to learning. The
investigation of such reforms differs from the
analysis of increases in spending because
reforms are—presumably—driven by a
particular vision of what must change. In
practice, however, some reforms make changes,
but what they do is unlikely to enhance
learning; others may have greater possibility for
enhancing learning, but be unable to effect the
changes required by their vision. Only those
reforms that do both stand any chance of
succeeding—and they may provide guidance
about how to use resources effectively.

Several examples provide a sense of what
such research might undertake. Odden and
Busch (1999) have reviewed the spending
practices of New American Schools. While
these schools have a core of funding practices
in common, including a principal and a teacher
for every 25 students, they also incorporate a
strong standards-oriented curriculum. Each has
a particular vision in addition that directs
marginal resources to particular purposes—for
example, funding for trips in Expeditionary
Leaming—Outward Bound; technology
coordinators and technology-related staff
development for Co-NECT; art and music
teachers and computer technologies for the
Modern Red Schoolhouse. These schools have
budgets within the normal range of schools, but
the speciﬁcities of resource use are governed by
particular visions of reform.

25

 

Similarly, the Puente project is an effort
to improve the performance of Latino students
by developing more constructivist approaches
to teaching English and writing in ninth grade,
using selected Hispanic literature to explore
issues of identity, creating smaller learning
communities of students within large high
schools, adding counselors to these learning
communities to help students understand the
requirements of the schooling system, and
using Latino mentors to provide other resources
to students. An evaluation (Gandara, 1998)
using a roughly equivalent comparison group22
has conﬁrmed the effectiveness of Puente in
enhancing college course completion and
college-going rates. The program’s additional
resources of about $500 per pupil were spent
principally on staff development to train ninth-
grade teachers in constructivist teaching
methods; on additional counselors; and on the
administrative costs of identifying mentors. In
this case again there’s a clear relationship
between the vision of the program and the
patterns of spending, and the elements of the
program reinforce one another; for example, the
additional spending on counselors does not
simply add resources to conventional
counseling, but rather adds counselors that have
speciﬁc roles for Latino student within learning
communities. Gandara documents another
effective (and anonymous) reform effort,
costing $578 per student or roughly the same
as Puente, that spent its resources very
differently, for additional teachers, summer
training, and classroom materials—making the

same point we do that effective reforms can use

26

their resources in very effective ways as long as
they follow a coherent vision of reform.

An evaluation of Career Academies in
California revealed that students in the program,
compared to a similar group of students, were
less likely to drop out of high schools and more
likely to graduate and continue to college. Of
the additional resources, 42 percent were spent
on additional teacher time, since academies have
smaller classes and additional time for teachers
to work with other teachers and employers. An
additional 40 percent of the incremental costs
were spent on employers’ representatives, since
creating internships and other connections with
employers is a critical feature of Academies
(Stern et al., 1989). Again, this reform allocated
its marginal resources to those practices central
to its vision of change.

These examples suggest the variety of
ways that resources can be used to create
improvements in learning. There’s no single
recommendation that emerges from these cases;
rather, there are various ways of spending
additional resources that seem effective.
However, all in these cases, spending is driven
by a particular vision. As the reform vision
dissipates, or is implemented unevenly, the
potential effectiveness of the reform vanishes.
For example, in the case of the AB 1724
reforms in California examined by Little and
her colleagues (1999), the additional resources
of about $155 per student were often spent on
core services in a low-spending state, were
sometimes subject to substitution as districts
redistributed non—AB1274 resources to other
schools, and were sometimes subject to district

restrictions on how funds could be spent—all

 

undermining the intent of focusing additional
resources on reforming practice. Under these
conditions, additional funding was spent within
many schools in bits and pieces, without any
unifying Vision or motive—and under these
conditions it’s unlikely that much improvement

can result.

Returning to Effective Schools

The research tactics already examined
look at cases where resources have increased,
and try to follow the use of resources into the
school and classroom, and hopefully to
outcomes. Another way of carrying out such
research would be to start with exemplary
outcomes, and then trace back to see how
resources are used. This is the tactic of the
effective schools literature—except that these
studies did not generally examine resource use.
In addition, the studies of effective schools
often did not look inside classrooms to see
what (if anything) was different in the teaching
practices of particularly effective schools; many
characteristics commonly cited—strong
principals, an orderly climate, frequent
monitoring of student progress—describe
schools as a whole rather than of classroom
practices.

One possible example of a different
study of effective schools is that of Miles and
Darling—Hammond (1998). They concluded
from the research literature that effective
schools require teachers knowing students well,
and cite the reform literature as supporting
common planning time for teachers to “create
new practices and engage in school problem—

solving.” They then searched for particularly

effective schools, as measured by above-average
and improving student performance, and that
also had mechanisms to enhance the abilities of
teachers to know their students
well—particularly by replacing non-classroom
teachers (like aides, specialist teachers, and
some administrators) with classroom teachers
to reduce class size—and to work with one
another. These were schools that used
resources to particular purposes, and their
practices clariﬁed that the particular use of
time—not just the amount of time on
task—may be important. Finally, unlike the
older effective schools literature, this study
observed in classrooms to see the instructional
effects of these changes. Unfortunately, the
causal relationship between teachers knowing
students well and the outcomes of these
exemplary schools remains unclear, because the
schools were selected to have both
characteristics, and the apparent relationship
between the two may be only the result of
careful selection. Still, the concern with the use
of resources to enhance particular teaching
conditions, plausibly related to learning
outcomes, is a step in the right direction. Future
studies could continue to examine exemplary
schools known to have strong outcomes,
controlling both for conventional resources and
for the family backgrounds of students, and
then search for evidence about how resources
are used.23

We note that the inference problems
evident in the Miles and Darling—Hammond
study are common in the discussion of
effectiveness and resource use. The most

common approach, particularly in journalistic

27

 

accounts like Jonathan Kozol’s Savage
Inequalities (1991), is to identify obviously
ineffective schools and show they have
inadequate teaching conditions— uncaring
teachers, inadequate textbooks, insufﬁcient
space, etc—and then to argue for additional
resources. But if the underlying problem is that
resources are not translated into positive
teaching conditions (e.g., from equation 2), then
no amount of additional resources will ﬁx the
problem.

Estimating Structural Equations

A ﬁnal possibility is that it might be
possible to estimate equations like (2) and
(3)—-or the more detailed equations in (4), (5)
and (6)—to replace the standard educational
production function in equation (1). (This tactic
continues to assumes that such stable
relationships exist, contrary to the argument,
from Murnane and Phillips (1980) and others,
that individual teachers uncover a different
relationship for each student.) The crucial issue
in doing so is to open the black box of
schooling—in particular, to estimate equations
for learning processes, rather than assuming
them away in reduced-form equations. Such
results would potentially generate more
information about which of many possible
teaching conditions contribute to outcomes, as
well as the process of converting resources into
effective teaching conditions, and information
like this might be the only evidence acceptable
to hard-nosed quantitative types. In addition,
once information is available on those teaching
conditions that matter most to outcomes, then

more refined accounting exercises of the kind

28

common in school ﬁnance would be
possible—for example, examining inequality in
teaching conditions compared to the inequality
of expenditures per pupil, and analyzing the
patterns in teaching conditions among districts
of different wealth, racial and income
composition. For example, it’s possible that
students in urban districts are rich in
expenditures per pupil or total resources, but
poor in effective or active resources.

In one study that estimates equations like
(2) and (3), Raudenbusch and his colleagues
used NAEP data first to determine that four
measures of resources—what we would call
teaching conditions—were effective in
enhancing NAEP math scores: the disciplinary
climate of the school, the presence of advanced
course offerings, the preparation of math
teachers in mathematics rather than some other
subject, and the emphasis on reasoning in math
classes (Raudenbusch, Fotiu, and Cheong, in
press). While they did not estimate equations
like equation (3), linking these teaching
conditions to resources like spending per pupil,
they did examine the allocation of these
teaching conditions and found differences in
access to resources by parental income and race
(Raudenbusch, Fotiu, and Cheong, 1998).
Whether these inequalities are greater or less
than the inequalities in spending per pupil is
unclear, though it’s plausible that some of these
teaching conditions—particularly instructors
teaching in their ﬁeld—are much more
inequitably distributed than money, especially
in some urban districts.

Similarly, Goldhaber and Brewer (1997)
used NELS data to estimate the effects of both

 

teacher characteristics (like experience,
certiﬁcation, and major) and teacher behavior
like control over discipline and teaching
technique, the use of small groups, questioning,
and problem—solving. While some of these
teacher practices led to higher math scores, they
found very little relationship between teacher
characteristics and teacher behavior—implying
that buying expensive characteristics like more
teacher education and experience did not
necessarily lead to more effective practices. In
addition, some of the teacher practices—the
percent of time teaching in small groups, the
percent of time instructing individuals, the
emphasis on problem-solving—reduced test
scores, implying either that these constructivist
practices lead to improvements that are poorly
measured in conventional test scores, or
possibly that the proponents of conventional
“skills and drills” are right. Thus the more
detailed these production function become, the
more it may be important to have an array of
outcome measures. The results clarify the
difﬁculty of converting resources R into
effective teaching characteristics TC (as in
equation 2 or 4).

However, in general researchers in the
US. are a long way from being able estimate
such equations. (Surprisingly, there appears to
be a greater ability to carry out this kind of
research in third world countries, where special
studies paying attention to conditions within
schools have been more common“) In most
US. data sets, information about teaching
conditions are sparse to non-existent; indicators
of student ability to benefit from instruction are

even more likely to be missing. In addition,

some teaching conditions are by construction
extremely varied and therefore difﬁcult to
measure: in student-centered classes, for
example, approaches to teaching vary from
student to student, and therefore it may be
necessary to measure teaching conditions at the
level of individual students, not teachers or
classrooms.25 These kinds of equations linking
teaching conditions to outcomes need to be
cumulative, examining the experiences and
resources students have had over the 12 years
of their elementary-secondary education, and
probably the consistency and continuity of
these experiences as well—rather than looking
cross-sectionally at the resources and teaching
conditions available in one particular year.
Outcome measures, which in theory encompass
a wide range of conventional and “authentic”
measures, and change scores as well as levels,
in practice usually get reduced to conventional
test scores (as in the NAEP results). And the
logic of linear regression is wholly inadequate
to educational processes in which several
conditions may be jointly necessary before
learning can take place. These are arguments
for the development of new data sets, created
with the information requirements of the
“new” school finance in mind, and potentially
using experimental designs rather than
continuing to use data of the kind now available
to estimate production functions of ambiguous
interpretation.

Each of these four strategies for research
has its own drawbacks. The first of them, the
investigation of “natural experiments” where
resources are suddenly increased, usually

provides little information on the effectiveness

29

 

of the practices changed. The second, the
examination of reform efforts, can make
comparisons only within a group of schools
trying to reform, and the link between reforms
and outcomes often remains troublesome. The
third, a slightly different way of carrying out
research in the effective schools tradition,
suffers from all the problems of that earlier
body of work. And the fourth, the effort to
estimate equations describing more carefully
what happens within schools and classrooms,
suffers from a variety of measurement and
logical problems, some of which may be
unresolvable.

However, given the inadequacies of the
“old” school finance, and of researchers’
understanding of the effects of different
resources, there’s not much point in continuing
to follow conventional research strategies.
Instead, a multi—pronged and overlapping effort,
with due attention to the weaknesses of each
particular approach, may be the best way to
address the question of how best to use the
educational resources for schools that have

always seemed in such short supply.

IV. Implications for Educators and
Administrators
Of course, schools don’t wait for

researchers to provide them with answers.
District allocations and school-site decisions
are being made all the time, for better and for
worse. The question is therefore whether the
perspectives of the “new” school finance

provide any ways for administrators—including

30

principals and others involved in site-based
management—to think about the effective
allocation of resources.

One conclusion from studies of
administrators is that they spend relatively little
time worrying about enhancing the
“productivity” or effectiveness of the
resources they have at their disposal. Boyd and
Hartman (1988) point out various reasons why
this might be true: uncertainty abut what factors
will increase learning; organizational factors,
including “loose coupling” and the lack of
control over teachers’ actions; and political
constraints on the ability of principals to direct
teachers, including the opposition of unions.
One way of summarizing these findings is that,
in the absence of real freedom to reallocate
resources and of clarity about how they should
be reallocated, harried principals and other
school-site administrators simply don’t have
the time to worry about it. This in turn implies
that district decisions about resources dominate,
wit “decisions” established by precedent and
only small and marginal changes possible.

Where principals have had more say over
budgets, they have become increasingly skilled
at responding to the incentives they face. In
England, for example, the introduction of grant—
maintained (GM) schools freed school heads
from local district authority, and enabled them
to allocate resources subject only to a board of
govemors—though with the requirement to
teach the national curriculum and to be
measured by tests linked to the national
curriculum. Under these conditions most heads
of GM schools have managed to practice some

selectivity, excluding special education students

 

and other pupils (including lower income and
minority pupils) who cost them more, either in
resources to enhance their performance or in
lower test scores. Schools have spent more
money on aspects of “show”——spruced-up
buildings and grounds, for example—since this
is important in competing for students. And
many heads have discovered the formally
correct strategy given the publicity surrounding
their passage rates on certain tests: the triage
strategy is to concentrate the best
teachers—whom heads know better than any
outsider or researcher ever could—on the
students just at the margin of passing, in order
to get them over the threshold; heads allocate
adequate teachers to those students who are
likely to pass on their own; and students with
little chance of passing are kept “warm and
dry,” with the least effective resources allocated
to them (Finkelstein and Grubb, 1998). The
same triage strategy has apparently emerged in
Texas, where the state’s accountability system
requires that 45 percent of students in all
groups (white, black, Latino, and economically
disadvantaged) pass the TAAS (Texas
Assessment of Academic Skills) test. Schools
have then concentrated their resources on the
students who perform close to the pass rate;
those who will pass without additional effort by
the school, and those who have no chance of
passing under any reasonable conditions, are
comparatively ignored.26 The point is not that
these resource priorities are the best ones,
certainly not for the lowest-achieving students;
the point is that school administrators can
quickly learn how to allocate resources

rationally when the opportunity and necessity

present themselves. Note that rational allocation
requires knowing who the most effective
teachers are—something that is achieved in
England, and a few American schools, by
administrators observing in classes constantly.
One advantage of site-based management,
then, is that with the right incentives it would
force school administrators to think more
carefully about the effective allocation of
resources—that is, about the allocation of funds
under their control (R) to teaching conditions
that are effective (TC). But agreeing about what
the incentives should be, then operationalizing
them, and finally providing the necessary
freedom to allocate resources are all politically
quite difﬁcult, particularly in bureaucratic urban
schools. One alternative (or complement) would
be for administrator preparation programs to
provide principals, other school-site
administrators, and district administrators with
the information and the perspectives to enable
them to think more carefully about the
allocation of resources, for example through
realistic projects and simulations. This would
be a marked advance over current programs in
which the allocation of resources seems to be
relegated to last in a long list of priorities.27
The tactic of relying on the judgment of
administrators to allocate resources effectively
is a specific version of the approach, mentioned
in Section I, of giving decision-making power
to the individuals thought to be best able to
make effective decisions. However, the ability
of administrators to make these decisions
matters little if they are not given any discretion.
We note that there have been contrary impulses

about the responsibility of principals and other

31

 

school-level administrators, just as there have
been for teachers: one impulse has been to
grant them greater authority (represented in
site—based management); the other has been to
constrain their actions tightly, to “idiot-proof”
schools by constraining the curriculum,
assessments, and the allocation of resources
tightly—and by subjecting schools to dire
consequences if they fail (as in “naming and
shaming”). Thus the question of the role that
administrators can and do play in allocating
resources is ultimately dependent both on the
capacity of administrators to make these
decisions, and on the state and district policies
that allow them the freedom to do so.

V. Implications for Public Policy
and School Reform: Creating

Complementary Reforms

Finally, what are the implications of the
“new” school finance for the policies enacted
by districts and states? If the “new” school
ﬁnance is a particular perspective at this stage,
rather than a set of concrete recommendations,
how can it be useful?

One implication is that a simple question,
or thought experiment, would be valuable
before many policies are enacted: What will
happen within schools and classrooms if a
particular change involving additional resources
is made? The popular idea of class size
reduction provides a good illustration. While
some teachers may be able to use smaller
classes to teach in different ways, others may

not have much idea about how to modify their

32

teaching in smaller classes—so some staff
development might be an appropriate
complement. The shortages of qualiﬁed
teachers that have materialized, especially in
urban districts, could have been foreseen, and
so a teacher recruitment component could have
been added. And the need for additional space
that has constrained so many California
districts was relatively clear from the outset.
Therefore the pallid results from initial
implementation, and the need to recruit and
prepare teachers as well as to support school
construction (CSR Research Consortium,
1999), could have been anticipated and
forestalled by reallocating resources within this
reform.

A similar thought experiment could be
applied to the current efforts that we (and the
British) call “naming and shaming.” Some
urban districts and many states have begun to
rank schools based on conventional test scores,
and publicized the low performance of the
“worst” schools—a process of shaming. These
are not necessarily the worst schools in any
sophisticated sense: they typically have the
largest proportion of immigrant and low-SE8
students, not surprisingly given the power of
family background, and “naming and shaming”
has often added to the humiliation of those
groups. Such policies have then provided some
“solutions” for low—performing schools
ranging from reconstitution—the threat to
replace all staff—to requirements in many
districts for low-performing schools to come up
with their own self-improvement plans to
additional funds—though these improvements
usually seem vastly inadequate to the task of

 

reforming schools. How precisely will these
policies change the practices within classrooms?
Will the demoralization among teachers
outweigh the impetus to teach more diligently?
Will the need to show improvement on test
scores lead to more narrow teaching to the test,
at the expense of the higher—order abilities that
so many educators and business representatives
have championed? Can schools without any
slack resources, or that are thought to be low-
performing, create their own improvement plans?
Will the shaming process motivate students to
work harder, or will it demoralize them too?
Unless these kinds of questions can be
answered unambiguously, it’s hard to see how
“naming and shaming” can lead to the
improvement of teaching conditions within
schools and classrooms—and these questions
have rarely been posed, much less answered.
The issue is not what answers that researchers or
advocacy groups would give to these questions;
the question is what answers policy—makers
themselves would provide, and whether the need
to answer such questions would cause them to
think about proposed policies more carefully.
Paralleling environmental impact statements, one
could envision Classroom Impact Statements
required to justify major policy enactments, in
order to focus attention to these issues.

A corollary is that the “new” school
ﬁnance asks policy-makers to think about the
complementary policies required to change
outcomes, rather than unitary changes. These
almost always would take the form of resources
plus some additional requisite—money for
computers plus resources for professional

development, class size reduction plus teacher

recruitment, the creation of smaller schools (or
schools within schools) plus resources for
construction costs and a vision of how teachers
and students will interact differently. The
current standards movement provides incentives
for improvement—as does “naming and
shaming”——but often without providing the
intellectual or ﬁnancial resources to respond to
these incentives. In many cases, resources plus
technical assistance maybe necessary, in order
to enable schools to implement reforms
successfully. And in many successful reforms
in Section III, a central vision—or vision plus
resources—was necessary.

To be sure, sometimes policy-makers do
think in these ways. A number of court cases
involving school ﬁnance have led to broader
legislative reformszg—implicitly (and
sometimes explicitly) assuming that reshaping
the allocation of money without reforming other
school practices would be insufficient.
Examples include the 1984 reforms in Texas,
the 1990 reforms in Kentucky (Odden and
Picus, 1992; Adams, 1994), and the 1991
reforms in New Jersey, in which legislators
were explicitly unwilling to provide any more
resources to urban districts without several
accountability measures (Firestone, Goertz,
Nagle, and Smelkinson, 1994). But legislators
do not routinely create legislation incorporating
several complementary reforms; often when
they increase resources for a particular purpose
there seems to be an imbalance between the
central policy and its complements—too little
professional development (or professional
development of the wrong kind), too little

technical assistance facing too many

33

 

implementation problems, too little vision, or a
vision that can be variously interpreted and
therefore can become diffuse and impotent.
Therefore the perspective of the “new” school
ﬁnance would try to establish a more careful
balance among the complementary elements of
a reform involving resources.

In terms of the instruments of policy,
however, we come to a central problem. The
existing state funding mechanisms for schools
(and public colleges and universities too)
allocate most of their money through general
grants in aid, without restrictions on local
spending; local revenue is raised through
property tax mechanisms, and the amount but
not the form of revenue is the main issues. But
the “new” school finance, more concerned
with spending resources in ways that affect
schools and classrooms, would constrain
funding—for example, by providing categorical
rather than general aid, or by providing funding
through constrained pilot programs designed to
replicate successful practices, or by providing
some unconstrained funding along with some
constrained funding (e.g., for technical
assistance or staff development) to make sure
that resources change classroom practices.
However, the approach of categorical funding is
contrary to the devolution of control to the
school level, and contrary to the tactic of
allocating principals more discretion over
funding so that they can better learn to allocate
resources to the most effective practices.

And so we see at least two contrary
proposals for “new” funding mechanisms that
are more concerned with outcomes. One is
exemplified by the work of Clune (1994),

34

Duncombe and Yinger (1999), and the reforms
in New Jersey: the ﬁrst step would be to
determine how much more funding is necessary
for high-need schools and students, and the
second step is to allocate these funds with
incentives (like elements of performance-based
funding) to spend these resources effectively.
This is a centralized or top-down approach, with
the central authority (district or state)
determining needs and creating incentives.
Implicitly, these plans start from the position,
from the school effectiveness literature and
other arguments, that adequate levels of
resources are necessary but not sufficient for
effectiveness. The ﬁrst task is to define what
“adequate” resources are, the subject of some
effort over the past few years (e.g., Minorini
and Sugannan, 1999; Guthrie and Rothstein,
1999). The second task is to try to promote the
sufficient conditions for effective schools,
usually by imposing some outcome
requirements measured by conventional test
scores (with all their well-known problems), or
by requiring reviews and plans like those in
New Jersey. The possibilities for incorporating
incentives to improve teaching conditions, such
as those mentioned in Section III, have not been
explored so far, and so these funding proposals
fail to link funding in any way to changes
within schools and classrooms.

A second, more bottom-up approach is
the proposal of Miles (1995) and Miles and
Darling—Hammond (1998), in which individual
schools develop their own reform strategies and
then find the resources necessary for them (see
also Odden and Busch, 1998). In many cases,

schools may be able to reallocate existing

 

spending, converting “inert” to “active”
resources—for example by replacing non-
teaching personnel by classroom teachers to
reduce class size and allow all teachers to know
their students well. Only after that would
schools search for new funding, either from
outside sources (like foundations) or from new
public resources. In effect, this approach allows
the school rather than the state legislature to
define what “adequate” resources are, and then
relies on the vision behind the reform to assure
that the resources are effectively spent. This
kind of reform—driven funding—~where schools
with particular visions and speciﬁc funding
needs apply for special-purpose funds—also
underlies the various local education funds
around the country, like the Boston Plan for
Excellence that provides mini-grants for
schools in the midst of reform projects (White,
1999). These funds tend to provide resources
for relatively speciﬁc reform purposes upon
application, so that funders can judge the clarity
of vision and strength of the reform proposal
before allocating any resources.

The top-down funding mechanism is
driven more by the need to provide funding and
the incentives for effective spending to all
schools. The bottom—up approach may be more
effective for those schools with clear visions,
but it doesn’t provide any direction for schools
that are too disorganized, or harried, or
internally contentious, to develop such visions.
One question is then whether some hybrid
approach is possible, avoiding the inequities of
the bottom-up or reform-driven funding
mechanism while still providing discretionary

resources for schools with clear visions, and

providing some incentives for reform for the
most disorganized schools. One possibility, for
example, might be to structure a three-part
funding mechanism:

1. A formula would allocate resources to
individual school accounts,
where—following the current logic of
adequacy in school ﬁnance—more
resources would be allocated to schools
and districts with higher student needs.

2. Schools could spend some relatively high
fraction of these resources by right,
though they would be subject to
performance standards and incentives, as

in top-down approaches.

3. The remaining fraction of funds in their
account (perhaps 10 percent?) could be
accessed only with a multi-year
improvement plan emphasizing spending
that would enhance classroom and school

conditions.

Schools with inadequate performance
might have higher amounts in this
“discretionary” account, or they might be
provided additional technical assistance; this
could potentially get around the problem of the
neediest schools being unable to apply
successfully for the amounts in their
discretionary account. Then equity would be
deﬁned by the total resources in a school’s
account, though actual spending levels might
differ from these amounts. If a school failed to
qualify for all the funds in its account, it would
be partly to blame for not receiving all the
resources it could, and it would have an

35

 

incentive to put together a coherent school
improvement plan.

Such a strategy places much greater
burdens on district or state officials to monitor
performance and evaluate improvement plans
than is now the case, and the feasibility of such
careful monitoring for all 88,000 schools in the
country seems remote, at least for now. But this
is the spirit of the “new” school ﬁnance—that
in the end resources will be much more
effectively spent if some fraction of existing
funds are reallocated to allow more careful
consideration, by individuals within a school as

”29 of how resources should

well as “outsiders,
be used.

Many specific mechanisms of state and
district policy mentioned in this section—the
consideration of how policy changes are likely
to affect school and classroom conditions (or
Classroom Impact Statements), the creation of
incentives for schools to spend resources well,
the specific procedures for school improvement
plans—require much more judgement about the
quality of schooling than is evident in current
debates about ﬁnance, which tend to emphasize
the technical details of funding formulas. But
that too is the spirit of the “new” school
ﬁnance—that policy-makers should start to
worry not only about the allocation of
resources, but also about how those resources
are spent. This in turn requires a certain kind of
politics to emerge, one concerned with the
quality of education rather than the division of

the spoils.

36

VI. Toward a New Narrative

for Spending

Throughout this century, a couple of
simple narratives have dominated the efforts to
fund schools. The dominant one is simply that
more is better, and that the solution to any
problem—whether evidence of ineffectiveness,
or efforts to include students denied full access
to schools—is to increase spending. A second
has been the concern with equity in spending,
particularly equity among districts within states,
creating a long history of efforts to revise
funding formulas and to harness the power of
courts, particularly in lawsuits following the
Serrano case. But the ﬂaws in these narratives
have now become apparent, as spending has
increased without eliminating the dissatisfaction
with public education and as efforts to enhance
equity in spending have failed to narrow the
disparities in outcomes. Evidently, some
different approach is necessary.

The “new” school finance is at least a
candidate for replacing these older narratives.
Without abandoning the current preoccupations
with spending levels and equity, it would add an
emphasis on effectiveness that is quite
consistent with the current interest in
accountability. It responds to the concern within
the history of American education for efﬁciency
(e. g., Callahan, 1967), one that explicitly links
resources with results. And by focusing on
experiences within schools and within
classrooms as well as resources, it is consistent
with the currents of reform over the past 15
years that have emphasized the inner workings

of classrooms and schools, the pedagogies and

 

cultures appropriate to enhanced learning.
There’s much to build on.

But ideas don’t come to dominate policy
unless they attract widespread allegiance, with
support from research (or evidence in general),
teachers and administrators themselves, policy—
makers, and parents. This is why we have
stressed the implication of the “new” school
ﬁnance for several different groups of
participants, in Sections [[1, IV, and V: only
when there is some consistency in perspective
and practices can a new vision emerge.

Of course, there are also barriers to any
novel narrative. At the moment, the nature of
politics is a serious barrier since the
preoccupation with dividing the spoils—a
politics geared to the “old” school finance—is
hostile to the concern with effective practice;
part of changing a policy narrative is changing
the politics that supports it. In addition, policy
narratives are often distinguished by their
simplicity, and it’s possible that a complex
narrative—like the “new” school ﬁnance, with
its emphasis on the multiple necessary
conditions for effectiveness—will prove too
complex and too varied to be widely accepted.
Finally, old habits die hard, and it will be
difﬁcult to introduce these perspectives to the
vast numbers of administrators, teachers,
policy-makers, parents, and researchers in ways
that make them stick, to prevent backsliding into
old ways of thinking. But the alternatives are
grim: without the kind of political change
necessary for the “new” school finance to
emerge, spending for education will keep
escalating (as in Table 1) without much

improvement to show for it.

37

TABLE 1 Expenditure per pupil (ADA)
In public elementary-secondary schools

Constant 1997-98 dollars

School year Current expenditure
1919-20 $453
1929-30 $819
1939-40 $1,020
1949-50 $1,437
1959-60 $2,065
1969=70 $3,494
1974-75 $4,261
1979-80 $4,733
1984-85 $5,307
1989-90 $6,343
1994-95 $6,440
1997-98 $6,624

SOUFCGZ Digest Of Education Statistics (7998), Table 169.

TABLE 2 Educational Expenditure as a Percentage of
Gross Domestic Product, 1995

Expenditures Expenditures for primary
for all levels and secondary education
United States 6.7% 3.9%
Australia 5.6% 3.7%
Austria 5.5% 3.9%
Canada 7.0% 4.3%
Czech Republic 5.7% 3.9%
Denmark 7.1 % 4.3%
Finland 6.6% 4.2%
France 6.3% 4.4%
Germany 5.8% 3.8%
Greece 3.7% 2.8%
Hungary 5.5% 3.6%
Iceland 5.2% 3.6%
Ireland 5.3% 3.4%
Italy 4.7% 3.2%
Japan 4.7% 3.1%
Korea 6.2% 3.8%
Mexico 5.6% 4.0%
Netherlands 4.9% 3.2%
Portugal 5.4% 4.1 %
Spain 5.7% 4.0%
Sweden 6.7% 4.5%
Turkey 2.4% 1.6%

Source: OECD (1998). Tables B1.1a, B1.1b.

 

Endnotes

 

’ NCES (1999), Table 94. These ﬁgures are for 1994-
95, when average spending per pupil in the country
was $5,989.

2 See, for example, the discussion of joint causality in
Marini and Singer (1988), particularly Mackie’s
(1974) “inus” condition where a variable is an
“insufﬁcient but non-redundant part of an unnecessary
but sufficient condition

3 In this review we will consistently refer to both
school-level and classroom-level effects. We often
assume that the two are complementary—for example,
that it’s easier for individual teachers to maintain
order, or high standards when those are priorities for
the school-but we will not develop this relationship.
For the critique that research on teacher effects within
classrooms and research on school effects are largely
independent, see Teddie and Stringﬁeld (1993),
Chapter 10.

See especially the articles in Berne and Picus (1994), a
volume that explicitly searches for the conditions
necessary for outcome equity rather than input equity;
several of the articles in Monk and Underwood (1988),
especially the contributions by Gamoran (1988) and
by Brown and Saks (1988); the articles in Odden
(1992); Odden and Busch (1998); David Monk
(1994a); King and MacPahil—Wilcox (1994); Barro
(1989). Even the interchange between Hedges, Laine,
and Greenwald (1994) and Hanushek (1994), which
turns largely on technical issues of how to carry out a
meta-analysis, ﬁnally concludes that money might
matter under some conditions—though these
conditions are yet unknown. More recently Cohen,
Raudenbusch, and Ball (1999) have distinguished
between “inert” and “active” resources, a distinction
that again clariﬁes that resources are likely to be
ineffective (“inert”) unless spent on certain practices or
accompanied by other conditions to make them
“active.” A long-ago foray into school finance by one
of us tried to distinguish equality of revenues from
equality of inputs, real resources, and outputs (Grubb
and Michelson, 1994, p. 6), where real resources are
precisely equivalent to the teaching conditions
described in section 11 below.

4

5 Public issues tend to be governed by policy narratives,

or easily-understood and widely-accepted stories, that
are themselves the results of long development. In
school ﬁnance, the dominant narrative is still one
about the efﬁcacy of expenditures, which in turn drives
the political pressures for more resources. The contrary
perspective, that “resources don’t make a difference,”
the result of the production function literature reviewed
in Section II, has not become a policy narrative
because it is not supported by any parents, educators,
or policy-makers. On policy narratives, see Roe
(1994).

40

 

6 See also Fuller and Clark (1994), contrasting the
“policy mechanics” in the production function
tradition with the “classroom culturalists”
emphasizing the socialization of children in
classrooms to various norms. They too call for
bridging this particular divide.

7 We use the British term “naming and shaming” to
describe the state or district policies that administer a
standardized test, identify the worst-performing
schools, shame them in the local papers, and then
require them to make some changes. The process of
“naming” or identifying these low-performing schools
is fraught with methodological problems, and the
remedies (ranging from state takeover to reconstitution
to pitifully small grants to self-improvement plans)
are invariably inadequate to the task. The only
effective feature of these policies seems to be the
shaming process. These policies, which have
burgeoned as part of the standards movement, merit
substantially more analysis.

8 The A.I.R. report is a good example of simpliﬁcation
in action. Reforms are rated on a 5-point scale similar
to that used by Consumer Reports; the text provides a
few additional comments on the quality of evidence,
but fails to provide any citations for those who might
want to check the data for themselves. It’s hard to
learn much from this volume except that the evidence
in favor of “promising practices” isn’t substantial.

9 Random—assignment studies are much more common
in job training but not necessarily more informative,
because they often fail to ask what characteristics of
programs are responsible for their success or failure.
For this critique see Grubb (1996) or Grubb and Ryan
(1999).

'0 Some consensus seems to be emerging on hybrid
methods drawing from both traditions (see the
National Research Council’s review of the reading
wars by Snow, Burns, and Griffin, 1998), but hybrid
approaches are immensely varied and probably
unstable.

” There’s a somewhat different literature examining the
effects of school spending by state on subsequent
earnings, but this too yields ambiguous results; see
Card and Krueger (1996).

‘2 Presumably this is true, though there was a
reassignment after kindergarten that was certainly not
random. Krueger (1997) has corrected for this problem
and conﬁrmed the effects of smaller class sizes, though
the effects are substantial (4 percentile points) only for
the ﬁrst year a student is assigned to a small class, and
drop to one percentile point for subsequent years.

‘3 It might be possible to develop more easily—measured
proxies instead of direct observations of classroom and
school practices. For example, Mayer (1999) has
found that self-reported measures of behaviorist versus

 

 

constructivist teaching practices are highly correlated
with observation-based measures. But such proxies
could be used only after they had been validated
through observation, and so the process of collecting
information on the way resources are used would be
much more difﬁcult.

‘4 The reformulation in equations (3) and (6) continue to
assume that there are measurable characteristics that
affect outcomes in stable ways. Murnane and Phillips
(1981), ﬁnding that teacher characteristics did not
affect learning, argued that effective teachers do not
have any common characteristics, except that they are
able to discover early in the school year the subtle
interventions, varying among students, that make for
effective teaching. Whether this means that
relationships like equation (3) are impossible, or
whether it implies that a particularly difficult measure
of TC is necessary, is unclear.

‘5 I have taken the idea of such an equilibrium from
Harkin and Davis (1996a, b). The idea is useful is
explaining why some classes seem to “collapse”; see
Grubb and Associates (1999), especially Chapter 2
and 6.

16 If two kinds of resources are jointly necessary, then
schools will be effective only when both are
present—but the linear additive functional forms of
most statistical work will find each of them
independently contributing to outcomes. Interaction
terms are not necessarily satisfactory because they
often generate collinearity and imprecise parameter
estimates.

‘7 The positive effect of teacher verbal scores or ability
measured in some other way is one of the only
relatively strong ﬁndings in the conventional
production function literature.

18 Often people talk about class size reduction as
valuable for allowing a greater amount of “one—on
one” or individualized instruction. However, if classes
are reduced from 25 to 20, and teachers spend about
half their time on whole-class instruction and the
remaining half on “one on one,” then a reform that
costs an additional 20 percent in resources for
classroom teachers increases the amount of
individualized instruction for every student from 18
hours per year (180 days x 5 hours per day x 1/2) to
22.5 hours. In small amounts throughout the school
day this might be quite effective, but only if teachers
are using individualized instruction in effective and
equitable ways.

‘9 In the ﬁeld of writing, a meta—analysis has concluded
that the presentational (or didactic) mode and the
conventional teaching of grammar are the least
effective (Hillocks, 1986); similarly, the National
Assessment of Educational Progress concluded that
writing proﬁciency is positively related to teachers’
use of the writing process (Applebee et al., 1994)

 

rather than grammar-based instruction. Knapp and his
colleagues (Knapp et al., 1993', Knapp et al., 1995)
examined math and English scores in elementary
classrooms with high proportions of low-income
students, and found that classrooms with larger
numbers of “alternative” practices—which they deﬁned
almost precisely as we describe meaning-making—led
to signiﬁcantly higher scores. Some speciﬁc practices
in the meaning- and student-centered tradition have
been found to be more effective, including co-operative
learning (Slavin, 1980; Walberg, 1986) and reciprocal
teaching (Palincsar and Brown, 1984; Brown and
Palincsar, 1989; Brown and Campione, 1994).

20 Teachers in reduced-size classrooms reported less time
spent on discipline, and more time addressing student’s
individual concerns; some other differences (like
discussing student-initiated topics and diagnosing
individual learning needs) were positive but not
signiﬁcant. The direction of changes were all
consistent with shifts toward more student-centered and
constructivist practices, though they were too small to
be of any real effect. But these averages conceal
substantial variation, and so the real question is which
teachers, under what conditions, changed their teaching
practices.

2‘ In a world of fungible resources, it’s often difficult to
detect what a particular source of revenue is spent on.

22 In all these evaluations the comparability of the
comparison group is an issue, since none of these
evaluations use random assignment methods.

23 Technically, schools with large positive residuals
from equation (1) would be selected. The error term in
this equation is a function of the error terms in
equations (2) and (3) and (depending on the functional
form of these equations) some of the parameters in
these equations as well. In general, then, these will be
schools that have either high levels of positive
teaching conditions given resources (high e), high
levels of outcomes given teaching conditions and the
family backgrounds of their students (high u), or
particularly large parameters connecting resources to
outcomes, from equation (3). Case studies of
exemplary schools deﬁned in this way would then
search for evidence about which of these characteristics
seem important. Having admitted that linear regression
is a poor technique to capture the causal mechanisms
in schools—particularly the joint effects of necessary
but individually insufficient conditions—it would not
make sense to consider these statistical methods as
anything but a rough guide for further study.

2“ Fuller and Clark (1994) summarize the effects of a
number of variables describing classroom
organization, pedagogy, school management—all
aspects of teaching conditions. See also Fuller et a1.
(1999), who were able to examine the inﬂuence of a
large number of teacher practices on literacy in Brazil
through the use of classroom observations.

41

 

 

25 See especially the discussion by Cohen,
Raudenbusch, and Ball (1999) of teaching “regimes,”
which are ways of adjusting teaching to the interests,
capacities, strengths and weaknesses of individual
students. Capable student-centered teachers therefore
change their practices for individual students, whereas
conventional didactic instructors have only one
prescribed regime—so the teaching conditions in their
classes are stable across students and easier to measure.
This idea is similar to the point of Brown and Saks
(1984, 1987), that teachers may allocate resources
differently among students within their classrooms.

26 Oral communication, Richard Lavine, Center for
Public Policy Priorities, Austin, TX. Apparently
these rational strategies are widely acknowledged
though they have not been formally researched.

27 See, for example, the treatment of resources in
conventional texts for school administrators. To take a
random sample, Hughes (1999) has only two pages on
funding; Speck (1999), while paying considerable
attention to the educational rather than managerial
roles of principals, has nothing at all about resource
allocation; Seyfarth (1999) and Drake and Roe (1999)
each allow a single chapter near the end of their texts,
and both treat resources as budgeting issues rather than
as educational decisions. For one effort to clarify the
funding issues for administrators, see Monk and Plecki
(1999). It may be that the neglect of resource
allocation simply reﬂects the reality that most
principals have little discretion, or don’t know enough
about effective allocation—but the inattention to
resources simply perpetuates the problem.

28 Litigation is a particularly crude instrument for
reshaping policy, since it can only say that certain
practices are unconstitutional. When courts have
declared current inequitable patterns of spending to be
unconstitutional, legislatures have responded with
more equitable distributions of resources. But courts
cannot order greater equity in effective spending or
active resources because that would take them too far
into the prerogatives of educators. The only solution
has therefore been to throw the decisions back to state
legislatures. In the few cases mentioned in the text,
legislatures have responded with more elaborate
reforms in addition to equalizing funding.

29 As a side note, we advocate classroom observations
by insiders and outsiders as the basis for school
improvement plans. Under the best conditions in
England, inspections of schools and colleges by a
combination of insiders and outsiders is used to
identify the “corporate” or institutional methods of
enhancing the quality of teaching, and then schools
develop multi—year improvement plans to implement
these reforms (Grubb, 1999 or forthcoming). Of
course, inspection can be used for punitive as well as
supportive purposes, so it must be carefully designed.

42

 

For some efforts to incorporate inspection (or school
visits) into accrediting visits, see Wilson (1999).

 

References

Adams, J. E. (1997). School ﬁnance policy and students’ opportunities to learn: Kentucky’s experience. The future of
children, 7( 2 )(winter), 79-95.

American Institutes for Research (1999). An educators guide to school wide reform. Arlington, VA: Educational Research
Service.

Applebee, A., Langer, J ., Mullis, 1., Latham, A., & Gentile, C. (1994, June). NAEP 1992 Writing report card. National
Center for Educational Statistics, Report No. 123-W01. Washington DC: US. Government Printing Office.

Austin, G. R. ( 1979). An analysis of outlier exemplary schools and their distinguishing characteristics (Paper presented
at the meeting of the American Educational Research Association). San Francisco.

Ballou, D. (1998). The condition of urban school ﬁnance: Efﬁcient resource allocation in urban schools. In W. J.
Fowler, Jr. (Ed.), Selected Papers in School Finance, 1996 (NCES 98-217).

Berne, R., & Picus, L. O. (Eds). (1994). Outcome equity in education. Thousand Oaks, CA: Corwin Press.

Boyd, W. L., & Hartman, W. T. (1988). The politics of educational productivity. In D. H. Monk, & J. Underwood
(Eds.), Microlevel school ﬁnance: Issues and implications for policy. (pp. 271—308). Cambridge, MA: Ballinger
Publishing.

Brophy, J ., & Good, T. L. (1986). Teacher behavior and student achievement. In M. Wittrock (Ed.). Handbook of
research on teaching, 3rd ed., (pp. 328-375).

Brown, A., & Campione, J. (1994). Guided discovery in a community of learners. In K. McGilly (Ed.), Classroom
lessons: Integrating cognitive theory and classroom practice. Cambridge: MIT Press/Bradford Books.

Brown, A., & Palincsar, A. (1989). Guided, co-operative learning and individual knowledge acquisition. In Lauren
Resnick (Ed. ), Knowing, learning, and instruction: Essays in honor of Robert Glaser. Hillsdale NJ: Lawrence
Erlbaum.

Brown, B. W. (1988). The microeconomics of learning: Students, teachers, and classrooms. In D. H. Monk, & J.
Underwood (Eds.), Microlevel school ﬁnance: Issues and implications for policy. (pp. 183-205). Cambridge, MA:
Ballinger Publishing.

Brown, B. W., & Sacks, D. H. (1975). The production and distribution of cognitive skills within schools. Journal of
Political Economy, 83 (3), 571-593.

Brown, B. W., & Sacks, D. H. (1987). The microeconomics of the allocation of teachers’ time and student learning.
Economics of Education Review, 6 (4), 319-332.

Callahan, R. (1967). Education and the cult of eﬂiciency. Chicago: University of Chicago Press.

Card, D., & Krueger, A. B. (1996). School resources and student outcomes: An overview of the literature and new
evidence from North and South Carolina. Journal of Economic Perspectives, 10(4), 31-50.

Clark, D. L., Lotto, L. S., & Astuto, T. A. (1984). Effective schools and school improvement: A comparative analysis
of two lines of inquiry. Educational Administration Quarterly, 20(3), 41-68.

Clune, W. (1994). The shift from equity to adequacy in school ﬁnance. Educational Policy, 8(4):376—394.

Cohen, D., Raudenbusch, S., & Ball, D.(l999, May). Educational resources, instruction, and research. Unpublished
paper, School of Education, University of Michigan.

Comer, J. (1996). Rallying the whole village: The Comer process for reforming education. New York: Teachers College
Press.

Cook. T. (1999, March). Considering the major arguments against random assignment: An analysis of the intellectual
culture surrounding evaluation in American schools of education. WP—99—2. Institute for Policy Research,
Northwestern University

CSR Research Consortium (1999, June). Class size reduction in California 1996-98: Early ﬁndings signal promise and
concerns. Palo Alto: EdSource.

43

 

Cuban, L. (1984). Transforming the frog into a prince: Effective schools research, policy, and practice at the district
level. Harvard Educational Review, 54(2), 129-151.

Cuban, L. (1990, January). Reforming again, again, and again. Educational Researcher, 19(1), 3-13.

Cuban, L. (1993). How teachers taught: Constancy and change in American classrooms 1890-1990. New York: Teachers
College, Columbia University.

Cubberly, E. (1905). School funds and their apportionment. New York: Teachers College.

Darling-Hammond, L. & McLaughlin, M. W. (1995). Policies that support professional development in an era of
reform. Phi Delta Kappan, 76 (7), 597-604.

Drake, T. L., & Roe, W. H. (1999). The principalship (5th ed.). Upper Saddle River, NJ : Prentice Hall.
Edmonds, R. R. (1979). Effective schools for the urban poor. Educational Leadership, 37, 15-24, 37.
Edmonds, R. R. (197%.). Some schools work and more can. Social Policy, 9, 28-32.

Ferguson, R. F. (1991). Paying for public education: New evidence on how and why money matters. Harvard Journal on
Legislation, 28(Summer), 465-497.

Ferguson, R. F., & Ladd, H. F. (1996). How and why money matters: An analysis of Alabama schools. In Holding
schools accountable: Performance based reform in education. (pp. 265-298). Washington, DC: Brookings
Institution.

Finkelstein, N., & Grubb, W.N. (1998, April). Making sense of educational markets: Lessons from England. Paper
presented at the American Educational Research Association. Berkeley, CA: School of Education, University of
California.

Finn, J. D., & Achilles, C. M. (1990). Answers and questions about class size: A statewide experiment. American
Educational Research Journal, 27 (3), 557-577.

Firestone, W. A., Goertz, M. E., Nagle, B., & Smelkinson, M. F. (1994). Where did the $800 million go? The first
year of New Jersey’s quality education act. Educational Evaluation and Policy Analysis, 16 (4), 359—373.

Fuller, B., & Clarke, P. (1994). Raising school effects while ignoring culture? Local conditions and the inﬂuence of
classroom tools, rules, and pedagogy. Review of Educational Research, 64(1), 119-157.

Fuller, B., Dellagnelo, L., Strath, A., Bastos, E. S. B., Maia, M. H., DeMatos, K. S. L., Portela, A. L., & Vieira, S L.
(1999). How to raise children’s early literacy? The inﬂuence of family, teacher, and classroom in Northeast Brazil.
Comparative Education Review, 43(1), 1-35.

Garnoran, A. (1988). Resource allocation and the effects of schooling: A sociological perspective. In D. H. Monk, & J.
Underwood (Eds), Microlevel school ﬁnance: Issues and implications for policy. (pp. 207—232). Cambridge, MA:
Ballinger Publishing.

Gandara, P., Mejorado, M., Gutierrez, & Molina, M. (1998). Final report of the evaluation of high school Puente.
Davis, CA: University of California, Davis.

Goertz, M. E., & Natriello, G. (1999). Court-mandated school finance reform: What do the new dollars buy? In H. F.
Ladd, R. Chalk, & J. S. Hansen (Eds), Equity and adequacy in education ﬁnance: Issues and perspectives (pp. 99-
135). Washington, DC: National Academy Press.

Goldhaber, D. D., & Brewer, D. J. (1996). Why don’t schools and teachers seem to matter? Assessing the impact of
unobservables on educational productivity. Forthcoming in Journal of Human Resources.

Goldstein, H., & Blatchford, P. (1998). Class size and educational achievement: A review of methodology with particular
reference to study design. British Educational Research Journal, 24 (3), 255-268.

Gray, J. (1990). The quality of schooling: Frameworks for judgements. British Journal of Educational Studies 38(3):204-
233.

Grubb, W.N. (1995). Coherence for all students: High schools with career clusters and majors. In W. N. Grubb (Ed.),
Education through occupations in American high schools. Vol. I: Approaches to integrating academic and
vocational education. New York: Teachers College Press.

Grubb, W.N. (1996). Learning to work: The case for re-integration education an job training. New York: Russell Sage.

44

 

Grubb, W.N. (1999). Improvement or control: An American View of British inspection. In C.I. Cullingford, (Ed.), An
inspector calls: OFSTED and the eﬁ‘ects of inspection. London: Kogan Page.

Grubb, W.N. (forthcoming). Opening classrooms and improving teaching: Lessons from school inspection in England.
Teachers College Record

Grubb, W.N., & Associates. (1999). Honored but invisible: An inside look at teaching in community colleges. New
York and London: Routledge.

Grubb, W.N., & Michelson, S. (1974). States and schools: The political economy of public school ﬁnance. Lexington
MA: Lexington Books.

Grubb, W.N. , & Ryan, P. (1999). The role of evaluation for education and training: Plain talk on the ﬁeld of dreams.
London: Kogan Page and Geneva: International Labour Ofﬁce.

Guthrie, J. W., & Rothstein, R. (1999). Enabling adequacy to achieve reality: Translating adequacy into state school
finance distribution arrangements. In H. F. Ladd, R. Chalk, & J. S. Hansen (Eds), Equity and adequacy in
education ﬁnance: Issues and perspectives (pp. 209-259). Washington, DC: National Academy Press.

Hanushek, E. A. (1997). Assessing the effects of school resources on student performance: An update. Educational
Evaluation and Policy Analysis, 19 (2), 141-164.

Hanushek, EA. (1986, September). The economics of schooling: Production and efﬁciency in public schools. Journal of
Economic Literature, 24, 41-77.

Hanushek, EA. (1992, February). Money might matter somewhere: A response to Hedges, Laine, and Greenwald.
Educational Researcher. 23:4, 5-8.

Hanushek, E.A., & Rivkin, S.G. (1997, Winter). Understanding the 20th century growth in US. school spending.
Journal of Human Resources. 32(1), 35-68.

Hanushek, EA, and others. (1994). Making schools work: Improving performance and controlling costs. Washington,
DC: Brookings Institution.

Harkin, J ., & Davis, P. (1996a). The communications styles of teachers in post-compulsory education. Journal of
Further and Higher Education, 20(1), 25-34.

Harkin, J., & Davis, P. (1996b). The impact of GNVQs on the communications styles of teachers. Research in Post-
Compulsory Education, 1(1), 97-107.

Hedges, L.V., Laine, R.D., & Greenwald, R. (1994, April). Does money matter? A meta—analysis of studies of the
effects of differential school inputs on student outcomes. Educational Researcher, 23(3): 5—14.

Hess, G. A., Jr. (1999). Understanding achievement (and other) changes under Chicago school reform. Educational
Evaluation and Policy Analysis, 21 (1), 67-83.

I-Iickrod, G. A., Chaudhari, R., Pruyne, G., & Meng, J. (1997). The effect of constitutional litigation on educational
ﬁnance: A further analysis. In W. J. Fowler, Jr. (Ed), Selected papers in school ﬁnance, 1995 (NCES 97—536).

Hillocks, G. (1986). Research on written composition: New directions for teaching. Urbana, IL: ERIC Clearinghouse on
Reading and Communications Skills and National Conference on Research in English.

Hughes, L. W. (1999). The principal as leader (2nd ed.). Upper Saddle River, NJ: Prentice-Hall.

Kerchner, C., and Koppich, J. (1993). A union of professionals: Labor relations and educational reforms. New York:
Teachers College Press.

King, R. A., & MacPhail—Wilcox, B. (1994). Unraveling the production function: The continuing quest for resources
that make a difference. Journal of Education Finance, 20(Sumrner), 47—65.

Knapp, M., & Associates. (1995). Teaching for meaning in high-poverty classrooms. New York: Teachers College
Press.

Knapp, M., Shields, P., & Turnbull, B. (1993). Academic challengefor the children ofpoverty. Volume 1: Findings and
conclusions. Washington, DC: U. S. Department of Education.

K0201, J. (1992). Savage inequalities: Children in America’s schools. New York: Harper Perennial.

Krueger, A. (1997, March). Experimental estimates of education production functions. Princeton N.J.: Princeton
University and National Bureau of Economic Research.

45

 

Ladd, H., Chalk, R., & Hansen, J. (1999). Equity and adequacy in education ﬁnance: Issues and perspectives.
Washington, DC: National Academy Press.

Lankford, H., & Wyckoff, J. (1995). Where has the money gone? An analysis of school district spending in New York.
Educational Evaluation and Policy Analysis, 17(2), 195-218.

Levin, H. M. (1994). The necessary and sufficient conditions for achieving educational equity. In R. Berne, & L. O.
Picus (Eds), Outcome equity in education. (pp. 167-190). Thousand Oaks, CA: Corwin Press.

Lieberman, A. (1995). Practices that support teacher development: Transforming conceptions of professional learning.
Phi Delta Kappan, 76 (7), 591-596.

Little, J. W., & Dorph, R. (1998). Lessons about comprehensive school reform: California’s school restructuring
demonstration program. Berkeley, CA: University of California, Graduate School of Education.

Mackie, J. L. (1974). The cement of the universe: A study of causation. Oxford: Oxford University Press.

MacPhail-Wilcox, B. (1986). Production functions revisited in the context of educational reform. Journal of Education
Finance, 12(2)(Fall), 191-222.

Marini, M., and Singer, B. (1988). Causality in the social sciences. In Clifford Clogg, (Ed), Sociological Methodology
1988. Washington DC: American Sociological Association.

Mayer, D. P. (1999). Measuring Instructional Practice: Can policymakers trust survey data? Educational Evaluation and
Policy Analysis, 21(1), 29-45.

Meier, D. (1995). The power of their ideas: Lessons for America from a small school in Harlem. Boston: Beacon Press.

Miles, K. H. (1995). Freeing resources for improving schools: A case study of teacher allocation in Boston public
schools. Educational Evaluation and Policy Analysis, 17(4), 476-493.

Miles, K. H., & Darling-Hammond, L. (1998). Rethinking the allocation of teaching resources: Some lessons from
high-performing schools. Educational Evaluation and Policy Analysis, 20(1), 9—29.

Minorini, P. A., & Sugarman, S. D. (1999). Educational adequacy and the courts: The promise and problems of moving
to a new paradigm. In H. F. Ladd, R. Chalk, & J. S. Hansen (Eds), Equity and adequacy in education ﬁnance:
Issues and perspectives (pp. 175—208). Washington, DC: National Academy Press.

Minorini, P. A., & Sugarrnan, S. D. (1999). School finance litigation in the name of educational equity: Its evolution,
impact, and future. In H. F. Ladd, R. Chalk, & J. S. Hansen (Eds), Equity and adequacy in education ﬁnance:
Issues and perspectives (pp. 34-71). Washington, DC: National Academy Press.

Monk, D. H. (1994). Policy challenges surrounding the shift toward outcome-oriented school ﬁnance equity standards.
Educational Policy, 8(4), 471-488.

Monk, D. H. (1994). Subject area preparation of secondary mathematics and science teachers and student achievement.
Economics of Education Review, 13(2), 125-145.

Monk, D. H., & Plecki, M. L. (1999). Generating and managing resources for school improvement. In J. Murphy, & K.
S. Louis (Eds), Handbook of research on educational administration: A project of the American educational
research association (2nd ed.) San Francisco: Jossey-Bass Publishers.

Monk, D. H., & Underwood, J. (1988). Microlevel school ﬁnance: Issues and implications for policy. Cambridge, MA:
Ballinger Publishing.

Monk, D., Nakib, Y., Odden, A., & Picus, L. (1995, October). The story of the education dollar: No academy awards and
no ﬁscal smoking guns. Phi Delta Kappan 77(2): 161-168.

Mosteller, F. (1995). The Tennessee study of class size in the early school grades. The Future of Children, 5(2), 113-
127.

Murnane, R., and Levy, F. (1996). Evidence from ﬁfteen schools in Austin, Texas. In Gary Burtless, (Ed.), Does money
matter: The eﬁ‘ect of school resources on student achievement and adult success. Washington DC: Brookings
Institution Press.

Murnane, R. J., & Phillips, B. R. (1981). What do effective teachers of inner—city children have in common? Social
Science Research, 10, 83-100.

National Center for Education Statistics (1999, May). Digest of education statistics 1998. NCES 1999—036. Washington
DC: Ofﬁce of Educational Research and Improvement, US. Department of Education.

46

 

Nye, B. A., Boyd-Zaharias, J., Fulton, B. D., & Achilles, C. M. (1993). The lasting beneﬁts study: Grade 6 technical
report (TSU-94-0030(A)11531380). Nashville, TN: Center of Excellence for Research in Basic Skills, Tennessee
State University.

Odden, A. (1998). Creating school finance policies that facilitate new goals. Philadelphia: University of Pennsylvania,
Graduate School of Education, Consortium for Policy Research in Education.

Odden, A. R. (Ed.). (1992). Rethinking school ﬁnance: An agenda for the 1990s. San Francisco: Jossey-Bass.

Odden, A., & Busch, C. (1998). Financing schools for high performance: Strategies for improving the use of educational
resources. San Francisco: Jossey-Bass.

Palincsar, A. S., & Brown, A. L. (1984). Reciprocal teaching of comprehension-fostering and monitoring strategies.
Cognition and Instruction, (1), 117-175.

Payne, C. M. (1997). “I don’t want your nasty pot of gold ”: Urban school climate and public policy. WP-97-8.
Evanston, 111: Institute of Policy Research, Northwestern University.

Payne, K., and Biddle, B. (1999). Poor school funding, child poverty, and mathematics achievement. Educational
Researcher. 28(6): 4—13.

Peterson, B., & Charney, M. (Eds). (1999). Transforming teacher unions: Fighting for better schools and social justice.
Milwaukee, WI: Rethinking Schools, Ltd.

Picus, L. O. & Wattenberger, J. L. (Eds). (1996). Where does the money go .7: Resource allocation in elementary and
secondary schools. Thousand Oaks, CA: Corwin Press.

Powell, A. G., Farrar, E., & Cohen, D. K. (1985). The shopping mall high school: Winners and losers in the
educational marketplace. Boston: Houghton Mifflin.

President’s Committee of Advisors on Science and Technology: Panel on Educational Technology. (1997). Report to the
president on the use of technology to strengthen K-12 education in the United States. Washington, DC.

Purkey, S. C., & Smith, M. S. (1983). Effective schools: A review. The Elementary School Journal, 83(4), 436—452.

Raudenbush, S. W., Fotiu, R. P., & Cheong, Y. F. (1998). Inequality of access to educational resources: A national
report for eighth-grade math. Educational Evaluation and Policy Analysis, 20 (4), 253—267.

Raudenbush, S. W., Fotiu, R. P., & Cheong, Y. F. (in press). Synthesizing results from the Trial State Assessment.
Journal of Educational and Behavioral Statistics.

Report of the National Education Commission on Time and Learning. (1994). Prisoners of time. Washington, DC: U.
S. Government Printing Ofﬁce.

Roe, E. (1994). Narrative policy analysis: Theory and practice. Durham, NC and London: Duke University Press.

Rowan, B., Bossert, S. T., & Dwyer, D. C. (1983). Research on effective schools: A cautionary note. Educational
Researcher(April), 24-3 1 .

Rumberger, R. W., & Larson, K. A. (1998). Student mobility and the increased risk of high school dropout. American
Journal of Education, 107 ( 1), 1-38.

Sammons, P., Hillman, J ., & Mortimore, P. (1995). Key characteristics of effective schools: A review of school
eﬂectiveness research. University of London: Institute of Education.

Seyfarth, J. T. (1999). The principal: New leadership for new challenges. Upper Saddle River, NJ: Prentice-Hall.

Shapson, S. M., Wright, E. N., Eason, G., & Fitzgerald, J. (1980). An experimental study of the effects of class size.
American Educational Research Journal, 17(2), 141-152.

Slavin, R. (1980). Cooperative learning. Review of Educational Research 50, 315-342.

Snow, C., Burns, M. S., & Griffin, P. (1998). Preventing reading diﬁ‘iculties in young children. Washington, DC:
National Academy Press.

Speck, M. (1999). The principalship: Building a learning community. Upper Saddle River, NJ: Prentice-Hall.

Stern, D., Dayton, C., Paik, 1., & Weisberg, A. (1989). Beneﬁts and costs of dropout prevention in a high school
program combining academic and vocational education. Educational Evaluation and Policy Analysis,
11 (4), 405-416.

47

 

Stern, D., Raby, M., & Dayton, C. (1992). Career academies: Partnerships for reconstructing American high schools.
San Francisco: Jossey-Bass.

Taylor, C. (1998). Does money matter? An empirical study introducing resource costs and student needs to educational
production function analysis. In W. J. Fowler, Jr. (Ed.). Developments in school finance 1997 (NCES 98-212, pp.
7598). Washington, DC: National Center for Education Statistics.

Walberg, H. (1986). Syntheses of research in teaching. In M. Wittrock (Ed.), Handbook of research on teaching, (3rd ed.).
New York: MacMillan Publishing.

White, K. (1999, June 9). Local funds playing larger roles in reform. Education Week 18(39): 1, 10.
Wilson, T. (1999, March). Foundations of the Catalpa school visit. Third Edition. Providence, R.I.: Catalpa.

48

EEEEEEEEEEEEEEEEE

\Iimim\gmmmmummmxmm

l'ﬁELI 53!:

POIiCV Analysis for California Education PACE
UhiVGFSitVO‘F California, Berkeley

and Stanford University

3553 Tolman Hall

Berkeley, CA 94720-1670

Telephone: (510) 642-7223
Fax: (510) 642-9148
http: //pace.berkeley.edu